Turquoise Energy Ltd. News #120
covering May 2018 (Posted June 4th)
Lawnhill BC Canada
by Craig Carmichael

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

Feature: Nickel-Nickel Battery: Should be better, cheaper than Lithiums. Next step, developmental production.
(See Month in Brief, Electricity Storage)

Month In Brief (Project Summaries etc.)

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

- Project Reports -

Electric Transport - Electric Hubcap Motor Systems
* Chevy Sprint Car, continued
* Reluctance Motor Designing Idea: Make it Flexible!

Other "Green" Electric Equipment Projects
* Carmichael Mill ("Handheld Bandsaw Alaska Mill")
  - potential competition? Nope! - Various constructions and trials - Cutting (a few) real boards - Superior band guides
* Proposed New Electrical Standards "RFC": a new standard Voltage, and Standard Connectors for 12 VDC, 38 VDC
  - New "38 volt DC" (+/- 15% = 33 to 43.7 volts) proposed as a "standard wiring voltage"
     - midway between 12 volt and 120 volt (3 times 12 volts; 1/3 of 120 volts)
     - 1/3 the current and wire size of 12 V
     - much lower line losses than 12 V, good for moderate power appliances as well as low power
     - highest 'safe to touch' line voltage
  - Revised CAT standard 12 volt plugs and receptacles (a little smaller, better connections and grip)
  - New HAT standard 38 volt plugs and receptacles (size is close to original CAT std., longer and stiffer prongs/blades)
* "The Indoor Vegetable Garden" - Year round gardening with LED Lights! (and other gardening) (Last update for the season?)

Electricity Generation
* HE Ray Energy - notes on magnetic saturation & grounding

Electricity Storage - Turquoise Battery Project (NiMn, NiNi, O2-Ni), etc.
* Nickel-Nickel Batteries - Further cell experiments - Cell construction and test - Water electrolyte cell - More tests
  - High Resistance Electrode can be solved by VERY VERY powerful compaction - NiNi Must Next Move to Developmental Production
  - Negative Current Collector Metal as a metal hydride for double energy storage?

May in Brief

More gratuitous beach pictures having nothing to do with the newsletter
Here's all we can see of the BC mainland from Lawnhill - very occasionally when it's really clear all the way across:
a few mountain peaks somewhere in the direction of Prince Rupert.
(They seemed clearer to my eyes than they do in the picture.
Some around the right hand arrow had glaciers but most of the white at the horizon is just clouds.)
Someone says they're not the coast range, but taller peaks farther inland.
Someone else says only atmospheric refraction would allow them to be seen around the curve of the Earth from this distance.
(It might be interesting to check out the trigonometry on those!)

Just a day or two after I posted the last newsletter saying how the beach was
usually mostly covered with a film of water and reflective, it was mostly dry, perhaps
as the snow had mostly melted in the mountains and the weather improved.
(I also started to realize that it's about the only beach with sand along this coast.
Most of them are just rocky. I got lucky when I bought this place!)

   I started the month working a bit more on the Electric Chevy Sprint car, which was getting to run pretty well but not very fast. (20-25 KmPH - but it looks like you could drive it for about 4 hours on a charge!) I started thinking more seriously about making a new reluctance motor. Back EMF is so low, and the rotor just a solid piece of steel (or steel laminate), that it can be spun up to very high RPMs, and the power just goes up with speed. 10,000 RPM would have the Sprint doing 100 KmPH. And if the "permanent magnet assisted" motor idea works as well as it seems to, the car would also have amazing range for its batteries.

   Then I got back to the bandsaw. It wasn't co-operating and I turned the wheels around so the front was the back & v.v., which moved them forward 3/4" and made more space for the band guides, lack of which had been the cause of some troubles. Now all had to be remounted and repositioned. It was still unsatisfactory. Later I decided to cut thin slots in UHMW PE plastic and mount them as band guides. They worked well at first but wore out quickly. It was worth a try! Then I saw a Woodmizer band mill and how they had done "railway car" band guide wheels. I'll try a variation on that next. I welded and turned a pair of these guide wheels with rims on the back to keep the band from pushing backward. That will eliminate the two troublesome rear wheels. But I only got one mounted by month's end.

13' Boards cut using the UHMW band guides; New "Railway Car" band guide wheels.

Nickel-Nickel Batteries

   Along with the bandmill, from mid month I experimented with new NiNi battery cells and electrodes, and then that became the main focus. Much was learned. The reason I've usually had low capacities is that the substances have been poorly compacted and so the electronic conductivity has been such that most of the substance isn't being used. The extreme pressures required for compacting can doubtless be had much the same way Edison got them: making very small cross section electrodes and feeding in a bit of material at a time from the end, then pressing it ...or walloping it with a sledgehammer or maul. But others have made various nickel hydroxide electrodes since since Edison and there are probably more elegant solutions. If graphite powder could be used as a conductivity additive it wouldn't need such extreme compaction pressure, but graphite causes strong self-discharge with this chemistry.

 - Plastic pill bottle bottom as housing               
- etched cupro-nickel sheet with tab terminal as minus current collector  
- compacted nickel powder and nickel mesh minus electrode        
- separator paper with "mud" electrode on it (extra paper bit to shield tab)
- coated graphite sheet as plus current collector (connection to top surface)
- (All wet with electrolyte, wax seal broken around graphite for disassembly)
   On the 15th I decided to try doing an improved test cell. Using the nickel mesh and nickel flake powder, then compacting them in the press at 10 Mg (megagrams or metric tons), seemed to make very nice negative electrodes. The other electrode wouldn't hold together at all. It stayed as powder until wetted, and then it became mud. Current capacity was pathetic. But chemically it worked! And with the oxalate being a great chelating agent as well as the electrolyte, there would seem to be no chemical reason it wouldn't last forever - unlimited charge-discharge cycles. What are the unique features?

* More mildly alkaline electrolyte (pH 12-13) using potassium oxalate and calcium hydroxide instead of potassium hydroxide. Besides hydroxide, oxalate is an ion that nickel and many metals are insoluble in. Finding some such salt with this rare property was a key.

* The electrolyte is dissolved in ethaline DES [Deep Eutectic Solvent] instead of water. Ethaline has high overvoltages before it starts breaking down. (DESes do however have lower current capacity - slower ion flow than water. The chemistry might or might not be coaxed to work in water, which has a lower oxygen breakdown voltage.)

* The more mildly alkaline electrolyte permits nickel to be used as a negative electrode. Nickel has more available amp-hours per kilogram (as opposed to theoretical values) than any other metal and is highly conductive.

* The positive electrode - what I'm using in the present tests - is a unique formulation of oxidized monel powder (nickel:copper alloy), nickel hydroxide, lanthanum hydroxide, and thiamin (from canned beans). The monel with the copper makes the electrode more conductive. Other formulations better than plain nickel hydroxide are certainly possible and are used by others, notably nickel with manganese, which forms mixed nickel manganates, of lower resistance and taking on various oxidation states in charge and discharge.

* For the positive electrode current collector, a thin conductive film of acetal ester doped with osmium (powder) coats a graphite current collector sheet. This prevents contact between the graphite and the electrolyte. In a less alkaline solution, the reaction voltage of the positive electrode is higher, and graphite will react to cause oxygen separation and self discharge. Every metal will also corrode away to oxide. This problem, solved by the osmium doped film, is what has kept this sort of battery chemistry from being created previously.


   Wow did it ever take me a ridiculously long time to figure all that out and put everything together at the same time into one cell! I had most of it by 2011 or 2012. Except for the ethaline DES [Thanks Leonardo Janus for bringing that to my attention!] - and thinking graphite simply couldn't be the big self-discharge problem - I might have had good cells several years ago, and I should have tried out the DES about two years ago. Going from my early and original oxalic acid idea to potassium oxalate was a pretty big leap too, but became a relatively obvious thing to do once I finally started experimenting with it. I presume that my 2.6 volt nickel-manganese cells should work great with similar formulations and electrolyte, too!)

   An important aspect of production will be compaction of the positive electrode powder. It seems it requires 16-20 tons per square centimeter - and that may be less than optimum. Other battery makers have solved this same problem in various ways. It would seem one way to get such pressures, needed only for an instant, is to drop heavy weights on a small area electrode compactor die rather than to try to use a press, which would need to be enormously heavy. I used a 6 pound maul to bash it with. It wasn't enough and I need to pound on a smaller surface area. (I may have a solution but we're into June and I have to end this newsletter.)
   But it does seem like this battery is ready to move at least to some sort of early production level. Then real uniform, manufactured cells can be tested and proportions, mixtures, techniques and construction can be tweaked up to maximum performance.

   Another interesting prospect is that typical metal hydrides are composed largely of nickel. Once all the nickel has charged back to metal chemically, if the alloy is specially formulated, it is possible that the nickel electrode could also become a metal hydride at just .1 volts more negative charge than the nickel redox reaction, and thus hold a whole further charge of hydrogen ions (protons) giving it double capacity at almost the same voltage. As the battery was used, first the hydride would discharge, then the nickel would oxidize. Whether a practical high capacity hydride can be formed on top of having the nickel chemical reactions however is a speculative idea, not a proven concept. It can be tried out once sealed cells are being produced - it can't be tried in unsealed test cells.

Nissan Leaf versus GM EV1 - another invidious(?) comparison

   A month or two ago I had wondered why the Nissan Leaf seemed to use substantially more energy per kilometer than was reported as being used by the GM EV1 two decades ago. The EV1 got more driving range from fairly similar battery energy - 160 miles (250 Km) versus 160 kilometers. Certainly for highway driving wind resistance is a factor, and the EV1 had the lowest wind drag of any production car ever, while one can feel it's fairly substantial in the Leaf at higher speeds and a headwind definitely seems to cause it to use more electricity per kilometer. (it doesn't get blown around by crosswinds the way the Toyota Echo does, though.)
   Somehow it had never occurred to me to wonder what they weighed. While the weight of the Leaf wasn't given in the manual, there was a "GVWR" label on the door panel. Subtracting "860 pounds maximum passengers + cargo" from the sticker's 4193 pounds makes the curb weight 3333 pounds. The EV1 was 2850 pounds, so 500 pounds less in spite of its somewhat heavier nickel-metal hydride batteries. So rear seats (the EV1 only had two seats), larger space overall and plushness come with a price. It's in electricity and range. If used Leafs weren't 1/2 the price of any other readily available production electric car, speaking only for myself I'd rather be able to travel to anywhere on this island and home again. Sometimes the extra space is useful; again just for me, rarely the extra seats. But if I still lived in Victoria where distances are smaller, the Leaf would be perfect except for rare longer highway trips.

   The Sprint's 1800 pounds or so (though with less than 1/2 the battery capacity) certainly makes it a flyweight. But the Sprint with the small forklift motor and fixed 8.9:1 reduction doesn't go fast enough for the street much less the highway. I took one more leisurely drive up Lawnhill Road in it at 20 KmPH (where I found and examined the Woodmizer band mill), and then transferred the insurance back to the Toyota Echo and insured my trailer as well, so that I could both reach Masset and carry plywood (etc) when I want to.
   Before I could use the Echo, however, I had to fix the sticking brake piston on the front right wheel. After I did that I took it for a test drive and discovered the the front left was also sticking. All that sand and salt on the highway last winter was certainly hard on the car! I spent the 27th fixing that, and since I just couldn't leave it with 3 wheels done, I replaced the brake shoes on the fourth, the rear right, as well, rather prematurely. (The front brake pads had lots left. Just those costly cylinders had to be replaced - about 260$ just for the parts to fix them myself!)

   On June 1st I drove the Echo up to Masset. En route I visited some people who milled lumber and who were very interested in the mill. They had read about the idea in my December article in the Haida Gwaii Trader magazine. Their blade mill had a 3/8" kerf so they were turning a lot of wood into sawdust to cut boards. So! Even people who already have sawmills are potential customers for the handheld band mill! (But they were out of the cheap salvaged 3/4" plywood I was looking for. Rats!)

