Turquoise Energy Ltd. News #79
  August 2014 (posted September 2nd)
Victoria BC
by Craig Carmichael

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

Month In Brief (Project Summaries)
  - Electric Hubcap Motors are more powerful than previously thought! - Electric Weel - Turquoise Battery Project - Write-ups & Website - Shopping & supplies - Aquaponics Greenhouse & LED Lighting - VA Wind Turbine?

In Passing (Miscellaneous topics, editorial comments & opinionated rants)
- Research and development tax credit programs - New video on financial system - Overpopulation & Epidemics

Electric Transport - Electric Hubcap Motor Systems
* Major magnet rotor improvement!
* "Bedini" unipolar motors?
* Huge Torque, Low RPM Electric Weel Motor-Generator Project
* Lightest magnet rotors
* J1772 car charging plug

Other "Green" Electric Equipment Projects (see month in brief: LED grow lights)

Electricity Generating (see month in brief: VAWT.s)

Electricity Storage - Turquoise Battery Project (NiMn, NiNi), etc.
* Improved plastic electrode pockets: two-piece 'boxes' with 'lids'
* Nickel-Nickel cell tests
* Lead-acid battery renewal update

No Project Reports on: Variable torque converter transmission, Peltier heat pumping, Lambda Ray Collector, Magnet motor, CNC Gardening/Farming Machine (sigh, maybe summer... 2015?), Woodstove/Thermal Electricity Generator, evacuated tube heat radiators.

Newsletters Index/Highlights: http://www.TurquoiseEnergy.com/news/index.html

Construction Manuals and information:

- Electric Hubcap Family Motors - Turquoise Motor Controllers
- Preliminary Ni-Mn, Ni-Ni Battery Making book

Products Catalog:
 - Electric Hubcap 4.5KW BLDC Pancake Motor Kit
 - Electric Caik 3KW BLDC Pancake Motor Kit
  - NiMH Handy Battery Sticks, 12v battery trays
& Dry Cells (cheapest NiMH prices in Victoria BC)
 - LED Light Fixtures

(Will accept BITCOIN digital currency)

...all at:  http://www.TurquoiseEnergy.com/
(orders: e-mail craig@saers.com)

August in Brief

Electric Hubcap family motors are more powerful than suggested previously

   The main concern with motor continuous power rating is whether the motor gets too hot at a given power. They're usually rated for the power they'll handle without gradually exceeding their maximum temperature. The materials of my motors (in particular the epoxy) are only rated for up to about 65°c which is much lower than steel body motors, but being highly efficient and having good cooling they make little heat.
   The concern for intermittent power rating above the continuous power rating is how much power they're wasting, since efficiency drops more and more as power is increased. Typically peak efficiency is at 10-20% of full rating. Usually the high end testing is ended when the power loss reaches 50%.
   Judging from the almost trivial motor temperature rises in the Electric Caik outboard at 1200-1500 watts in actual operation on the water, it appears that's a low percentage of potential power and so I've been rating my motors extremely conservatively. They could probably do four times the power I've been saying for short periods without overheating (and without exceeding 50% loss), and maybe as much as double for longer if not indefinite periods. At the risk of maybe going overboard the other way, that would be: 12/6 KW (Caik), 18/9 KW (Hubcap) and 48/24 KW (Weel).
Very roughly assuming efficiency losses of 50% and 25% at such high levels, output power would be 8/6 HP, 12/9 HP, and 32/24 HP. Such figures might put a new complexion on what people decide they can use Electric Hubcap type motors for.

   Proper testing for which I have had neither the proper equipment nor time to do so far would obviously be valuable.
I've been wanting to buy a commercial BLDC motor controller that won't blow up if I drive them to the levels indicated above, in lieu of beefing up my own controllers to 500+ amps capacity without failing, which they've hardly managed 1/5 of so far. So far, I've been putting it off until I (presumably) get my annual SR & ED tax credit from Canada Revenue Agency.

Electric Weel Motor/Generator

Large and Small:
The newly made stator end piece for the huge torque, 28", 24(?) Kw Electric Weel motor,
compared to the size of the outboard's 9", 6(?) Kw Electric Caik motor.

   In the early part of the month, I was molding a piece per day, occasionally two, for the giant Electric Weel motor, and fastening the pieces made in July together with polypropylene strapping and the left over epoxy. Towards the end of the month I had two of the three body pieces completed, the stator end piece, and the center piece with a thick rotor side rim to safely enclose the rotor and stop any magnets that might potentially come loose from flying out with potentially deadly force.
   In spite of good intentions, the one additional mold piece needed for the rotor end cover isn't made yet. (Two pieces from the 3 piece mold for the stator end cover can be used for both ends.)
   This huge motor is certainly a lot more work than the small units, with 24 moldings of body pieces instead of 3, then they have to be epoxied together, plus a much larger rim that needed three epoxy mixings instead of one. It's taking over 200 dollars of epoxy.

   In a discussion, I also came up with a great way to improve attachment of the magnets to the rotors: cut slots through the rotor plate just inside from each magnet, and wrap the epoxied polypropylene strapping right around the magnet and rotor plate from inside via the slot to ouside around the rim. If I also balance the rotors carefully, I think the Caik motor should be safe at at least 3000 RPM this way, whereas I've been reluctant to take it above around 2000 so far. This will be a major improvement to the whole family of motors.
   For the Electric Weel specifically, I got the idea to use lexan plastic for the huge (26" OD) rotor instead of steel to lighten it. Only outside where the magnets need steel backing will have a steel ring. Even with the rotor lightened by having a 1/8" main body with a 3/16" outer ring instead of being thicker throughout, the motor with a steel rotor will weigh about 100 pounds (...that's the exact figure I came up with in estimating it), and the plastic one can drop 8 or 10 pounds off that. Pounds count for handling and installation, and in an EV it improves range that little bit.

Electric Weel body parts so far, in assembled position - now it
just needs the top (rotor) end cover... and all the other parts.

   The first Weel is already spoken for for use as a large, low RPM generator for a novel floating river hydroelectric project by Rick Linden of Coastal Geosciece Research Corp. (http://www.coastalgeoscience.ca/ - a website even more in need of update than mine!) Being made specifically as a generator for lowest RPM, each rotor magnet will alternate polarity and it can't be used as a motor.

