Turquoise Energy News #137
covering October 2019 (Posted November 5th 2019)
Lawnhill BC Canada
by Craig Carmichael

www.TurquoiseEnergy.com = www.ElectricCaik.com = www.ElectricHubcap.com

Month In "Brief" (Project Summaries etc.)
  - House - New Chemie Batteries - Ground Effect Vehicle - Two Lithium Battery Charger Circuits; AC and DC Sources - Internet! - Planetary Gears!: Miles Truck & Chevy Sprint

In Passing (Miscellaneous topics, editorial comments & opinionated rants)
 - Gardening, Growing Barley, Processing Grains into Flour With Common Kitchen Appliances - Hair Preservation with Recap - Continuing Fatal Financial Flops - Small Thots - ESD

- Detailed Project Reports -
Electric Transport - Electric Hubcap Motor Systems
* Ground Effect Vehicle (R/C Model)
  - Ultimate Key to Longitudinal Stability and Pitch Control: a Canard - Modern Fabrics and Paints
* EV Transmissions: Off-the-Shelf Planetary Gears, More. (Miles Truck & Chevy Sprint)

Other "Green" Electric Equipment Projects
* Working with the Handheld Bandsaw Mill (& Alaska Mill)

Electricity Generation
* 5 Blade Windplant
* My Solar Power System: - Monthly Solar Production log et cetera - 8 months! Notes.

Electricity Storage
Turquoise Battery Project (Mn-Zn or Ni-Zn in Mixed Alkaline Salt electrolyte)
* Electrolyte to Depassivate Discharged Manganese Oxides? - Air Leaks? - Bleach - Nickel Oxides Electrode - Making NiOOH - Better Technique - Paste Electrodes? - 3D Printer Insulated Surround
-- Other Battery Related Projects --
* Honda NiMH Hybrid Car Batteries (installed in solar system)
* Two Lithium Battery Charger Circuits: for AC and DC Power

October in Brief

   I note on November 1st that this is the third anniversary of my viewing of the house at Lawn Hill on Haida Gwaii where I now reside. While it's an odd layout, I certainly lucked out getting this particular spacious, well built and still pretty new place, with its big workshop and a storage room the same size under it, two garages, on five acres and with a lovely sandy ocean beach.

South side of house with new 5-blade windplant, porch,
livingroom bay window, last section of spruce log
to mill (debarked, with metal roof pieces cover ...and
some remaining bits of clutter from the 2nd last section),
small bedroom windows, back door of garage and workshop,
solar panels on roof and on lawn, 'double' greenhouse.
The white painted wall behind the garden gives
more much needed light to the plants near the wall.
(There's a wire fence. The deer haven't figured out they
could go up and across the porch into the garden.)

And here's the north side as seen from my neighbor's yard.
(My fence with rotting posts blew down in a high wind.)
[...and what was with my camera that day? Fog on the lens?]

   I think my angels must have worked overtime to get it for me because I was somewhat clueless at the time and thought I wanted a much cheaper but older and inferior "fixer-upper" place 15 Km north of it. I would have had far more money left over, but I would have been spending my time fixing it up instead of doing more valuable projects. It would have been notably farther from town and is so low lying as to be in jeopardy if - when - sea levels rise. (What was I thinking?) But it had just been sold and this place was the only acreage left on the east coast where I wanted. From viewing it, I knew it was nice, but it was just a quick look and I didn't have all the details. Once here I gradually recognized how ideal a place it is to live and work.

   On October 7th and 8th I finally did a last polish on the application to the "Charge the Future" battery challenge and sent it off. Not that it was very polished. In January we'll hear whether those who have been awarded some funding includes Turquoise Energy.
   According to the information on line, they are expecting funding recipients to just be starting with a concept for a better battery, and hoping the recipients will be able to use the funds to make a working prototype by the end of the project. I am starting with essentially working prototypes after already working on the R & D for 12 years unpaid, and I hope to use the funding not only to improve them via more R & D, but to start actually producing batteries by the end of the project. I hope they are thrilled with that prospect. After that, if manufacture and sales can be made viable as I hope, it should fund itself and even provide some funding for other projects.

New Chemie Batteries

   I'm not very ready to manufacture batteries yet, but every month brings improvements. The lagging item now is better positive electrodes. At this point I don't care about the exact chemistry. Nickel oxides electrodes work even if nothing else I try does. The other components all seem to work well except for the warping, ever-leaking cells, and that's just a matter of getting and installing the proper software for the new 3D printer - or maybe even just turning up the extruder temperature another 10 or 20 degrees.

Nickel oxyhydroxide electrode on a graphite
 foil current collector in a 3D printed "test
            size" battery cell case.
   At the start of October the only cell that had worked and recharged reasonably well was the nickel oxides-zinc one in alkali, using nickel oxide electrode chunks taken from a commercial NiMH cell. The difference between it and any other nickel-zinc battery was the jelling of the zinc that should make it very long cycle life instead of very short. So it would be something of special value to produce, even if none of the other experiments and ideas were to work out.
   But before the end of the month I had made a cell with my own gelled nickel oxides electrode, and proven that it works with chloride salt in the electrolyte as well as with just hydroxide. The amp-hours is low, and this must be the nickel "+" side's problem, because the zinc (even the very same electrode) has worked better with other "+"es. I am confident of being able to make further chemical improvements as well as to get decent overall performance from whatever chemistry is used.

   The chemical improvement I really want to get working is the nickel manganates positive electrode. It should have higher currents than nickel oxides, more amp-hours, and recharge properly, which plain manganese oxides don't seem to do. The other item is the calcium-copper oxide coated copper positive current collector. If it works it should allow substantially higher currents than graphite based current collectors. Those items will be the November (and maybe December) battery research focus.

Ground Effect Vehicle

   But at last I got to some other projects besides new chemie batteries this month. I came up with yet another exciting new idea for the ground effect vehicle: a front canard replacing the rear elevator of all previous ground effect craft. Flying so low, a ground effect craft has great need of being able to respond instantly to gain and loss of ground effect lift as it goes over waves, without gain or loss of altitude. A rear elevator can only respond indirectly by lowering the tail even more than the wing is already dropping. That's not okay when flying so close to the water.
   The canard can itself gain or decrease lift at the bow almost even before altitude is gained or lost, keeping the height steady and level. I visualize a microcontroller operated canard, so the human operator only has to control the throttle and steering controls. Liftoff and setdown will (maybe) be handled automatically as the speed is increased or dropped. It is said that ground effect craft have been restricted to flying over pretty calm water. With a canard I expect this one will be much more seaworthy.
   Perhaps I should have thought of this much sooner - after all the only airplane I've ever flown was a Vari-Eze with a rear propeller, swept back wings and a front canard. (Okay, I only kept it level and on course for a while while the pilot smoked his pipe. In 1985.)

   And I even did a little actual work on the R/C model. I made and mounted the canard and made the rear elevator (which now will just be a fixed rear end of the wing - maybe adjustable for trim). I'll seal all the fabric with Minwax "Polycrylic", which model airplane makers have been raving about recently as a superior replacement for nitrate/cellulose dope - and even for products specifically made to replace that dope.
   The position of the canard (pretty much the only place I could have put it at this point in the construction) seems ideal: not only does it deflect air up and down as it is lowered and raised, but it deflects it right over or under the main wing, magnifying its lift changing effect.

  And on November 5th before posting this, I realized that a better place for the ducted fans would be to mount them on the canard. I'll make that change. Better and better!

  The potential to really open up the BC north coast, isolated islands and other hard to get to places is more there than ever.


   I installed my windplant the day before a huge storm. I wired it up after finally figuring out what to do, on a day with enough wind to test it. I used the MPT-7210A boost charge controller that had once been running my DC solar system. It looked good and spun well, but the output wasn't a lot. In the storm if things had been configured it might have hit 100 watts.

   On the evening of the 21st a chair blew off my porch, the power went out almost all night, and a section of my fence blew over. The posts had rotted away at ground level. Wind! Great - more work! I decided to leave it down. Firewood!

   But the next day I finally came up with a plan for connecting the windplant into the DC power system to charge the batteries: with an MPT7210A boost solar charge controller. It could take the lower voltages from the windplant and boost them to charge the 36 volt DC battery system. It was about right to handle the range of voltages I was getting from the windplant.

The new windplant components on the solar power board wall:   
Boost 36 volt charge controller; circuit breaker on input, 3-phase  
diode bridge from 3-phase windplant, and circuit breaker on output
to batteries. (I think it says 20.16 VDC on input from windplant via 
the diode bridge, 37.65 V out, .43 A, and 16.10 W output to batteries.
   In moderate wind the next day I wired it (23rd). It put out 10 to 20 watts. Then it sat motionless for the rest of the month and beyond. There were maybe 5 or 6 days in October it would have done much of anything. Getting occasional moderate and high winds and being in a "normally windy" place are two different things. Masset gets more wind, but tidal power is by far the best bet for this island.
   The front of the house roof was the best available place (failing putting up a huge tower above the trees), but windspeeds there were only half what they were on the beach at low tide away from all obstacles. Since power depends on the cube of the wind speed, 1/2 the speed means just 1/8th of the power. In "no wind" there's no difference.
   This also meant my worries about it over-revving in a storm were doubtless unfounded. It didn't need any high wind protection. It was rated for 160(?) Km/Hr winds anyway, and 200 Km/Hr winds would only be 100 owing to its sheltered location.

   The one thing that might drastically perk up its performance in the lower winds it does get would be a venturi duct, but building that would be too much of a sideline at the moment.

The venturi duct idea.
Capturing twice the frontage area into the same size propeller means
double the wind speed at the propeller and hence 8 times the power.
(To actually capture twice the wind the opening must be more
than double the area to account for inevitable friction losses.)

   In conjunction with testing the windplant I added a switch to the kitchen hot water heater. The third element can either be shorted out by the switch, or in series with the other two to reduce the power demand. In the DC system, that makes it either around 250 watts or 160 watts. In the low winter sun, I wish I could turn it down even more! (Hmm... a DC to DC converter could do that by lowering the voltage.)

Two Lithium Battery Charger Circuits - AC and DC Sources

   Since my NiMH batteries weren't perfoming well in the solar power system, and since I had some lithiums, I gave thought to how to effectively charge that type without risk of some of them going over-voltage, since they have almost no tendency to balance each other.

   Lithium open circuit cell voltages are very specific:

* under 2.8 volts is deleterious
* 3.29 volts or less is pretty low or discharged
* 3.32 volts is fairly well charged
* 3.34 volts and above to 4.2 is fully charged
* above 4.2 risks failure

 Once they are charged they can go to very high voltages without drawing any appreciable current, but they are at risk of failure if they go over 4.2 volts. I didn't want to use the power wasting voltage shunt boards that can be placed on each cell to keep it below 4.2 volts. (I have some from the Swift, but some of them are quite corroded.)

   I started by thinking of charging each cell separately with a 3.4 to 3.6 volt source. Then I realized one could charge two cells in series at 6.8 to 7.2 volts without any risk of the better one rising to the 4.2 volt maximum limit, since the low one will never be under about 3.3 volts. That would cut the number of chargers in half - much better. Then I determined that even if three cells in series are charged at up to 3.6 maximum volts each or less (10.8 volts maximum), none will go overvoltage. When 12 volts (four series cells) is reached, the danger of imbalance causing overvoltage on the stronger cells begins, and it gets worse from there as the number of cells and voltage increases.
   So instead of charging a 36 volt bank with one charger, one can charge it as four 9 volt banks of three cells, at 10.5 volts or less, without any "Battery Management System" ("BMS"). (The generally recommended 3.6 volts per cell is the very maximum for three cells in series - 10.8 volts total: a little less is better. To go any higher than 3.6 per cell, only two cells can safely be charged in each string.)
   (Nov. 3rd... Hmm, actually for trickle charging (eg, solar) 12 volts/4 cells in series per charger would be all right at under about 3.35 to 3.50 volts per cell, total 13.4 or even up to 14.00 volts.)

Three 100 amp-hour, 12.8 volt lithium batteries in parallel.
The two 5 amp chargers each charge a 6.4 nominal volt, two
series cells section up to 7.2 volts: [0 to 7.2], and [7.2 to 14.4].
Details in detailed report under "Electricity Storage".

   To do it from an AC line power source is pretty easy: get four adjustable power adapters and put each one on three of the cells in series. The DC outputs are isolated from the AC and from ground, so the power can float at whatever level the cells are at. I found the 3 to 12 volt adjustable power adapters I bought drew current from the load when unplugged, so I had to add isolation diodes so they would only charge and not discharge the cells, and I adjusted their voltage up by .65 volts to compensate. I just did the front 12 volt section of the car with two chargers each doing 2 cells in series.

Four DC to DC buck converters adjusted to 10.05 volts each, charging a nominal
36 volt lithium battery as four 9 volt sections. Note the diodes soldered on in
many places: see circuit diagram under "Electricity Storage".

   To do it with DC from a solar charge controller proved to be a different kettle of fish. There don't seem to be any adjustable or 'right voltage' isolated output DC to DC power adapters. I contrived to make a "totem pole" arrangement with non-isolated DC to DC buck (voltage reducing) converters, but I had to try several variations, and they ended up needing 7 isolation diodes, on both inputs and outputs, and if the charging voltage is insufficient, the top bank won't get charged. (If that insufficient charge voltage rises as the lower ones charge, the top ones will charge last.) In a variation, if the voltage to the top set of cells is too low, one might use a DC to DC voltage boost converter instead of a buck converter for that top set. (or even 2 or 3 or all sets - a wide range of charge voltages could thus be accommodated.)
   In trying to do it with the PowMr solar MPPT charge controller, thorny problems cropped up and I finally realized it simply couldn't be done with a charge controller that gets its own power off the batteries on the same connection that is also its charging power output (almost all of them). A battery can't charge itself when there's no solar coming in, but the voltage has to be there and they will try, which can only run them down. However the MPT7210A boost charge controller draws its power from the solar panels instead. It should work with that one - I hope - but I didn't get to actually trying it so far.


   On the 27th was an event which will be most helpful: the fiber optic internet was at long last hooked up! That was supposedly to have happened shortly after the time I moved up here from Victoria. It took over 2 extra years, but it's here at last. That still doesn't give the whole island a truly good connection to the mainland, but that's supposed to be coming too. Already it's far better than the "connection: poor" pirate WiFi from a very slow satellite link. (I went to a page and waited for it to reload... it seemed awfully slow. Then I realized it had already reloaded in the time it took me to glance back at the window after clicking the mouse. What a change after typically waiting 15-45 seconds for it to maybe load a page!)

Handheld Bandmill

   I was still, very gradually, milling my last spruce log in October. (As of November 3rd I'm finally cutting into the last 12 foot section.) I had lately noticed the bandmill seemed slower than usual. But why? It finally turned out to be the new smaller V-belt pulley on the saw - the band was actually going too slow. And I had had about enough of re-welding washers onto the "railway wheel" band guides, so I ordered some alternatives. See the 'detailed report' under "Other Equipment Projects" for more.

New "yoke roller" and "thrust bearing" band guide

   I am planning to make and post a video of a take-apart of my fine bandsaw mill to help those who would like to build one themselves. After I'm finished milling those last big lumps of spruce!

Planetary Gears! Miles Truck & Chevy Sprint

   I found something I've somehow never managed to find before (but probably available all along): planetary gear gearsets already fully assembled -- ready to use drive components. (at "Anaheim Automation") To say the least they weren't cheap. But I had decided to replace the big, ugly, lossy transmission in the Miles truck with a simple efficient planetary reducing gearbox. A helical gears unit promising up to 97% efficiency seemed like just the thing. I knew planetary gears must be more efficient but I was surprised that it was so high. There's the "ultra efficiency" I'm always after for EVs.

    GPBS90 Planetary Gearbox
   I worked out that a 5 to 1 reduction unit should be about right for the Miles mini cargo truck, with the rear differential contributing a further ~ 2.16 to 1 for a total of 10.8 to 1 from motor to wheels. I picked a model to order and have been in touch for details. Having the Curtis programmer I can increase the maximum motor RPM a little, and with the extra efficiency 10.8 is surely a slightly reduced overall reduction ratio. It still wouldn't turn the truck into a highway speeds vehicle, but it would be up to at least 60(?) Km/Hr instead of 40 with less waste of battery power over distance, and so it would be much more useful than at present.
   I don't have enough lithium batteries at present to do both the truck and the Sprint to desired capacity, but if I put 2/3 of them into the truck for 7200 watt-hours I'd have a useful range vehicle. (IE, it could at least make it into town from here. It would surely be useful around town.)