   In Masset I met up with Lawrence of Driftech Mechanical who had built an electric tricycle I had heard of earlier. He was closing up shop for the day and we went to his house and he showed me the tricycle and demoed it. It was a far more impressive machine than I had expected from the name "tricycle", one that could burn rubber and was well capable of fantastic acceleration to highway speeds and beyond. It would 'blow away' gas cars. It had a Volkswagen manual transmission and a 48 volt forklift motor (longer than mine in the Sprint and probably 80 pounds instead of 50), being run at 72 volts with 6 lead-acid batteries that gave it about a 20 Km range.

He reached the far end of the block in a flash and was turning around before I could even walk a few yards to the road to get a picture, and I didn't get a good one. (A video would be better. I didn't even think of that!)

   In getting ready to finish this newsletter I copied the month's pictures onto a USB memory stick. All seemed in order. Just as I was about to delete them off the phone (AKA the camera), my brother phoned my cell phone number, by accident. He said his phone was in his pocket until he heard me say "Hello". I lost the USB connection to the computer when I answered so I didn't delete the photos. This proved to be fortunate as they hadn't transferred properly. Over half the pictures either weren't there or had internal errors in their data and were missing major sections. There were no indications anything was wrong until I tried to view them. Our guardian angels work in creative ways!
   I copied them again from the phone onto another USB memory stick... with pretty much the same results. For whatever reason, I had to copy a few at a time instead of all at once.

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

Fish Freezing Tip

   Fish meat in the freezer gets "freezer burned" quickly. It's been said that one shouldn't try to keep frozen fish longer than about 3 months. But when I was a teenager our family froze a whole fish in an ice cream pail full of water. About a year later we dug it out and ate it. It was still perfectly good and fresh tasting.
   Here is some halibut I was given. I cut it into steaks and separated them with some plastic bags so I can (hopefully) separate them out of the container of water (ice) when I want to eat one. I confess it might be hard to get them apart. But they'll last a long time in the freezer!


You know you're out in the boonies when you go into a cafe called "Angela's", and it's not a chain; it's owned and run by a woman named "Angela".

Emergency Instructions:

Yellow alert: Get the heck out! Step on it!

Red alert:      Stop and wait for it to turn green again.

   "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 thought of... 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 and consistency 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.

Electric Transport

Chevy Sprint Car - Forklift Motor & Fixed (8.9:1) Reduction

Misc. Improvements

   On the 5th I put the rear view mirror back on. The tow truck driver had sheared it off on a tree when he pulled it out of my yard when I was moving. I glued the 3 plastic screw sockets back on with methylene chloride, reinforcing each one with an extra piece of ABS plastic, also stuck on with methylene chloride. (It was lucky the mirror's plastic was amenable to softening/melting with methylene chloride.) Then I tapped the metal on the door, bent or ripped by the screws, back to its original shape and screwed the mirror back on.

   As the car seemed to be running well I got up the nerve to use the programmer to raise the motor current limit to 250 amps instead of 200. I drove across the acreage once, and then went on the highway to the acreage next door. It was a considerable difference. It got up to over 25 KmPH instead of 20. This time it sounded like the motor was winding up some and it should have been about 2500 RPM. I drove through the two-driveway yard to turn around. But on the way back it didn't go much over 20 because the slope was a bit up instead of down. From that speed it went up my driveway hill with seemingly good margin, unlike the first trip when it barely made it back up. It's still nothing like the speed and power one wants on the street, let alone on the highway. So far it's a relief just to go out and get back without getting stuck out on the road somewhere. Should I try the full 300 amps? Hmm!
   I neglected to bring a current clamp meter for the battery current and I didn't check the motor current. (Not a very methodical experimenter, am I? But then, one can't be looking down at meters much while driving. I totaled my very nice white Tercel wagon (bought new in 1986) in 1993 and suffered substantial injuries by looking down for a moment too long at the wrong time. PS: Seatbelts save lives!)

   But the next day, the 6th, I upped the motor current limit to the full 300 amps and tried the same paths (twice to get more meter readings), reading the battery current clamp-on meter. It still didn't get up to 30 on the highway - somewhere over 25, and a little faster but still under 25 on the way back. Really no better than when limited to 250 amps. (If it was running well, the Electric Hubcap motor would have at least got it up to 30!) Battery currents were up to 130 to 180 amps in acceleration or uphill, with just one reading somewhere well over 200. But on the highway at "top" speeds they were down to about 53 amps going (=2.7 HP, down slope) and 83 amps on the way back (=4.2 HP, up slope). That meant that the motor simply didn't want to rev up any higher with a 36 volt supply. (At least then there was no worry then about it over-revving and flying apart!)
   With 3 sets of cells in parallel, the Sprint would be easier on the batteries than the Swift was. 180 amps (6840 watts @ 38 volts, 9.1 HP) even if it was continuous is only 60 amps from each battery cell, where the Swift could easily draw 120 amps continuous from its one set. For the same 10000 watts (100 amps @ 100 volts) that the Swift used to drive along the highway, the Sprint would only draw 265 amps (265 amps @ 38 volts), or 85 amps per cell instead of 100. But I think that the Sprint would use less power with the less lossy transmission... if it could attain highway speeds. (Sorry about the confusingly similar car names, "Swift" and "Sprint". I didn't name them.)

   If one took the median power, about 70 amps, what might the range be? To leave a very minimum reserve one might use 250 amp-hours out of the 300  available. (supposedly available - these aren't new batteries) 250/70=3.57 hours of driving. At 25 KmPH, that would be about 90 Km, using just over 100 WH/Km. That's not very impressive given the low speed. A better motor would seem to be in order.

   If I had used the 4 to 1 transmission reduction I first thought about instead of the 8.9 to 1, top speed would have been about 55 to 60 - but only if it would start moving, which it might not if facing uphill, starting from a pothole, on soggy lawn, etc. To go faster with this motor and controller the car would need that variable transmission or multiple gears to keep the motor speed down at higher travel speeds... or a higher motor voltage.
   Going uphill on a last leisurely drive up Lawnhill road before transferring the license and insurance back to the Echo seemed to show it probably wouldn't maintain highway speeds going up grades, regardless of gearing, with the 36 volt supply.

   There are several problems with higher voltage. First the motor controller is 36 volts. I'd have to buy another one for 48 volts. Second, that would still only take it up to 35 KmPH - nothing like highway speeds. Third, would it be over-revving? Probably it would be fine at 48 volts/35 KmPH, but it was after all supposed to be a 36 volt motor, and I hadn't done it any favors busting the fan rim trying to take it apart. On top of those things, fitting the extra batteries would be a nuisance, fitting four solar panels would be even harder, and I had just bought the 36 volt to 120 VAC inverter to run from the car in case of power failures. Plus I'd rather stay at 36-40 volts than up it to 48-54. 72 volts would take it to 50 KmPH, but that's getting quite dangerous and the motor really might break up at 5000 RPM. I had grounded the system to the car frame, assuming the highest voltage wouldn't be hazardous.

   I'm pleased that it runs and seems reliable, but when I couldn't get it going any faster, I rather lost enthusiasm for the project as presently configured. As from the very start, it needs a variable transmission or a faster motor. (I also wanted to get the bandsaw mill working.)

Reluctance Motor: What's Available?

   I thought that perhaps there might be a higher RPM forklift or other suitable sep-ex motor available. But even if it was rated 6000 RPM - a very high RPM for a larger size motor - that wouldn't be highway speeds. 10000 RPM continuous would be a good rating to shoot for. But at that speed there would be 16 times as much centrifugal force on the rotor as at 2500 RPM.
   It all makes me think more about reluctance motors with their very low back EMF and solid steel (or steel laminates) rotor, that can do such high RPMs safely and efficiently. (And surely it would be even better if it was PM assisted.) Since power is torque times speed and the torque is still high at high RPMs, quite a small motor should have the power for a car if it can be cooled adequately. The rotor makes no heat, so only the stator coils need cooling. An adequately cooled reluctance motor should give the desired 9200 RPM for 100 KmPH. If a motor was also permanent magnet assisted the currents should be substantially reduced, reducing the current and heating loads, and hence the motor size - or at least the cooling requirements - could be further reduced.

   Then again, it's been 3 years since I was doing the reluctance motors. Has someone else invented this particular wheel since I had been working on it? Times do change. Time for a web search!
   A web search on the 7th came up mostly with studies and theory about reluctance motors. There were a couple of line voltage reluctance motors available, 1800 and 3300 watts at 120 and 240 volts. They came with their own motor controllers. They only went to lower RPMs - in the 3000s range. There was nothing that looked applicable to power transport. And I saw nothing about PM assisted reluctance motors. The field would appear to be still wide open.
   Once again, with development funds much could be done probably resulting in a commercial transport motor and controller relatively quickly. With one person part time one might be getting good results just as commercial products started to emerge elsewhere. It doesn't matter much at that point whether yours is better or not; the main market is passing by.

DC to DC Converter Idea Again

   Setting the HE rays idea aside, I started thinking that if the dual inductor (or other such device) could be used as an effective pulse transformer, it could go the other way: the 20 turn coil could be the input and the 5 turn coil the output. That should reduce the voltage to 1/4 and multiply the current by 4 times. That would have the same effect as putting 4 times as many winds in the motor coils, without changing the motor. 15 amps from the controller's field drive would become 60 amps into the motor field - enough to at least get the car to move. Of course, that would be if it worked as an effective transformer with high currents. I wasn't seeing that when going the other way, from the 5 to the 20 turn coils. The 120 volts DC output built up only gradually, and it died the instant the light bulb was turned on.
   That doesn't necessarily mean the idea wouldn't work with the right transformer. And diodes that would take the current - perhaps low voltage alternator diodes.

Reluctance Motor Designing Idea?: Make it Flexible!

   In my 2015 reluctance motor experiments I cut or had cut rotor shapes to suit my ideas at the time. Now I think, why not just use a relatively thin solid rotor disk like the ones from Princess Auto ("7.8 inch Brake Disk with 1 inch shaft") and bolt pieces of steel to it to try out various shapes and sizes of "salient poles"? They would just have to have bolt holes to line up with the bolt holes in the disk. That should make it easier to try different things out to discover the most optimum patterns.

Other "Green" Electric Equipment Projects

Carmichael Mill ("Bandsaw Alaska Mill")

Handheld Bandsaw: potential competition? Nope!

   I checked out "handheld bandsaw", which device someone had mentioned. Had I reinvented a wheel that was already being made and sold commercially?
   But they proved to be an entirely different thing from a lumber mill. The band twisted around 90° so it cut downward instead of horizontally with the same saw orientation. They could only cut less than 5" wide by 5" deep before the work would hit the frame. A demo I saw showed one cutting only unspecified metal bars and tubes, not wood. (If the pieces were steel, it was far more impressive than if they were aluminum. What a crappy video not to say, when they were blabbing throughout, jokes and trivia!) If it was used for cutting wood, it couldn't do much more than crosscut 4"x4"s to length. But they cut freehand, so one would surely get straighter cuts from a skillsaw. The blade bands were only about 35" in length and they said theirs was 17 pounds.
   It could be a very nice metalworking tool to have - potentially better cuts than an angle grinder with a zip disk, without the expense and bulk of a typical metal bandsaw. But it was nothing like the "Carmichael Mill" in construction, purpose or potential use.

Band Guide

   I got back to the mill on the evening of the 8th and got part of the other adjustable band guide done. I noticed that one "ski" had been crooked all along because a hole wasn't in quite the right place. I filed it out a bit so the left assembly could be straight. I finished it on the 9th.


   Then I tried out the saw but adjustments were "out of whack". The bearing holders being bolted into slots they can be moved as desired, but of course each thing that can be moved becomes an adjustment point that has to be carefully aligned. A production saw would eliminate most of the adjustments with fixed position parts, cut and drilled exactly by CNC so everything would fit perfectly without adjustments. Only the vital adjustments - the band guides, band tension (allowing band replacement) and perhaps one wheel tilt - would be adjustable. As I was trying to adjust things while attempting to cut a piece of wood, it started to rain. That ended that!