   As I think about this project, I wonder if it has the potential to replace river dams with more environmentally benign floating structures. One thing working against this idea regardless of the efficacy of the generators themselves, is that dams usually are used to store winter rain water or snow melt for the dry summer season. But it could probably harness river and stream water that wouldn't otherwise be harnessed at all. A hydro dam magnifies and harvests nearly all the energy at one point in the stream, or perhaps it should be said, of the section of the stream that the reservoir backs up to. Each floating unit won't harvest a major portion of the flow, but they can be placed up and down a river to harvest it at as many points as are desired or seem useful. The aggregate just might rival a showpiece dam project. Or it might complement it because if a dam holds back water for summer, floating units downstream will also have water flow to work well in summer.

Turquoise Battery Project

   With the 3D printer repaired, on the 4th I printed a bunch of electrode containers... each pair improved a bit over the previous as I saw flaws. But leaking electrodes from July's batteries, which I finally abandoned for that reason, got me thinking about an idea for improved plastic battery electrode 'pockets'... something like little boxes... where the sides of the box slide over each other, so that no matter the exact thickness of the electrode materials, there'd be no gaps around the edges. If it seemed helpful, I could even glue the edges together. (Fearing the usual paper deterioration I'd rather not glue prototypes and then be unable to open it to replace the paper, but it's doubtless the way to go for production.) But somehow I hadn't had time to get it done.

   On the evening of the 11th after dinner, I finally decided I just had to get this together. Working around some frustrating idiosyncrasies of the 3D printer - actually mainly of the printing program - I had a workable pair done by about 1:30 AM. (It occurs to me I should download a new version of "Pronterface" in case there have been any improvements that would help.)
   I made a nickel negative electrode that crumbled, then another one, and put it in one of the new boxes. By then the end of the month was approaching and I decided to leave it until my conductive carbon black arrived to make a better positive side from. Supposedly that should have been shipped on the 24th, but it hadn't arrived by September 1st.

Electrode boxes that I hope should serve to prevent leakage
of electrode substance into body of battery cell

Write-ups & Website

   From the 12th to 17th I was busy writing up info Canada Revenue Agency ("CRA") wanted to have preliminary to a technology review and inspection of Turquoise Energy's facilities, the first since their initial visit over four years ago. There was a deadline for submitting it. I finally mailed off about 40 pages of info on the 17th, a little about finances but mostly about the projects claimed last year.
   Having done a sort of an 'overview' writeup on each project, I plan to use the material to update the Turquoise Energy website. I've put more about the tax credit funding subject, and its stormy history, in "In Passing".

   And I finally started a rewrite of the Turquoise Energy web site on the 31st. A couple of weeks previously, Jim Lawrence came over and went into web page CSS coding with me. CSS 5 has new features, created since he had done up the website for me in 2011 in CSS 3, which make it far simpler to do what he did then. In several frustrating hours I had the commands worked out to create similar article frame borders in two columns to what he had done then, but which were then so complex they got messed up every time a change was needed. I kept the upper part of the page with the menus and links, and the footer part with icons and links to twitter, facebook, etc, and the appearance of the main section but with the new CSS 5 coding. Of course, after three years most all the projects are much farther along and needed or still need major revisions. It'll take a couple more sessions before I can upload the new main page, and then the other pages can follow one at a time.

Shopping - Supplies

   I spent the whole day of the 21st shopping for various supplies, in particular for the Elecric Weel motor, and a cheap flux core wire feed welder to help with building a new 'box' for the variable torque converter transmission. I find the considerable time I have to spend either visiting stores or sitting on the computer looking for and ordering things this somewhat frustrating, as well as seeing the money go out, when I want to get building. But one can't build without parts and materials. I got a great deal at a tent and awning store on 1.5" PP strapping for the rim of the rotor compartment (the immediate concern) - a 60 or 70 yard roll of white that the clerk said had been 'kicking around a long time' and had become a bit discolored, for about 1/3 of the regular by-the-yard price.

Aquaponics Greenhouse - LED Grow Lighting
Access to the greenhouse is through the garage wall for security.
   Making the greenhouse dragged out through the whole summer, since I didn't put much time into it. Finally I put on the last outside "coroplast" panel on the 29th. There's still work to do on it, but I now have an enclosure.
   It seems the federal government would love to see food being grown in the arctic in greenhouses, since it costs a fortune to have it flown in for workers and residents in the far north. I got a 5 meter RGB - LED strip light from Jim Harrington on the 23rd, and it seems I've sort of been volunteered to work on ways and means for the lighting. It can be cut into sections or bent around, and it even has a remote control. I also have the blue and red plant-grow wavelength LED.s I purchased myself for flat panel grow lights.
   Victoria isn't the frozen north, but it's certainly a challenge to try and keep a greenhouse warm and lit and grow food here in the winter, so maybe things worked out here could be applicable. If I'm to keep tilapia fish for aquaponics, the water certainly needs to be kept warm, and so does the greenhouse to grow plants. The tilapia tank (to be) is an apartment size fridge (insulated!) that holds water nicely if it's lying on its back, as I found when I first saw it that way lying on an old pickup truck. The idea for an insulated fridge or freezer I got from "cold weather aquaponics" on youtube, where others are also working out the challenges of growing food in cold climate greenhouses. If I'm able to contribute anything new or special to that work, it'll probably be in the LED grow lighting area and maybe renewable energy supply. (Perhaps I should contact them...?)

   The adhesive backed lights could simply be stuck to the underside of a metal strip -- eg, aluminum angle irons. These would let all the light shine down from above the plants, while the vertical part placed at the front would shield the LED emitters from peoples' eyes. This much sounds simple enough.

   Aquaponics is apparently an excellent way to provide food for people from small spaces with very minimal inputs in supplies: fish food, water (but it's mostly recycled), energy and labor.

Vertical Axis Wind Turbine

   I have again been thinking some about the VAWT idea. It has occurred to me to make a very miniature version on the 3D printer. Or I'd make a larger "real" version with the top and bottom pieces (and maybe a center piece?) to hold the blades cut precisely on the CNC router.

   And I ran across an interesting so-called "bladeless" turbine. Of course it isn't bladeless, but the blades are inside a housing something like the one I envisioned a while back to "aim" the wind at the blades better, but still more elaborate and allowing more efficient wing-like blade shapes.
   Of note, where I would orient this unit with a vertical axis and a vane so the housing would pivot about the shaft axis to face into the wind, the authors had it horizontal. In that case the entire unit would pivot including the generator at one side.
   It should work about the same either way, but with a vertical axis the generator can be in a fixed position underneath with no need for slip rings to wire it.