   And then, here I've been fiddling around all this time with planetaries from auto transmissions trying to do the variable torque converter in the Sprint, and trying to figure out how to fit the components together outside of a proper case, with much trouble and limited success. But starting to think of the cast housings of the ready-made units, why could I not make my own suitable enclosed housing, out of pieces of pipe big enough to enclose the entire gear and shafts, and turn internal and end bearing plates, and a mounting for a chain sprocket, on the lathe?
   On November 1st I happened to visit someone who had scraps of pipe in his shop and got a piece of heavy steel pipe that would be ideal for the housing. And with custom turned end pieces enclosing it, I could put proper gear oil inside. Suddenly the planetary gear + centrifugal clutch infinitely variable torque converter seemed much closer and more promising.

Free Outboard

   Earlier I had looked at the "refuse transfer station" for a pipe. Instead I found discarded outboard motors. I took a smaller one for a potential electric outboard conversion. Just in case. Sometime. Maybe. I also 'scored' a big chunk of 1/2" thick stainless steel plate. That could be good for battery electrode punches and dies.

Not bad, but where is its hat?

In Passing
(Miscellaneous topics, editorial comments & opinionated rants)

Gardening - Growing Barley
 - Processing Grains into Flour With Common Kitchen Appliances

   In the spring I planted barley. It was one of the few things outside the greenhouse (besides potatoes and peas) that did well in the clouds and cool weather this northern summer. It probably wasn't much more than 50-60 square feet in all (ie, something like 7' x 8'). I threshed the heads in a blender. That broke up most of the chaff from the seeds. Then I winnowed them in a mixing bowl, shaking the bowl and with a hot air hair dryer blowing the light chaff out. There were still husks on some. I rubbed the grains between my hands and blew again, but it didn't help much. Now what? I suppose I could grow barley sprouts, which are said to be very good for you - and for which I had originally bought the seeds.

   Anyway I ended up with 270 grams of seeds, over 1/2 a pound. If I want there should be enough grow a much larger patch next year. (The original seeds will be about 4 years old.)

   To complete the grains processing plot, while I haven't used the barley, I have ground quinoa into meal or coarse flour, and wheat grain into coarse flour, with a coffee grinder. Less grain in the grinder makes a finer grind. Grinding a bigger batch longer, even a lot longer, doesn't make it fine. (I've just made a nice applesauce cake with the wheat flour. I'm going to try planting winter red wheat today (Oct. 10th). Done.)
   I often add quinoa flour to bread in place of some of the wheat flour. (Too much makes the bread rather crumbly - I'm just now trying 1/2 cup out of 4 total cups of flour instead of a whole cup, so a hamburger won't fall apart. Yes, that's better.)

   So there it is: grains can be grown in the garden and processed into flour with common household small appliances. There's no need to buy anything more special than these three common items. Of course this was just a "test" and one would want a much bigger plot to get a really useful amount. (The quinoa gave the most grain from the same small plot.)

   And here are some gratuitous pictures of some of my greenhouse produce, which did surprisingly well, eventually, after the all-cloudy July we had.

Greenhouse Produce mid October
Beans, peas, cherry tomatoes, cherry hot peppers...
and a foot long cucumber that cleverly hid behind 3 leaves.

And "coastal star" romaine lettuce seeds, cabbage, hot red peppers, leek,
 cauliflower, cherry tomatoes, pole beans and a more reasonable cucumber.
(Only decent cauliflower I've ever grown! it was delicious.)

A single cabbage, left to grow for a second year, filled a greenhouse bed.
Earlier it sprouted several branches and grew enough seeds for a whole
farm. In the fall it grew some funny clusters of leaves and two more
very nice (if smallish) heads of cabbage. (with a third one maybe coming.
The fourth, farthest right in the cluster, is actually a separate
cabbage, planted this year, which has yet to make a head.)

After getting three lovely big boletus mushrooms at the edge of the
highway in early October (and cooking and freezing them), now all I
find is a couple of wormy ones,
...and oodles of deadly poisonous amaneta muscaria

Hair Preservation with Recap

   I started putting a towel on the pillow on my bed, and I pulled it down to cover my scalp. This kept my head warmer at night, which evidently should help prevent thinning hair and baldness. It didn't always stay when I rolled around, and usually got lost somewhere before morning. But it was much better than nothing. (Now I think I know where the term "night cap" came from. Before good house insulation and heating, probably lots of people wore them in bed. Probably something like a light tuque or beanie.)
   Later in the month, I tried putting a pillow up on edge, above and behind my regular pillow, sort of between it and the wall, so the top of my head goes against it. I used a down/feather pillow - smaller and easily shaped. That seems to work pretty well.

So now to recap, to fight thinning hair we have:

- tuques, beanies, caps and hats for wearing in cool and cold temperatures (indoor and out)
- extra pillow 'overhead' or a nightcap for sleeping in cool to cold bedrooms
- frequent shampooing/showers, leaving shampoo on scalp for a couple of minutes or more (and some shampoos are evidently better than others)
- at least a daily hair brushing for more scalp stimulation than using a comb
- once a month spraying or rubbing of ethyl alcohol on the scalp to kill any deleterious bacteria, perhaps clean out pores and hair follicles, and perhaps cause some vaso-dilation. Ethyl rubbing alcohol and vodka have both been used.

   Vaso-constriction of scalp blood vessels, usually from cold scalp, is evidently the biggest culprit that causes hair follicles to become "dormant" and stop growing hair. But there seems to sometimes be some sort of bacterial or perhaps mite infection component or cause, too, and or possibly just accumulation of dead skin in hair follicles. (I think I had some sort of mite infection once - little scabby bumps here and there all over my scalp. They could be scraped off with a fingernail. Daily shampoo eventually cleared them out.)

Continuing Fatal Financial Flops

* Well... The US "Fed"s incredible volume of money printing didn't end after a few days or on October 11th as they promised. Instead the amounts are growing. (along with those of the ECB, Japan and China. In fact one or more of these institutions has always been printing money since 2008, serially. Now it's all together.) Now they're saying it may be a "permanent" fixture of the financial system. What kind of fairness is it that the banks and the banking system give themselves free money while the rest of us get more and more poor? Where are the bailouts for the average person? Where are the low interest rates on our debts? These everlasting, ever growing central bank balance sheet expansions can't end well.
   Be prepared for some big eruption of the financial system. It has for some time now been strongly advisable to pull some cash out of your account to cover at least a couple of weeks' groceries, gas and so on. And if you've become used to paying for everything with a debit card, you may be astonished to discover how fast cash evaporates from your wallet these days. A couple of hundred bucks can easily be gone in a day - even in one gas fill-up for some, or just stocking up a few groceries.

Small Thots

* Energy Equivalencies [Below: Canadian dollars, BC residential electric rate ~12 ¢/KWH, imperial gallons.]

 - Diesel is said to contain 3.65 KWH/Liter when burned in an efficient generator. Diesel should then cost about 43¢/liter to be competitive with power grid electricity.
 - It is said that a boat outboard motor burns about 1 gallon of gasoline per hour per 10 horsepower.
 - One horsepower is about 750 watts, so 10 HP for an hour is 7.5 KWH. Per liter, that's less than 2 KWH, when burned in an outboard. So to be competitive a liter of gasoline should cost under 24¢. (Probably cars are somewhat more efficient than outboards, so they added road tax.)

* Only 41% of US corn has been harvested instead of the average of 61%, and now record cold temperatures and snow in the continental USA are finishing it off. With this, combined with many other crop and food losses globally, global food scarcities seem inevitable. Prepare.

* Some say China has culled 130 million pigs owing to African Swine Fever. The USA has 75 million pigs. Others put the death estimate much higher, at up to 300 million pigs - half of the total population. Some say this won't end until 70% of the entire Chinese pig population has been wiped out. There is no vaccine and no cure and it is contagious. Just kill the pigs and try to keep the remaining herd healthy. Pigs in China have become seriously overpopulated in attempting to feed the human overpopulation. It is now spreading across Asia and Europe. Will it hit the Americas? Some estimate 1/4 of the world's pigs have died.

"This is a global plague unlike anything we've ever seen before." -- Michael Snyder, zerohedge.com

   Not to underplay the seriousness of that situation in and of itself, the parallel to the human situation is unnerving: we are likewise overpopulated. A disease that preys on humans and spreads like African Swine Fever certainly could have the potential to wipe out 2/3 of the human population (especially a malnourished population) before a vaccine can be developed, and these days it won't stop at any national border or (with air travel) ocean.

* I solved the mystery of the red spots on my legs. They would seem to be from mites that live in the spruce or spruce bark in the wood shed. Apparently they think human legs (or at least mine) are a good substitute and they contrive to 'get on board' when I bring in firewood. I bring it in in big garbage pails because they don't drop bits and the handles are high up so you don't have to stoop down to pick them up - great for when you have to set them down in front of the door to open it. If I spray the buckets off with water frequently, and don't leave them out in the wood shed, and rest the pail only against my knee and not my thigh as I carry it, I get very few of them, and only on the knees.
   Witch hazel remains by far the most effective way to get rid of them once acquired, one wetting of the skin where it is usually being enough.
   When I've finished milling the spruce (soon!) and run out of spruce firewood (not so soon!) I'll gladly switch to burning alder.

(Eccentric Silliness Department)

* I've always liked to call a bunch of crows a "crew" because they go strutting about, inspecting and pecking at things, like workers on a job site. Recently I've seen a couple of what I could only describe as "crow conventions" there were so many of them. Or maybe they were "crow trade shows".
   But someone told me a group of crows was called a "murder of crows". A little bizarre! I guess people only notice crows when they're cawing angrily. According to that analogy, these must have been "massacres of crows" or perhaps even "genocides of crows"!

* Conpare:  to make an unfavorable comparison.

* Discover: to remove the cover to see what's inside.

* The cougher coughed upon the coughee. No one else wanted it after that, so he added the tin of coughee to his coughers. Having thus "won" it, he had little incentive to quit coughin' for next time.

* Counting on Puns:
owe - won - too - free - fore - fife - sicks - savin' - ate - nein* - ton - all even - doesn'
   *nein: the number of words of German I speak.

   "in depth reports" for each project are below. I hope they may be useful to anyone who wants to get into a similar project, to glean ideas for how something might be done, as well as things that might have been tried, or just thought of and not tried... and even of how not to do something - why it didn't work or proved impractical. Sometimes they set out inventive thoughts almost as they occur - and are the actual organization and elaboration in writing of those thoughts. They are thus partly a diary and are not extensively proof-read for literary perfection, consistency and completeness before publication. I hope they add to the body of wisdom for other researchers and developers to help them find more productive paths and avoid potential pitfalls and dead ends.

Electric Transport

Ground Effect Vehicle (first the R/C Model)

Recap of Design Details

   In case there are new readers and in general, I'd like to go over the whole project and put everything that's been so gradually "accruing" in one place.
   I first learned of ground effect craft in March 2017 from a video 'suggestion' on youtube. From there I found some more info and soon found a video of a 2011 radio controlled "catamaran" style one, which seemed impressively stable -- flying just over a paved road. It seemed to have spawned a few more similar models also to be seen on youtube. What I gradually came up with as highly valuable design features were these:

1. The catamaran shape. Two hulls prevents spilling air off the ends of the wings, increasing the lift efficiency. The single wing between the two hulls can be smaller, and longer front to rear than it is wide. The height of the ground effect depends on the front-back length of the wing(s), so this should give the highest-up ride for its size. In addition, it's an easy shape to dock and to maneuver at low speeds. (I made the hulls flat bottomed (yet with a "step") because I know it takes less power to get a flat bottom boat planing. We shall see if this proves better for takeoffs.)

2. The central "dorsal fin" vane instead of a rear tail. A negative 'feature' of ground effect craft in general is that with little or no banking for turns, tighter turns are difficult. The catamaran in particular tends to simply fly diagonally on its original course when turned, before gradually straightening out. The center vane should counteract this tendency without banking. (If it is tall enough it will automatically cause banking into the turn.)

3. The special wing profile. The designer of the profile I've adopted for the 1/4 scale radio controlled model, John Ryland of Ryland Research, says that unlike typical airplane wings, the suction lift from above is more to the stern, while the compression lift beneath is more forward, and that this improves longitudinal stability. (His little model seemed quite stable as he changed its weight distribution. Computer modelling assisted with determining the profile.) Whether the lift to drag ratio is as good I'm not sure.

4. The forward "canard" elevator -- this month's design change: see below. The canard near the front can respond positively to changes in pitch - a dipping or rising nose - virtually as they start to happen. As the front loses lift going over the trough of a wave (it's suddenly "higher" over the surface and so the ground effect is reduced) and the nose starts to dip, aiming the canard up adds lift and directly pulls the nose back up to maintain level flight. Vise versa with a wave crest.
   A rear elevator can only respond to a falling nose by pushing the tail down, which will level the craft but with a loss of altitude. To regain the altitude, the tail must be pushed down further to get the nose up. So the tail loses even more altitude. That's okay in an airplane even at 100 feet. It's not okay in a ground effect craft at 1 or 2 or feet. It might hit the next wave crest.
   Previous ground effect craft have worked poorly except over calm water. I expect the front canard will make the ground effect vehicle considerably more "seaworthy". I plan that it will be computer controlled for instant response and automatic trim, for a smooth, level ride without effort from the operator.

5. Ducted Fan propulsion. A ducted fan provides the maximum static thrust for takeoff from water with the lowest horsepower. My original idea was one fan - perhaps a gasoline engine in the full size vehicle.

6. Multiple Ducted Fans. Having left and right propulsion greatly improves maneuverability when docking. I hope to use electric propulsion not only in the model but (probably) in the full size vehicle, and that would allow reverse as well as forward thrust for maneuvering. There are new aircraft designs with many small electric ducted fans, and this might serve well in the full size vehicle as a safety factor - one fan failure of 6 or 8 fans won't end the flight in the middle of the water somewhere, necessitating rescue.

   A Burt Rutan Vari-Eze two person airplane with front canard is said to get 40 miles per US gallon at 150 to 180 knots. That's a very efficient aircraft. A ground effect craft should be even more efficient. Thus a battery that would fly some other plane 30 minutes might last on the ground effect vehicle for 60 or 90 minutes, making electric propulsion more practical than for high-flying aircraft.

Ultimate Key to Longitudinal Stability and Pitch Control: a Canard

   Strange how thoughts develop. As I was getting up on the morning of the 10th my wandering thoughts crossed by the name "Faeroe Islands", which had been mentioned in an ad on youtube. I wasn't quite sure where those were (somewhere around Scotland), so then I thought of the Orkney Islands, off the north coast of Scotland, where powerful floating tidal flow units are already generating megawatts. Scapa Flow, the harbor in the middle between the several islands, had been the home base of the main British fleet in both world wars.
   Early in World War Two a daring U-boat commander had sallied into Scapa Flow underwater in a shallow, narrow channel between two islands, thought to be unnavigable, and with a torpedo sank a battleship in the midst of the fleet. No one at first understood what could have caused the explosion in this "safe" place where the larger channels were all mined against just such an incursion, and the submarine got away the way it came.

   Then I thought of the bow planes, the "dive planes" on a submarine that gave it such vertical maneuverability. Then I thought that if there were "bow planes", front rudders, on a ground effect craft, it could be more stable. But those would be something else sticking out the sides, making docking the craft more difficult. (It's already the one concern I have with the two 'offside' ducted fan motors.) Then, why should those "bow planes" be separate? Why not have an elevator on the front of the wing? As the front of the wing rose up or veered down, its motion or even the tendency toward such motion could more easily and quickly be compensated for by an elevator directly pushing the front of the wing up or down. The rear follows the front. An elevator at the back can only more indirectly respond to the rise or fall of the wing that has already happened.
   Our imagination is always limited by what we already know, and airplanes we know all have a certain familiar design -- they just don't have control surfaces on a leading edge anywhere. Even birds have similarly shaped wings and a rear tail/elevator. Instinctively, control surfaces at the rear of an airfoil have negative feedback. They will "return to center" if the control stick is released. A front control surface, if hinged at the back, would prefer to jam full up or full down. But the similarity of the airplane to the bird is deceptive. The bird can minutely maneuver and aim its whole wing. It can move it forward or back to change its center of lift. A front control surface, wherever it is hinged, would seem to be just what is needed for the ground effect craft.