   I returned to the chase the next day (10th). When everything seemed to be in order I set a fatter beam on the sawhorses and brought out the saw. I had already tried cutting this beam from both ends with the cuts going off, so the saw would have to make its own new path without simply trying to follow another. In adjusting I found that the band in the wood could be seen to have veered down or up by looking at the band guide on each side. I had just a little trouble with the right side but plenty with the left. The spring didn't seem to be strong enough to force the guides to move to cut more downward. I soon went to the spring drawer and found one that I could cut a suitable piece off of. It fit just inside the other and I worked them both into position and got the the adjustment bolt back through inside them.
   The direction of adjustment seemed counterintuitive and had to be thought about. The pivot hinge was behind the bearings, so if the blade went down below, lowering the aim brought the guides closer to the cut. But lowering the aim meant the cut would veer downward even faster. It had to be aimed up so further cutting would bring the band back up to where the guide wheels were. And then, which way did one turn the bolt to achieve that? I'm sure I got it wrong more than once, and I had to back up and cut another slot that went the right way. Since the adjustments weren't very close to start with, it took a while and a few tries to get it going well. I stopped and checked every few inches or a foot.
   It was easy to tell when it was running well: the motor and band turned more freely, and the saw bounded ahead into the wood. When it was going off, it got harder and harder to cut, and the band would get hot. (Sometimes, if the band didn't have enough tension, it would cut a catenary, the shape of a wire between two telephone poles, instead of a flat board. This time I seemed to have it tense enough to avoid that... for 6" wide. What about 16"? Well, everything in its own time!) At one point, the band came off the wheels. I thought I'd have to stop there because that's hard enough to adjust even in the shop on the bench. But I persevered, tightening and loosening pillow block bearing bolts and pushing the bearing blocks around with a screwdriver, until finally the band stayed on when the saw was running. Eventually, adjusting the guides quite a lot for a while but gradually getting them better aligned as I went along the board, I came to the far end. I had cut a real 12.5' spruce board, 1.5" x 6", usable for coarse purposes or if planed down thinner until it was flat. Now, if only the next one would go much faster!
   The top of the beam was too hilly to use as a flat surface for the next cut. I looked around for a 1" x 6" to put on top but I didn't see anything suitable. So that was enough for the day! After I put everything away I looked for sawdust. If I had used a chainsaw mill I could have raked up a small mound of chips, I'm sure. But there was just a bit of fine dust in the gravel. I didn't even see it at first.
   On the 11th I got it out again and cut a couple more boards. I was starting to get it better adjusted and by the second one it cut more or less straight rather than digging down or rising up. But there was still a bit of catenary and it still went up and down a bit here and there. These might be improved by more blade tension. And even when cutting well the band was still getting hot. I could have cut much faster if I hadn't kept stopping to let it cool. (Once the temper is gone from the teeth, any blade is toast.) Pushing the bearing blocks with a screwdriver seemed like a poor way to try to tension it both because it was hard to get good tension and because it had to be done on both sides of a wheel, making it hard to keep the wheel aligned.

Now What?

   The next items to tackle for improvement were:

* Get the right rear band guide back on (it didn't seem to fit right with the new adjustable top-bottom wheels)
* water drip to cool the blade
* some sort of screw system to tension and align one of the wheels.
* Front and back board glides instead of the left and right skis.

Imperfect Wheels

   I thought about how to do band tensioners on the 13th. Suddenly it occurred to me to try something. Sure enough, when I turned the wheels to different points of rotation, the band got looser and tighter. For all my efforts, the plywood wheels weren't entirely even and concentric all the way around. That doubtless explained some of the unpredictability of results. It would to be hard to fix with confidence without making new wheels, preferably of cast aluminum alloy, turned to machine precision.
   I thought again of the wheels on the meat cutting bandsaw. They were very different from other bandsaw wheels. A rim at the back stopped the bands from going too far back, and the relatively flat outer rim was just slightly too narrow: the tips of the teeth were always just off the front, touching nothing.  I thought this design might well be the best. If they had been 10" instead of 8" I think I'd have pulled them out and tried to use them in the mill. They were cast from steel rather than aluminum. I wondered if I could make anything like that? Given my lack of success at casting an aluminum boat propeller blade a few years ago, I decided that casting them was out. And then, as is so often the case, they couldn't quite have fit onto my lathe to machine down.
   What about just trimming the present wheels nearer to perfection? Perhaps right on the saw, with a file or some sort of scraper? For the powered one, just remove the band and turn on the motor? The other one might be trickier. I took off the band. (I managed to run my hand into the blade a couple of times and cut a knuckle a bit - what else is new?) I stuck a chisel next to the rims and turned, shaving off bits of wood. It seemed to be mostly just the driving wheel was a little off center. In other words, the center of the wheel on the saw was somehow slightly different than it had been on the lathe when I made it. It was off enough to explain it all. I had sharpened along one edge of this chisel as well as the tip, and I used this edge to scrape the high area down as I turned, while resting the sharp end of the chisel on the saw body as a lever so it wasn't "freehand". At first I turned the pulley by hand, but that was too tedious and I used the motor.

   Again, I started thinking that with a bigger lathe I could have completely assembled the wheels on their axles and turned them true, hopefully preventing this later misalignment. I've started to realize that many things I've done - or wanted to do but couldn't - would have been simplified or made possible if I had a bigger diameter lathe. So many rotors and things have been 10". The lathe has a 10.0" gap, which is at best minutely too small to turn any item down to 10". And it's hard to get a 10" piece mounted and on even if it doesn't hit the frame. If it was 11" instead of 10 and the 'gap' area of that diameter was about 3" longer, it would make a tremendous difference. It's such a small amount - maybe I should take the angle grinder to the body of my lathe and trim it down a bit!

   When I was 'done' I wasn't happy with the alignment - the band was riding too far forward on the wheels. But I took it out and tried to cut into a new cant. I got a few feet in before it went out of line. There still didn't seem to be much tension, either, and I suspected that was another reason it wouldn't stay in line. Back in the shop the tension still seemed to depend on the position - ugh! Well, maybe I'll try making the tensioners and it can be sufficiently tense everywhere? And maybe I'll take the wheels apart and see if there's any way to make sure they run more centered.
Band Tensioners and Aligners:
Loosen the bearings a bit and screw both end bolts to press against them.
Then tighten the bearings up again.
   On the 14th I made the tensioners. I had been wondering how, but on this morning I simply walked up to a pail of scrap aluminum and picked up a 1/2" thick by 1" wide bar. (Gosh, I had a piece that thick?) I cut off two pieces 2.5" long. I drilled and tapped one end, off center (5/16" - 3/16" in 1/2" thick), for a 1/4" adjustment bolt. Then I drilled two holes for 5/16" carriage bolts to mount the pieces on the saw "backbone". The pieces were mounted solidly in place with the head of the 1/4" bolt on the end touching the bearing holders for the undriven wheel. With the bolts on the bearing holders loosened, unscrewing the bolt pushed the bearing holders along the carriage. This arrangement was able to put considerable pressure on the bearings and wheel, which could be aligned so the band ran as desired and tightened to provide substantial band tension.

One of the (now two) adjustable band guides

  The adjustable band guides stuck out farther than the originals, and that was part of the trouble getting things into proper alignment. I had had to leave off one of the bearings that stopped the band from pushing backward because there just wasn't room to put it on. Now I realized that if I turned the wheels around, the way they were mounted on the shafts, they - and the band - would sit about 3/4" farther toward the front of the saw. That would leave more room to properly position the band guides. So I realigned the band to run in the new position and (after much adjusting, with better tension) it seemed much better.
   Then all the band guides needed moving and adjusting. The rear stop bearings needed to be mounted more solidly as the one still there had been bending out of place. The mounts needed modified shapes for the new band position, and to be made from thicker aluminum. I did that on the 18th and tried it out. It just wouldn't cut straight. I couldn't make one decent cut in the cant. No adjustments seemed to help. I took it back in the shop and did some more adjusting and setting up, but that was all the time I had. Surely the blade couldn't be dull after cutting only a few boards. Had it hit some rock in this last piece of wood? Again, if that was common, the whole idea started to seem a little impractical. I was starting to get pretty discouraged.

   The next morning (19th) I decided to give it one more try, and then try a new cutting band/blade. It occurred to me that it had come off a rear wheel more than once, and that the two rollers would have pinched the teeth and reduced or taken out the set - the way teeth stick out to each side a bit so the cut is wider than the blade behind. That might explain the poor cutting. But then I noticed that the "skeg" or "side ski" that the saw buts up against as the blade pulls it sideways had a mark on it. The blade had also more than once come off the rollers, and it had hit the skeg. A steel blade cutting steel?: that all by itself explained why the blade was dull and the saw wouldn't cut properly! So I put on a new blade. And I took off a bit of the skeg with the angle grinder so it wasn't so close to the blade, and cut at a bit of an angle so hopefully (no guarantees) if it did hit it again, the back of the band would hit instead of the teeth. These of course are things one learns making a prototype that is actually used, so they will be built into the production model to eliminate potential trouble points before there there are hundreds or thousands of the mills in use.
   I tried it again and went about 8 feet though the 12 foot cant. It seemed to be going nicely. I was using about a 9 foot 2"x6" as a flat top board and I had to stop and move it to the far end. After that the saw wanted to bind up for no apparent reason. But I got there and when it came out the far end the band sprang forward and came off the wheels. I could see that it had more and more of a "cup" in the board toward the end after a big knot or after moving the 2"x6", both at the 8 foot mark. I decided that next time I would cut something down so I was milling 6 or 8 feet - I just might make a few decent boards instead of one cut that never finished well. Aside from the mill itself, the wood had some ugly dark spalting in it and I really had to start getting it cut up - better I had already done it! Valuable lumber is going to waste.
   (20th) I tried cutting again but one could see by looking in the end of the cut that within inches it was starting to bow and it just got worse. The best reasons I could think of for the were (a) that one side of the band had been dulled and (b) that if the band was pressed hard against the back guide bearings on each side, it could bow a bit in the middle and then it would naturally cut up or down in the middle even while it was cutting straight at the edges; hence the "cupped" shape. From experience so far, if it's cutting well it isn't pressed against the back bearings. The thing that presses the band against the back bearings is if it's not cutting well. If that's happening right from the start of the cut, it probably means it's dull.
   Which meant I had one new band left to try out and see if I could get good results. But first, I really wanted to change the "skis" to front and rear cross-slides like chainsaw mills use. The "skis" weren't keeping it very steady. I also moved the left side out a couple of inches to 8" width first so I could use a 2"x8" for a top board and second so I could see if the cut a couple of inches away from the guides was straight rather than rising or lowering.

   Some thoughts: The more I fiddle with adjustments on this, the more I think I like the meat cutting saw's exact shaped, "no drifting backward or forward" rimmed wheels. Surely setting up the guides and everything should be much simplified. (Of course... I haven't tried milling wood with it. Maybe I should.)
   On the topic, when ripping wood with a regular bandsaw, when the cut starts to go off a bit, one simply moves the front of the board left or right, thus re-aiming the angle of the wood going into the blade. Simple! But that's for thinner boards and it isn't an option with any band mill. There the band has to cut perfectly in line with the direction the saw is moving, so the alignment is much more exacting. And being above a wide horizontal cut, one readily can't see if the direction is going off soon enough to make such tiny corrections anyway.
   People around here say that band mills are always finicky and that those who "get serious" about milling lumber soon ditch them and get a dimensional (2 circular blade) mill. Those "serious" people aren't the target market here. Like the Alaska mill I think this could be a fabulous product for milling "in situ" and not having to move big logs, and to provide a really economical way to make smaller amounts of lumber from a tree or a few small or special trees. If I can perfect it to that point...
   Then I found that even a family with a dimensional mill was interested in my mill because it wasted so much less wood than their thick kerf blades.