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

R & D Tax Credit Programs (Canada, history)

   In July I got some notice from CRA of several pages, saying that they wanted to do a finance and technology audit on Turquoise Energy's work. Being in the middle of projects I ignored it for 3 weeks, then read it and discovered there was a 30 day response time limit. From about the 10th to the 17th I was busy writing up the info they wanted preliminary to an inspection of my facilities and got little done on the actual work beyond molding 2 or 3 more sections of the big Electric Weel generator. I finally mailed off about 40 pages of info, a little about money but mostly about the seven projects TE claimed for last year. When I first submitted SR & ED investment tax credit forms in 2010, someone from CRA came to make sure I was doing what I claimed, and this is the first time since then they've wanted to do an inspection. I think that the Turquoise Energy News monthly reports being on line has been considered sufficient proof for quite some time.
   Notwithstanding the time it took me to get this together, seemingly the same as if my 20000$ tax credit claim for 2013 was for a million$, Canadian taxpayers should be pleased that the SR & ED tax credit program, based on a percentage refund of resources already invested by companies in R & D, is doubtless a pretty tough program to scam. And it's about the only government program open to those doing quite speculative inventive work. Most funding sources forget there's "D", development, in R & D, and it's much harder to get funding for creating new products than for "R", research.

   I remember the early 1980s... 1984 or so (the fermenting, enthusiastic early days of computing) when "Scientific Research Tax Credits" ("SRTC"s) were frist introduced and Frank Hertel scammed 70 million dollars of taxpayer money with vague, then unworkable promises of remote power meter reading, and Fitsch(sp?) of 'Fitsch Research' got 35 million. (Both fled to Venezuela with their 'winnings'.) Any number of smaller players also scammed millions, while those actually doing R & D couldn't get funding except through these high profile companies in trying to make themselves look legitimate.
   Hertel had a history of bankrupt companies behind him in Germany, as anyone who bothered to read his resumé and check out his references would have discovered. An astute prospective employee came back laughing after his interview and said the whole thing was an SRTC funding scam. Where was the government's "due dilligence"?
   Fitsch research tried to hire me in late 1984... not when I applied, but some months later. I got an interview when, as it turned out, they were about to be audited to justify the lavish amounts bestowed upon them. They wanted me to start immediately and go to California and design IC memory chips. I had good experience designing products using IC chips, not actually designing the chips. And it all seemed too rushed. Where were they when I was free and had offered my talents 3 or 4 months before -- and why hadn't they interviewed people in good time? A month or so later Fitsch was in the newspaper as the sheriffs were removing the furniture from their offices. I expect I'd have never got a paycheque and would have been left to hitchhike home from silicon valley.

Another Documentary on the crooked money system: Century of Slavery: History of the Federal Reserve

   In a similar vein to Mike Maloney's educational  Hidden Secrets of Money Episode 4  now comes James Corbett's  Century of Slavery: History of the Federal Reserve, 'an intense 7 months of work' according to Corbett. It's been released free at CorbettReport.com and on YouTube.com, and there's a full 90 minute version or a condensed 55 minute version (which is the one I watched). Corbett goes into more detail about the specific history of banking and money in the USA than Maloney in this very watchable and informative documentary. Both producers stress that the best way to help yourself prosper and keep from being inexorably bled by the deliberately complex and confusing ponzi game financial system, is to educate yourself about how it works.

Financial and Economic Collapse and Epidemics

   As economic conditions worsen and the financial system implodes, and with worldwide travel, it becomes more and more likely that some of the various serious diseases in the news - such as MERS, Ebola (evidently it's airborne, BTW), Dengue fever, and various virulent forms of flu - are likely to start sweeping across populations weakened by overcrowding, poverty and shortages of nutritious food. With a still growing global population (even tho that growth has at last slowed considerably except in a few countries), and grossly inequitable distribution of wealth, a crash of global systems and with it the "population bubble" has in fact become inevitable.
   In the middle ages black plague, coming "out of nowhere" via rats brought from Asia in ships, gradually killed over half the population of Europe in the only recorded century so far of global population decrease. (It was said that while no one was immune, the "lower and meaner" sorts of people were disproportionately affected. We might suspect that many of these were less "cultured" groups with poorer nutrition and lower standards of hygiene... such as it was back then.)
   In 1919 and 1920, at least Germany and parts of Europe were starving, and an influenza epidemic killed more people (tho worldwide) than the devastating first world war (fought mostly in Europe) that it followed. Antibiotics now available can mitigate bacterial diseases, but viral ones like flu still can still sweep through populations, as we all well know. Flu vaccines are apparently not very effective, and also vaccines have been implicated as a major cause if not the major cause of autism. (An interesting study showed the great flu epidemics were also statistically linked to schizophrenia: it was noted at a British psychiatric institution that schizophrenia was commonly grouped into very specific age groups, which correlated to mothers being apparently 4-5 months pregnant at the time of a flu epidemic. Since the correlation wasn't noticed until decades after, when the patients were adults, it wasn't determined whether the pregnant women had to contract the flu, or merely be exposed to it at the wrong time.)
   Vast tracts of the interior of my large province of British Columbia were planted with nothing but pine trees. A usually minor pest, the pine beetle, started multiplying and spreading until it infested the whole province, killing vast pine "forests" (plantations?) provincewide. (I expect most farmers can tell us the hazards of crop monoculture.)
   Starfish have been in the news a couple of times in recent years now. First it was because they had been multiplying in numbers to the point where it was feared they would utterly destroy all the world's coral reefs. But this overpopulation bloom has now brought a plague on the plague of starfish, and they have been lately in the news again, this time for dying off en masse of some disease that rots them away.
   When any species has grown in numbers too large for the available resources to sustain (in the human case, farming and land for all purposes), epidemics strike. Instead of a leveling off of population with barely enough for everyone, there is disease and a population collapse. With seven and a half billion people in the world today many vast areas are overpopulated compared with what current societal and technological conditions can support. Diseases will soon have a golden opportunity to wipe out maybe 2/3 of this mass of people to produce a second century of global population drop, and a very substantial one. I won't say the planet can't support this population. If seawater were desalinated and the deserts turned into gardens, if resource distribution and population distribution were equitable, it might. But would the population then cease to grow, or would it continue increasing until it was unsustainable at an even higher level?

   As it is, the three cosmic core values (quality of life, growth and equality) are out of reach for too many people. The reduction of pressure from our species will be a welcome relief to the Earth. The lesson taught by the global devastation will probably not be lost on the remaining populations everywhere. They will probably decide that having children in uncontrolled numbers isn't a right, and maybe that improving all the human races by selective breeding to reduce the proportions of "challenged" children and increase those of the more intelligent and physically fit, only makes sense. The Universal Father loves every person equally, but he doesn't need for large percentages of them to continue being born with various built-in genetic problems that detract from their personal growth potential and from building the utopia that every inhabited world should become.

