Turquoise Energy Ltd. News #29
Victoria BC
Copyright 2010 Craig Carmichael - July 1st 2010
  *** Happy Canada Day! ***


  June in Brief (summary)
    * This month: Torque converter, solar cells, Ni-Mn batteries
    * Democracy for Canada: On-line referendums

  Editorial/Review: Documentary Movie The Corporation - the pathological pursuit of profit and power
    * The dark side of society, the economy and politics today.
    * ...and even with all its fine insights, it doesn't touch on what goes on to deliberately keep us using as much oil as possible!
    * Attainder looks more and more appealing.

  Electric Hubcaptm ("EH") Car Drive System Project
    * 3rd party extols the virtues of axial PM (eg, EH) motor format
    * Motor fixed. Motor & controller working fine.
    * Supermagnets: protect the neodymium... even from dry air!
    * Coffee can jig for getting metal hub past magnets to assemble rotor
    * "Bare-bones" EH motor weight: 37 pounds (17 Kg)

  Mechanical or Magnetic Torque Converter Project
    * The new magnetic radial design.
    * Magnets edge-on are much better than face-to-face!
    * Swiveling magnets permit one-way force.
    * 2" magnet edges (over 30 pounds(?)) are much better than 1" (15 pounds)
    * Arms meet & magnets stick - arm swing limit stops needed... made.
    * Tests reveal all the desired properties except one... the force is much too light.
    * Installing enough supermagnets to provide enough torque would seem to be impractical.
    * Clock escapements again: this idea, which I keep harking back to, is probably what'll work best, but making one that'll take and transfer the forces without busting...
    * Next design figured out: Lots of small nylon escapement 'anchors' (however many are required!) each with small force should add up to large force.
    * Recap: the several designs since this project commenced.

  Nanocrystalline Solar Cell Project
    * Metal grid rear electrode: no more loss of light than 'transparent' tin oxide?
    * Glaze mix 9 is probably fortuitously close to an 'ideal' clear nanocrystalline borosilicate glaze mix.
    * Tartrazine (common yellow food dye) seems like a great dye to use.

  Turquoise Battery Project
    * MnO2 discharge with H2O2(?)
    * Sealed Ni-Mn test battery
    * Measured very low electrode conductivities by voltmeter (large voltage gradients through electrode thickness during charge): this, not chemistry, explains many of the problems I've been having all along.
    * Seemingly electrode compaction is insufficient for good particle contact; even more is required.
    * 12 ton hydraulic press purchased for electrode compaction.

The Lead-Acid/Sodium Sulfate Battery Longevity/Renewal Project
    * Sodium Sulfate 'battery treatment kits' For Sale
    * (Fairly) comprehensive info is now on the web at:
    * According to local battery pro, the best way to access batteries with glued-on lids: drill 1/4" holes in the 'solid' side of the lid, plug them with heat gun glue. ("Rubber stoppers always leak sooner or later.")

Newsletters Index/Highlights:     http://www.TurquoiseEnergy.com/TENewslettersIndex.html

Construction Manuals for making your own:

* Electric Hubcap Motor
(latest rev. 2010/02/xx)
   - the only 5+ HP motor that can easily be made at home?
* Turquoise Motor Controller
(latest rev. 2010/05/31)
   - for the Electric Hubcap. (Probably there are commercial controllers that would work, too.)
* 36 Volt Electric Fan-Heater
   - if you're running your car on electricity, you'll want a way to defog the windshield and keep warm.
* Lead-acid battery longevity treatment - "worn out" battery renewal procedure.

all at: 

June in Brief

   This spring Turquoise Energy Ltd. has received several thousand dollars from Canada Revenue Agency's Scientific Research and Experimental Development Tax Credit (SR & ED) program. 35% of what the clean energy projects have actually cost me in the last two years has been reimbursed. It's not the kind of salary most experienced professionals would expect or that I ever received for software developing or just working as an electronics tech, but it's most welcome to an independent inventor. Most major inventions, the ones that really change the world, are "pre business stage", and no one wants to invest in them or fund them. SR & ED is also the only government funding program that really seems to fund the "D" in "R & D" as far as I can discern. And now being 55 I deferred my homeowner taxes, a major relief. I'm out of debt, I have money to spend on supplies and tools this year, and I even got a real TV and have signed onto cable for a while! (As for replacing the 1984 car...)