   And come to think of it, on a 'canard' aircraft the elevator is actually in front of the wing. Those fly quite nicely - in fact I've actually flown one, a Vari-eze that won the 1979 Oshkosh airshow for "best homemade aircraft of the year". (The only thing I've ever flown. Admittedly I only held it level and on course while the pilot had a smoke.) Since I already knew of it and had experience with it, perhaps I should have thought of this design much sooner and skipped the submarine, but that's how the thought stream progressed! A youtube video says the Vari-eze gets 40 miles per [US?] gallon at 150-180 knots. If the ground effect craft is even better, we should expect amazingly low energy use.

Vari-eze, a popular homebuilt front canard, rear swept wing, rear propeller aircraft.
(designed by Burt Rutan long before he made the "space plane" that won the "X-Prize".)

   I can visualize a front elevator, probably controlled by computer, responding in real time to the aerodynamic forces that would cause the craft to rise and fall, a tendency to "hobby horse" in large wave crests and troughs. Inertial sensors can detect the slightest change in motion when the nose of the craft starts to rise or fall and the control would compensate. It could maintain an ideal height with little variation. Passengers would enjoy a smooth flight even over a stormy sea. (Even I wouldn't get seasick! ...except maybe in big swells over larger areas, which it would probably have to follow up and down.)

   Here then would seem to be the ultimate key to longitudinal stability of ground effect craft! This missing key is what has kept them restricted to smooth water, and from being safe and practical in real world conditions. Even an intrinsically less stable craft design could doubtless be made to work quite well, again, particularly with a microcontroller doing the up-down trimming.

   The next question was whether or not I could add some sort of front control surface or canard to my half-built model. If it made more lift at the front, it would just be necessary to move the weight (mainly the batteries) forward in the craft to keep the center of gravity in line with the lift. Or it could be made "lift neutral" in center position.
   I decided a 'canard' between the two hulls looked doable, and would be best and most stable. The idea of a front elevator with a rear hinge as part of the main wing somehow is unappealing. The previously planned rear elevator, somehow not made yet, could just become a fixed part of the wing (or for the model could be "trimmed", adjusted to different positions, to see where it worked best for any insights that might be gleaned for optimizing the full size vehicle). I looked at the model. It could either be a very narrow canard (front to back) very near the front, or it would have to be raised above the level of the wing. Would a narrow one have enough control surface? The whole thing could pivot. And it would be well to the front, giving it maximum up-down aim leverage. I decided that was what I would try.

   When people see it, the front canard on a 'catamaran' ground effect craft will seem so natural -- that is the "obvious" way to build them. How can it be that no one has thought of it before? Wright brothers first aircraft had a front canard, and there has been a century of not-very-satisfactory ground effect craft. And yet, look at the circuitous route my thoughts took to suddenly "get it", 2-1/2 years after first hearing of ground effect craft myself and even having already had personal experience with canard design aircraft.
   Well! (Of all the nerve!) Having mentioned "canard" and "ground effect" in an e-mail or two, on November 4th ever-ready Google/Youtube came up with a video suggestion of a very large radio controlled model ground effect craft with a canard from November 2010 that it had never shown me before even in searches. The maker said it was made lightweight and went so slowly so they could observe it better and keep radio control over it without having a high speed chase boat. The planned heavier human carrying version "will take off" at 120-140 Km/Hr. Call me chicken but my intended cruising speed is around 80-100. (The intended human carrying craft was to be the same size as the 7.2 meter long model but was apparently never built, for the usual reason: no funding or partners to be had for a promising new design. There were some Japanese characters in the video title, but it also included "Fugure-8 Flight of WISES Model". With the spelling it should be easy to find.)

   His design predates (by only months) the first "catamaran" model shown (so far) on youtube, but it had wingtip features to accomplish the same thing. It looks quite viable, but I like mine better. And I too am wondering about keeping the model within a good range for radio control. At least mine is 1/4 scale, so it won't take up quite so much of the small lake I intend to try it on. His model weighed 57 Kg. The 1.23 x .76 meter bare body of mine so far is just 1.35 Kg, but it needs more fabric and more paint and hull decks, and when the motors and batteries and electronics are added it'll be lots heavier.

   Once again, the lack of progress on this project seems to have miraculously worked out to allow yet another important and beneficial idea to improve the final design. At some point, it has to get built and flown, but in the meantime the more improvements that have come to light, the more confident I am of getting a fabulous, efficient, stable and practical craft for rapid transportation over water!

Modern Fabrics and Paints

   Having used up a can of varathane spray paint on the wing fabric and yet it still wasn't airtight, I decided to simply try some other paint. I used white 'rustoleum'. One coat (on top of all that varathane) seemed to make it pretty much airtight. and white. It also had the opposite effect of cellulose dope on canvas: instead of the featherweight polypropylene fabric tightening and stretching, it loosened and went slack. It sagged between the supporting pieces.
   This suggests another technique should be used. Next thing was to take a big piece of fabric and lay it out flat on an even bigger sheet of polyethylene plastic. Then it would be cut and glued (still using epoxy, I think) to the airframe in flat sheets. (Luckily there don't seem to be any compound curves.)
   ...or could the paint itself glue it to the model? That might be worth a try! (Actually it seemed to "melt" the polystyrene and I ended up glad I had done no more.)

(20th) I cut and shaped pieces of wood for the canard. As I did so I thought of paints again. I looked up "model airplane dope" on line and found that there are more modern alternatives. There's one called "Eze-Dope", and on a seller's website some model builders were saying how great it was, without odor and fast drying. (No wonder the hobby store I was in a while back didn't sell aircraft dope any more.) One customer commented to that effect in 2014, but then he added a note in September of this year, barely a month ago. He said after reading something in a model aircraft magazine, he had tried MinWax "Polycrylic" and found it was even better.

"-----Updated 9/17/2019 After reading an article in Model Aviation, I decided to try Minwax Polycrylic (satin finish) thinned with water to 33% Polycrylic. One coat with a wide (1" brush) gave me better results than I got with two or three coats of Eze Dope. See photo. So, while Eze Dope gave me a good enough alternative to nitrated dope as a tissue sealer, Polycrylic is a decidedly better one."

   It was nice of the store selling "Eze-Dope" to let him post that, years later! (I suppose they didn't know.)

   I looked that up and some common stores had it, but they all wanted the customer to pick it up and wouldn't ship it. Then I found that "Home Hardware" had it. There was one of those in Masset. That was 85 Km away, but at least driving there didn't involve ferries and motels and 1000 $. They didn't have any but would order it. It seemed it would take until the end of the month to arrive. Nothing is fast when it has to come across BC and then catch the ferry to this island.
   To finish this, I drove up to Masset on November 1st and picked it up. Then I went down the street to "Co-op Home Centre". They had it sitting on the shelf - no need to have ordered it at all! Oh well, Home Hardware was 50¢ cheaper. The co-op also had the elusive #10-24 threaded rod I had previously been unable to find.

(22nd) I cut and shaped pieces of wood for the rear elevator, which will now simply be pinned into one position before each flight. Different positions will be tried along with different weight distributions front to rear to see what works best.
   I found a 2017 youtube video about fabric and coverings by a model aircraft builder 'of 40 years experience'. He himself was trying Polycrylic for the first time. He did three different fabrics and three different fabric fillers. Here was someone of long experience with fabric aircraft materials, and he was more than a little impressed with Polycrylic. It would seem it must be the best! He said he used a heat gun to tension the fabrics but I don't know if that was before or after applying the finish.
   He also gave the weights per yard of his fabrics. I measured and weighed a piece of the light PP cloth and found it was 20% lighter than his lightest fabric: 28 grams per square yard versus 34. And being PP, it was probably stronger too.

   At some point I started gluing fabric onto the canard and the elevator with ordinary white carpentry glue, which was suggested in the same video. It dries a LOT slower on styrene foam than on wood!

(31st) I finished the canard and elevator, and I mounted the canard on the craft with a 1/8 inch dowel that ran right through everything from the left edge to the right. Now it just needed the radio control servo hooked up to the radio receiver... or to an inertia sensor and a computer control to automatically adjust the canard for level flight even in wavy conditions. But that can come later.

Another profile

With wing recovered November 2nd
(I was going to do the starboard motor mount properly,
but my mini milling machine was again "on strike".)
(Egads, what's with my cell-phone camera today? "foggy")

   I am very pleased with the position of the canard. It could hardly have worked out better if it had been planned from the beginning. It extends the lifting surface almost to the front of the hulls, which will in itself improve the ground effect. It also seems to give some meaning (besides floatation in water) to the hulls extending so far forward from the wing. Now the whole craft seems like a lifting body.
   And its position just in front of the main wing gives it another effect: If it is aimed up, not only does it increase its own lift and raise the nose, but its slipstream will deflect more air under the main wing, increasing the main wing's lift. If it is aimed down, it deflects air over the main wing, decreasing lifting pressure beneath and increasing pressure above, which decreases lift. I expect the altitude control to be very responsive.

   If there's any tendency to tip sideways and dip one hull into the water, as some of the 'catamaran' models on youtube had, the canard for the full size vehicle can be broken into left and right "aileron" sections - which will again both probably be computer controlled.

   A late thought came to me on November 5th: The ducted fans might be mounted on the canard, instead of sticking out the sides. They'd still be vertically in line with the wing. And the air thrust would be aimed up or down with the canard. It could be aimed under the wing for "inflation" of the largely trapped airspace, making for a lower takeoff speed. The canard would be even more effective at maintaining desired attitude and altitude. The height of the canard with respect to the wing height could be optimized for best effect. (least drag?) And there'd be nothing sticking out the sides to impede docking. For the model I'd need to redesign and rebuild a suitable and stronger canard and pivot - and preferably on the model the height of the canard pivot should be adjustable to find where it works best. Doable! I think I will! It may be proceeding at a snail's pace, but this project just keeps getting more and more exciting!

EV Transmissions: Off-the-Shelf Planetary Gears, More.
(Miles Truck & Chevy Sprint)

   Once before, years ago, I had looked on line for planetary gears, but I don't remember finding much, and later the ones I had found were gone. They were pretty costly and came without cases. I got almost the same thing from auto transmissions much cheaper. Now I looked again and found some already fully assembled in cases from "Anaheim Automation" - ready to use drive components. To say the least they weren't cheap. But they were no doubt just the thing.
   So here I've been fiddling around all this time with planetaries from auto transmissions and trying to figure out how to house them and fit the components together outside of their intended transmission case, with much trouble and limited success. Was it worth 1000$ to get one ready made? If I knew exactly what I wanted, probably. If I got one and it turned out it really needed a different ratio, if I ended up ordering multiples, it could get expensive really fast.

   But starting to think of cast housings, could I instead make my own suitable enclosed housings, out of pieces of pipe big enough to enclose the entire gear and shafts, and turn internal and end bearing plates on the lathe? Now if I had a real R & D budget I wouldn't even consider such a thing - I'd just order one ready made to try out, from those who are already all set up and experienced with making them! OTOH this way I'd get a custom unit to my requirements.
   I decided the thing to do would be to order a stock one for the truck to see how it's made (to get any useful construction tips) before I install it, and then make the custom one for the Sprint.

I lucked out November 1st visiting Lawrence of Drifttech Mechanical in his garage. He had a selection of pipe pieces, and we found a piece of thick walled pipe that when I got it home looked perfect. Boring out the inside just slightly on the lathe will make the ring gear fit in perfectly. And then it could butt up against the unbored part inside - a solid end stop. Perfect! Next: a bearing for the planets assembly (odd size - 1.20" or 30.5 mm). While not having everything figured yet, it looks like I should turn two ends with centers to hold a bearing on each side, perhaps from aluminum plate. But it needs some way to hold the drive chain sprocket.

   Anaheim Automation speaks of the 'reduction ratio' from the input to the output shaft, from 3 up to very high numbers. It would seem then that the 'input' shaft (actually a recess/tenon to fit an input shaft) is the sun gear, the 'output' shaft is the planets assembly, and the case between is the ring gear.

    Planetary Gearbox GPBS90
   I had recently decided on the Miles truck I should use a simple 5 to 1(?) reduction planetary of that configuration, simply replacing the present bulky and doubtless lossy transmission with this small part. These ones boast 'up to 97% efficiency', and anywhere near that would of course be excellent - I'm sure much better than the hacked standard transmission now in the truck. Interestingly the helical gear models were also rated as being "up to 97%". I have instinctively always assumed they would be lower because of the sideways pressure. Having the Curtis programmer I can increase the maximum motor RPM a little, and with extra efficiency I can surely reduce the overall reduction ratio a little (not a lot - 10%?). That still wouldn't turn the truck into a highway speeds vehicle, but it should be up to about 60(?) Km/Hr instead of 40, and with less waste of battery power over distance, which would make it much more useful than at present.

   I don't have enough lithiums at present to do both the truck and the Sprint to desired capacity, but I could put 2/3 of them in the truck for 7200 watt-hours and have a useful range vehicle. (IE, it could make it into town from here. Coming back... well, it wouldn't be 'empty' when it hit town, and the small capacity should charge faster. And with two solar panels on the roof, always charging...)

   For the Sprint of course I want to do the variable transmission. But a complete, enclosed planetary gear still seems like just the thing for that. The whole case can be made to rotate to give effect to the desired motions. A five times faster output to the special centrifugal clutch component should get it spinning nicely - 1/5 torque and 5 times speed for good centrifugal grip. (I hope that's not too fast.)

(Nov 1st) I checked the reduction ratio of the truck's differential by jacking up a back wheel. For 2 turns of the drive shaft, the wheel turned about 1.85 turns. If both rear wheels had been turning equally, that would have been .925 turns, or .4625 turns for one turn of the driveshaft, or 2.16 to 1 reduction. That's not far from the 2.4 I was figuring. The 5 to 1 planetary would then make it 10.8 to 1 reduction from motor to wheels. That's probably just about right. A 6 to 1 planetary would make it 13 to 1 - 20% more torque but 20% lower top speed. I think I'll gamble on the higher speed (since the speed is at the bottom end of "useful" anyway) and order the 5:1 unit.

Other "Green" Electric Equipment Projects

Handheld Bandsaw Mill (& Alaska Mill)

Instruction Video Planned

   Perhaps of particular interest to those who are interested in making a mill like this one, I have decided to do a video of taking it apart and reassembling it. That's in lieu of making an actual "how to" instruction book. I hope to do it in November or December after I've finished milling my last spruce log. (How can it be I'm not finished yet?!?)

Working with the Mill and Latest Mods.

   I finally finished the second last 16 foot cant in our lovely indian summer up until mid October. Then rain, drizzle and more drizzle brought cutting to a halt. Was the last section of log really going to sit out another winter? It was rotting! In spite of the metal roofing over it, the bark seemed to carry moisture all around - the top was as rotten as the side down at the ground. On the 24th, again stymied, I went out in the drizzle and de-barked it with the sharpened, square-ended shovel. I should have done that last year!

   [Warning: digression 2 paragraphs =>] Somewhere I lost my peevee. I couldn't find it anywhere. The log sections were so heavy I had used a jacks-all (sp?) to roll them. I considered that I might have carelessly left it lying and rolling a log onto it, but I thought that if I had done that I'd be able to see it underneath. Only when I moved the last bottom slab of the 16 foot section did I find it again. The log was so heavy it had pushed the peevee right down into the moss and dirt so it was virtually underground. I also found I had hit it with the chainsaw in trimming the log, but just a couple of slight nicks.
   I still remember when I bought that peevee about 1980. I saw it in Capital Iron in the garden tool section in a bin with shovels and rakes. I was thrilled to find one for sale and although it cost 4-5 times as much as a shovel or rake, I grabbed it. Oh boy! As I pulled it out and started walking a store clerk stared in astonishment. "Are you actually going to buy that?" he asked. "Well yah!" I couldn't believe my luck finding one; he couldn't believe anyone would want one. I hadn't seen it before so I don't think they could have had it long, but they never got another one in. It has grabbed, pried, levered and rolled a lot of logs and heavy things over the years. Well, I'm really digressing here, to little cause...