Electric Hubcap Motor Systems - Electric Transport

Magnet Rotor Mechanical Improvement!

   My magnet attachment setup for the magnet rotors was the best I could come up with at the time. But all along, I've been somewhat uneasy about it. The polypropylene strapping is strong stuff, but keeping it attached to the rotors and the magnets glued to it, all withstanding high centrifugal forces at higher RPM.s without something delaminating, has been something of a challenge. It was worst with the first attempts at the Electric Caik rotor, where the slick coatings on the new magnets let them slip right out of the 'sleeves' at a very low RPM, and even after sanding them to give the epoxy a better grip and re-doing the rotor, I decided I didn't want to run it above 2000 RPM when I had originally intended 3000.
   I've thought of a new pattern or two for the strapping that would cover the outer end so the magnets couldn't slip out. I haven't made a new rotor since, so this never got put to the test. But the whole thing, epoxy, strapping, magnet and all, could still delaminate and pull away from the rotor with enough centrifugal force. Some way to definitely attach the inner end of the strapping so it absolutely couldn't pull off with any force less than required to rip the very strong strapping itself would be much better.

   In considering the magnets for the Electric Weel motor with its two-piece rotor on the 11th, I finally thought up a truly better answer: make a slot through the rotor just above the inner end of each magnet. The strapping can be pushed through the slot and wrapped right around the magnet on the rotor vertically (from inner slot to outer rim), overlapping itself on the back side, and of course all epoxied into place. Even a loose magnet couldn't slide out, and the strapping would have to tear to pull away from the rotor at the inner end.
   With this arrangement, which I will adopt for all, the rotors will be much more robust and able to withstand considerably higher RPM s without flying apart. Maybe I'll have the Electric Caik motor doing 3000 RPM or more after all... and maybe even get the 14' aluminum motorboat up planing!

   After over 6 years of motor development, here was a place to fix a remaining real design weakness that's been nagging my thoughts! Counting moving the bolts farther from magnetic fields so they don't get hot, that'll be two new improvements, and of course the motors will be better than ever!
   Rotor balance didn't matter much under 2000 RPM, but it does anywhere much above that. One can figure out where the imbalance is by putting the rotor on an axle and setting it on the axle between two perfectly level rails. It'll roll until the heavy side is down. Weight would then be added to the top side or removed from the bottom.

Unipolar "Bedini" Motor-Generators

   There's "efficiency" from 0 to 100%, but with some systems such as heat pumping, there's also "coefficient of performance" which yields energy performance substantially in excess of the input energy. Could there be such a thing with motors as well? Could my 95% efficient motors be radically improved? John Bedini seemed to have demonstrated it in the 1970s with his unipolar motors.
   The more I considered it, the more I liked the idea. With a regular motor, when a permanent magnet is between two coils, the coil behind pushes with an energization the same polarity as the magnet (eg, both south), while the coil ahead is magnetized to attract the magnet (ie, north). But the iron in the coil itself also attracts the magnet regardless of polarity, even if the coil isn't energized. If only the coil behind is energized to push, the coil core ahead still pulls on the rotor magnet, and regardless of polarity. The motor has magnetic "cogging" and wants to jump to certain points of rotation where magnets line up with coils.
   Furthermore if the coil isn't energized, it can be used to generate electricity just where the magnetic attraction pulls the rotor, and also the collapsing field from when it was energized can be recaptured. With light mechanical load, Bedini seemingly had it generating as much electricity as it used, charging a second battery even while being driven from the first. The rumored Japanese e-bikes I heard of last month, evidently using unipolar motors, apparently have far greater travel range than others.
   One could make a motor controller to do the Bedini 'trick' with bipolar magnet rotors - but with double the MOSFET drivers in order to activate each phase independently north or south. With unipolar rotors, they only need to repel one polarity so they only need to be activated in one direction, making the controller much simpler instead - with half the MOSFET legs (3) instead of double (12) of the common BLDC controller.
   The "Y point" of the coils is tied to the battery plus voltage ("B+") instead of left floating. The three phases of coils are then actuated independently just by pulling the other side to ground. So the 3-phase bridge changes to 3 pull downs, eliminating the high-side mosfets - half the heat-making power transistors and the complexity they add.
   Tying the "Y point" to B+ has another benefit: when the magnetic field collapses after powering the coil, the voltage reverses. With the Y point floating at 1/2 the supply voltage, this voltage doesn't get higher than the battery voltage unless it's greater than 1/2 that voltage already (and even then perhaps it just pushes the Y point around a little), so some of the resupply of power back to the battery is lost. With one side of the coil at B+, all unused energy (above the .5v reverse diode drop) is returned. Furthermore, a 'regularly' configured motor has to go fast enough to generate more voltage than the battery in order to passively recharge it. But with one end of the coil at B+, the coil's back EMF would generate electricity above the battery voltage whenever it's off and the polarity is right. (Do I have that right?  There seems to be no ground reference for the coil voltage at this point. This may be where Bedini charged a second battery - because the circuits didn't work out for charging the same one supplying the motor. To charge the same battery might need some sort of charge pump or an isolated output DC to DC converter. This needs more thought.)

   A complication with the control arises with the magnet sensors. With alternating north-south magnets, the hall sensors can be pretty much relied on to switch back and forth at the center points where the field crosses zero. There are unipolar hall sensors, and even with unipolar magnets, south and north fields alternate if the sensors are strategically placed, but the fields are unlikely to be even lengths. I can see having to go back to the optical sensors I was using before I found out about Hall sensors. Those of course will have to be mounted in the rotor compartment rather than more conveniently in the stator compartment. If I adopt lexan plastic rotors, a new idea mentioned in "Electric Weel Project" below, they could perhaps be painted with the appropriate stripes, and the LED.s and sensors be placed on opposite sides of the rotor.