Torque converter side of motor rotor, with 5 pivoting arms.
The arms match 5 supermagnets on the inside of the drum driving the car wheel.

   Having again conceived what appeared to be a promising design, I went to work on the torque converter, with a pause to fix and test the motor itself after two nickel plated magnets came unglued at high speed and flew off in May. The rim magnet and even more the swiveling drive arm pieces were somewhat intricate, and with five of everything, consumed many hours in the making and adjusting. By the 18th, it was ready to test. The test showed it seemed to be doing all the right things, but the forces created were surprisingly weak. I decided that rather than multiply the mass of supermagnets by an amount that would be hard physically to fit in, it would be preferable - again - to try something else.
   That something else will be something along the lines of the "anchor" clock escapement mechanism, but using multiple small nylon "anchors", spaced in 12 divisions around the rim of the rotor. This time, weak forces will be expected, but they can be multiplied until they are strong in total. If 3 or 4 or 6 anchors don't provide enough force (likely), I'll try 9 or 12. If that's still insufficient (less likely), there's 18 or 24 in two columns. If that's still not enough (hopefully unlikely), I'll make heavier ones. For the motor to turn vis a vis the drum, all the anchors must swivel back and forth, rattling in and out of "slots" and "teeth" in an aluminum ring around the inner rim of the drum.
   As with most or all mechanical torque converters, the operating principle is that these pieces will swivel effortlessly at low speed, but put up more and more resistance to changing their direction of motion as the speed increases, in accordance with the square of their speed, which is dependent on the speed difference between motor and wheel.

   I managed to put in just a bit of time on the DSSC solar cell reflectors, painting underglaze and glaze and firing a few tiles and melt-quenching them. Results looked very good in effect, though my ingredient mixing and technique could be improved to provide a more uniform result. It may be possible to make a yet better nanocrystalline, clear borosilicate glass glaze, but the characteristics seem great and currently I think beating the "glaze mix 9" I did last month would take considerable doing for perhaps marginal improvement. I picked out toluene as a potential electrolyte solvent, and tartrazine (yellow food dye) as looking like a good photosensitive dye. Its maximum absorption centers on 427nm, the blue wavelength of maximum solar energy.

Two tiles with zirconium silicate white underglaze and two thin layers of glaze, melt quenched both times.
Third one is without underglaze. The glaze has pulled microcrystals out of the porcelain near its base.
One (with underglaze) and only one glaze layer, showing many microcrystals. The leftmost piece looked akin to it before the second layer was done. Some paintbrush strokes are still visible.

   And I spent a few days on nickel-manganese/salt batteries. I learned a few things and I think working batteries of exceptional energy density are close. A better electrode compactor seemed to be needed, and I spent the last couple of days of June looking at hydraulic presses, which were much cheaper than I'd imagined. I bought a used 12 ton one - a common 12 ton hydraulic jack in a heavy steel frame - at Barclay's exchange for just $135 (with a couple of easily replaced loose pieces missing). This would seem to be the way to go!

L: Manganese 'negatrode' with copper mesh electron collector after (apparently insufficient) compaction.
R: Ultimate goal - to give something not so much bigger than the front case the stuff
to make it roughly the useful energy equivalent of the lead-acid battery behind.

   And I'm thinking: here's July coming, the nicest month to be down by the sea, and I haven't put together the wave power unit! Ah well!

   I also had another exciting idea in June, unrelated to clean energy. Democracy works by participation, but opportunities for the majority to participate seem to be strictly limited to deciding who to have as your legislative representative, which choice - "by the way" - will also determine something else entirely: which partysan politician gets to inflict his own personal agenda on the whole country or province.
   These leaders like to claim that the silent majority "have given us a mandate" for this, but there is no way to find out what that silent majority actually wants. It is silent because it has no voice. Instead, the most vocal and insistent, or the most powerful or influential, are heard, and their clamouring voices are taken to be public opinion.
   I now plan to set up a web site where anybody can start a petition about anything - local, regional, provincial or national, and every applicable Canadian who cares to can vote on line by ranking the available choices.
   (The choice ranking vote ("The 1, 2 vote", "STV", "automatic runoff") is the only single-ballot system where the voter is completely free to vote exactly as he feels would be the best outcome, second best, etc, from any number of choices. It wins hands down over the primitive "X" vote system. The outcome is not prejudiced by any number of available choices having any number of variations on the themes - there's no "vote splitting", and no "strategic voting": no "choosing the lesser of two evils" in hopes of preventing a completely contrary result.)