It's all in the Band Guides! UHMW Plastic Band Guide Blocks

   I thought about it overnight and by morning (21st) I had a new idea: ditch the complexity of the rolling guide bearings and their mounts, and make band guide slots in solid blocks of slippery UHMW polyethylene plastic. With the cooling water system (until then just stopping often), they should stay cool enough. The sides probably wouldn't wear out in any hurry. If the back of the band dug through the plastic until the teeth hit the guides, they would simply cut the edges of the plastic slot and not be dulled. In fact, I could extend the fronts of the guides to the point where the band couldn't come out of the slots until it also fell off the wheels. That would keep it from hitting the metal skeg or anything. But if the back of the band did tend to dig in and wear out the plastic, I could put in a metal pin. If it was slightly loose, it might even spin in the slippery plastic. Or it could be shaped specifically to do so with a narrow "axle" at each end. The back guides would be in line with the side guides - much better - and the band couldn't go over the edge of the rear guides and slide back (and be forced crooked), the way it sometimes did with the rear bearings.
   Instead of hinge pin plates for adjustment, I would simply make the back a shallow arc or a very broad triangle. To aim the band up, loosen the lower mounting bolt and tighten the upper one to shift the angle, and vise versa. (No springs to compress themselves and throw off the aim!)

   I made these guides that morning from 1/2" thick UHMW. They certainly simplified the saw. I looked at the "skis" and how to change them, and one more time said Ugh! and left them. I set the guides 8" apart instead of 6" and used a 2"x8" as the guide board. I had bucked my 12' cant at the 5'/7' mark where a big knot was. It would probably break there anyway. (It did - it just fell apart.) I started at the short end. The 5' cut went great - smooth and easy, the saw gliding through. So I decided to cut the rest. I started having adjustment problems, mild at first but they didn't seem to correct by adjusting. When I was 3/4 done, I noticed that the saw band was overlapping the back of the wheels, almost ready to come off. It had started about in the middle and the slots as cut wouldn't let it go back farther. Not surprisingly the back of the band had gradually deepened the slots. The teeth wouldn't be damaged, but the slots were widened out by the teeth and once the band was back in the middle they weren't guiding it much. So it definitely needed some metal at the back, and now new slots. I finished the cut. The last half of the board was cupped but not as badly as some.

   Another item that's been bothering me, maybe related to the cupping, is that with the hinge behind the guides, when one adjusts the guides they not only change angle, their position moves up or down. That automatically takes them even further out of alignment with wherever the band is positioned in the wood. I now occurs to me that if the hinge pin is above (or below) the front of the guide slot instead of well behind it, the slot will move back and forth a bit, but it will stay at its same level. I bet that would be a great advantage and reduce the 'cupping'. And now that my guide is plastic, I can easily drill a hole and put a hinge pin through it. (As for a piece to hold that hinge pin and a screw to do the adjusting... hmm!)
   But then again, probably the best thing to reduce the cupping is more band tension. I'm not sure how far to go on that with my bolted-on plywood wheels, but I expect higher tension is the main remedy.

   On the 24th I just rigged up the same UHMW band guides to work again. On one I put in a 1/2" spacer to get the thin part of the slot to the blade; on the other I turned it upside down and cut a new slot. On both I drilled and tapped holes and put in 1/4" bolts as "pins" at the rear to stop the back of the band from digging into the plastic and lengthening the slots. After a bit of cutting I put a couple of drops of oil on the band where it went into the pins, seeing how it was metal rubbing on metal.

UHMW Slot Band Guides on Right and Left side of the saw.
Cutting was quite straight once aligned, until the band wore the slots wider.

   After mashing a finger trying to reposition it (now a partly black nail to match the other hand),  I cut into the same heavy cant I was cutting before. For a couple of feet things went well, but then it started veering up. This time I was listening for the motor to sound more labored and it didn't get far. I adjusted it down, backed up 6", and tried again. It cut a new slot and again went well for a couple more feet. I adjusted it down again and this time it continued to cut pretty well. By the end of the 13' board it was bowed up just a bit, but it wasn't a bad board. A couple of times the motor sounded more labored but I couldn't see anything wrong. It turned out to be big, hard knots in the wood.
   I went to cut another one but the band fell off the front of the wheels before it started cutting. On examination, damp sawdust had built up on the plywood wheels where the teeth were, and that made the wheels wider there. If it's not cutting, the band centers on the widest point on the wheels, so it moved forward. Another lesson learned: for a production model the wheels had to have slots in them for the sawdust to fall out - either radial or axial. or something... or maybe the usual urethane 'tire' on the bandsaw wheels would shed sawdust. (even damp sawdust?) One would think the thin guide slot would scrape it off the band before it reached the wheel, and it probably did get some of it. For the prototype with plywood wheels it seemed I would simply have to scrape the sawdust off the wheels after each board.
   I put the band back on and cut another board. This one went pretty well with only one small adjustment, still on the leading side and still in the same direction, more downward. There were shallow dips or bulges where hard knots were, but much less cupping and otherwise not bad boards.
   I cut a third board with no further adjusting except the height to make it thicker. But on the fourth board, the cut didn't start well. On examination, the band guide slots had worn somewhat wider and weren't guiding the band very well. It was to be expected I suppose, but if they had lasted even a day's cutting that might have been good enough. The saw cut the best while they lasted - the best results so far. It was worth the try.

Simplifying the Cooling Water Plan

   Something I noticed when it was cutting well was that the band didn't get very hot. Instead of having a continuous flow of water squirting on it, I started thinking more in terms of just having a wet sponge contacting the smooth part of the band at the top of its travel. That would probably keep it pretty much cold. It might be a large sponge, or a small reservoir on top of a small sponge feeding water to it, with maybe a funnel to fill it. That simplifies the construction and the operation (no valve to keep turning turn on and off!) and the amount of water needed for that won't add any weight to speak of to the saw. Operation would probably consist of making sure the reservoir wasn't empty.

Woodmizer Mill Type Band Guides

   My previous rollers in line above and below didn't seem very good because they sometimes pinched the blade, if adjusted too tightly or if a bit of sawdust went through. Then it didn't move freely. But if not quite tight together, the band could aim itself up or down quite notably - partly because the wheels weren't very wide. At one point I had to mount two rollers offset a bit because the holes weren't quite right. Rollers above and below that weren't quite in line seemed better. They could even bend the band a bit and it would only hold it straighter. But it still was a less than perfect grip. The bearings were still narrow.

   Backing up a day, on the 23rd I had driven the Sprint up Lawnhill road and there I chanced to see near the road, and of course examined, a "Woodmizer" bandsaw mill. The band was heavy and wide, much thicker than any of mine. The band guides were each a metal(?) block below and a wheel above. The wheels could be adjusted for gap to the blocks (blade thickness), or the entire assembly could be pivoted with adjustments at the back. The 2"(?) diameter wheels were the width of the blade minus the teeth and had "rail car" wheel rims on the back to stop the band from going backward. That way it didn't need other wheels farther from the wood as back stops. That would be better, and also similar to but better than the meat cutting bandsaw with rear rims on the main wheels, since the guide wheels are closer to the cutting.
   So... How about that design? Rollers the full width of the band behind the teeth with projecting rims as the "backstop" made a lot of sense.  It's probably the best way to do it. I still thought slick UHMW bottom blocks on springs would be the way to go for the blocks. (But maybe the other way up, with the wheels on the bottom and sprung blocks on top.) Used that way the plastic pieces could do a lot of cutting before being worn out, and the guides would present little friction to motion. And there wouldn't be any manual adjustment for blade thickness.
   Two bearings with a washer between them was the width of the blade less the teeth. A large washer behind that followed by a third bearing might make the "rail car" protruding rim. But it turned out there was a problem. If the inside of the large washer was the diameter of the axle (a 5/16" bolt) it wouldn't turn with the bearings. If it had a larger center so it contacted the outer turning parts instead, the nut against the unsupported centers of the bearings caused them to jam. I didn't fancy trying to get two spacers of such exact thickness that they would work, even if that was possible.

   It appeared it would be necessary to turn a housing with a rim to fit over two bearings as the shaped wheel, rather than using the 'naked' bearings as a wheel. I tackled these on the 26th - an afternoon and part of an evening of buying a pipe 'coupler' fitting that looked like it was the right size, then cutting it in half, turning off the hex grip part and welding washers onto one side, then more machining on the lathe to fit in the bearings et al. On one side of each wheel the bearing can easily be popped out. On the other it's a pressed-in fit. I was pretty pleased with the result.
   Perhaps ironically they have the same problem as using 3 bearings and a washer: if a bolt is tightened to hold the wheel, the bearing centers get pressed together. But here I put very small and short pieces of pipe between the two bearings on both wheels so the centers are held at the right distance - actually just a little long for a tiny bit of free end play.

   Then what about the other face of the band? My first thoughts were the plastic on springs, but the bearings offset a little from those of the first side had worked out well earlier and started to seem like a very attractive way to do it. The 'rail car' wheels would of course be innermost, closest to the wood being cut. The offset bearings/wheels a little farther away on the other side would simply press the band tight against the inner wheels. They didn't even have to be very wide, although 'width of the band' like the inner ones would probably be best. One of them could in fact be used to adjust band tension - without affecting band alignment like the double tensioners pressing against the bearing mounts, mentioned above. And if the tension was was set with a spring it could compensate and keep the band tension relatively stable even if the wheels were lumpy such as if sawdust got under the band at the wheels or whatever.

   I got just one wheel mounted on the saw on the 29th. It seemed good. I asked a mechanically inclined neighbor with a sawmill come over and look at the mill. He seemed to like the offset guide wheels idea.
   He also thought the band was moving too fast and that I should slow it down. I suspect he's right. It seems ironic that I can't get a motor that turns fast enough for the car project, and I can't get one that turns a slower speed for the mill. I can see if any place in town has a 10" lightweight aluminum pulley. That would slow it down 20% over the 8" pulley. Meanwhile I adjusted the pulley on the Ryobi skill saw out a turn for probably a similar speed reduction.

Proposed New Electrical Standards: A Low Voltage Standard and Standard Connectors

   Okay, I don't make the rules, but I am going to offer suggestions... How it would be if I was "in charge" and in the absence of others' thoughts on the subject. I am open to other ideas as well (particularly extensions to the ideas), so perhaps this writing is also an RFC (Request For Comments).

   Adoption of the proposed standards would not take anything away from anything or anyone compared to what exists now. It would only provide focal points for manufacturers to better support those using low voltage DC line power covered by the specs for whatever purposes - off-grid, marine wiring, third world...

Standard Voltages (12.6 VDC, 38 VDC, 120/240 VAC)

   There are many possibilities for using various voltages for various things. In the lower voltage ranges 12, 24, 36, 48, 60 and 72 volts DC have all been applied, with the first four being fairly common for "off-grid" homes and for boats. (Note: The voltages are generally a bit higher than the nominal figures.) The trouble is, for each voltage, a different set of appliances is needed. And a different set of plugs and receptacles should be used so that appliances will only be plugged into an appropriate power outlet - and almost none are defined for any voltage. Almost everything has to be miscellaneous cutom connectors or hard wired.
   Currently there are high voltage AC wiring standards in place. 120 volts or 240 volts at 60 Hz are the most common, with 240 used in some countries and 120 in (at least) North America. UK (at least) uses 240 volts at 50 Hz. I have no proposals for changing any of those. (Along with specifying tolerances for new standards I will note that "120" volts has been variously specified as "110", "115", "117" and "120" volts, and that figures both higher and lower may be found depending on supply and loading. Also for a pure sine wave, the peak voltage of 120 VAC RMS is plus and minus 171 volts, and that value also depends on the quality and tolerances of the supply. Many AC supplies aren't pure sine waves and may have various high frequency spikes and harmonics in them. "Modified sine wave" inverters aren't even close to pure. Nor is the frequency figure, 50 or 60 hertz, always exact. I mention these things simply as a reminder that AC power also has its tolerances and exact figures can't be depended on.)