Huge Torque, Low RPM  Electric Weel  Motor-Generator Project

   On the 8th I got back to this and I wrote the G-Code (mostly by cutting and pasting from the lower piece) and made the 3rd piece of the stator-rotor division mold with the CNC router. Again the parts would be in eight 45° pieces to be fastened together into a complete ring. On the 9th I finished tidying it up and drilled and threaded holes for bolts to clamp it together. I let the first piece harden up well before removing it, thinking it might just pop out easily, but I almost wrecked the mold getting it out. After that I waxed the mold.
   On the 10th & 11th I made a piece each day and 2 on the 12th. I finished on the 19th, skipping the odd day because of the time spent to write the CRA info. At the same time, I used the epoxy that squeezed out of the mold to start gluing the outer end (8 pieces molded in July) together, strengthening the seams with PP strapping on the outside. It's hard to do more than one in a work session because the epoxy in the mold has to set. Also it takes a while to clean out the mold and get everything ready for each one. It's far more work to do this large motor than the small ones!
   When I do a second unit, I can use the different molds and maybe do 2 (or even 3?) pieces at a time. (if I buy a bunch more 4" C-clamps.) But I took this one slow to see how the molds worked out - I might (and did) want to change something.
   Also probably worthy of note is that owing to some of the parts turning out a bit dry of epoxy, I'm changing the ratio of 4 epoxy to 1 polypropylene cloth, to 5 to 1. That's by weight - the light cloth is still over half of the bulk.

   On the 21st I needed supplies in order to proceed, and spent the day shopping. But I got a great deal on 1.5" PP strapping for the rim of the rotor compartment - a 60 or 70 yard roll of white that the clerk said had been 'kicking around a long time' and had become a bit discolored on one edge, for about 1/3 of the regular by-the-yard price.
   The next day (22nd) I decided I'd do the outer part with 2" strapping instead of 3", and alter the form for the top. Now the entire rim would be even height, and the cover piece would be the full motor diameter and made with a lip just inside of the rim to 'lock' it into position. This was in fact the reason I hadn't made the top mold piece(s) yet - just in case the design changed before I got there!

   I didn't have enough 2" strap and bought 15 yards at regular price. Then I made a big wooden ring to fit inside on top of the center body ring, to wrap the epoxied strapping around to make the outer part of the rim. (The wooden form was also modified - lower - because of the rim change.) I made it out of 2" boards screwed together. These would be waxed and a liner strip of 1/16" thick polyethylene plastic would be wrapped around the wood, to keep the epoxy from sticking to the wood. In the worst case, unscrewing some of the boards from each other should free up the wood from the rim after the epoxy hardened. (Wasn't necessary.)
   On the 23rd I epoxied the center body ring pieces together. The main trick to this was to make sure the 24 sets of coil buttons matched and the coils fit on both sides. With the stator end ring already made, later adjustments would be difficult. Luckily PP-epoxy has just a bit of flex. The other trick was to get them even so nothing stuck up to where the magnets rotate. I bought some plastic clothes pegs. I ended up just using the angled pieces as shims to adjust heights to evenness, but they were probably the best thing I could have used.

   On the 25th I waxed and then set up the wooden ring on the motor body center ring piece (I had to wait for Monday to buy some PE plastic to line the outside edge). I painted epoxy onto the outside of the body center ring and onto the 2" wide webbing, and wrapped two layers around the outside. Since the center ring's outside edge is 1/2" thick, the webbing stuck 1-1/2" up above, forming the outside edge of the rotor compartment. After a few hours I pulled the wooden ring out of the middle.

For wrapping the inside of the rim I "spooled" the strapping within, then started painting it with
epoxy and pressing it against the now solid outer rim to form a thick solid wall.

Electric Weel axial flux Motor or Generator composite body, 28" diameter:
Stator end is under midsection with thick rotor compartment rim (white).
(The pieces are held at finished spacing by unwound coil cores, not visible.)
Material: "Advanced" molded composite of epoxy, and polypropylene cloth ripped
into 6" square pieces or 2" & 1.5" polypropylene strapping or "webbing" (rim).

   On the 26th I used 390 grams of epoxy in three mixings and 7 layers of white 1.5" strapping wrapped around the inside to thicken the rim to 1/2". (it was a bit more - I'll do 6 layers next time.) I tried to paint fast, but the day being unusually hot and sunny for Victoria BC, first little tub of epoxy started to smoke and set up before I had finished. I salvaged the brush by putting it in the freezer. For the second two, I moved out of the sun, and I kept the plastic tub of epoxy in a slightly larger plastic tub of ice water. I tried to make a little video of Electric Weel Motor Construction  as I went, but without a camera person I couldn't show doing the actual work - I'd have gummed up the camera with epoxy.
   When I was done I set the piece in the sun to set. With the black plastic outside and white inside, the sun facing side got quite hot and was soon set, so I turned it around. A while later it was solid, and I sanded off some sharp edges with the belt sander. It had been an all day project, with breaks for breakfast, coffee, and later to lop off some blackberry canes that were threatening access to my carport.

   Despite good intentions, I didn't get to laying out and making the mold piece for the rotor end cover pieces in August. The way I did the rim, two of the three pieces for the stator end can also be used for the rotor end, and only one new piece needs making.

Lightest Magnet Rotors? - Electric Weel Total Weight Estimate

With a 26" diameter rotor, a 5/16" thick steel rotor would be horribly heavy. I've thought of various ideas for making it lighter. The way I've done the first one is with a 1/8" steel disk, with a 3/16" ring at the outside where the magnets are, welded together and making 5/16" thickness to carry the magnetic fields. I then had pieces of 3/16" disk left over: an 18" disk that I had planned to use for the center of one end of the plywood body, and a ring about the right size for a bicycle wheel magnet rotor but only about 1.5" wide when I have all the 2" magnets and coil cores.
   Before that, as an earlier TE newsletter or two show, it first occurred to me to make the whole thing as a frame, welded together. 'Spokes' with transverse cross section would give it more structural strength against warping and bending. That was actually my first plan, but as I put it together, I was losing confidence in the strength of my structure at higher RPM.s, and at the same time, I found the CNC waterjet steel cutting companies and did the above mentioned design.

   Now it occurs to me that a lexan plastic rotor, perhaps bolted to a steel center disk to connect it to the shaft, would probably be lightest. This would then have a steel ring on the outside where the magnets go. I may even cut off the welds on the present rotor and keep the 3/16" ring but replace the center disk with lexan.

Weighing the parts (# = pounds), I find:

Coil end of body - 8 #
Ring part of body - 3 #
Coils - 23 #
Shaft (short), SDS bushing, bearings, bearing holders, seals - 12 #?
Magnet rotor (metal) - 27 #

Attempting to estimate the total weight, I get:

Total body - 25 #?
Total Wiring - 26 #?
Rotor (metal) with magnets - 37 # ?
Shaft etc. - 12 # ?

Total: 100 # even.