   Although the petitions would be 'unofficial' unless and until the system is some day endorsed and taken over by our governing institutions, anyone who wants to know what the public really thinks about an issue can simply go to the web site and look it up. It is also likely that the number of voters will be a good indication of how strongly people feel on the issue. The more who participate the more respect the system will gain, and it'll be easy to participate - just go to the web site, read and decide. It will be much harder to push an agenda if it's known that the great majority don't support it. Those who agitate ceaselessly to get their own way will be shown to be the small minority they are. When an unwelcome developer threatens to sue a city if it doesn't do as he wishes, city council will have much more power to stand up to him if they can say "Sorry, the citizens have rejected your proposal." Conversely, it may be found that things they are reluctant to endorse may be shown to be quite acceptable to the public.
   There will be a lot involved in setting this up, and I won't get it going this year if I start it on my own. Then there will be a lot of publicity needed so that people know about it and how to use it. There are also serious issues of cheating (ability to wrongly cast multiple votes) versus having a vast collection of data on people so their voting eligibility status is known, and maintaining the security of that data.
   But I trust I won't be acting alone on this once people start to grasp the tremendous potential of having such a system in place.

The Corporation Documentary

   I watched a documentary video, The Corporation - the pathological pursuit of profit and power. (Mongrel DVD. It looks from an insert like it's also a Penguin book, by author Joel Bakan.) I watched it in a few short sittings: it was too much to absorb, and much of it was too disturbing, to take in all at once. I think it's a "must see" if you want insight into the real workings of the economy and politics - into a good part of why things are the way they are and perhaps why we're gradually losing our freedoms, inching for many decades towards being police states. First the various ways in which the actions of corporations accurately mimic the behavioral profile of psychopaths are discussed, one at a time. Then there are various specific subject areas and cases, some really shocking.
   Did you know that as well as being "Rah, rah!" for Mussolini and Hitler, big business almost turned the USA into a fascist state in 1935? They didn't like Roosevelt's "New Deal". Think of the world we'd live in today if they had been successful! The rather surprising betrayal of the plan to congress by the ruthless general they picked to be dictator derailed execution the planned coup, for which half a million soldiers or disgruntled veterans had been earmarked.
   Things long considered to be basic human rights such as drinking water and saving crop seed have been privatized at the behest of corporations to be sold for profit to the people that should already own them. But even the psychologically fine-tuned persuasions of advertising to get children to nag parents to buy things they won't want (not to mention adult targeting) don't escape the documentary's purveyance.
   Our society benefits greatly from trade and industry, but the dark side of the way it's run today is horrific. A corporation is a "person" in the eyes of the law, yet it never grows old and dies, and if it kills, steals, bribes, kidnaps or destroys the environment, the worst it faces is a fine - a calculated business expense. So it rolls on and on, an invincible, unaccountable, psychopathic juggernaut.
   My one note is that for all its revelations, the video doesn't seem to recognize what now seems plain to me: that big oil, the automakers, and the battery companies are all owned not by random profiteering stock market swingers, but by one group: a clique of ruthless gangsters who work deliberately and unceasingly for a century to kill all alternatives to gasoline. They are rotting our society from the inside.

   Identifying the problems is the first step. Determining good solutions is harder. The Russians revolted against the Czar only to find they had raised a new and worse form of dictatorship, Bolshevism.
   What are good solutions to the many unaccountable, rogue corporations and their owners who so control and manipulate our economies and dominate our policies and politics, gradually destroying the planet and working so much harm to human freedom, safety and security? Near the end of the film, some cases and solutions are discussed.
   It turns out government has indeed reserved to itself the power to dissolve corporations - it just never does it, regardless of the current provocation or the company's history of continuing lawlessness or harm. And in California, even regardless of a citizens' group formed to demand that one particular state oil company be dissolved. The attorney general wouldn't do it. There seem to be too many politicans who've gained power by representing big business interests rather than the real national interest.
   People have taken legal action, sometimes successfully, against the most flagrant abuses, and in one case shown it came down to actual revolution and bloodshed - but it's all in the film.