   Today there is only one common lower voltage standard: 12 volts DC nominal, with no specifications as to tolerance and no proper standard for plugs and receptacles or much else. It's a good voltage for low power items, and it's electrocution safe for the un-electrically-savvy. But for long runs and higher power levels the wiring has to be really thick and the currents are really high - 10 times the cross section and current of 120 volts. And a 2.5 volt drop in a wire drawing current loses 20% of the power to the appliance.
   The only change I would propose here would be to formalize the voltage as being 12.6 volts DC plus or minus 15%, rather than 12.0 volts +20% to -10%. Either of those voltages and tolerances takes it from 10.9 volts from a lead-acid battery when low to 14.5 volts while charging, as well as covering the voltages commonly found in "12 volt" NiMH and lithium ion batteries from about 14.2 volts (both types under charge) to 11.0 or 12.0[?], when getting low, under moderate loads.

   I would suggest that the lack of appliances on the market made to operate anywhere between 12 and 120 volts, not even LED and other light bulbs being common, is due to the lack of a standard: the markets for electrical goods for 24, 28, 36, 48 and other voltages are too diffused by the many possibilities. Selecting any one of these as a standard in place of the rest would provide a focal point. One voltage in particular seems to offer the best combination of advantages.

   I would propose creating just one 'standard' voltage in between 12 and 120 volts: 38 volts with a tolerance of plus or minus 15%. That is to say, from 33.0 volts to 43.7. It's right in the middle: three times the 12 volt standard and one third of the 120 volt standard. Significantly it's also the highest voltage that's safe enough to touch, with electrocutions from this voltage or less being vanishingly rare. But it's a better choice for wiring than 12 volts if even moderate loads are to be powered, with 1/3 of the current and wire cross section. And a two volt drop in the wiring under heavy load is a 5% line loss instead of 16%.

   The 15% tolerance covers three 12 volt lead-acid batteries in series discharged down to 11 volts each, or while being charged at up to 14.56 volts each. (same as "36 volts +20% to -10%") So the spec in fact has virtually the same minimum and maximum voltages as would be specified for a "36 volt lead-acid battery". "38 volts +/-15%" also better describes the voltages that will be found with (12) lithium ion cells or (30) NiMH cells. It's pretty easy to set up or install, for example, three 12 volt batteries and three 12 volt chargers, eg, for off-grid or marine applications, or, as I've been doing, for an electric car.

   Plus or minus 15% is a pretty loose tolerance, but it seems necessary owing to battery variations - different chemistries and different states of charge. It also allows better for voltage drops in the wiring. Specifying the center voltage and tolerance gives manufacturers defined figures to work and comply with in their designs.

Other Potential Standards

   Of the other potential voltage standards, 25 ("24") volts isn't a very big step up from 12. If more power is required than is practical with 12 volts, my feeling is that one might as well step a little higher.

   As discussed 38 volts (AKA "36" volts) is more of a difference and still pretty safe to touch. (Note however that solar panels or other power sources to charge all the batteries in series for this line voltage may potentially put out up to almost 60 volts open circuit and are getting into the hazardous range.)
   Conceptually "3" times 12 volts perhaps seems awkward, neither 2 nor 4 times 12 - an uneven multiple. It's a prime number: 3*1 or 1*3. Three batteries can't be arranged into a square or rectangle. I submit that such thoughts are mainly psychological stumbling blocks and should be given little or no consideration in picking a voltage standard. 12 volts was also an arbitrary standard. Some old cars and other equipment used 6 volts. 8 volts - four lead acid cells - might at one time have been deemed more "multiplicable", more conceptually appealing, than 12. But 12 volts was chosen, and it is convenient that the "mid range" voltage selected is a multiple thereof.

   An argument could be made for choosing 50 ("48") volts over 38. But the wiring is only somewhat lighter (25%) than for 38 volts. DC is more hazardous to touch than AC and around 50 volts (+15% is closer to 57.5 - almost half of 120) is starting to get into less safe territory, and "don't touch" and "shut off first" precautions and less "nonchalance" become advisable. Solar panels to power a 50 volt system would be over 75 volts open circuit - hazardous even in dry weather. There would likely be an annual electrocution death rate statistic associated with 50 volt wiring systems that wouldn't be there with 38.

   There are still countries where electrical expertise is a rare commodity. Better training and qualifications for workers understanding electrical precautions apply even more strongly with line voltages above 50 - and doubtless having even higher source power voltages. And DC is more hazardous than AC. It's surely better to go with the well known and common 120 VAC if potential higher than 50 volts is required.

Power (Solar Panel) Considerations

   Solar panels today are commonly being made essentially for 12 or 24 volts (with working voltages around 18 (36 solar cells) and 30 volts (60 cells)) But if 38 volts was a standard, new panels would surely soon be made to suit, perhaps 45 working volts with 90 cells, or they might be made to use with two in series, eg, 22 volts each. (Unfortunately two 18 working volt panels isn't quite enough voltage - see Electric Sprint, Solar Charging topic in TE News #119.)
   Safety: Two 60 cell panels in series - 60 working volts but over 75 with no load - is a definite hazard, as a friend of mine was shocked to discover when installing them. Being still alive, he changed his mind and put his four panels all in parallel to stay under 40 volts instead of 80. Even a 90 cell panel - around 57 volts open circuit - would be getting up there and hazardous if it's damp. With a 50 ("48") volt standard, the definitely dangerous 75 volt figure for panels would apply.
   Other sources such as windplants can put out higher voltages than expected and need their own considerations. Even one for a 38 volt line may put out hazardous voltages in a good wind. Shading solar panels and stopping the propellers of windplants may be good installation or maintenance safety precautions, depending on... everything. But batteries are always live. Best their voltage, distributed to all the wiring, not be hazardous to touch!

Standard Connectors: Plugs and Sockets

       Original CAT Std. 12 VDC plugs and sockets in 4-plex wall plate,
                  uses regular house wiring boxes, cables, etc.
   The other major problem with lower voltage standards is that no "standard" connectors have appeared except 12 volt "cigarette lighter" plugs and sockets and my own little known CAT Standard 12 volt plugs and sockets based on the popular AT type 12 volt plug-in fuses and sockets. Everything else is a custom plug and socket or hard wired.

   Using "car cigarette lighters" as a standard 12 volt connector is... well, bonkers. My "CAT Standard" plugs and sockets written of in previous issues (and with 3D printer designs for plug & socket housings and wall plates already uploaded to "thingiverse.com") is obviously a big step up.
   In doing drawings for this article, and with so little of it in use so far, I decided to optimize the "CAT Standard". The plugs and sockets seemed a little larger than necessary, and the plug blades a little thin and a bit too short. So the spec now has blades 1 mm thick x 6 mm wide x 8 mm long, 2 mm closer together. One further change is adding a current spec. The "regular" size would be for 15 amps or less. For up to 40(?) amps, the pins would have the same separation at the closest point but they would be wider, 8 mm instead of 6 mm. Thus a 15 amp appliance with narrow pins could plug into a larger 40 amp socket, but a 40 amp appliance with wider pins couldn't plug into a 15 amp socket.
   The thicker blades mean abandoning the cheap Pico .205" blade sockets (good riddance!) that don't hold the pins very solidly and bend if the blades are a little thicker than spec or get flexed or twisted a bit, and instead making better ones, of any design. I made the blade spec a little longer than AT fuse blades (8 mm) because plugs have wires on the end and must take more stress. Nothing "yanks" on a fuse and tends to pull it out of its socket. Is it still "CAT" standard - Connectors based on AT fuses? Well, those were still the pattern prototype. Maybe it's the Revised CAT standard, the "RAT" standard?

   The 38 volt plugs and sockets use the same system but the closest spacing is increased a bit (by 2 mm) so 12 volt appliances couldn't plug into 38 volt sockets and vice versa. That actually makes it very close to the original CAT standard. Shall we call it the "HAT" Standard for "Higher voltage CAT Standard? (With a nod to Dr. Seuss.)

(There's a blade spacing and narrow width negative blade combination that would allow appliances that are able to run at either voltage to plug into either type of socket. That's probably an advantage. There's one possible bad combo with the setup: 38 volt, 15 amp items could plug into 12 volt, 40 amp sockets. They (presumably) wouldn't work right, but nothing can ever be plugged in backward, and no 12 volt appliance can be plugged into a higher voltage socket, which are the likely situations to damage equipment.)

   The diagrams below show the dimensional specs.

   38 volts with 15 amp plugs and sockets allows appliances of up to about 500 watts (2/3 HP or small hotplate burner) maximum with common #14 AWG house wire. 40 amps allows about 1400 watts (just under 2 HP - large hotplate burner) with #10 wire. Standard wall receptacles for still higher power levels are not presently in my contemplated specs, but they would likely be just the same design with even wider pins/blades/prongs and sockets. (And maybe longer blades and deeper sockets.) Probably hard wiring or large in-line connectors is more appropriate for very heavy loads (like my 300 amp car motor). (or in a building, using 120 or 240 volts.)


   At some point I'll get out the 3D printer and make printable shell/housing designs for these connectors and sockets. I have the idea for printing the front face down and having front-back joining housings. They'll look better than the side-side joining CAT shells I designed before, with 'perfectly' flat faces of a single piece. (Then I suppose I should do the click-lock types, and a "cigarette lighter" adapter for the 12 volt type. Modified wall plates will also be needed for mounting the modified design sockets.)
   The pins/blades are simple enough. Receptacle ends will need doing, since they're now a different size than AT fuses. For a real, optimized standard, plugs need to be longer and stiffer than AT fuse blades & receptacles, as there is some mechanical stress from pulls on the wires.

Indoor Vegetable Growing With LED Lights (and other gardening)

   When the tomato plants really started growing near the start of May, they were soon into the LED light 'bulbs' above. I propped the right hand legs of the light table onto the dolly they were on, but that solution only added 3" and was short lived. By the 15th I had to move them out to the greenhouse. Then the spinach went to seed and also grew up into the lights.
   This points out that the LED indoor garden product, if there is to be one, will have to have lights that are easily adjusted up and down, preferably in more than one section, to cover seedlings well or to allow for tall growing vegetables. (A neighbor is planning to hang his grow lights for seedlings from the ceiling on pulleys.)
   Somebody gave me a small coffee plant in a pot. I put it in the LED garden. I'll see how it'll do indoors with the LED lights, but the weather around here is nothing like where coffee is grown.
   As the lettuce and spinach end in the March box, I'll shut down the indoor garden for the summer. The greenhouses are good enough... except that slugs keep coming and eating tender young seedlings. I haven't had that problem indoors! Hmm... slug bait...

   If LED Indoor Gardens are to be produced as a commercial product, I think I would team up with a furniture maker. A lot of the work would be much like making furniture: drawers or cupboards underneath for storing gardening supplies, supports for lights above and other wooden or 'prefab' parts. (No particle board. It doesn't withstand getting wet, and that would be almost inevitable.) The electrical, LED lights and electronics would be added on to these cabinets once assembled.

   I bought fruit trees in Victoria just before I moved here and planted soon after I arrived. So this is their first complete spring. I bought two supposedly "compatible" (according to the nursery) apple trees, and a "compatible" pair of pear trees, which are intended to cross pollinate each other. On both fruits, one tree flowered and finished before the flowers came out on the other. It remains to be seen which, if any, bear fruit. I may need to find more trees... and somewhere to put them. Of the two apricot trees (self pollinating) that I got a few months ago, the one in the pot looks okay. The other now has a lot of dead branches. I was concerned about the one in the pot not getting enough water and watered it frequently, whereas I was less concerned about the one I'd planted and apparently neglected watering it. But of course its roots are still largely in the soil that was in its pot and haven't had a lot of time to grow out more. Oops.

Electricity Generation

HE Ray Energy

   I had pretty much run out of ideas by the end of April. Then on the 5th it occurred to me to try placing a light bulb itself near the coil to see if any RF energy around the coil would light up the bulb.

   Also, it seems to me that it is almost surely necessary to magnetically saturate the core of the coil or transformer being used. Normally one does not want the magnetization to go so far. Saturation is supposed to be the limit. The currents suddenly start to rise when the core is saturated. In this case and on this small scale, that "EMP" is needed. That may require not only higher voltage but a stronger pulse with higher current than I've been using.