   That's no doubt quite light for such a motor, at the bottom end of what I thought might be expected, and it'll certainly be easier to handle than 150 or 200 pounds. But the plastic rotor would shave a few more pounds off it.

(Let's see: area of circle = pi * radius^2
  = pi * 13"^2=530.9 sq". Area * .125" thick = 66.4 cubic".
 The outer 3/16" ring is about 10-7/8 I.D. so 371.5 sq". inside has no metal.
 530.9 - 371.5 = 159.4 sq". * 3/16" thick = 29.9 cubic"
 66.4 + 29.9 = 96.3 total cubic".
 rotor 27# / 96.3 cub" = .28 #/cub".
 ring 29.9 cub" * .28 #/cub" = 8.38 # weight of ring
 8.38 # - 27 # = 18.6 # weight, of main rotor piece, removed.

   Add the weight of the lexan and center hub for it, about 8 pounds, and the total weight reduction of the motor is about 12 pounds for a total of about 88 pounds. That's probably worth it, although it's still winch territory rather than "by hand" to install. And it's a little less steel to rust. OTOH there's now only 3/16" thickness of steel to carry the magnetic fields instead of 5/16". Since I don't know the optimum thickness anyway... I'll just ignore that!

   It would be nice to eliminate the steel altogether. I could then make the rotors by layering lexan type plastic (for the inner hub), glued with methylene chloride, and having the CNC router precision cut everything including rectangular holes for the magnets. But that's probably not feasible magnetically.

   When I went to Industrial Plastics they had some pretty imposing prices for the lexan. It looked like it might be a couple of hundred dollars. I wandered around to their off-cuts bins and shelves, where, as I remembered, they had quite a lot of scraps of something called "impact modified acrylic plastic" - probably as good as lexan - in pieces cut at odd angles for some job that was never completed: all 7.50$ each or two for 10$. They weren't big enough for this large rotor... but from the larger sizes two of them glued together with methylene chloride would be, with half the rotor from each piece. I got about 14 sheets (70$), all the ones that would work, adding to 4 I had bought earlier for flat panel LED lighting. The seam between the pieces will have another piece a few inches wide backing it, also melted to the main pieces with methylene chloride, eliminating possible weakness at the seam. In addition, the 'lexan' will be thickened around the hub to several thicknesses with a key slot on each side to secure it to its axle, and without an SDS bushing, thus almost eliminating the weight of that hub. (Hmm... under a pound, it seems.)

   I plan to cut the 'lexan' rotors with the CNC router machine. This cutting includes the center hole with the key slots, the outer rim, rectangles that the magnets will fit into (thinnest rotor and most secure magnets!), and slots for the PP strapping. With the CNC router and all cut in one G-code program and one operation, everything will line up "perfectly" with itself. The seam cover and other hub pieces will be done separately and all melted together (with methylene chloride as always) while mounted on the axle to prevent any possible misalignment.

J1772 Car Charging Plug - other safer charging plugs?

   With the 3D printer finally working again, on the morning of the 12th I printed out a plastic shell for a J1772 plug, the plug all those EV charging stations use, for the Mazda. I don't know whether the charging stations will kick out at the slightest imbalance between the two lines, or whether there's enough slack that it will work if I put half the chargers on one line and half on the other, using the ground as a neutral. But I intend to try.
   A safe outdoor charging plug that doesn't turn the power on until it has a secure connection is a good thing, but it seems to me dysfunctional to make the "standard" for it so that it's apparently unusable by the majority who have 120 VAC chargers, simply owing to it being missing a neutral pin. (I don't think you'll find a 240 VAC battery charger in a "box store" at a good price in North America unless it's a dual voltage one, will handle any voltage from 120 to 240, or is a special order... much less a model specific charger for your particular e-bike or whatever.)
   I had a hard time finding out the pin lengths and diameters. One site implied that the large pins were 3.6mm diameter. I gathered the small one just might be about 2.5mm. It didn't seem like much to go on. I didn't get it done.

  In the absence of any agreed practical 120VAC EV standard, I keep trying to think of a way to make a safer 120V EV plug-in. The best I've come up with so far is a regular socket (with a ground fault detector-breaker of course), but one where the last couple of millimeters of insertion of the plug pushes a button in the socket that activates a relay that turns the power on. The main problem I foresee with this is that people can plug the plug end in first, and then their extension cord is live before plugging in the vehicle end. Also the extension cord can easily be stolen while the vehicle is charging unattended. If on the other hand this switched socket is itself on the end of a (four wire) cord of sufficient length, that would greatly reduce the likelihood that people would use an extension cord.
   Perhaps the simplest arrangement of this sort would be a cord with a shortened blade for the live line connector. Then the ground and the neutral would make connection first, and the line only in the last couple of millimeters. I could see a grounded plate over the end of the cord as well - or on the socket - surrounding but (definitely!) not touching the live line. If there were any arcs in dampness, they should go to the metal plate rather than the fingers.

Electricity Storage

Turquoise Battery Project

Electrode Frames - Better 'Pocket Electrode' Design

   Near the start of the month I made some plastic electrode holding frames. I kept coming up with minor changes while printing them, and over some hours came up with quite a few (image to left).

   Then on the 6th or so I got an idea to make two part electrode frames that fit together like a two piece cardboard box, one's edges fitting just inside the other. The rear side would be solid plastic (replacing separately cut ABS pieces), with the terminal supporting tab as a piece of that. If the sides were a little shorter than the thickness of the electrode, pressure on the faces would clamp it together 'perfectly', and there would be little chance anything could leak out except through the porous face - with the thick watercolor paper covering it. A thinner briquette wouldn't have a hollow space to expand into, and a thicker one wouldn't create gaps at the sides. If experience shows any stuff still manages to get out the edges, they can be glued together.
   This hearkens back to nickel-iron 'pocket electrodes', but in plastic - something I've wanted to attain for quite a while, but only now at last found this good design for. One difference is that the porous front face won't be strong enough to retain expanding electrode material without bulging, so pairs of electrodes will still have to be clamped securely together. (Even Edison made pockets of round perforated "pencil" metal tubes so they couldn't bulge.)

    On the night of the 12th I worked through some frustrations with 3D printer quirks in designing the 'boxes', and printed 2 of each part before 1:30 AM.