   On the sunny 26th I squared up the last cant with the Alaska mill, and on the 28th I got three 2" by 6"s cut. But it had seemed like slow going for quite a while. And the saw pulley got very hot. My link belt started slipping, melting and breaking, and finally I had bought a regular V-belt and now put it on.
   Finally I thought "I wonder if it's the 1 inch pulley I put on?" I got the old one and put it back on. Sure enough, it cut faster. It was like getting my good saw back after using a crappy one!
   The 1 inch pulley going to the 9 inch made the band speed actually too slow. The skillsaw motor never labored much - instead the belt just slipped. Looking at the 1" pulley, I saw that it had kept getting so hot that the belt would slip, and the combo of heat and rubbing had polished the inside slick and smooth. That's surely why I had to tighten it so much and then the hot link belt kept breaking (melting?) where it slipped.

   The old one was a variable pulley. I'd estimate it was set to more like 1-1/2" to 2". Now it seems more like the motor is working - hopefully not to the point of overheating and burning out. I may get a selection: 1.5", 1.75" and 2.0" pulleys, and try them out to find which is "optimum".

   And I found the saw getting some nasty vibration, to the point I thought something had gone wrong. But it seemed to be the V-belt setting up a resonance, in spite of being pretty tight. I put the link belt back on and it stopped. It also seemed to have more power - which is the reason for using link belts in the first place: typical V-belts are pretty inefficient. (I wonder why they don't make them all grooved like the ones for variable transmission pulleys?)

   At the same time, I had ordered and received two "idler bearings" and two thrust bearings whose inside diameter matched the outside of the idlers. These replaced my "railroad car wheels" with the rim being a pulley welded to a custom machined part. (I got them at vxbbearings.com, part numbers:

"CY36L - 1-1/8" Heavy Duty Yoke Rollers",
"12473 - TC1828 Thrust Needle Roller Bearing 1-1/8" (ID) x 1-"?/?" (OD).)

   I've only put one set on so far. It seems to work well enough, although the loose thrust bearings rattle around some. The "yoke rollers" are perfect - a good diameter and a good width for guiding the band without touching the teeth, so the guide rollers don't start flattening out the 'set' to the teeth. (I didn't order runner washers for the thrust bearings and had to make do. The last thrust bearings I got, the matching washers came with them. In my experience with my "railway wheel" welded-on washers, they gradually wear out anyway and have to be replaced - a real pain when you have to weld them.)
   It seems to me now that I should have found thrust bearings to fit onto the 5/16" bolts. They would go behind the "yoke rollers". Then I would make a "shoulder washer" 5/16" x 1-3/4" to put in between. with another washer behind the thrust bearing against the bolt head. That should be pretty solid.

   On the 31st I finished off that last 16 foot cant. Now the new boards need to be stickered and added to the considerable stack under pieces of sheet metal roofing to dry. ("Stickered": Cut small pieces of wood ("stickers", meaning "spacers") and place them across the rows of boards to make an air space between each layer of boards so they can dry faster and better. Hey, I don't make up these terms, I just use them. "Cants", hmpf!)
   Then there are a couple of 16 foot edge slabs that probably have a couple of boards in them. Or should I just cut them into firewood and get on with milling the last 12 foot section of log, the base of the tree?

Electricity Generation

5 Blade Windplant

(5th) I got out the wind meter and tried out some spots. As I had surmised, the front corner of the house roof seemed to be the best spot if I wasn't going to erect some tall tower. I got readings of about 1.8 to 2.2 meters per second. (Of course the wind is always coming and going. To most fairly compare any two spots, one ideally needs two anemometers and a second person on the cell phone or 2-way radio.) It seemed best at least 4 or 5 feet above the peak. As I walked along the peak toward the other end of the roof, it got lower, no doubt mainly because of the tall trees too close to the other end of the house.
   The top of the driveway where the wind funnels in at ground level (at least for our typical southeasters) was a close second, just 2 or 3 feet off the ground. Figures might have been 1.7 to 2.1. Everywhere else had too many wind shadows - mostly of trees.
   Then it occurred to me to go down to the beach. In the access path there was again that wind tunnel effect, and I got readings up to maybe 3.5. Down on the beach itself I was getting up to over 4 - even a 4.8. (I should probably have walked more toward the water, too, to be sure I was out of the wind shadows of the trees.) Theoretically the strip beween the highway and the beach is my property. Theoretically. Hah! Even if I put the windplant in the path, and somehow got the power across the highway, everybody uses that path to drive their "quad" to the beach. A small windplant set up there would surely be vandalized by annoyed people or stolen.

The front corner of the house roof it is! This also would give the best chance to try out a venturi to funnel in the air, too, if I got around to it. Except for the driveway, but that would be in the way.

High Wind Protection

   Occasionally we get some very strong gales here, with winds over 100 KmPH, maybe up to 120. With wind power increasing at the cube of the wind speed, this had been bothering me. I wished people had adopted the Piggott tail vane (rudder) design where a high wind would swing the propeller away from the wind, as far as edge-on instead of facing directly into it.
   On the 9th I suddenly had the idea to mount the whole windplant on a pivot with a weight at the bottom. If the wind was strong enough to overcome the weight, the whole windplant would tip over. The stronger the wind, the farther it would tip. Since the windplant pivoted on its own base, with a slip-ring for not twisting the wires, the tip-over hinge - if the whole thing was to be kept simple - had to be pointed in one direction (or perhaps could be bidirectional) rather than being equally applicable regardless of wind direction. But the strong gales here come from the southeast. We've only had 2 or 3 bad ("windplant wrecking force") gales since I've lived here, all from that direction. For those it would work.

   I also figured out how to attach something to the mounting flange provided. It looked like the only possibility would be to weld a piece of pipe to it. But I found a 1/2 inch stainless NPT inside threaded pipe coupler that seemed to be a perfect fit for the bottom mounting flange. It was indeed such a snug fit that I had to laboriously pound it into the hole in the flange with a hammer. Perfect! - it will never come loose.

   On the 10th as it was getting dark, it had become quite windy. I measured at the top of the driveway where it funnels in and where I have done wind power tests. I got readings of 3-4 meters per second. Then I went across to the path to the beach and got 4,5,6 meters per second. On the beach that climbed to 7,8,9. As it was low tide (and with an almost full moon so I could see to come back), I went way down the beach, far from the trees, and got readings of 8,9,10,11. Lifting the 7 foot pole as high as I could (making it maybe 12 feet?) sometimes added about one to the reading, sometimes not. While I wasn't going to go up on the roof in the dark, I did climb up the ladder and stick up the pole. From this less than ideal location I got readings up to 5.5. Since the power is proportional to the cube of the wind speed, that's 1/8 of the power down at the bottom of the beach. I'm starting to see the real reason windplants are so often placed just offshore!

(16th) Skipping ahead in the narrative, in it all I think I found a wholly better way to protect the windplant in high winds. It works for the same reason that the propeller starts up so slowly, and gains power as it gains speed until it is whizzing around with high energy: It starts up catching only the air near the blades, but as it speeds up, it starts to catch all the air in the whole area swept by the blades, catching more and more power.
   Usually one aims to find the ideal speed, loading down the generator just to where it puts out the most power, but without slowing it down too much so it loses wind - to the 'maximum power point'. If the wind is too strong, either the high power available overheats the generator, or else it 'spills' power and instead the propeller speed becomes excessive, and the unit may fly apart.
   But in there I found that if the output is shorted, the propeller doesn't spin fast at all, even in high winds. It only drifts around. If the prop isn't spinning, it doesn't catch all the wind in the circle, just that near the blades. So, the short may make a little internal heat, but nothing like when it's spinning, and certainly there's almost no centrifugal force.

   Why don't people utilize this as the ultimate high wind safety mechanism? It seems dead simple. A microcontroller control would be made to sense that the power output (or voltage of current), or the generator temperature, is too high, or the frequency coming from the output is too high (over-revving) and instead of just switching in another dump load to utilize more power and make more heat, or running out of options and letting it over-rev, short circuit two or all three outputs together, perhaps with a solid state relay or two of them. That will bring the unit almost to a stop and the wind will mostly just go right by it.

Actual Assembly

   On the 11th I finally got a little done. I took a 7 foot spruce 2 by 4 that I had milled and cut the ends diagonal (rain runs off), then I ran my electric hand planer around it and rounded it off everywhere. Spruce isn't the best choice for outdoors, unpainted, but that's what I had.
   I picked a foot long piece of pipe with an outside thread on one end to go into the fitting pounded into the mounting flange. I took an old 3-wire extension cord and cut the plug off it. I slipped it through the pipe. Then I soldered the three short wires sticking out of the windplant to the wires of the extension cord, then pulled up the pipe and threaded it into the coupling in the mounting flange. The cord now came out the bottom of the pipe. (If you followed all that, you don't need to look at the picture!)
   Then I cut a couple of pieces of 1/8 by 3/4 inch steel bar for two straps. I sprayed some rust paint on them and then put them in the kitchen oven for an hour (turned on for 5 minutes for heat) to dry. With all in place I drilled holes and bolted the straps and the windplant to the wood.
   Then I found that if the wind was blowing from the wrong direction the propeller blades would hit the back of the 2 by 6. I had to disassemble it and cut a piece out the length of a blade, making it a 2 by 5 or thereabouts near the top.

   When I picked it up I discovered that the windplant was so heavy compared to the 2 by 6 that the balance point was only a foot from the top; in fact, above the bottom of the propeller. It seemed the high wind hinge would have to be as high up as possible, and it would still need some more weight screwed to the bottom end of the 2 by 6.

   On the 13th I got back to it. I made two aluminum flat brackets/plates to hold the hinge pin and bolted them to the fixed 2 by 4 upright 'pole'. I cut pieces and screwed this to the edge of the roof.
   I screwed a piece of 2 by 6 on horizontally at the bottom of the windplant's 2 by 6 and a diagonal to stiffen it - something like the sketch.
   I plugged in a plug with its wires shorted together to prevent the unit from revving up much if there was a breeze. But it was calm.

   Then I got two neighbors to come and help me raise it. We discovered that if the plant pointed "backward" and the pole hinged over, the propeller blades would hit the 2 by 4 upright. I had thought they might. But the pole should never tip up unless the propeller is pointing the other way, so presumably it should never happen.

   It was up!


(14th) Early in the morning it was blowing a storm. I got up and took down the ladder. (Last time there was a wind it had blown over and smashed a glass ashtray that had been sitting for ages on the porch railing. I'm glad it wasn't the patio doors!) I unplugged the short, and the windplant spun up to high speeds and the "tilt" mechanism was being activated. It would tilt back and then come down again with a bang that went through the whole house. (Well, I did get it from "BangGood"!) When I plugged in the short, it slowed way down, and stayed tilted up. (I didn't want it to tilt this easily. It needs a heavier counterweight.)
   I connected a cable with wire nuts and some aligator clip leeds to the 36 volt kitchen water heater. The voltage between two phases was about 16 volts AC and the current was around 2.75 amps.
   Then I got out a 3-phase bridge rectifier and wired in a 3-prong plug. I soldered the DC output to the same cord and connected it to the water tank. It didn't seem to give much more: 16 to 18 volts, 2.5 to 3.0 amps. So it would have been delivering 40 to 55 watts to the water heater. There were lots of sparks when I connected and disconnected aligator clip leeds.
   Then I took the wind meter and got readings in the driveway and around the house of 4.5 to 7 meters per second. Up over the roof at the windplant would surely have been in the upper part of that range if not a little higher.

   I had the feeling it was just a bit overloaded. With a lighter load, the prop would have spun a little faster, and maybe delivered actually more voltage and hence more current. A little later I came back and it was just 5 volts. I looked and the unit was spinning only slowly. Definitely the load should have been reduced, and reduced more as the wind lightened. I disconnected one wire from the rectifier. The speed went back up. The DC voltage went up to 12-15. Then the wind seemed to pick up a little. I reconnected the phase (this time with an aligator clip - I was getting tired of soldering!) and found the voltage up to 19 and current over 3 amps. That made to 60 watts.
   If I disconnected the water heater the open circuit voltage ran up to over 50-55 volts DC. Yes, dragging it down to under 20 was surely an overload. Eg: 40 volts at 2 amps is more power than 20 volts at 3 amps.
   (This brings up that if I put the third heater element in the water heater in series with the other two, it would be less load. This would be useful for the solar on dull days, too. I just need to put in a switch: all three elements in series (~150 W @ 38 volts), or short one with the switch and have two as it is now (~240 W @ 38 V).)

   How did that power compare with the theoretical? Say 7 m/s, (times 3.6) is 25 Km/hr. I worked out last month that 20 Km/Hr should be over 60 watts. (25/20)^3 should then be 120 watts. But as I say, it was probably overloaded and not running at its "maximum power point". And was the wind really doing 7 m/s? It varied moment to moment. At least the power was the right order of magnitude. It would however heat the water a lot faster at 120 watts than at 60. (In the clouds and rain the 1000 watts DC solar was only managing 25 watts, so I wouldn't have turned the water heater on as the batteries were already low from leaving it on too late the previous night. Some decent batteries is obviously a big part of the solution for that.) 40-60 watts was better than nothing, although it would take several hours to heat the water. The storm showed every sign of lasting that long. Well, all very fine for a test and a day's hot water.

   Then I decided to rewire the hot water heater. I put in the switch so it could be switched between 2 heater elements in series and 3, by shorting one or leaving the switch 'off'. That took the better part of 2 hours, but the gale seemed still about the same. With the switch open, readings were now around 2.7 amps, 22-24 volts. So it lost .3 or .4 amps, but gained 5 or 6 volts. 2.7 times 23 is 62 watts. I think it was spinning faster.

Wait! I can measure it! I connected the meter on "Hz" to the AC direct from the unit:

2 water heater elements in series (heaviest load) - 70 Hz [around 55 watts?]
3 in series (somewhat lighter load) - 75 Hz [over 60 watts?]
No load connected; Open circuit   - 100 Hz

   I don't know what propeller RPM those figures translate to, but one can see the relative speeds depending on loading. That gives us means for figuring out where the maximum power point is. It's probably found by loading it down to somewhere between 75 and 90 % of the free running speed.

   The gale continued without let up all day. At some point in the early afternoon I found winds up to 9 m/sec (32.4 Km/Hr) and voltage around 25-26. That would have been over 70 watts. (I have no doubt that the winds were considerably stronger in the open, like down on the beach or just offshore.)
   As it was getting dark the plant started banging around on the safety pivot. It had finally heated up the water to where the tank shut off, and it had no load. I felt the water and decided to turn it up higher anyway, so the tank came back on.
   The power might have been 15-25 watts more if the windplant hadn't kept tipping back. Spilling wind was supposed to be for still higher winds. But I wasn't about to go up on the roof in the gale and pelting rain to add another piece of counterweight. One could lose one's balance on the roof in the wind. Or if one's ladder blows over there's no way to get down. (Well, of course I would find some way to tie it up.) And in general it would be most unpleasant. The banging of the windplant against the pivot was somewhat disturbing tho.

   The next morning was calm and partly sunny.

   I had heard in the evening that the winds had been supposed to be "up to 100 kilometers per hour". Hmm... It was certainly a howling gale, but even if the beach and on the ocean was double what it was near my house, that certainly didn't match what my anemometer was reading. If it doesn't get a whole lot worse than that, and it was only giving under 100 watts out of 1200 rated, I don't think the tip-over safety device is needed. I decided to just screw it closed, and did so. Soon I realized that shorting the output was a better and simpler way to protect the unit.

I had heard that molded plastic propeller blades aren't very stiff and will bend a little in high winds.
But this is ridiculous!
...of course, it's all a camera trick.