   On the evening of the 27th I hooked up the coil to the car again. I disconnected the motor, so the "armature drive" was open circuit. In that condition I found the field drive would only put out the programmed "minimum field drive" current regardless of stepping on the 'gas' pedal. But if I hooked it up the car would start to move, so I didn't.
   With 2 magnets around the coil and the same 40 watt incandescant "candle" light I got 12 volts instead of 11 (one magnet) or 2.6 volts (no magnets). That was with 2 amps of "field drive" from the motor controller. With the programmer I started adjusting the "minimum field current" upward. Anywhere from 3 amps to 5, the voltage stayed between 21 and 26 volts, with the higher currents giving the higher voltages. Further increase to 6 or 7 amps of current didn't raise the voltage. The motor controller shut itself off before I got to 8. At 4 amps and up it was around 25 volts. I expect the coil was probably saturated anywhere above 4 amps.
   The coil hummed and the filament of the light glowed dimly. Nothing seemed to get warm or hot to speak of. Two loose bulbs I placed with their bases in the box to see if they would light up with RF energy did nothing. (Maybe they need an antenna?) The coil hummed just as loudly with or without an output load, depending only on the current driving it.

   With the light turned off the DC output voltage rose slowly toward around 120 volts. But turning it back on immediately dropped it back to 25 volts. Where did the 25 volts come from?  It was certainly less power out (25 V * .16 A = 4 W) than was going in: 36 volts * 4 amps = 133 watts. But if it was coming through from the source simply by transformer action, why did it take a couple of minutes to build up to 120 volts with no load, and yet it could maintain 25 volts with a load? And what did the magnets have to do with it?
   Later I tried a 150 watt bulb. This time the voltage dropped to 12 and the current rose to 350 mA. 12 * .35 = 4.2 watts: same power. Then two 40 watt bulbs in series made for 29 volts, .11 amps; 3.2 watts. The filaments barely glowed. (One was so dull I had to close the door to the dull, cloudy daylight to be sure I was seeing it.)

    I really should take the oscilloscope out there as well as just a voltmeter and ampmeter, and check waveforms. There might be some narrow spike to 120 volts, and it might be just spikes, the leading and trailng edge of the pulses, being transformed and coming through from the source. And it just might be worth trying more magnets, but I don't like to get very many in close proximity - I already mashed one finger this last week under a cant of wood that was heavier than I thought.

   Finally on the 30th I was reading something about early radio and I remembered that at least some of the HE ray receivers had to have a ground connection. There was definitely something to try!

Electricity Storage

Nickel-Nickel with Oxalate Battery Chemie

Cell Construction

   Nickel-nickel seems like a very good battery chemistry. But thoughts kept coming to me that they will somehow be better than I expect and substantially better than lithium. Call it intuition. On the night of the 14th I had a dream of riding in an electric car. My brother who was driving told me he could drive for two days without recharging because of the nickel-nickel batteries. Could they really be so much better? There are some hypothetical possibilities. What if something like lanthanum nickelate forms in the positive, taking the nickel from valence 2 to 4 and so moving two electrons per nickel atom instead of one? [That might be La2(NiO2)3... or something along those lines] Or something to do with the oxalate? What if hydrogen can be stored in the cupro-nickel negative, giving it 'double capacity' - hydrogen storage plus the nickel to nickel hydroxide reaction?
   It seems likely to me that they're much better, but I can't prove it to others unless I make can good working cells and get the better measurements.

   Which brought me to the next question: could I actually make a battery cell that worked, lasted and was practical? I had recently noticed that while the laundry room had almost no space to work in, the wide bathroom counter in the next room had some free space at the right end... a small space to do lab work! So on the 15th I decided to do what I've been meaning to do for over a month now: take a few hours and make a "real" nickel-nickel cell, or at least a better test cell. I did it like this:

1. I found a pill bottle just the right size for electrodes I could make in a small round electrode compactor I had made who knows when. I cut the bottom 1" off of the bottle to be the body of the cell.

Negative Electrode

2. I cut a circle of cupro-nickel sheet metal to fit, with a long tab to go up one side on the inside, as the negative terminal on the top of the cell at the edge. (no holes to leak except on the top!)

3. I etched the sheet metal circle in (weak?) ferric chloride for about a minute. This etches away copper faster than nickel, and leaves a fractally rough, nickel rich surface at the micro or nano scale. This gives it much more reactive nickel surface. It looked dark and dull instead of shiny.

4. I cut two square pieces of nickel foam (~95% air). The corners could bend and fold over to make them round. The two were 1.6 g of nickel. (This would make a good conductive interior for the electrode. One might make the electrode without the nickel foam. Nickel powder flakes by themselves should be conductive enough. I'll try that later. Meanwhile... every advantage I can give it!)

5. I measured out some nickel flake/powder; an amount that looked good which weighed in at 6.15 g. I put one screen/mesh into the compactor then half the flake, then the next mesh, then the rest of the powder. Then I folded the corners in over the powder. It sort of made a mesh basket, so I added one more piece over the top, which would have been about .6 to .7 g. I thought the electrode might end up 2 to 3 mm thick.

6. I took it out to the hydraulic press and pressed it to 11 Mg (11 metric tons. YAY for a heavy duty press with a gauge on it!). It wasn't very even - .75 mm at one side up to 1.35 mm opposite, average about 1.05 mm. But it held together like a single disk of lightweight sintered(?) metal and it certainly looked like a great battery electrode. It weighed 8.1 g. In theory 8.1 g of nickel should give 3.86 amp-hours. Hopefully it would give 2 or more. Obviously an electrode 3 times thicker would give 3 times the amp hours without increasing the weight of the current collector sheet of cupro-nickel behind. The main limit to the thickness is how much it reduces the maximum current available, which includes charging current and hence speed to charge the cell.
   Under 40 power magnification the compressed powder surface looked like gray sand with grooves in it. The grooves were from ridges on the compactor die's surface, which wasn't very smooth. Where the nickel mesh was also looked like sand, but with shiny parts that looked sort of like someone had dribbled water on the sand in places, which had beaded up.

   The backing metal was 11.4 g. If the bottom of the cell was flat instead of a shaped pill bottle, I could see making thicker electrodes and just using something like lightweight cupro-nickel foil or screen under it for a current collector. Or maybe even just the fine mesh with powder impregnated, no backing sheet. Since I had some powder and mesh left over (and I had to do something with the powder - once taken out, powder wasn't going to go back in the can just in case it might have become contaminated) I put in two layers of mesh and the powder, then decided it was too thin and added a scoop more powder. This time surely I would have a thicker electrode! Well... after a 10 ton pressing it was about 1.15 mm thick and weighed 10.2 g. This theoretically had almost the energy storage capacity of a NiMH "C" cell. (A real 5 amp-hour, 85 gram one, not a 2.5 amp-hour "AA" cell in a "C" cell case!) What would it really be?

   The electrodes are 5.2 cm diameter, which is 21.24 sq.cm so 10 tons is 471 Kg/sq.cm. That's probably a pretty good figure - a little less than I've heard was used for iron in some nickel-iron cells (620 Kg/sq.cm(?)).


7. I cut a circle of "graphite foil", with a bit of a gap at one edge for the negative tab to get past. Then I got out the osmium doped acetal ester and tried to paint it on. It beaded up. I didn't like that, so I got a piece of "flexible graphite", cut it the same, and painted it too. On this it smeared and didn't bead up. Later I did a second coat on both. It seemed mostly to take on the foil the second time around. (How come it seemed okay on the previous piece for the test cell? Did I wash that one first or something?)

8. I got out what was probably the best of the nickel-hydroxide-monel-lanthanum-thiamin powder I made all those years ago. Here was a bit of a quandry. I had little confidence the loose powder would hold together the way the negative ones had. On the other hand I was afraid to add anything to 'glue' or 'gel' it, lest it cause self discharge. Finally I shrugged and used the straight powder. I put in 16 grams and compressed it again to 10 tons. (That's probably still less than a match for 8 grams of nickel in the negative. So what? If it works, yay!)
   It came out about 2 mm thick. I didn't actually measure it because it seemed too fragile.


9. Next I cut a circle of heavy paper as an insulator between electrodes, and another strip to run up the side with the negative tab. The circle was a bit bigger because I didn't want anything to short. It should seal all the way around. I wetted it with Brita filtered (~=distilled) water so I could fit it around the edges, going up the sides a bit... then remembered I was using DES and there shouldn't be any water. I could shape it then wait for it to dry, or make another one. I made another. It didn't quite cover around one edge so I cut another little wedge to fill it in.

10. I mixed some ethaline DES and potassium oxalate and poured a little in.

11. I went to put the electrode in, but it was so fragile it slid off the compactor bottom piece onto the counter and became powder again. Apparently all 10 tons in the press had done was break up the lumps and make new ones. The only option seemed to be to pour the powder into the cell, so I did that. It didn't all wet, so I added some more electrolyte. That turned the powder into mud. Perhaps it would "form" with charging and discharging? That is something some new electrodes do and it's part of making some cells. (It didn't.)

12. I put the flex graphite piece on top, and found a couple of large nuts and bolts to weigh it down a bit.


   I hooked up my crappy new lab power supply that didn't eliminate the need for external meters for both voltage and current, and set it to 1.7 volts. The cell would only draw 1 mA. I hoped it would go up as the paper started to saturate, but it only went down, to .25 mA after 10 minutes. Pressing down to connect the powder better didn't seem to change it much. It didn't seem to be holding a charge, which I thought was probably some powder got past the separator paper and made a bridge between electrodes.
   I also discovered that the power supply acted as a small load when set to "off" or actually turned off. It was perfect for making one think the battery had a self discharge problem it didn't have.
   I took it apart again. I made up a new separator paper. This time I wetted it with water and bent up the edges to make a bit of a dish. I think that's the thing to do - have a paper "basket" up to the top of this electrode. Then I left it overnight to dry, but I didn't get back to it on the 16th.

   On the morning of the 17th I reassembled the cell, this time with the original separator sheet on top of the new "dish" sheet. (I used the original sheet as a "platter" to hold the mud electrode together while I picked it up and put it in.)
   It didn't seem very promising. It was initially around .45 volts. If I put the large, heavy nut on top the voltage started dropping. Somehow it was making a bit of a short. I took the cell apart, found nothing, and put it back together. The large nut still made it self discharge. I used a smaller, lighter weight that was more in the center without coming near the edges. That seemed to be the answer for the time being.
   When I put on charge at 1.7 or 1.8 volts, it only charged at .6 mA. That didn't seem very promising. Nothing ventured, nothing gained, so I left it on charge. I checked a few times. It gradually improved over the day. By night it drew 2.5 mA with 1.7 volts from the power supply. And when the charge was removed it took quite a long time (3/4 hour?) for the voltage to drop to 1.3 volts. "Short circuit current" through the DVM started out at 50 mA (dropping rapidly), up from under 5 in the morning. When put back on charge after being shorted, it drew upwards of 15 mA, gradually dropping to the 2.5 mA figure again. It still wasn't much to speak of, but it was far better than when first connected, and apparently improving by the hour. I set the charge down to 1.6 volts, which dropped the current down to 1.5 mA, and left it overnight.

   Before I quit I belatedly checked some resistances:

* The (other) compacted nickel powder electrode was pretty much a short circuit. Whatever factors were preventing amps of currents, the negative electrode certainly wasn't one of them.
* The graphite foil spare current collector with the osmium doped film was about an ohm
* The doped flexible "gasket" graphite was under an ohm
* Undoped pieces of the foil and gasket graphite weren't much different than the doped ones, which indicates that the film must be a pretty good conductor. It depended more on whether the test leads were pressed softly against the material (1 ohm) or the points were jabbed into it (under 1/2 ohm).

   What was still unknown was the electronic conductivity of the positive electrode powder. The resistance was likely - and later proved to be - very high since I couldn't get it to stay compacted and I couldn't add graphite or conductive carbon black.