   The idea of attainder has grown on me. Why should it only be people working for progress - inventors like Rudolf Diesel (murdered for his engine), Churchill, Roosevelt, the three Kennedy brothers - that have to look over their shoulders, hit the dirt when a car backfires, and wonder when their life might suddenly be violently snuffed out? Why not also the most successful criminals as well, who often perpetrate such violence, and who are so rich and powerful that they seem to be above ordinary law? It would seem that only such a direct act by the peoples' chosen representatives can terminate their ever growing domination of society and commerce, and return their vast stashes of well-laundered but essentially ill-gotten gains over to the public treasury.

   Perhaps fittingly, though the film features many American corporate insiders and critics (including corporation presidents and a surprisingly humble Michael Moore), the end credits show it's a Canadian production with public financing from a number of Canada's leading sources.

The Electric HubcapTM Vehicle Drive System

Commendation of EH type motors

   Here's a commendation of the Electric Hubcap (EH) type of motor from a company (Apex Drive Labs) making one of rather similar configuration. I downloaded this a year or two ago and just ran across it again. Two points of interest are first that their gap from coils to magnets is, like on the EH, unusually wide - about 1/2 inch, and second that they use two stators to drive one rotor, providing almost two motors in one thicker package.

The coils of this motor are "U" shaped, with the two ends
meeting an inner and an outer magnet on the central rotor.
(Unfortunately, no motor weight was given, so there's no
basis for power to weight ratio comparison with the EH.)

   Most radial flux motors have hundredths of an inch flux gap. The .5 inch gap (which any "regular" motor person would think must be a typo) means the useless static pull from the magnets to the stator is far less for the axial flux type, which leads to lower bearing wear, no gradual loss of magnetism in the supermagnets, and higher efficiency.
   I suspect the efficiency improves especially in larger power sizes - I've heard it said that around 10 HP up the PM motor becomes less efficient than the induction motor. Inefficiency in the larger sizes must surely stem from the tremendous static forces between the magnets and the stator -- in radial flux designs. It seems likely Apex derived the wide flux gap from mathematical modeling. I got there by experiments that showed it worked best, without at the time knowing others were using a similar gap.
   Apex's having two stators - one set of coils on each side of the rotor - helps to compensate for the low magnetic flux of electromagnets compared to supermagnets. It might almost be said to be two motors in one. At the time of the article, they planned two direct drive motors, on left and right wheels, for a small Neighborhood Electric Vehicle (NEV). Evidently by the watt specs they have almost four times the power of the EH, so almost 8 times the power with two motors - 40 HP. This is far more power than is required for a NEV, but without a torque converter or gears the huge size would be needed to attain sufficient torque to get the car going up a hill.

   "It is widely recognized that axial flux permanent magnet (AFPM) machines usually have higher torque densities and efficiencies than their radial flux (RFPM) counterparts. Its pancake shape geometry and high torque capability make AFPM motors a preferred choice for direct-drive systems. In recent years, many different topologies of AFPM machines have been developed and reported [1].

   My comments on this: Yes, higher torque, but for direct drive of the wheels, it's still not commensurate with what's needed except with very oversize motors. Hence gears, or better, an efficient torque converter, allows a much smaller motor. (Since there's no ultra-high permeability soft magnetic materials or superconducting coils to raise the electromagnet flux and hence the torque.) Yes, there can be "many different" axial flux topologies, but I'll bet the Electric Hubcap is by far the easiest one to make, having standard auto and trailer axle parts for its main structure and common boxed strip nails for coil cores! The parts are all readily available (...no "by prescription only" or otherwise banned or controlled items) and any motor repair shop could easily make them.

Repairs to Magnet Rotor

The rotor, motor side: - new flat head screws help to "clip in" the magnets so none fly off again.
Note also how solidly the lug bolts for the torque converter arms are pounded in: at least one of the heads is visibly flattened and dented. One of the four lug bolts for the axle hub had to be removed to fit all five for the torque converter.