(16th) There was a fair breeze and I connected the windplant to a 250/500 watt electric heater. It read 22 volts, 3/4 of an amp, for a grand total of around 16 watts. The other (250W) setting didn't work. A 400/800 watt heater didn't work at all. A 1500 watt heater stalled the windplant. The wind picked up and the figures became 27 volts, .9 amps: 24 watts. (I couldn't get a stable Hz reading.) I caught a gust at 35 volts, 1.12 amps (39 watts). I could feel just a bit of heat from the heater. Did that make up for having the patio door cracked open to run the wire in? Probably not!
   I started to get the feeling I'd be lucky to ever see much over 100 watts out of this 1200 watt rated unit. The winds just don't get that strong. But just twice the wind speed is 8 times the power. Maybe if I could put it down on the lower beach away from the trees? Maybe if we start getting those infamous northern hurricanes like on the Atlantic coast?

   Later in the day (after going into town) I reconnected it and saw it hit 42 volts and 1.6 amps = 67 Watts. An average might have been 35 V, 1.3 A for 45 W. The heater wasn't hot, but it was warm. Big waves pounded the beach. Well, at least it was something.
   In the evening I tried driving the hot water heater again but the wind wasn't as strong and it was entirely too much for it. The blades virtually stalled and it was just a couple of volts at 1/3 of an amp - less than a watt. Back on the heater it went up to around 25 volts and .9 amps - over 20 watts. If it had had just the right load in the gale on the 14th, and if the "tilt" switch hadn't been tipping it away from the wind, it might have been over 100 watts after all. Nothing like 1200.

Permanent Connection

   I didn't want to use the 24 volt charge controller supplied because my batteries are 36 volts. But if I could count on the unit putting out over about 45 volts or so, perhaps I could simply connect the rectified output to the PowMr solar charge controller in parallel with the solar panels? If its voltage was higher than the solar, the controller would use it instead.
   OTOH if I had ordered the 12 volt windplant, perhaps I could have used the MPS7210A boost converter to bring (eg) 15 to 30 volts DC to up near 40 volts to charge 36 volt batteries. The voltages work out better. Well - so far - I'm not planning on getting another one. If I do I'll try the 12 volt unit. Now that I see this one can give over 50 volts it seems like a good idea.

   The only thing those wouldn't provide was a "dump load" in case the windplant was spinning too fast and the batteries were already charged. But the little 24 volt charge controller that came with it didn't have a dump load either. Perhaps for that, on sensing 55(?) volts it could connect to a water heater in an uninsulated tank in the greenhouse? I don't suppose that would do much for the greenhouse. Keep it from freezing?

   The best option would doubtless be a microcontroller based controller that would decide whether to charge batteries and or variably turn on dump loads (maybe 3 or 4, one or more as needed, eg 40/80/160 W or 30/60/120/240 W), roughly maintaining maximum power point output, moment by moment. I'm not willing to create such a unit in the midst of so many other projects.
   And on the 14th in the blowing wind, I wasn't willing to do much outside work at all. I certainly got it put up on the right day, calm while working then a gale the next day to test it!

(22nd) I finally came up with a plan for connecting the windplant. It doesn't use the 24 volt charge controller that came with the windplant:

1. Run the 3-phase AC from the windplant to the garage with a 3-wire extension cord.
2. Rectify to DC with a 3-phase bridge rectifier. It seems to be under 35 volts open circuit in a good wind, and it certainly will be under load.
3. Connect it to an MPT-7210A 'boost' DC to DC charge controller to put out 39+ volts and charge the batteries, in parallel with the solar panels and their PowMr charge controller.

   This doesn't do the high wind protection by short circuiting the windplant output. That would have to be handled separately.

   I did the above on the 23rd. It was a windy afternoon (and getting on by the time I had it done). Readings were around 10 watts. One could hear the windplant speeding up and slowing down a bit as the MPT unit looked for the maximum power point. It was probably getting it about right, not loading down the propeller to a stall, yet not letting it spin up too high either. Considering it was made for solar and not wind, that's pretty good.
   Later the wind picked up a bit and it was doing 15 watts or a bit more. Still a disappointing replacement for solar, of which on this dull day in the later afternoon, 1000 watts of panels were only yielding 40 watts. In fact I had been unable to get the kitchen water tank up to temperature. I had been turning it on, and then off again as the batteries dropped.
   I went for a walk on the beach - I've been falling off on getting exercise - and gaining weight! The rain stung my face and it was apparent that there was much more wind there. ...double the speed, as I measured previously? Let's see... double the wind speed is 8 times the power. So 12.5 watts would have been 100 watts. That amount of energy would certainly be useful!

   I decided that if the windspeed at the wndplant was only going to be half the speed that the weather stations were saying, I didn't need to bother with protecting the plant from high winds. They just wouldn't get that severe where it was.

Venturi Concept (one more time)

   I couldn't put the windplant on the beach where the highest winds were for several reasons, including that it would be too much work. The other obvious thing to do to increase the output power even in lighter winds would be to make a venturi duct to feed wind to the propeller at accelerated speeds. Inlet and outlet openings just 1.5 times the diameter of the propeller would feed it from 2.25 times the area.
   A 2 meter diameter front and rear would have over twice the cross section of the 1.28 meter windplant propeller diameter; in fact 2.44 times. Allowing something for friction losses, if this forced twice the wind through the venturi, would it not have to travel twice as fast at the narrowest point, where the propeller would be located? and thus have 8 times the energy? Would that not be much better than simply making the propeller double diameter, which would only be 4 times the energy?

   I would note that:

1. For a generator the flaring exit is more important than the flaring entrance. (For an airplane duct the entry is the more important.)
2. A rounded entry/exit shape (see detail) improves performance, drawing in or sucking out air more effectively.
3. The best profile for both entry and exit section of the venturi tube could no doubt be worked out with a computer simulation. I don't want to wrap my head around that, but if anybody does, please feel free!

   Here is the idea, with high wind protection: a moving top that would blow open to shed wind.
   I may never build it because of having too many projects, but it's the concept.

My Solar Power System

The solar equipment panel in the garage (batteries below).
* 3 1KW grid-tie inverters - the AC cords are the outputs.
* PowMr 50+V input buck solar charge controller (39.5V output to NiMH dry cells battery).
* New, right: MPT7210A boost solar charge controller with 3-phase diode bridge, from windplant.
With the surface mount circuit breakers (which enclose all bare wire ends)
I wouldn't again make a circuit breaker box with panel mount breakers.
The 30x60mm, 24W LED panel light that lights this image from above-left
is plugged into the 36 volt HAT connector socket on the breaker box.

Month of October Log of Solar Power Generated [and grid power consumed]

(All times are in PST: clock 48 minutes ahead of sun, not PDT which is an hour and 48 minutes ahead. DC power readings - mostly the kitchen hot water heater - are reset to zero daily, while the others are cumulative.)

Date  House solar KWH(Grid+DC), +Trailer Roof solar KWH = day total KWH made [power co. meter readings] weather, usage...

Sept. 30th 96.93+.48, 836.90 -> 9.32 KWH [Finish car chj; BR Heat; 67944@18:30; Bath] sunny.

October 1st 97.69+.25, 837.43 -> 2.54 KWH [Car 60 Km, chj@3800W; 67969@20:30] heavy clouds & rain. I found the (ground fault) circuit breaker to the trailer blown, so the 4 panels on the trailer were off-line for an unknown length of time.
2nd 98.89+.43,  838.66 -> 2.86 KWH [BR Heat; 67980@20:30] Cloudy.
3rd 103.50+.54,  842.07  - 8.56 [67990@19:00] mostly sunny
4th 108.03+.36,  845.72  - 8.54 [85Km,Chjcar@3800W; 68013@18:30] mostly sunny
5th 111.88+.37,  849.04  - 6.74 [trailer heat; 55KmChj@3800; 68034@19:00] Sun & clouds
6th 113.33+.45,  850.23  - 3.09 [68049@19:30; trailer heat 250W(on from now on for winter)] Cloudy.
7th 116.73+.61,  852.41  - 6.19 [68064@19:00] AM Sunny, PM some clouds & rain.
8th 120.91+.80,  856.37  - 8.94 [68078@19:00] Sunny! (Frost early AM - well, it is October! Fairly warm PM.)
9th 123.41+.56,  858.67  - 5.36 [55Km,Chj@3800W; 68104@18:30] AM sunny, PM mostly cloudy
10th 125.97+.44, 861.53 - 5.86 [68120@18:30] some sun.
11th 126.81+.35, 861.53 - 1.19 [85 Km,chj@3800W; 68150@20:00] Overcast. Trailer plug got pulled out previous night, unnoticed until I went to check the solar output. In other news... car went 88 Km, getting 7 Km/KWH, and recharging at 90% efficiency would have used about 14 KWH recharging. Add 6 KWH for trailer heat, that's 10 KWH for everything else. The same, 6+10 the day before but without the car. Definitely consumption is rising for the dark, cold winter, just as solar production plummets! The solar hot water wasn't hot this day, but it was certainly draining the batteries when I turned it on. I worked on the mounting for the windplant.
12th 129.22+.42, 863.89 - 5.19 [55Km,Chj@3800w; 68173@19:00] Some sun and some rain. So today ~9 KWH to recharge car, 6 for trailer heat and 8 for everything else (including bedroom heat at night).
13th 130.85+.57, 865.91 - 4.22 [68191@19:00] Some sun but more clouds.
14th 131.21+.00, 866.17 - 0.62 [68211@18:00] Horrible, simply horrible! Howling gale & rain. Dullest day! Winter has struck!
15th 133.00+.47, 867.72 - 3.81 [68227@17:30] A bit of sun, then cloudy, some rain.
16th 134.33+.39, 868.98 - 2.98 [60Km,chj@3800W; 86253@18:00] Bit o'sun, rain, clouds and wind
17th 135.50+.43, 870.00 - 2.62 [68269@18:00] Started sunny, but soon turned to rain & clouds
18th 137.08+.47, 872.15 - 4.10 [68289@20:00] A little sun, & some rain. 0 KWH for trip to Masset in Toyota Echo... but 52$ for gas - ouch!
19th 139.39+.43, 874.42 - 5.01 [Laundry; 55Km.Chj@3800W; 68316@18:00] G'day
20th 141.07+.40, 875.95 - 3.61 [15Km,chj@3800W; 68342@20:30] Mostly overcast.
21st 141.66+.10, 876.43 - 1.17 [55Km,chj@3800; 68367@18:00] Overcast
22nd    1.07+.73, 877.64 - 3.01 [Power fail most of night; start charge car (2KWH?); 68379@17:00; finish charge car] Mostly overcast, sprinkles.
23rd    1.62+.34, 878.12 - 1.36 [68405@18:00] wet, dull, windy. Connected windplant - helped by a few watts.
24th    1.92+.23, 878.35 -   .76 [68425@16:30] Wow! Wetter, Duller, no wind.
25th    3.78+.54, 880.13 - 4.18 [68443@17:30] largely sunny but a bit of clouds and rain.
26th    5.96+.47, 882.56 - 5.08 [55Km,chj@3800W; 68466@20:00] Sunny!
27th    7.96+.48, 883.87 - 3.79 [35Km,chj; 68492@23:00] Sunny! (Plug came half way out from the trailer)
28th[PwrFail]+.42,886.57- 5.62 [68505@17:00] Sunny Again! Est. house: 2.5 KWH - blown breaker reset meter)
29th    0.87+.43, 887.18 - 1.91 [68526@16:30] Overcast.
30th    1.47+.32, 887.69 - 1.43 [laundry;55Km&chj; 68560@17:30] overcast.
31st    2.98+.24, 888.72 => 2.78 KWH [68579@17:30] mostly cloudy, a bit of sun in the afternoon.

Solar: House+DC, Trailer  - total KWH [grid power meter reading(s)@time] Sky conditions
01st    4.00+.19, 889.33 => 1.82 KWH [68598@18:00] light overcast.
02nd   5.41+.56, 890.50 => 3.14 [55Km.chj.car; 68627@17:00] cloudy AM, sunny PM.
03rd    7.20+.42, 891.87 => 3.58 [68648@19:00] Mostly sunny.
04th    8.08+.39, 892.65 - 2.05 [68666@17:30] cloudy. WHAT IT'S THE 4TH ALREADY?
05th    8.79+.11, 893.28 - 1.45 [68684@20:00] cloudy.

KWH-  # of Days (Oct)
0.xx  - 2
1.xx  - 5
2.xx  - 5
3.xx  - 5
4.xx  - 3
5.xx  - 6
6.xx  - 2
7.xx  -
8.xx  - 3
9.xx  -

Monthly Tallies: Generated KWH [Power used from grid KWH]

March 1-31: 116.19 + ------ + 105.93 = 222.12 KWH - solar [786 KWH - used from grid]
April - 1-30: 136.87 + ------ + 121.97 = 258.84 KWH [608 KWH]
May  - 1-31: 156.23 + ------ + 147.47 = 303.70 KWH [543 KWH] (11th solar panel connected on lawn on 26th)
June - 1-30: 146.63 + 15.65 + 115.26 = 277.54 KWH [374 KWH] (36V, 250W Hot Water Heater installed on 7th)
July  - 1-31: 134.06 + 19.06 + 120.86 = 273.98 KWH [342 KWH]
August 1-31:127.47 + 11.44+91.82+(8/10)*96.29 = 307.76 KWH [334 KWH] (12th panel connected on lawn Aug. 1)
Sept.- 1-30: 110.72 + 15.30 + 84.91 = 210.93 KWH   [408 KWH] (solar includes 2/10 of 96.29)
Oct.  - 1-31:  55.67 + 13.03 + 51.82 = 120.52 KWH - solar [635 KWH - from grid]

Egads, here I am again with a late newsletter I'll have to work out the monthly totals another time, again!

6  month total March 1 to August 31: 1643.94 KWH made; [2987 KWH consumed from grid]

Things Noted

* On September 30th over 9 KWH was generated. No daily figure in October hit '9' again even on the sunniest days as the sun sunk down toward the horizon with long tree shadows reaching everywhere and the days getting shorter. It's a long way down from 17 and 18 KWH on sunny days in June (and with only 11 solar panels then instead of 12) or the several '15's in August. (All-cloudy July we won't talk about.) In spite of that, collection in September was 2/3 and in October was 1/3 as much as in summer - not a lot, but not trivial.

* After the 10th the collection figures became rather abysmal. Obviously the panels need to be standing up at a much steeper angle (than 15° from flat) toward the south. I may remount some or all of them.

* The time of day I went out and got the day's solar collection readings got earlier and earlier as the days got shorter.

* We can be sure the shallow slope of the south roofs along with the lowering sun angle to reduce the sunlight hitting the collectors combined in reducing the collection. I plan to mount a couple of panels on a pole at latitude angle to catch more in the winter. I dug out the old satellite TV post, which had a big concrete footing, dug a new hole for it in a sunny spot, and got my neighbor to move it there with his bobcat/forklift. The cement footing was so heavy the back wheels lifted off the ground a couple of times. (Perhaps I'll put the panels from the lawn on it? Or the two old 208 watt, 54 cell panels from 2012? No one seems to want to buy them at any reasonable price. I'm offering them at a bargain, but I'm not going to give them away, so I might as well put them up as have them sit in my cellar?)

* Except when it was fully sunny, I kept turning off the grid ties from the four 250W panels to feed the DC system with the hot water tank in the kitchen. They didn't need to be off except when the water heater was on, but who could remember to keep flipping them off and on? On dull days the batteries couldn't be kept charged even with them off, and the water heater couldn't be brought up to temperature.

* 50 watts in dull overcast is just 2.5% of the rated 1000 watt panels' output. Accounting for the poor angle of the panels, that's about as dull as full sunlight out by Ganymede & Jupiter (4%). It's still brighter than at Titan and Saturn (1%), but it could be down there by December.

* With a starting point of 6 KWH per day to keep the trailer warm, electric consumption from the grid shot way up over summer levels when it might only be 6 KWH total and any more was supplied by the solar. Added to that even on a "nothing" day now were bedroom heat, and then I turned on the almost 200 watts of LED lights of the indoor LED garden near the end on the month for 8 to 10 hours a day to grow romaine lettuce, carrots, onions, and spinach for winter, and to keep the coffee bushes happy. I brought in tomatoes that were in pots in the greenhouse, but they didn't look good and died. (I think they were suffering from drought ever since I went on holidays in August. The tomatoes still in the greenhouse planted in the ground were still going, if not producing much.) And on days when I drove somewhere, consumption rose further. There was no point trying to charge from solar.