   The next morning it had improved a little more. All the currents had risen only a bit, but left off of charge and dropping through the 1.3 volt area, self discharge had dropped from around 5 mV/minute to 2.0 and was still lower when it was down to 1.25 volts (1.25 mV/minute). Instant short circuit current was 75 mA. I came up with one new metric for testing: "Short circuit current after 10 seconds shorted", which read 29 mA. Currents were read with the cell fresh off charge at 1.6 volts. They were always lower if the cell sat until it dropped to 1.3 volts before testing. After being shorted the voltage would jump back to over a volt in a couple of seconds, take 10s of seconds to rise up to near where it had been, then minutes to reach whatever voltage it was going to reach.
   I reduced the charge to 1.5 volts. I was hoping performance would continue to improve, but currents stayed similar and only the self discharge continued to gradually reduce. On the morning of the 19th it still held 1.327 volts after an hour (60.0 minutes - this is how testing batteries can occupy large amounts of time!), and was dropping by 1.4 mV/minute.

Sealing the Cell

   Something I wanted to try quite soon was to at least somewhat seal the cell. Would anything change if air was kept out? (With reaction voltages above nickel's, the negative electrode discharges when oxygen gets in. Would that be happening with nickel?) The graphite sheet covered the top... maybe just drip some candle was around the edges? I tried that and it was quick and easy. (What, candles have a purpose other than burning buildings down?!?) Heat glue would probably have worked well too. (The candle wax was easier to remove later.)

By early evening the cell retained 1.295 volts for an hour, with the self discharge down to ~1.6 mV/minute in that region. By 1.23 volts (almost 2 hours), it was less than 1 mV/minute. I don't have an "official figure" for the expected cell voltage. Eventually it seemed to be somewhere around 1.35 to 1.4 volts. That was somewhat higher than I had expected (~1.25 volts). Around 10% more energy - encouraging!

Water Based Cell

   Something else I wanted to try was using potassium oxalate in water as electrolyte instead of in ethaline DES. Obviously it was working in ethaline, but currents should be markedly higher in a water base if it worked. Since I wanted to continue testing this cell, and having already made two nickel powder/mesh electrodes, I decided to make another cell. (I must have made that compactor to fit some ABS pipe or something - what and where was the right one?)
   If it didn't work, I could take it apart and let it dry out, then add ethaline and have a second cell like the first.  I got to it on the evening of the 19th. Aside from the doped graphite foil for the positive current collector and the 10 gram nickel negative electrode, I used 20 grams of "+" powder, and I used nickel-brass for the negative current collector per the idea a ways below.
   Instead of starting charging from 1.6 volts at 1 mA it was about 6 mA. But it didn't seem to work. It discharged itself quickly as soon as the charge was removed. Why should that be, when the nickel-manganese cells with the same positive (and double the overall voltage) more or less worked? It made little sense. I was concerned that there was a short, but it didn't discharge to zero, only to .47 V or so. Then I thought of the way the film had beaded up on the graphite "foil". Maybe it wasn't coated? I would take it apart and redo the separator paper and use 'gasket' graphite. This time also I would drip in some candle wax anywhere that looked even slightly likely to cause a short through the separator paper, notably around the negative tab.
   But still the original graphite "foil" had worked in the cell in March, or I wouldn't have had success and continued experimenting. But the film hadn't beaded up on that one when I was painting it.
   (I didn't get any farther on this.)

Old Cell

   On the evening of the 20th I put the old cell back on charge. The next morning the self discharge had gone down still further to about 70-80 seconds per millivolt, maintaining over 1.32 volts for over an hour. But the short circuit current was also down again, to about half of what it had been at the best point. (15mA after 10 seconds.) My thought is that it needs more electrolyte - ethaline, KC2O4 or both. Quite probably the electrodes had absorbed some of the oxalate. Or perhaps the ethaline was escaping and evaporating or some had been absorbed. But I hadn't measured it and I didn't know how much was needed or how much to use in the first place. I would have to scrape off the wax and add some more of each.

22nd: 1.5 hours 1.303 V; -1.35 mV/minute; 10 seconds shorted: 16 mA.

Summarizing the reduction in self discharge:
day - mV/minute - drops to 1.3 volts after _ hours:
17th: - ~-5 mV/m - .75 hrs
18th: - -2.0 - 1

Then I started measuring "cell open circuit voltage after 1 hour disconnected from charge" instead of "hours to discharge to 1.30 volts", and the "millivolts drop per minute" was also measured after one hour:

19th: -1.6 mV/min; 1.327 V @ 1 hr
21st: -1.5; 1.323 (est)
22nd: -1.5; 1.339 V

   I decided to take the cell apart, add or change electrolyte, and redo it. The idea of dripping in wax to seal any suspect potential leaks between electrodes promised better reliability. First, I thought I'd run a load test and chart it in order to have a good comparison with the redone cell. It wasn't very good - my usual milliamp-minutes instead of amp-hours, and that with just 500 ohms for a load - around 2 milliamps. Anyway there was nowhere to go but up!
 Out of sequence, here with the other figures are figures after overnight charging with an improved plus electrode from June 1st, AM:
 10 seconds shorted: 40 mA. (starting from about 65 mA - it didn't drop half as fast as with the previous electrode)
 -1.1 mV/min; 1.298 V @ 1 hr
   By evening:
 .82 mV/min; 1.402 V @ 1 hr
 Short circuit current was down, but starting from 39 mA it was still 35 mA after 10 seconds.

Remodeled (Old) Cell

   I did it on the morning of the 23rd. It seemed plenty wet so I just sprinkled in more CaO and KC2O4. But I used a new separator paper, which I made into a deep enough basket to contain the electrode paste, which came off the old paper in clumps of mud which were easily mashed and molded to fit.
   It started at around .9 volts. If it had been all charged before, why didn't it come back to around 1.3 volts when reassembled? This seemed to suggest that only a portion of the paste had been being charged, perhaps even a small portion. That might be the explanation for the low performance. But how would I make it more conductive, when it was a powder that wouldn't compact in a press and adding graphite or 'conductive carbon black' made it self discharge?
   I put several weights on top (this time that didn't seem to cause any shorts) and got higher charging current than before, but not by a lot. (eg, 2.5 mA instead of 1.5? - all depending on time and state of charge.) Then I started putting a bit of my own weight on top, and this helped, and further, the currents didn't go all the way back down when I stopped pressing. I pressed here and there and soon had it doing 15 mA at first and 5 mA continuous for a while, a bit later dropping to 3.5. It wasn't the two or more orders of magnitude improvement I was ideally looking for, but it was heading toward one order. But poorly connected grains in the electrode would also explain it: a connected grain gradually discharges into an almost unconnected and uncharged one, which gradually transfers its charge to another uncharged one... In fact, having a large percentage of the grains poorly or not connected probably explains everything: self discharge (slow as it is), low currents, low energy storage, very gradual reduction of self discharge as more of the poorly connected grains gradually become charged, no distinct drop-off voltage point during discharge. It seemed to be a good formula for an electrode, but it isn't a proper electrode unless all the particles in it connect together electrically.

   I was afraid of forcing a break in the separator paper or in the osmium film in the graphite, and didn't want to push on it too hard. If the film was broken self discharge would go up with no evident reason. (Perhaps it was already broken or scratched somewhere and that explained the self discharge it did have. But the poor powder connections would equally or better explain it. I suppose I could try out a new doped graphite sheet?)

   How then to make a positive electrode that works well? It might just boil down to manufacturing technique. Perhaps if the powder was wetted with something and compacted in the press the electrode could hold together compacted until it went into the cell, and the particles would maintain electrical contact with each other? Then whatever was wetting it could be allowed to evaporate before final assembly. And the cell could be assembled with a rigid shell, and then liquid wax could be injected under pressure above the top electrode, pressing it down. Or maybe some sort of binder would work, but I hesitate to try anything that might cause chemical problems.

   Some hours later I remembered that with a new dry separator sheet installed and the old saturated ones removed the electrolyte might be a bit dry. I hadn't sealed the edges yet so I simply poured on a bit more ethaline DES. That seemed to up the current from 3.5 mA to 5, which rose slightly over the next hour (as it soaked in?) to 5.8 mA. There was still a lot of room for improvements before it would be doing amps and amp-hours.

   After I had slopped in a little more ethaline, the self discharge went way, way up. After supper I investigated and found another gremlin in the works. Without thinking much about it, I had put zinc coated nuts or washers on top of the top graphite sheet to weigh it down and to connect to it. That was okay as long as it was dry, but once wetted with electrolyte, zinc is a ~ -1.25 volt electrode. So I had a strongly negative electrode shorted to the positive current collector! That explained why I'd had to use a smaller bolt the first time: it wasn't causing an internal short: it was wet around the edges then too.

   At some point I measured the pH as about 12-13 when I had the cell apart. That's the expected pH for water with calcium hydroxide in it, but I wasn't sure about in ethaline. Very much seemed to work just the same as in water.

   When I put it back together on the 23rd I couldn't stop rather serious self discharge, even by adding a second separator sheet. A short seemed unlikely and I had wiped off the excess electrolyte from the top. I suspected my pressings on the top to compact the powder had scratched up the osmium doped film on the graphite.

High Resistance Electrode can be solved by VERY strong compaction

   On the 30th I took two small, thick pieces of nickel-brass sheet. I slipped the larger sheet under some of the dry positive electrode powder, and set the smaller on top. (With wet powder, one would be measuring ionic conductivity - a false reading for conduction of electrons.) The powder might have been a couple of millimeters thick. Poking both meter leads through the powder to the bottom sheet confirmed I had a short between them, that there was connection to and across the bottom sheet. Then I moved one lead to the top sheet. Just touching it the resistance was off scale. With some pressing it came down to tens of megohms. Pressing heavily on the lead onto the top sheet, thus pressing powder together, brought it down to the upper ones of megohms, eg, 6 megohms. I used three pieces of different sizes for the top, from about two square centimeters to 1/2 square centimeter. I thought that if I pressed on a smaller one the powder would scrunch together better and give a lower reading, but they were all very similar. If I pressed in one corner the powder could be felt crunching up, and the reading would go down to lower ones of megohms until the top and bottom pieces touched at the corner and it became zero ohms. With less powder, perhaps under a millimeter, readings in the lower ones of megohms were had, but still over two megohms.
   Somehow the particles have to be brought into much better contact with each other and with the current collector sheet, but without pressing very hard on the separator paper or the coated graphite current collector. The chemistry is working but the mechanical aspect needs to be solved. No doubt others have solved it, somewhere. Is there something that could be added to it? Does it need far higher pressure - tens of tons per square centimeter perhaps? - to compact it into a cohesive "briquette"? Does the positive side need to be a thin film electrode?

   I decided to try high pressure. I put some powder on a sheet of stainless steel, and put the 1 square cm piece of nickel-brass on top. I put it in the press and pressed it to 5 Mg or 5 metric tons. The powder formed into a sheet, but it was very brittle and crumbly. So I tried 10 tons, and it was a little better. Resistance however was still in the megohms. I cut the sheet in half for 1/2 a square cm and tried 12 tons That got the resistance below 200 kilohms. Then I tried cutting it in half again, now for 1/4 of a square cm... actually still more like 1/2 because the nickel-brass was squashing out thinner... and pressed it to 10, again maybe 20 tons per square cm. Somehow resistance was back up into the M ohms. But the tests were getting ambiguous because the nickel-brass was squashing and the thick piece of stainless steel now had a substantial dish where the pressing had been. From the one instance of lower readings it was looking like the idea might just work.
   I cut two pieces of mild steel for upper and lower plates, the top one about .5 sq.cm. I pressed with 10 tons of force. I made several tries with 6, 8 and 10 tons. If the top piece stuck to the piston of the press, good readings couldn't be had - M ohms again. Probably some loose powder from around the edges gets in when I try to re-place it on the pressed area. When it stayed in place, various much better readings could be had, 100 K ohms even down to tens of ohms. 8 tons seemed to work at least once, but 6 didn't. So it might appear that the pressure needs to be 16 to 20 tons per square centimeter.