Nickel Manganese Tests

   I put in a zinc strip during a load test to check electrode voltages against. I had assumed that the manganese had been discharging most of its energy over a few hours. Thus the rapid voltage drops in a load test would be due to lowering Mn negode voltages, and the drop slowed somewhere under 2 volts because the zinc conductivity additive would start to discharge, turning the cell to Ni-Zn.
   That wasn't what I was seeing with the zinc strip, so I continued the test down to lower voltages. Now it appeared that it was the posode that was rapidly dropping voltage during the test, until the voltage was under about 1.6 volts. The Mn stayed about .4 volts more negative than the zinc strip. Below about 1.6 volts, the negode started losing voltage - much more slowly than the previous voltage drops - while the posode steadied. I think it went from the nickel-manganate and nickel hydroxide voltage down to the manganese dioxide reaction voltage, and then down to the Mn2O3 voltage. When it was down to a volt, the negative was almost down (up?) to the zinc voltage. (But I also suspect the zinc strip was also changing voltage over time.)
   I have some ideas about this, but basically I'm not sure what will happen. So far I'm not seeing any further improvement in the self discharge, but I'll continue the deep discharge tests to low voltages. They may well be making some beneficial changes, especially in the posode.
   The next test (5th. 15:00 PM) showed little improvement in the self discharge (30': 2.518v), but my thoughts turned back to the borax which would probably form some borohydride in the negode. I dumped in about 1/4 teaspoon - twice what I'd put in before. Now something came back to me: the first electrode had been slowly improving. With more zircon (11%), the second one had started out better, but didn't improve. I would have changed the electrolyte unless it looked clean, which is unlikely. I don't remember adding borax the second time. No borax, no gradual improvement!

   I note that the little bits of white (KCl?) and green (CuO?) stuff that were exuding from the electrode terminals grew into fluffy crystal forests after the addition of the borax. (I've always seen polyethylene, and glass, as hydrophobic, but somehow even the glass marble on the lid becomes salt encrusted.) By the 8th, the cell was definitely holding higher voltages longer at first. But over an hour and more, it was losing considerably faster. It was as if there were two causes of self discharge. One was decreasing, but the other was increasing. I opened the cell.
   The electrolyte was black! Perhaps the same chemical reactions that were reducing the self discharge of the negode were contaminating the electrolyte and causing the second type of self discharge? But this could presumably be solved by changing the electrolyte. I put the electrode assembly in water for a while to help dilute the electrolyte held in the electrodes, and did so. The crystal forest on the lid quickly disintegrated in the water.
   Then I realized that there was only one cable tie holding the electrodes together instead of two. It was nearer the top, and the bottom had swelled up and was releasing black electrode powders. This was the last battery made before I got the 3D printer working again, so the electrodes were merely wrapped with watercolor paper, which had broken open again, and some PP nonwoven cloth that didn't have enough fibers to prevent leakage. The electrode pockets are looking more and more attractive!
   I used a little more paper and 3 layers of fat macramé cloth between them and stuffed it all together again. But I'm starting to think more definite results might need to wait for a new cell made with the glued 'box' pockets described above. (I only wanted to run a test here not get into more battery making work but the morning has gone - I was trying to work on the Electric Weel motor!)

Nickel-Nickel Cells

   The nickel-manganate positive has almost double the voltage in salty electrolyte that it would have in strong alkali. Choices for a really good negative side are more difficult: iron, cadmium, zinc and metal hydride all eventually succumb to gradual chemical changes and deteriorate, or even short out the cell. Some other metals with good reaction voltages, eg vanadium and chromium, have various soluble states and will rapidly dissolve. Two especially promising metals that look like they should last 'forever' are manganese and nickel.
   The reaction voltage of manganese is so high it's hard to make it work, as the ongoing thread of these newsletters shows. No one else has even got it to charge and hold a charge. I've accomplished that, but a gradual self discharge has prevented practical cells so far.

   Nickel metal has the unique attribute that it won't oxidize in pH 14 alkali even at a positive voltage, so it's never been considered for an alkaline battery negative. Instead it's been used for non-corroding current collectors -- which is the reason alkaline batteries became popular. But (as I finally understood) it will oxidize at any lower pH, so it can be used as a negode in salt solution, at least with a graphite based current collector even if no metal works. Like Mn, Ni should last approximately forever in a negative electrode to make 'perpetual' cycling batteries. The chief drawbacks of NiNi over NiMn are higher cost and over a volt lower reaction voltage, making cells only around 1.1 volts nominal. On the other hand, this voltage is so low self discharge should be no issue, and sealed dry cells should be practical.
   There's another factor: The current drives of my Ni-Mn cells seem rather low. Whether this is just my constructions or inherent in the chemistry is unknown at this point. It might be that Ni-Ni would have a higher current capacity, perhaps even a much higher drive, than Ni-Mn. This might mean that Ni-Mn would be great for low current devices, but not for high loads such as electric transport. Ni-Ni might turn out to be as good as or better than nickel-metal hydride for either dry or wet cells, and would be far easier and safer to make at home or in small production.

   At any pH from about 8 to 12-1/2, a cell with nickel in both electrodes should charge to about 1.25 volts and last an indefinite number of charge-discharge cycles. (I'm not sure where the Ni(OH)3- shown at pH 13 and 14 comes from - it doesn't seem to apply in today's alkaline cells. I also don't know how pH.es below zero and above 14 are obtained.)

The first thing I did, on the night of the 3rd and the morning of the 4th, was to make the electrode frames. Hopefully these could drastically cut or eliminate the oozings of electrode materials into the electrolyte.

   The next question was just what form the nickel electrode should take. With such a low reaction voltage - about .1 volts less than hydrogen - there should be no need for overvoltage raising ingredients.
   Should there be a metal to improve conductivity, and if so, what? Did nickel need one, or would pure charged nickel metal itself be plenty conductive? Zinc electrodes generally have no conductivity additive and have very low internal resistance. I couldn't think of a metal I'd want to use except maybe copper. If I used that, perhaps monel would be a good form. But copper has a lot of soluble reactions.
   Then again, with such a low reaction voltage, would carbon black or graphite be suitable? And if so, what about using graphite felt? And would a graphite foil current collector need doping?
   I decided to go for nickel oxide (nano particles) for the nickel substance, and graphite felt with a graphite foil current conductor for the first try.

The usual battery reaction to NiOOH valence 3 isn't shown in this chart,
Instead its voltage of .49 is shown as going to NiO2, nor is a valence 6 shown anywhere else.
This may be more errors in the charts I was using all this time until I discovered Pourbaix diagrams.