Electricity Storage (Batteries)

Turquoise Battery Project: Long lasting, low cost, high energy batteries

Electrolyte to Depassivate Discharged Manganese Oxides?

   Somehow since the beginning, probably 2008, I had thought I understood that no state of charge of manganese oxides was soluble or an insulator, so I assumed MnO2 could always be recharged. I finally read something that indicated otherwise, that one state of valence 3 - either MnOOH or Mn2O3 (doesn't matter which one) - became passivated and wouldn't recharge.
   This explained otherwise puzzling low capacity and loss of capacity in so many of my cells! Using discharged MnO2 from old alkaline dry cells, if it wouldn't charge up this explained why I had such low initial amp-hours, and why they just got worse after 2 or 3 charges.
   But I couldn't believe there was no way around it. The Polish researchers that mentioned the problem were avoiding it by using potassium sulfate electrolyte, which (as I surmised with oxalate) moved two electrons at once and so took the MnO2 (valence IV) straight to Mn(OH)2 (=MnO, II) with no intervening valence "III" state.

   I thought I had a success when I added some potassium oxalate to the first cell on the evening of the 3rd. It improved the voltage and current drive. In fact, the cell would now discharge only very slowly below 1.7 volts open circuit, instead of below 1.6. The improvement also in voltage under load from less than a gram(?) of this salt was quite surprising.
   But again once the cell was discharged it didn't seem to get it all back again. By now I figured most of the Mn in the electrode was at valence 3 and somehow passivated. With a 20 ohm load, the cell was down to about 3 minutes before it went under 1.0 volts. Pathetic! In the original first test of this cell, while the voltage had been under 1.2 most of the test, it ran for almost exactly 2 hours down to 1.0 volts. And here one suspects that already much of the MnO2 was in the MnOOH state - explaining the low amp-hours of all my cells with manganese oxide positives.

(4th) The other thing to try to get the MnOOH to recharge was sodium. I added 1/2 gram of sodium oxalate after lunch. I'd have added more but it isn't very soluble. That was probably already more than would dissolve. (I was a little reluctant to try sodium chloride because all I had was "sea salt" or "iodized", neither of which, by definition, was likely to be pure.) I put the cell on charge at 1.9 volts. The sodium oxalate didn't seem to do much, which wasn't surprising considering how little soluble it is.
   So I added 1/2 a gram or so of table salt. (pH remained at 13.) This seemed to have an effect. Upon charging, the discharge times before dropping through 1.0 volts very gradually started to lengthen: 3 minutes, 4 minutes, 5+ minutes (times 3 tries), and then 6+. With the cell only drifting lower slowly from 1.8 volts, it also didn't draw much current at 1.9 volts and so it took a long time to recharge between tests. (If it's discharging at 60mA and only charging at 8... I wasn't even sure I was allowing enough time between discharge tests.
   Then I thought to add another 1/2 gram of NaCl. By then it was late. That was it for the 4th. I left it on charge overnight, seeing it wasn't charging much anyway.

   Another thing worthy of note: with not sealing the filler hole, the cell didn't start leaking again, and likewise the other one sat idle without leaking.

(5th) In the morning I disconnected the charge and let the cell sit an hour. It only lazily drifted down from 1.9 volts, saying it couldn't have been drawing much charging current. It was still at 1.83 when I connected the 20 ohm load. This time it lasted 7.5 minutes. That was again an improvement, tho not a large one. It had had all night on charge, so it seemed it was the cycling that improved it rather than just the charging. That was about 6.5 mA-H. It was going to take a lot of cycles and a long time to get to an amp-hour at this rate!
   But the important thing was the direction. Where everybody else apparently has only had manganese oxide electrodes becoming more and more passivated - as per all my previous results - it was gaining; depassivating. Having a sodium component to the electrolyte seemed to be the key.
   The day's discharges lasted 7.5, 7+, 8 and 8.5 minutes. Very gradually longer. And the charging was so slow I wasn't always sure I was putting more in than came out during the last discharge. I thought of raising the charging voltage from 1.9 volts, but it wouldn't be good to be charging MnO2 to permanganate, KMnO4. (Considering how long it took to drift down from 1.9 to 1.8 volts when the charge was removed, was it already charging to permanganate? Yet it fell to about 1.55 volts almost instantly when a load was connected.)

   Might we then presume that if the mix started out charged to MnO2, the presence of the sodium would prevent it from passivating in the first place? So, if instead of putting the (mostly discharged) MnOx electrode directly into the cell, putting it in some bleach first would chemically charge it to MnO2, whatever state it came out of the old cell in - and electrically passivated or not.

    Bleach (NaClO), on giving up its oxygen to oxidize the Mn to MnO2, would doubtless turn into salt, NaCl. Now NaCl is an intended ingredient of the electrolyte, so any remnant after draining and rinsing was probably okay. (Did it even need rinsing? How pure is the bleach? I guess it needs rinsing in case of unreacted bleach.)
   I decided that rather than form electrode bricks and bleaching them, I would just bleach the powder and lumps straight out of the old dry cells, in plenty of bleach. That would hopefully ensure that it would be starting out as fully charged MnO2 when I made the bricks. And it would mean that the cell would start out with both electrodes already fully charged. Perfect!
   (Let's see... I read in somewhere where MnO2 was bleached in NaClO, but they also added KOH to the mix. What was that for? Perhaps just to ensure it was alkaline. Wouldn't want "Mn++" ion dissolving out in acidic solution! First I thought I should add a bit of KOH too. Then I remembered the dry cell mix is already full of it. So just diluted bleach then!)

   Well, I started out in 2008 thinking that while starting without knowing too much might cause me trouble, but without having the same preconceived ideas everyone else had, without "knowing" battery details that everyone who was educated in the field did, I might come up with something new. It certainly didn't think it would take almost 12 years (not that I hadn't already come up with, at least, charging Mn to metal in water for a highest voltage negative electrode), but here is something. I thought MnO2 electrodes were said to be short lived because they get overcharged to MnO4- and dissolve. It didn't seem to me this was insurmountable. But everyone in the field probably "knew" that MnO2 passivates on discharge and that that is the cause of its short cycle life. And so they passed it over in their search for a long lasting positive electrode. I didn't know, so I tried it. And in spite of all my misunderstood problems with it, I've now found a way to make it work. And maybe the nickel manganates idea, too.
    (Well... it soon seemed it wasn't really working after all.)

Air Leaks?

(6th) After charging all night, the cell was still drawing over 7 mA, whereas after the previous night's charge it had been down to almost 3. When the charge was disconnected, the voltage dropped pretty quickly into the 1.7 volt range instead of drifting very slowly down from 1.9. A load test ran for a mere 3 minutes before dropping through 1.0 volts. What was different? Why had it deteriorated like that?

   A possible answer was staring at me. I had decided to top up the liquid with 1 cc of water. And I had left the filler cap off, that little bit of modeling clay. Was enough air getting into the cell through that tiny filler hole to reduce its previously improving performance? Theoretically oxygen would spontaneously discharge the zinc. AFAIK it shouldn't affect the MnO2 side. But if it was the MnO2 side that was low on capacity, why would reducing the zinc side a bit - or even a lot - cause any noticeable performance drop?
   I placed the cap back on (careful not to make it really airtight and cause pressure buildup and leaks), and put the cell back on charge.

   Well... It didn't help. Apparently MnO2 by itself really isn't a good rechargeable electrode substance. That takes it back to trying to get good results from nickel manganates.


   I did a bit more reading. Another researcher thought that the discharged manganese oxide "sloughed off" the electrode, electrically disconnecting it and so making it unavailable. That was different than "becomes an insulator". Does somebody know the actual reason and others are ignorant like me, or is everybody just guessing?
   And yet this makes a certain amount of sense: If MnO2 discharges to Mn2O3, an oxygen atom has left the space, so what's left is less dense. If on the other hand it discharges to MnOOH, just a hydrogen has been gained instead and this wouldn't apply. Mn3O4 loses still more oxygen, but the final product Mn(OH)2, one expects would be hardly different density than MnO2.

   Rather than simply abandoning a cell that didn't seem to charge, I decided to put in a little bleach and see what happened. (1/2 a cc.) That was after all what I now wanted to do with all the MnO2 before putting it into batteries. As it was now late, I left it idle overnight. It didn't seem to have any notable effect. I put it back on charge.

   I had to confess that I couldn't seem to get the manganese to recharge to MnO2 by any simple technique or salt I had tried.

Nickel Oxides Electrode

(14th) That being the case, and having had (so far) poor performance trying nickel manganates, perhaps it was a good time to go back to something known to work and proceed from there? That would be nickel oxyhydroxide in pH 14 alkaline solution. But even there, I had used the bits of NiOOH electrode from a commercial dry cell. It would be going another step to make my own electrodes. A pretty safe step, one expects, but it needed to be verified.

Making NiOOH

   Supposedly one can 'pre-charge' Ni(OH)2 to NiOOH with bleach. But I looked at the Japanese patent again and saw they also added sodium hydroxide to their bleaching solution. Why? Of course!, if the nickel will only stay charged at pH 14, then the oxygen would bubble off without the NaOH. (I remember seeing a nickel electrode bubbling from a decade ago. Was that from bleach at neutral pH and then the oxygen bubbled off, or from hydrogen peroxide to DIScharge it?) But they said they washed it with water afterward, which would have put it in neutral pH water. Perhaps my theory is off-base. Nevertheless I proceeded to do it their way.
   I only had 6% bleach, so I added less NaOH, too. Next they heated it to 60°C and added the Ni(OH)2, and kept it heated with agitation for an hour. My God!... I actually bought a lab electric hot plate with magnetic stirrer this summer - I figured I'd want it for something! I could actually do it, the same way, instead of doing some workaround stirring it on the kitchen stove! I did so.
   I poured in 20 grams of turquoise-green Ni(OH)2. It started to darken immediately and was soon black. Presumably it was already NiOOH, but I decided to heat it and stir it for an hour per directions. (What if just a little had turned black and it made it all look black? A bit of conductive carbon black had turned a whole jar of bright yellow paint jet black.)

   After the hour I had trouble filtering the NiOOH (presumbly gamma form NiOOH) out of the liquid. (Next I must get some proper lab filters and filter papers.) But I found that the black precipitate slowly settled to the bottom. Then the fluid could mostly be poured off without losing much. I then refilled the beaker with pure water and repeated. I figured the bleach and hydroxide would be pretty diluted. Anyway bleach would become salt (NaCl) and sodium hydroxide was also good electrolyte. I couldn't see that they would do much harm if there was only a little left. (Slightly nervous about the salt.)
   Then I put on some heat under the beaker and left it quite a while at 60°C. It got a lot drier. I turned it off for the night, but dried it some more the next morning (15th). It started to get a bit crumbly. It weighed 37 grams. I had only put in 20 grams of Ni(OH)2, so the rest must still be water? I added over 2% (.5 g) of Sm2O3 and over 5% (1.25 g) of conductive carbon black, having decided to assume it was still about 20 grams of NiOOH. (The Japanese patent said "acetylene black", but I think that's just a different name for "conductive carbon black".) Then I added a few drops of the dishsoap 'sulfonates gel'.
   The next step was "mixing for 1/2 an hour in a kneading machine." I didn't have one, so it was in a mortar and mostly with a teaspoon - the pestle mostly just made it jam up against the glass, and it was too stiff to squash it from there, so it couldn't do much kneading.
   It still wasn't dry dry, but had the consistency of a stiff modeling clay. I separated some fibers from the carbon fiber cloth and put them into the bottom of the compactor die. Then I crumbled some bits of the NiOOH mix and dropped them on top. Then two more layers of fibers and crumbs. I took it out to the press, but I didn't get much pressure up before curlicues of paste started oozing out every crack. Perhaps if needed to be drier? Then again it was already a pretty solid paste. It was about 4 mm thick.

Turquoise nickel hydroxide converted to black nickel oxyhydroxide.

   I hoped to not have to make a new zinc electrode and case, so I took the first case and managed to slice it open with a knife. I decided to go with something that worked - at least it did with manganese, but it has a lower reaction voltage than nickel - and made a graphite foil current collector. Instead of just a strip, I cut a piece to fit under the whole electrode, still with a 12mm wide tab to the terminal. I scoured the contact surface with scotchbrite, then for good measure scratched it up with sandpaper.

   I took two chances with experimental things at once. The manganese oxides electrode had charged and held its charge (as much as it was going to) with the dishsoap gel in the mix, so I added that. Then, a cupro-nickel current collector would have been sure to work, but only in pH 14 alkaline electrolyte. If I used graphite and it worked (presumably!), I could later try changing the electrolyte from alkali to lower pH alkali-salt and see if that still worked, or if that made the higher voltage NiOOH spontaneously discharge itself. So hopefully one cell would do two experiments - nickel in alkali and then nickel in salt.

   I removed the piece of separator cloth because its fibers stuck out. Perhaps that's part of the reason it leaked? It still had the separator grille, which should be plenty, and even that was jelled with agar for added protection against shorts. I put everything in place and closed it up.

   Initial voltage was under a volt. I used an eyedropper to add potassium hydroxide until it came up into the top chamber. then it was still only 1.12 volts - didn't seem like much for a "pre-charged" cell. I followed the wires and it was connected to the power supply, which was unplugged. If it's not on, it acts as a load instead of a supply. With that disconnected, the voltage slowly rose past 1.3.  Then I got the charge power connected, and the cell just jumped to 1.9 volts drawing only 20 milliamps. Ugh! What could be the problem?
   It smacked of an electrode with poor conductivity. I had a feeling I should have dried the nickel electrode further and compressed it much harder. But why should it need charging anyway? The nickel side was pre-charged to black oxyhydroxide, and surely the zinc side wasn't far discharged?

   After a couple of hours of this slow charge, the cell would still only put out 130 mA if shorted, dropping to 80 after 10 seconds. With a 20 ohm load the cell immediately dropped under a volt, delivering around 40 mA. Worse... a puddle began to form. The cell leaked. Ahrg!
   Rather than try to fix the leak immediately, I decided to disassemble the cell, let the nickel electrode dry out better, and then compact it better in the hydraulic press.
(16th) Rather than wait for the electrode to dry, I had only used half the NiOOH mix. I made another one with the other half. It seemed much drier than a day earlier, notwithstanding that I had plunked a little plastic cup over the top. It crumbled more easily into smaller pieces.

Better Technique

   After again doing two alternating layers of carbon fiber and electrode mix, I had the thought that it might be better to put some carbon fiber right into the mix, and "knead" it all in one. If it was well mixed, the fibers should spread out in three dimensions, and they would be better bonded to the electrode material even before compacting. So I put some into the remaining mix and mixed it all up as best I might without going to a lot of trouble. It seemed better. The fibers weren't moving around free but had mix clinging to them. And then I didn't need to use the tweezers to spread the fibers apart and make layers - just put the mix into the die.

   In the Polish rechargeable manganese-zinc battery making project they had used some sort of advanced and extremely light "graphite foam" (I forget the name). I suspect the graphite fibers are just as good. And they're cheap and (with this technique) easy to work with.

   But dry as it seemed, in the press the mix surprised me again, oozing out around the edges as soon as there was much pressure on the punch. I would say it took more pressure this time before that happened. But my little pieces of plastic/cellophane (to keep the electrode from sticking to the metal parts) were soaked and everything just around the brick looked quite wet. And it was still 5 mm thick - I was hoping for about 4 or less and I wanted to put a few tons of pressure on it to make sure the particles were well connected together. I peeled off the pieces and set the brick on edge and on the woodstove to dry further.

   Before bed I put it back in the die and pressed it to 9 tons -- and down to 3 mm. That was more like it!