   Unless the electrodes can somehow be hammered into form with heavy blows, effective cells much bigger than button cells aren't going to be made without a very special production setup to get that sort of pressure over wider areas. And then they'll have to be treated very gently or they'll break up again. But it does appear that it can be done, that practical working cells are possible. After all, others make nickel hydroxide electrodes, albeit for pH 14 alkaline cells, with various mixes that don't always include graphite - they must have at least similar challenges. Later I remembered that, yes, Edison had had the same problem in his tall, flooded nickel-iron cells, to the point where he made the plus electrodes the size and shape of a pencil in perforated nickel-plated or nickel metal tubes, repeatedly loaded from the top end with a few more grains which were each pressed in with tons of force. In the less alkaline cells metal isn't an option unless it's perfectly coated with the conductive film. Any break in the film and the metal would corrode away.

   On the 31st I thought of the jeweller's rolling mill. These are made to flatten or shape strips of softer metals such as gold and silver. Perhaps somehow I could contrive to have it compact a thin strip of the powder? It could be done in more than one pass, perhaps with textured rollers so it was only flattening part of the material at a time. Could it bring enough pressure to bear to accomplish the task? Even nickel-brass, being harder than regular brass, was a challenge for it. But maybe, with a thin enough strip?... but only if I could figure out a way to feed the material in and not have it break up in the process.
   Then I went over to my neighbor's for a chat. I noticed he had some heavy rollers for heavy equipment and so I explained the problem and the idea. He said rolling would tend to spread material and thought it would be easier to get the pressure if I put a heavy pin inside the same cylinder as an axle and pressed on it. Then I thought I might just get that instantaneous pressure if I hit the pin with a sledgehammer or maul instead of using the hydraulic press. He said to put something inside like cardboard or tinfoil so it would release after I had done something like that. (It turned out not to need it.) He said that cylinder was spare and so I took it. (The pin wasn't a spare, unfortunately.) It was 38 mm inside diameter, making it 11.34 sq.cm surface area - about half the size of the previous sets.

   I checked out a trailer stub axle I had and it seemed to be a pretty good fit, a bit too short, if I chopped the thin end off it. Then the widest part would be the bottom and it was pretty much just the right diameter - slightly undersize. The shaft I'd be hammering on was a little thinner yet, so it wouldn't get stuck if it mushroomed out a bit wider. 'Too short' just meant using a bit longer "plug" in the bottom. Now... what did I have for that plug? Not finding anything that looked like it might work, I went back to the neighbor's and I got a very large bolt. The shaft was too small, but even the small sides of the hex head were too large.  It must have been very hard steel, even harder than the stub axle, because I ended up taking hours to turn the head down to size on the lathe, and I stopped to sharpen my carbide tool about 4 times. But as the squashed metal in my previous pressings had shown, harder is better - as long as it's not brittle!

   Then I took it outside and measured up 20 g of the powder. This made about 8 or 9 cc. To this I mixed in 2 cc (.8 g) of pottery supply "Veegum" (a mixture of bentonite clay) as a 'glue' or binder. I poured it into the cylinder with the shaped bolt underneath - I'd made it a close fit and it would lose little powder around the edges. To make a long story short, pounding on a relatively hard part of the gravel driveway with two heavy blocks of steel under the "compactor unit", it finally felt like I was hitting something solid and didn't bounce much. (It bounced badly on a big "solid" tree stump.) I walloped it pretty hard but I didn't want to risk missing the pin or hitting it on the edge. It was enough that the "tool steel" axle soon had dings and a bit of mushrooming on the end. I threw in a bit of metal and measured through the powder to the outside metal. It seemed I had it down from over ten megohms to under one. It was disappointing, but at least it was an order of magnitude improvement. Another factor was that it was about 2.5 to 3 mm thick. Thicker electrodes have more resistance. I miscalculated in doubling the powder, when the surface area was half. Since I used twice as much powder, added the Veegum and the electrode was half the area, I shouldn't have been surprised it came out so thick. I also added a few drops of Sunlight dishsoap when I was compacting it to help glue it together - not much; it was still pretty dry. This one held together well which I attribute partly to it.
   The lathe had left a lot of grooves on the critical end surfaces, too deep to polish out. Since they weren't mirror smooth some compacted powder ended up stuck to the dies. And of course there were ridges on the compacted electrode where the grooves were. (I filed the dies down and re-polished them on June 1st. A few deepest grooves weren't entirely eradicated and they still didn't have that mirror smooth surface, but they were close enough.)

   I put the new electrode into the same cell, removing the original one. Charging current was around 4 mA. That was disappointing but better than 1 or 2 - and it was from a 3 times thicker electrode with half the surface area. If it had been the same dimensions it would probably have been around 10 mA: the order of magnitude improvement expected from the lower, but still much too high, electrode resistance.
   And of course it would help to make a new cell with diameter to match the electrode so it couldn't mush out around the edges.

   For a while the self discharge was awful. Since I had painted the conductive film on both sides of the graphite sheets, I flipped this one over. That cured the problem. (Doubtless the regular cure would be "do not disturb" once the cell is assembled.) The Veegum and Sunlight soap didn't seem to be any problem as this electrode worked better than any other so far.

   If I do the same thing with a still smaller die, I should be able to get more pressure per square cm when I hammer it. Or maybe I should have a bigger hammer - 10 pounds instead of 6? Or a better system - a 20 pound weight in a track coming down squarely on top every time, with the die on a heavy, solid base, might do the trick. In fact, just setting the die on a big rock instead of on gravel might make the difference. (It didn't.) Perhaps for test cells one could have a pipe to control the fall of the weight, and drop it down the pipe onto the die. If reaching up and dropping the weight into the pipe doesn't initially do it, get a longer pipe and a ladder to drop from higher up.
   It just has to get to that threshold pressure of 20 tons per sq.cm, even for an moment, over a much broader area than .5 sq.cm. (Or else do something like spit out tiny electrode "pellets" at a rate of at least several per minute, which would be placed side by side until an electrode area was filled. Ugh! Or Edison's "pencil electrode" system, but making rectangular shapes? That sort of "end feed" layout certainly reduces the cross section that needs to be pressed. Get it down to 1/2 sq.cm - even if it's 10 cm tall - and a hydraulic press can manage it. But the die had better be stronger than mild steel!)

   But on the morning of June 1st the cell had had overnight to charge and so was working better, so I continued the experiment. I opened it and found the plus electrode - the whole electrode and not bits of mud - was stuck onto the graphite. Since the compaction was only marginally better I attribute it holding its shape to the Veegum and dishsoap. I added more electrolyte so it was pretty soaked instead of a dry cell. Currents immediately doubled. After a while charging current was down to 5 mA, but that was after doing some charging around 8 mA for a while, so the electrode was getting better charged faster. (And it's still better than 1 or 2 mA.) Short circuit current started out at around 65 mA and didn't drop off as fast as before after 10 seconds (still 40 mA). By evening everything was even better with better figures than with previous electrode - except that currents were lower. However, the drop in current over ten seconds during a continuous short circuit was much lower too. It just kept putting out. The drop in self discharge may be the improved electrode, the reduced surface area, or just fewer scratches or gaps in the doped film coating the graphite. (Figures are given above, along with those of the previous electrode.)
   A discharge test in the evening was still low capacity, again demonstrating that I still didn't have enough compaction of the electrode, but this time it held its voltage better under load, and when the test was ended it recovered its voltage much more rapidly. It showed rather similar voltages to NiMH cells, maybe even a touch higher. I thought NiNi would be about .1 volts lower, nominally "1.1 V" instead of "1.2 V", but that didn't seem to be the case. The load test started out at well over 1.3 volts, and the voltage dropped off rapidly once it hit about 1.15 volts. On the morning of June 2nd the voltage was still 1.368 after 3 hours off charge, showing that whatever the self discharge is, it's not inherent since that was at least an order of magnitude lower. And that's with an unsealed cell where air might reasonably be expected to cause self discharge. But it was definitely time to quit experimenting for now and get this newsletter of May's D & R out.

NiNi Must Next Move to Developmental Production

   It certainly appears that the last challenge to making working cells is to greatly reduce the very high resistance of the powdery positive electrode material, by very high pressure compaction. That resistance perfectly explains why my cell, though working chemically, performed pathetically, and this was supported by the somewhat improved performance of the new electrode with somewhat better compaction. The challenge now is not to invent the nickel-nickel battery but to invent a practical way to produce good positive electrodes. (I think I've figured out a workable way for a few test cell electrodes as I do a final edit of this newsletter.)

   Preferably production after that is something that should be approached in a reasonably big way. Not only the compaction but many or most of the processes need to be automated to prevent production bottlenecks. People are definitely going to want millions of them once they become available and known. They will want to switch from lithium types, which are being manufactured in great quantity today. Electric vehicle manufacturers will be chomping at the bit for them, as will "off grid" sites and marine battery users. It will not be possible to satisfy global demand from even a very large, high production factory.

Negative Current Collector Metal as a Hydride for double energy storge?

   Usually to make a good metal hydride, big chunky atoms like lanthanum are alloyed with nickel to achieve the maximum inter-atomic spaces in the crystal for storing hydrogen ions - protons. But might cupro-nickel have some amount of proton storage capacity? NiMH cells are pH 14 alkaline, in which nickel won't oxidize, so no reaction of the nickel to oxide occurs. Either the nickel-lanthanum alloy stores protons or it won't work. But in fact, in such hydrides the hydrogen ions can be packed in more densely than the density of actual liquid hydrogen. Apparently theoretical storage capacities upward of 1000 AH/Kg can be attained. I have little idea what is actually attained in practical commercial cells, but from the weight of a 10 amp-hour "D" cell, I'd guess it's probably under 300 or so.
   The nickel-nickel cell operates at a somewhat lower pH where nickel will oxidize to hydroxide, giving us the 477 AH/Kg of Ni storage figure for the nickel exposed to the electrolyte. But the interior of the backing sheet isn't exposed to the electrolyte. If in addition to being a backing sheet it can store protons, its storage capacity would be much increased.
   But owing to the nickel also reacting, lanthanum would quickly corrode. So my thought is to add some tungsten, lead or bismuth instead: other big chunky atoms with low reaction voltages (I haven't checked out which ones might work as to reaction voltages, much less as to other characteristics) that won't corrode like lanthanum with its very high reaction voltage. Even if it doesn't attain very high values like 300 AH/Kg, any amount just adds more storage capacity to the nickel's reaction energy. Maybe it can attain nickel's full theoretical 477 AH/Kg or higher. OTOH I really know little about metal hydrides. Only that for good ones it gets complicated with cobalt and other rare earth additives to keep the lanthanum from oxidizing.
   If hydrogen/proton storage energy can be added to nickel reaction energy, it will be a heck of an energetic electrode!

   With the nickel reaction being slightly lower negative voltage than hydrogen storage, the cells can be vented without issue. But cells with hydride storage have to be sealed and withstand a certain amount of pressure. Air (oxygen gas) getting in would turn the protons spontaneously back into OH- ions and H2O - self discharge.

   In 2008 I bought some nickel-brass (AKA "nickel-silver" or "German silver" - Cu:Zn:Ni eg, 65:18:17% - contains no silver) sheets for current collector plates. Then I realized that zinc as a negative electrode gradually deteriorates. Zinc has a soluble state ("zincate" ion) that seems to apply at any pH, and the atoms migrate while it's in that state. In a battery this state is short lived and the zincate turns into zinc oxide, but owing to it migrating, the electrode loses active substance and grows "dendrites" (slivers of zinc) that often short the cell through the paper. (Cadmium, under zinc in the periodic table, does it too if evidently somewhat less, and many NiCd cells end up shorted.)
   So I switched to monel (Cu:Ni eg, 63:37%) or other cupro-nickel (eg 70:30%) to avoid zinc. But this might have been a mistake. First there's only ~18% zinc in nickel-brass and second if we don't charge the cell to the higher reaction voltage of zinc, it will become zinc oxide and stay in that form. If the zinc did migrate out of the electrode, it would leave the surface pitted and rough - just what an electrode wants, to have the most surface area in contact with the electrolyte. I doubt enough zincate would migrate into the paper to short out the cell. If some zinc made it to the positive electrode it would simply remain there as oxide.

Haida Gwaii, BC Canada