Nickel Negode

   Working out amp-hours, pure nickel (discharging to nickel hydroxide) would have 914 amp-hours per kilogram. This high figure surprised me, given that nickel hydroxide as a positive is just 289 theoretically... and realistically maybe 1/3 of that. But why should it have? Zinc is 820 and manganese is 976. Nickel is between them in atomic weight and all of them move 2 electrons per reaction. Nickel discharged to hydroxide in the negative side, counting the "(OH)2" mass, would be 579, or nickel oxide (NiO), my planned starting form, would be 718. If the voltage seems low, at least that's good amp-hours! The NiO should charge to Ni metallic nano particles and then discharge to nickel hydroxide and never be NiO again.
   It seemed there was nothing else to add unless it was a percent or two of "vee-gum" - a bentonite clay to "glue" the briquette together a little better. But better than what? Zinc electrodes were flimsy, but nickel ones might be fine with nothing, and it would have the felt anyway.
   On the evening of the 4th, I put the felt in a small jar, filled it with powder, tapped and refilled until the felt seemed saturated. The triple felt layer was about 1.75g, and the nickel powder in it was about 10.5g - so about 7 theoretical amp-hours. Then I dripped in some Diesel-Kleen, Sunlight dishsoap and water and pressed it to 10 Mg (mega-grams - tons), 625 Kg/sq.cm and left it in the press for a while. When I took it out, unlike most electrodes it was somewhat flexible instead of brittle. This may be from using too much liquid. Some oozed out in the press.

   My [pottery supply store] nickel oxide was black, indicating that it was non-stoichiometric, which means that some nickel atoms might be at valence 3 (or 1 or 0?) rather than 2 and the crystal structure isn't entirely regular. This is probably an advantage in conductivity over green (stoichiometric) nickel oxide... unless it has deleterious impurities. I chose it over my turquoise colored nickel hydroxide also because it was more dense. I thought the fluffy hydroxide might charge to rather loosely packed metallic particles, reducing conductivity. However, whatever form it starts as, it will all discharge to hydroxide. Maybe I should go back to adding thiamine for a better size balance between charged and discharged?

Posode (with graphite powder)

   Next I pressed the posode. Since the graphite didn't seem to be the cause of the self discharge, I used some powder with graphite that I still had in a jar. (When I get the fine conductive carbon black in a month or so, I think it should at least double the current capacity over graphite powder. For now I just want an electrode that works.) I pressed it to 8 Mg. It was only 5g of the lighter powder in the felt, and it compressed to a thinner electrode. This is backwards, since it needs at least 1.5 times the active material to match the negode's amp-hours, even in theory. If it gives 1/2 an amp-hour I'll be thrilled. Oh well. "Works" is the main point here.


   I left the electrodes to dry overnight and on the morning of the 5th I singed them and then assembled the cell. This time I remembered to paint the calcium hydroxide layer onto the posode current collector sheet. I didn't however do the osmium doped acetaldehyde. This was probably a mistake, as the initial current capacity seemed very low and the cell had to be charged quite slowly - 30mA and it was almost at 2 volts. This was also probably from using straight nickel oxide with no metal in the initial negode. (Why don't I think more often to put in  zinc strip to help see which electrode is doing what?)

   Putting both electrodes in the little frames made the work easy. The frame edges took the place of wrapping the paper around the briquettes, so just little squares of paper (~41x44mm) were needed for the active faces. These were placed in the frames in advance with one edge folded up a bit, then the briquette dropped in. Behind each one goes the current collector, and then a heavy piece of plastic to hold everything stiff. A cable tie wraps it all up into one layer cake assembly. A bit of RTV silicone glues the graphite foil terminal tab to the plastic tab piece.
   Next holes, slots, are made in the lid of the jar for the terminal tabs. When they are pushed through, They're glued with heat glue or RTV. If desired, a round hole is put in the lid for a filler and inspection hole. This is covered with a small glass marble (from Michael's Crafts store floral arrangements section). But the hole isn't vital. The jar can be filled first, and can be unscrewed and opened, with the electrode assembly attached to the lid, not to the jar. (Good for changing the initial electrolyte if it gets dirty, too.) A pinhole for emergency venting might be nice - otherwise I'm sure something will give to release any dangerous pressure buildup without a blow-up.


   After a while and the usual internal leakages of materials - and self discharge somehow just as bad as the manganese in spite of the much lower voltage - I gave up testing the cell. I decided to assemble a new one using the new plastic electrode 'boxes' that would hopefully last. On the 19th I made a briquette with monel pressed to 10 Mg - no graphite felt. It was so crumbly that I dropped it back into the mortar for remixing, and got out the Veegum (a bentonite clay), intending to add a few percent. On the 23rd I finally got back to it. I added around a gram of VeeGum. (somewhere in adding and subtracting and forgetting I managed to confuse myself as to just how much it was - .75 or 1.75g.) Considering the crumbling briquette and thinking how hard monel is, I pressed it this time to 15 Mg instead of 10.

   Before making the posode, I decided to wait a few days for the conductive carbon black I ordered, which was to arrive just about the end of the month. But it still hadn't by September 2nd.

Lead-Acid Battery Renewal

   The second battery I renewed (of three seemingly identical size 27 batteries) behaved rather differently from the first in the Mazda. Of course lower voltages are to be expected from higher pH, but they both got about the same electrolyte. They would both eventually charge up to full voltage, but the first one, as soon as it was used, would drop down to about 7 volts. Under load it would drop to 6, 5 or even 4 volts, but it stayed about the same while driving several miles. The second one only went down to 10 volts. But over about 3 miles the voltage dropped more and more, until the voltage display was winking out at about 2.5 volts under moderate load and it may have been going to zero or even lower. I tried not to stress it too hard, but to drive anywhere I kept finding it failing on the return trip.
   For a while, neither of them seemed to improve markedly over time - the first to get to higher voltages, nor the second to either go higher or to last longer. (If it's deteriorating further it would be no surprise.)
   Finally the second one started going to higher voltages, and one day (25th?) it stayed high longer, but I drove too far that day (owing to a rather long detour) and it rapidly went down to nothing again after a certain point. I stopped and disconnected it on the way home since I was doubtless damaging it. But now the voltage (with no load) stayed up higher longer, and on the 28th I reconnected it to see if it was better, the same, worse, or essentially shot. It didn't seem to last as long. On September first, with the car having been plugged in and no driving the previous day, the battery started out at just 10 volts, notably lower than it had been doing. It dropped fairly rapidly and I had to disconnect it again. I can only assume the pulse charger hadn't kept it up once it got it there. It may have decided the battery was bad and quit or something.
   It will be interesting to try the third battery, which I haven't refilled yet.

Victoria BC