(17th) I pried apart the cell. It was surprisingly easy compared to the first time. I dumped out the old nickel 'brick', which seemed more like mush, and put in the new one. I cleaned all around the edges and did everything possible to ensure there were no leaks. It was a fiasco! Not only did it leak like a seive, but there was almost no current - a milliamp or two on charge, and just 13mA discharge. This obviously wasn't going anywhere.
   I decided I simply had to print a new case and pried it apart again. (Getting easier each time.) Black goop was dripping from the electrode. It wasn't completely mush, but it had swelled up and the surface had loose particles that were washing off. I had compacted it well. The Japanese patent hadn't mentioned it, but obviously a nickel oxyhydroxide electrode needed some sort of binder to hold it together.
   CMC gum (sodium carboxy methyl cellulose) or VeeGum (a 'bentonite clay'), might be typical. But I was using the dishsoap as a gel? Of course the dishsoap was mostly liquid. It had to be pretty dilute. Probably with just a few drops, which quantity I had really selected so it wouldn't add too much liquid to compact the electrode, I wasn't adding anything like enough of the sulfonates gel.
   I put the substances of the two electrodes back into the mortar and then added a whole cc of the "Lemon Fresh Sunlight". Then I "kneaded" it for a while by hand (mortar, spoons, pestle) to mix everything together - this time including the carbon fibers as well.
  Then I left it all to dry out again. I had lots of other things to do. In the evening it was much drier and could be crumbled into small lumps to drop into the die, but again it surprised me and oozed out around the mold when I tried to compact it. At about 5 tons I glanced down and saw that the punch was all the way to the bottom, with most of the material having oozed out.

Paste Electrodes?

   I had certainly "kneaded" the material very well in this process. It gave me the idea that something like this (but automated) was the way to do the kneading - the thorough mixing and blending of the electrode substances. I scraped the material back into the die and pressed it again, but this time only to 2 tons. At that pressure it didn't ooze out. But the brick now looked much the same as with 9 tons: it was only about 3 mm thick. Apparently then it was "fully compacted": the compaction of the material was in the kneading. And the graphite fibers were well incorporated and probably pretty well dispersed within the electrode.

   Then I started to think that maybe that was a better way to make the electrode anyway: knead the substance and then extrude it under pressure into a mold. I have heard the name "paste electrode" before (along with "powder" and "sintered" electrodes). Obviously the moisture content would have to be carefully controlled - it has to be almost dry, not have enough moisture that one is trying to compress a liquid. And yet enough for it to flow, if only under great pressure, from the kneader/extruder into the mold.
   With this technique I could see filling molds for large electrodes - for my "full size" cells - without excessive pressure. Maybe 10 tons or less instead of increasingly huge as the area to be compacted was increased. Much more practical!

   Later I remembered that I should be able to measure the volume and weight of the electrode and calculate its density. IIRC a nickel oxides electrode should be around 1.8 to 2.2 g/cc. If it was below that, it wasn't compressed enough. That would be a good check on compaction for next time.

3D Printer Insulated Surround

(18th) I cut some extruded styrene foam sheets to make a housing so the air around the 3D printer could be heated for printing ABS. I wasn't really sure that's how I wanted it, so I didn't glue the pieces together. I stuck in a few nails to hold it in place. The "front" wasn't much, but it left me room to insert the SD card with the designs, and blow hot air in with the hot air gun.
   Then I printed another battery case. The front (top of cell), nearest the heat gun, lifted. But the back (bottom) stayed put. Apparently I actually had the front too hot, as it was very soft and pressed back down to the touch. Of course, after 15 or 20 minutes into a 35 minute print, one is reluctant to experiment with different settings, or seeing if the heat gun really needed to be on continuously.

   It also occurred to me why the old RepRapPro printer might have printed ABS so much batter than the new one: It extruded coarser filaments. They would probably have been stiffer when the next layer was printed on top, holding the shape better. (A coarser 3D printer was better?)

Just for whatever, here's a large boat printed by a really big 3D printer.

Next New Cell (Nickel-Zinc with graphite current collector)

(20th) In the evening I finally got back to it. The NiOOH electrode was only about 3 mm thick and the space for it was over 4, so in order to keep it from swelling up I put in two extra separator grilles, cut down a bit to fit. Then I put the old zinc electrode on top and put the case together. This time I went around the outside edge with a strong magnifying glass. I found some spots that might not have been sealed and a couple of definite gaps. Apparently I needed to push the pieces together strongly instead of lightly after applying the solvent/glue. I redid it until I was satisfied that at least it didn't look like there were any leaks. I just put in a couple of eyedroppers of water to keep the electrodes moist because it was late. I had noticed the nickel electrode turning a light color around the edges as it dried. That probably meant the NiOOH was revertlng back to Ni(OH)2.

(21st) I filled the cell with 20% KOH. It started out at just 1.1xx volts, but was very gradually rising. That didn't seem like much for a supposedly pre-charged cell. I left it 1/2 an hour and it rose to 1.740 volts. Hey, that was more like it! But in another 1/2 hour it had dropped to 1.730.

   In spite of all my efforts to make a highly conductive electrode, when shorted it only put out 60 mA, dropping rapidly to 40, and only drew 4 mA on charge, rising to 1.92 volts without resistance. I upped the voltage to 2.0 instead of 1.9. That at least got it up to 13 mA. I decided to leave it there.
   In a bit I shorted it again and got 170 mA. Maybe it wasn't a writeoff after all! Again it's puzzling that a cell that's supposedly pre-charged should need charging before it would do much of anything.

   Then I checked the pH. It was only 13 instead of 14. With my already weak solution, plus the already soaking with water zinc electrode, plus a couple of cc I had added to keep the plus side moist, the electrolyte was too dilute. I started adding flakes of potassium hydroxide with tweezers, seeking out the smaller flakes that would fit through the tiny filler hole. It was tedious. The pH strips finally started looking a little darker... 13-1/4? The charging current went up from 13 to 15 mA. 13-1/2? 13-3/4? Current was up to 17 mA. When shorted it put out 250 mA. It started staying above 1.8 and then 1.9 volts for a minute or two when taken off charge. This was at least starting to seem more like it. Time to try a longer charge, an hour or more, and see what would happen.
   I checked my KOH solution in the jar and found it wasn't much over pH 13 either. I wanted to leave 'room' for adding salt to the electrolyte later, so I didn't want it too concentrated. OTOH the first test was about making sure 'traditional' nickel-zinc with pure alkaline electrolyte worked with my own (jelled) electrodes.

   After 3 hours of charging the cell was starting to look pretty decent. Taken off charge it almost immediately dropped to about 1.895 volts... and drastically slowed there. With a 20 ohm load it took almost 4 minutes to drop to 1.0 volts. After another 2 hours charging current (at 2.0 volts) was up to about 30-40 mA. Could there be an internal short? Nope. Taken off charge it sat just below 1.9 volts. When shorted the cell hit .47 amps and was still at .26 after ten seconds. With a 20 ohm load it took 6-1/2 minutes to drop to 1.0 volts.
   It certainly didn't behave as if it had been 'pre-charged' before assembly. Oh well, it was working! Of course it wasn't good enough yet and I hoped and prayed it would continue to improve.
   Most gloriously, the piece of tissue underneath the cell remained dry.

(22nd) The 20 ohm load time (with cutoff at 1.0 volts) rose from 4 to 6.5 to 8 to 10 to 13 to 18 minutes on successive load tests. Voltages under load gradually rose and the voltage snapped back more swiftly and strongly after each test. It seemed like awfully slow improvement from under 5 to 17 milliamp hours, but at least it was going in the right direction and continuing that way. It held more voltage longer when taken off charge, but it still seemed to lose a lot of charge over a few hours. That too is doubtless improving. But will it eventually get up to at least 200 milliamp-hours and hold most of its charge overnight, with nothing funny happening to kill it? It has a long way to go. And certainly far more charge has been put into it than comes out of it, so it would seem it only improves with cycling, not simply charging.

(23rd) Discharge time for running a load seemed to have plateaued at under 20 minutes. But things weren't entirely static. Momentary short circuit current finally rose from 1/2 amp to 3/4 and I saw a .82 before bed, and longer shorts (after 10 seconds) held .25, .29, .33 and then .35 amps. Charging current had previously stopped falling at about 20-30 mA with a 2.0 volt charge. Now it didn't fall below 40, and even 50 by night. To get less I had to turn it down to 1.9 volts. With a 1.9 volt charge, discharge lasted a minute or two less than with 2.0. How well would it hold charge overnight? I hated to look.
(24th) As I feared 10 hours later it was down to 1.698 volts. I dumped out the electrolyte (a dribble or two) and filled it with distilled water. After waiting a bit, I dumped that, and repeated for 3 total rinses. Whenever I dumped the cell, quite a bit of black powder (+trode) came out. In the last minor variation of the cell I had left an open space at the top. (It was easier to put in the current collector sheet and tab without damaging it.) I guess electrode powder was breaking loose there, but some of it was fine enough to have come through the three grilles, too.
   Next cell: narrower space for the electrode so it only needed one grille, print the grille at a bit higher temperature, make a little piece to block the gap, and maybe strengthen the dishsoap gel a bit more. (Even tho it's probably a source - or the source - of the impurities causing the self-discharge.)

   I strengthened the KOH solution a bit, adding 4 grams of KOH to what would have been 57 grams of water and 11 of KOH. 57+15=72; 72/57=1.26 sg. Still not very strong per most batteries, and still only a little over pH 13. But I left it like that. Performance wouldn't be great, but if it held charge better, I could then say "it works".
   Then I would try adding salt(s) without again draining the cell.
   After 2 hours of charging, I could tell it was better. Current had dropped to 30 mA instead of 40, and when I disconnected the charge it took 10(?) seconds to drop from 2.017 to 1.900 volts instead of 2. Then it got worse again, as I could see from the same effects, and the charging current went back up to 40. I could see bubbles inside. No doubt the gas was carrying up particles. Furthermore it had somehow developed a leak. Was it the manipulations, or had I put the plug in the hole too hard and it built up pressure?  I ran a load test with 33 ohms instead of 20. Aside from lasting 33 minutes instead of 17 or 18, it seemed to be pretty similar to previous tests.
   Overnight it held to 1.742 V for 5 hours (up to pee anyway), 1.734 V for 9, 1.727 V after 11, and 1.715 after 15-1/2 hours. Better than the previous night but nothing like great.
   Then again, I was attempting to make a "typical" nickel oxides electrode, other than gelling it. One thing I didn't do was to briefly torch the surface of the electrode to 'sinter' the surface particles together. That might have eliminated most of the 'crumbs' coming loose and probably still some swelling of the electrode, which both doubtless contribute both to poorer performance and to increasing the self discharge.

(28th) In the evening I got some methylene chloride and smeared around the joins. then I added a little electrolyte. It still leaked. It seemed to be in a lower corner on the seamless side. In the morning (29th) I smeared some of the solvent around everywhere, concentrating on that corner. Then I eyedroppered in some 20-20 electrolyte. (100 cc water, 20 g KOH, 20 g KCl - SG 1.4.)
   When I hooked up the meter it read 1.742 volts steady. Not only was the salt not causing self discharge, but the leaking cell had held good voltage overnight without a big charge beforehand. I connected the charge at 2.0 volts.
   The cell still leaked(!), this time in another corner. Probably the solution is to increase the extruder temperature for the 3D prints of the cases so the strands of plastic fuse together better. And to download and install the proper printing "slicer" and G-code maker for the new printer. (Cura? IIRC.) This should be more feasible now because on the 27th my new internet connection was hooked up.
   Late on the 29th I was rereading this article and was reminded from near the start of the month that potassium oxalate had not only worked but had upped the current drive with the MnO2 electrode. Would it similarly help with NiOOH? As the cell was leaking anyway, and being charged, I added in a dropper of KC2O4 solution. Unlike with manganese dioxide, it didn't seem to make any difference.
   The next time I tried seeing how long it held charge, it stayed above 1.76 volts overnight, over 1.74 for over 10 hours, and 1.70 for 25 hours. If not satisfactory, at least that was improvement.

   The next electrode was going to be a mix of new Ni(OH)2 and MnO2 from dry cells. Now I can either torch (surface sinter) another purely nickel oxides one, or do both at the same time. The first is more illustrative of the torching effect. The second, two experiments at once, is faster progress toward what is expected to be the desired outcome. I suppose the first is the way to learn more.

   The better a cell performs, the more tedious the tests become. I start to think more and more of a microcontroller unit for automatically cycling cells over and over and logging the performance.

Honda NiMH Hybrid Car Batteries

(17th) I had mentioned many moons ago (May, TE News #132) the idea of arranging the Honda hybrid car batteries I had acquired at least 3 years ago if not 4, into five rows of six tubes each, with each tube of six cells having a nominal 7.2 volts.
   That would put 6 parallel columns of 5*6=30 cells each in series, giving a nominal 36 volts, the same as my other cells. It got around the awkwardness of each welded tube being 7.2 volts instead of 6, and hence being unable to make any lesser voltage work out with other cells and other configurations.
   If we assumed the cells were 6.5 amp hours, then the six columns would total 39 amp-hours. 39 AH * 36 V = 1.404 KWH. If it actually worked that was probably substantially more storage than my even older Tenergy NiMH "D" cells. I had my doubts. If some dry cells were way down in capacity, might the others not be just about as bad? That was a good part of the reason I hadn't bothered. But after all they were a different manufacturer. Maybe they'd be better?
   But back then, I had 3D printed some supports/spacers that would enable doing that, along with some for my other battery tubes. I had simply never put them together. I had however charged up the batteries. They were now at the bottom edge of 'won't corrode away', about 1 volt per cell (just 31 volts for 30 cells in series).
   And recently I had run across something I needed: #10-24 threaded rod. I had some when I moved up from Victoria but had lost track of it, and then I had found no one sold them here. (I could get 1/4" or larger, but no one seemed interested in bringing in any that was smaller. I had printed the plastic pieces for #10 and would have had to redo them for 1/4".) It had seemed like a stupid little thing to bring an almost done project to a stop, but it did.

   Now I cut some threaded rod into four 10 inch lengths, cut four pieces of wood for it, and put the mountings together. Then I wired them up. I disconnected the other batteries from the charge controller and connected it. By then it was evening of a dull day, and the PowMr controller was only putting out 20 watts. The voltage rose from 31 alarmingly fast and hit the programmed 39 volts well within an hour. That sure didn't say much for the capacity. Less than 20 watt-hours from about 70 pounds of batteries? I think my own marginally successful battery experiments worked better than that!

(Oops, I didn't get a picture of the white printed plastic supports!)

                        Piles of NiMH Dry cells - none of them with a lot of capacity
   Perhaps the best that could be said was that at last I had done it, and got the batteries off the shop floor, out from underfoot. And seemingly it reinforces the view that old dry cells seem to dry out and don't work well any more, so that if one wants lasting cells, they should be fillable. I wonder if it's worth harvesting the nickel oxide material out of old NiMH and NiCd dry cells? What I've found so far is that their cans are much thicker and tougher than those of MnZn 'throw away' cells. (to take the pressure of overcharging without bursting.) Opening one with the pipe cutter was hard slogging. One would want some automated process.

Two Lithium Battery Charger Circuits: for AC and DC Power

   My NiMH 'D' dry cells having turned more or less to crap (very low capacity) after just a few years, I wanted to put my lithium batteries into the solar power system, but they just aren't self-balancing. One cell will be fully charged first and its voltage will rise and rise. Another cell in the string will refuse to rise above 3.3 or 3.4 volts even when charged. Cells held above 4.2 volts are in danger of failing, perhaps catastrophically. Hence all the "battery management systems" and "excess voltage shunt boards" to keep them happy.

   What seemed to be needed - seemingly to me a no-brainer - was a separate charger for each cell, about 3.6 volts until charged, or about 3.35 volts for continuous daily solar trickle charge. Each charger would sit bolted on top of its own battery in place of the power wasting voltage shunter boards that often adorn them. What was needed for a string of cells in series was chargers with a common input but isolated outputs, so that regardless of input the outputs could be offset at the different DC levels. The input could be AC or any level of DC, as long as the charger could put out its maximum current until the voltage hit 3.6, and then didn't rise above that.
   For just 3.6 volts, an isolated output, current limited DC to DC converter suggested itself. I couldn't find any adjustable (or 3.6 volt) isolated output DC to DC converters on line. That would seem to indicate that no one has done it. What was the problem? The only reason I could think of was that isolated output converters were too expensive. When one can buy a non-isolated, adjustable DC to DC converter board for 2$, it didn't seem reasonable that an isolated one would be prohibitive. "36 volts" nominal (actually close to 40 volts) usually means 12 cells in series, so 12 converters. (If you really want just around 36 volts, use only 11 cells. But that seems unsymmetrical.) Surely they couldn't cost more than 15$, and that price would be just 180$ for what ought to be the best possible charging system for my 36 volt solar or 36 volt electric Sprint car. Two wires would run in parallel to all the chargers.

   Did I really want to make isolated output DC to DC converters? Maybe but not badly enough to divert from other projects!

Charging From AC Line

   Or else, perhaps an AC to DC power adpater, regulated at 3.6 volts or regulated and adjustable over that range. The DC outputs are always isolated from the AC. (Except one lead-acid battery charger I used once. The grounded "-" terminal caused sparks in my application.) There seemed to be lots available that were 3 volt, 4.5 volt, 5 volt and up, but charging a lithium cell is very specific. It has to be 3.5 to 4.1 volts, preferably 3.6 to 3.8 or (my best guess) 3.35 volts for continuous trickle charge. I found some on Amazon... or at least they said so. They appeared to be just the thing - chargers for tools with single lithium cell batteries. The cheapest one was 15$. It said 3.6 volts at the top of the page, but the picture of the unit said "9.0 Volts" in tiny print right on it. This does not inspire one with much confidence of getting what is expected. The others started at 20$ and went up. That seemed like a lot and I didn't look at them. (Perhaps some might have offered quantity discounts?)
   I went to AliExpress and found some for 12$ that were 'adjustable 3 to 12 volts, 5 amps', with an adjustment knob and an LED meter voltage readout. I was pretty sure 5 amps was the best current rating I would find. That left only the question of what happens when the battery is drawing 5 amps at say 2.9 volts because its charge is getting low, and the charger can't get it up to the set 3.6 volts? Does it just limit and put out 5 amps, or does it do funny things - like quit because of "overloaded output"? It does say "battery charging", which should imply that it will work in constant current mode, but nothing actually says so. You often get frustratingly little information on line. I decided to try them out and ordered 13 - an extra in case of trouble.
   AC power adapters aren't the ideal thing for solar charging with 36 volts DC, but they would be good for wall charging the car. Except of course for the length of time it would take to charge 300 amp-hour batteries at 5 amps if the car was in daily use. 15, 20 or even 30 amps would be better. Once again, with nobody else seemingly wanting to charge that way, the only way to get that economically would be to make one's own. At least 12 units * 3.6 volts * 5 amps is only 216 watts, so it would be easy to charge with solar panels - even one 305 watt panel on a sunny day - and a 120 VAC inverter.
   This would need two power bars, one with 8 sockets for the 24 volts of batteries at the back under the tailgate.

(22nd) While I was thinking about them I went to top up the spare lithium batteries, one 12 volt battery at a time on the old lab power supply. At 14 volts one of them didn't draw current - it was still charged from a year ago or whenever I last did it. On the next one one cell was charged and rose quickly to 4 volts. It was at one end, so I reconnected the alligator clips to the other three and set the voltage to 10.5. Then one of those three rose up higher than the other two, and I went for just the end two at 7.0 volts.
   The voltages are very specific. At that point it occurred to me that if one charged two cells at a time, at 7.0 or 7.2 volts, then even if one of them refused to rise above 3.28 volts (typical 3.29 to 3.32), the other cell would have at the most 3.72 or 3.92 volts on it. That's well within the safety range which extends up to 4.2 volts. The units I ordered are adjustable from 3.0 to 12 volts.
   So charging lithium cells in series in pairs at 7.2 volts would still be safe - better and simpler than a "BMS" - and there would be half the number of chargers (still 5 amps) and half the wiring and less interconnecting of paralleled cells. For 36 volts, only 6 chargers instead of 12. (or 12 for the 72 volt Miles truck if I put lithiums in it.)

   Still later I realized that even with 3 cells in series, if charged at 10.7 volts or less, none of them would rise above 4.2 volts. (barring a failed cell or other disaster of course.) That reduces the number of chargers required for 36 volts to 4.

   Of course I'd love to put my own ?-zinc batteries in everything and maybe eventually I will, but EV.s draw a heck of a lot of current when you step on the electron pedal. Lithiums supply it.

(29th) The 5 amp, 3 to 12 volt power adapters arrived in the morning. (Fast shipping!) I plugged a couple in. My only initial complaint was that all the adjustment is below 1/2 way on the control - anywhere on the whole upper half just gives 12.5 volts - and it's hard to get within 200mV of the target; even harder to get within 10. But somehow I got there, close enough.
   I thought about how to put them in the Sprint car. With 12 volts under the hood and 24 at the back it would be hard to divide as 9-9-9-9 (nominal battery voltage - really 9=10.5 to 10.7; 6=7.0 to 7.2). I decided that rather than run another wire from front to back to do that, I would just use two adapters under the hood and 3 in the back: 6-6-9-9-6 volts. After all they didn't cost much. (I could have used 4 in the back and made them all 6 volt.) There would have to be a 120 volt cable from front to back anyway.
   I cut off the "12 volt power adapter" plugs and crimped on some ring lugs to screw under the battery bolts. While it might be nominally nice to be able to unplug the chargers instead of wiring them in solid, I've never found those plugs to be very reliable. If any one charger doesn't work right, one set of cells won't be charged when you come to drive somewhere or otherwise use the batteries. That's just trouble.
   I checked the first one and found the meter to be a little off. Not badly. Them I short circuited it to see the 5 amps. It put out 16 amps and the connection at my thumb got hot! I hope they don't destroy themselves charging a low battery. Perhaps it's 5 amps at 12 volts and 10 amps at 6 volts? Of course, I had it at 0 volts, shorted, which is an unusual condition.

   It should be noted that once the cells have charged to the set voltage, they are still drawing current and there's no way to tell when they've finished charging, or their exact state of charge. So I plan to put in an AC power meter on the supply line. The watts being consumed should drop to "not much" once all the cells are fully charged.
   It should also be noted that if batteries are in parallel and there's chargers connected at, say, the 9 volt level, all of the paralleled cells must be connected together at that 9 volt level.

These things were removed when putting the new chargers in:
the previous (and inadequate) 2 amp charger and the power-eating shunt boards that prevent each cell
from going over 4.2 volts by drawing enough current and burning enough power to reduce the voltage.
(Theoretically there should have been 12 of them, one on each cell, but I didn't have them all installed.)

Replacing this are the two power adapter chargers, and an AC power meter to tell when the cells are charged.

"Car alternator" isolation diodes soldered to copper plates, to prevent the chargers from discharging
the cells when unplugged. Note: In order to solder to the body of the diodes,            
 the zinc coating has to be filed off.                                   

   When I went to connect the first one to the car batteries, it occurred to me there was a dirty trick they could pull: to draw current from the battery when they weren't plugged in. Sure enough, it drew about 35 mA. Can't have the batteries being perpetually discharged whenever they're not being charged!
   So each power adapter needs an isolation diode.

Plan For DC Solar Charging (36 Volts, 72 Volts)

   With 'only' 6 regulators needed for 36 volts, I came up with "a cunning plan" for charging from DC using non-isolated DC to DC converters. By 'cunning' I mean unorthodox, perhaps risky. The charge controller itself replaces the top DC to DC converter/regulator, so it only needs 5 instead of 6.
   (I was going to show the circuit diagram here, but I think I'll only show the successful one.)

   On some consideration, the lithium cells I've used are fully charged at about 3.33 volts. 7.0 volts for two is 3.5. Often 3.6 is recommended. However, If I'm using them for solar, they'll get trickle charged all day. For continuous daily charge, it would be better that they get to 80-90% fairly quickly and not be pushed hard up to 100% and beyond. I decided 3.35 volts would probably be sufficient - so 6.70 volts, carefully adjusted, and to be raised a bit if measurements showed some cells weren't getting a full charge. So the voltage from the charge controller would be 3.35*12= 40.2 volts. If the cells are all charged, there should be virtually no current flowing. If they aren't charged, they can draw pretty heavily with the charge voltage just a small bit higher than the cell voltage.
   Such a low charge voltage says also that it should be safe to charge with three cells per voltage regulator, at 10.05 volts. Even if two cells didn't want to rise above 3.28 volts, that's 6.56 volts leaving only 3.49 on the other cell. That's well within 'safe' range. Now we're down to just 3 converters plus the regulated charge controller for the top 10.05 volts, to manage 36 volts!

   (Here's when I realized one could charge 3 in series...) In fact, now that I think about it, even at 10.5 volts over three cells, and even with two cells 'stuck' at 3.28 volts, the third cell will be just under 4 volts. Gosh, the 12 power adapters I ordered could safely do grid charging for both the 36 volt car and the 72 volt truck, 4 and 8 of them!

(24th) I decided to hook up 12 of my 14 - 40 amp-hour lithium cells in place of the now pathetic NiMH ones in the solar power system. That's only about 1500 watt-hours, but if they still have full capacity or near to it, they should be a lot better. On dull days with 1 KW of solar producing only 40-60 watts, it was getting tedious trying to get the 36 volt kitchen sink water heater up to temperature without running the NiMH batteries down too far, even turning the water heater from 250 to 160 watts. (I didn't actually connect them to the house, but I did make and test the charger circuit - or rather, 3 or 4 versions of the charger circuit until I had one that worked.)
   I had 5 higher power DC to DC converters with adjustable current limits. 3 of those plus the PowMr itself could divide the 40.2 volts into the required 10.05 volt segments, charging 3 cells each, without any further need for any "BMS" ("battery management system").
   My one reservation was that the stupid PowMr can easily decide that a system over 40 volts is a 48 volt system instead of 36 (it's happened to me!), and set the top three batteries on "incinerate" and "explode". How I wish this most vital of all system parameters was a manual setting instead of "auto detect"! I decided to set it to 42 volts instead of 40.2, and use four of the DC to DC units for the four sections. That would only give an input to the top one of 1.8 volts above the output. but it should be enough. And if the PowMr went crazy and upped the ante to 56 volts, the top unit would still output 10.05 to the top three cells, even tho its input voltage would be 25.85 volts instead of 11.85. That would at least theoretically be safe. (OTOH it's way over the voltage limit for the bottom charger. Maybe I should put in a fuse in case it blows.)

   I had also purchased 5 DC to DC down converters with LED voltage displays. That would be nice, but they didn't have current adjustable current limits nor good heatsinks. (Later I ordered some 20 amp DC to DC converters - even without a display, those would be the best.)

   In looking into the current limiting I feared it might be done with a small sense resistor in the ground line. That would mean one can't actually tie "Vin-" to "Vout-" and ground them both. But in stacking them in "totem pole" arrangement to charge batteries in series, there is no way to avoid tying them together.
   So I looked up the datasheet for the XL4016 DC to DC converter chip that they all seem to use. "Vin-" and "Vout-" are connected together in all the "typical application" circuits shown. But the higher power, adjustable current limiting ones I had had something labelled "R010" connected to Vout-. And if I passed a current through from Vin- to Vout-, there was a small voltage, 10s of millivolts, between them.
   I decided just to go ahead anyway. The bottom one, connected to "Vout-" of the charge controller, could be set to a specific current. The others should follow suit. Right? Right? Like I said this is an unorthodox plan, intending to eliminate more complex and specialty lithium "BMS" systems. There are a number of interacting variable and I think the way to see if it works is just to try it.

   I got everything ready, but I didn't actually connect the wires, because when I put one on, an LED lit up. I decided to make the final connections with everything in place and ready to connect to the solar power.
(26th) I took the first battery out to the garage and discovered that the black wire was on "+" instead of "-". This 12 volt set of four was connected mirror image to the others. I un-wired everything and turned it around. As I did that, I suddenly realized there was a flaw in the whole plan: it was fine when it was charging, but when there was no solar power, the top set wouldn't be charged. But the other three were connected to the batteries, and they would use the upper cells to keep charging the lower ones!

   Then I thought of tying all the "Vin+"es together and direct to the charge controller. As switching DC to DC buck converters, they didn't care much what the input voltage was. If it was 42 volts and each one was 10.05 volts, then the top one would be supplied by 42-30.15=11.85 volts, the next one down would be 42-20.10=22.9 volts, the third would be 42-10.05=31.95 volts, and the bottom one would have the full 42 volts.
   That raised a different problem: the "absolute maximum" voltage rating on the XL4016 chip is 40.0 volts. But I didn't see much way around it: the bottom converter would have to be run at a little over its maximum rated voltage. I really should write the manufacturer (of the chip itself) and explain why their maximum rating is just a little low for proper implementation of 36 volt systems. If they could modify the design slightly (or even just guarantee it) for up to 45 volts, it would be a whole lot better.
   But this brought a different problem. Again if the charge was ON everything was fine. But if it was OFF, disconnecting the power, the voltage would go down to zero and the voltage on the "Vin+" of at least the top 2 units (whose "ground" was at 20.1 and 30.15 volts respectively) would be negative. At least the top unit seemed to short its output and short circuit its three cells.
   I considered a number of options with steering diodes, none of which seemed to quite work. A diode to each converter sounds perfect, but the PowMr charge controller gets its own power from the batteries. The diodes would keep it from powering up. Any attempt to bypass the diodes would bring back the original problem.
   Well, it did say it was a risky plan. I'd much rather have four isolated output converters - like the AC to DC power adapters.
   Come to think of it, even that has a conceptual difficulty: with the power adapters, either the chargers are plugged in and the AC power is on, or not. If it's not, obviously they don't charge. With the solar power, the DC solar output voltage is there whether or not there's any power from the solar panels, but they'll still be trying to charge as long as the voltage is there. Batteries can't charge themselves. That's the problem: that the PowMr gets its power from the battery as well as charging it. When there's no solar it's impossible to have power to the controller supplied via its output, and yet somehow have no power from the controller output. What is needed then is a solar charging output that is ON only when there is charging power.
   The MPT-7210A charge controller gets its power from the solar panels instead of from the battery. So it should work. Then the isolation diodes strategy would work.
   Too bad the solar panel voltage is a little too low to charge the top cells. And that if it was much higher, it would blow the DC to DC on the bottom cells. Otherwise one could simply use the DC to DC converters as charge controllers. Hmm... hmm... How about, 3 buck converters for the lower 3 banks, and one boost converter for the top 3 cells? Straight from the solar panels 32-42 volts. (Via a circuit breaker, and all diode isolated, of course. Fuses on the outputs?) Of course, those wouldn't have anything that finds the maximum power point on the panels and adjusts its current draw accordingly. So it is best to have a real solar charge controller.

   I decided I would have to use an MPT-7210A, then I set about wiring things up, with diodes to isolate Vout+ from the battery. There was still another surprise. Even with only Vin- connected, when Vout+ of converter 2 was connected to Vin+ of the lowest converter, that one's LED came on. Apparently even if nothing else is connected, Vout+ will clamp to Vin- by internal diode action, which means it will still provide power to the unit below. So all units still had to be powered direct from the main charge controller, which again must have no voltage if it isn't providing charging current.
   But since Vout+ clamped if taken negative to Vn-, it was a sure bet the Vin+ would as well. So all but the lowest converter would have to be powered through a diode. I got out 3 more car alternator diodes - the last ones I had - and wired up the new power line with 4 branches.
   Finally, this configuration worked! For testing I didn't get out an MPT7210A; I only connected it temporarily to the PowMr while the sun was shining. I had to get the PowMr output up to just about 42 volts to get the top segment to charge. It would seem that if there is insufficient power available, the bottom banks will charge first and the top one only once the voltage is sufficient because the lower ones have charged. Not the most desirable effect, but as long as they can be charged up fully when the sun shines and they are charged, it should work.

4 adjustable DC to DC converters as charge controllers for each 9 volt segment.
Note the "car alternator" isolation diodes (a) on each orange wire at the cell connection,
and (b) on each red power wire connection except the lowest voltage one, at the power adapter.
(& note how I'm standing on the other side of myself to keep my feet out of the way.)

   Here is the final circuit diagram:

 I still think four isolated output DC to DC converter units would be best... where are they? I thought China made everything.

Haida Gwaii, BC Canada