   Drilling holes in a magnet worked, but was difficult, and each hole broke out a clump as the drill neared the bottom. I turned the magnet over and used the divots as the recesses for the flat head machine screws. (This was the magnet that the nickel plating came off of.)
   I soon spray painted this assembly well with rust paint to prevent corrosion of the rotor and the magnet iron and the vulnerable neodymium. Someone tipped me off that the neodymium degrades from oxygen even in dry air, turning into non-magnetic oxide powder, so I spray painted all my supermagnets, which are mostly scratched up from use on various rotors one time or other. (So that's why the older ones always seem to look worse and to be weaker than I remember!)
   I glued the magnet on, then drilled and threaded the holes in the rotor. The threading tap broke off in the last hole - impossible to do anything with, so I put one screw on the outside of the magnet.

   It's also worth noting that I use a big tin can with the ends removed (and a slit cut in the edge so it'll fit in) when I put the axle hub on or take it off. (I've been doing this for some time but have previously neglected to mention it.) Otherwise, the hub always ends up latching itself onto a magnet with considerable force, making the job very difficult and rather hazardous. Even at that, the hub has to be worked into place onto the lug bolts. (The tin can, though it's also attracted, is thin metal and much easier to manipulate around the magnets.)

Motor Weight

   I worked out the weight of the bare EH motor if it was made now with the rather light 6129 disk brake rotor disks, which seem (without having tried them yet) as close to ideal as will be found using ready-made parts. It would weigh about 37-1/2 pounds, 17 Kg. (Parts cost might be around $300 to $350. "Bare motor" means: no mounting straps, no cover, no torque converter specific pieces... except two trailer hubs and the 6" flanged trailer axle... it would only need one hub and perhaps a short axle if made for use without a torque converter.)

Axle (6" long, 1-1/16" diam, with flange) - 1.6 (all weights in Kg)
4 bearings, axle nut - .37
2 hubs (length turned down to both fit on axle, diameter turned down to fit in 6129 rotors) - 2.6 (est)
2 6129 Disk Brake Disks @ 3.0 - 6.0
9 coils @ .46 - 4.14
12 magnets @ .125 - 1.5
4 lug bolts for rotor - .15
other nuts and bolts - .14
Wires, plugs, Hall sensors - .5

  Total 17 Kg, 37.5 pounds

   This weight means little to the car overall. And with the link pins, it will pivot when the car wheel hits a bump so it's largely not "unsprung weight". Still... lighter is better.
   Potential places where weight could be further reduced:
* Replace the stator hub and its brake rotor disk with a single flat steel disk, with a short pipe coupling welded to its center to mount the bearings. That might save about 1.25 Kg (2-3/4 pounds).
* Replace the 1-1/16" axle with 1". This would save only a bit, and since the flange is needed on the end, it seems impractical to replace the trailer axle with a 6" long, 1" diameter bolt.
* Drill holes in the rotor disks and cut out some unneeded metal in the spaces between coils. The holes would also act as ventilation holes, but otherwise, I'm guessing it would be a fair bit of work for minimal reduction. (You don't want the rotor metal much thinner, and be careful where you cut - the metal completes magnetic circuits between coils/magnets.)
* IF nanocrystalline ceramic cores can be produced, they would be lighter than metal by around 3.5 pounds. This is the biggest potential reduction: 9% - IF.

   If there's going to be a torque converter, replacing the magnet rotor assembly - disk and hub - might make it harder to build with only a marginal weight reduction, so I haven't mentioned that idea. On the other hand, the trailer hub has to be turned down on a lathe anyway.

Mechanical or Magnetic Torque Converter Project:
Torque Leverage Without Gears

   When I first screwed magnets into the aluminum drum, I didn't concern myself with north-south polarity. Either polarity would attract the steel arms, right? Then I realized that the steel arms as the rotor spun would be magnetizing themselves first in one direction, then the other, and it takes energy for that to happen - a needless inefficiency! When I remounted them on the outside, I set them all the same way.

First I mounted the magnets on the inside of the rim.
Then I realized that the 12" aluminum pan was practically non-magnetic,
and that an extra 1/2 inch of radius could be gained by putting them on the outside.
It's now about as big as the 13" car wheel allows -
any larger and a flat tire would cause damage, with the magnets hitting the road.

   On the evening of the 4th I finished making the new pivoting arms to match the diameter of the large aluminum drum. I did them by bolting new pieces onto the original (2nd set) arms. Two piece arms were necessary anyway because they were now longer (2 & 9/16" from pivot point to tip) than the distance between adjacent pivot bolts, so full length straight arms couldn't have been threaded on.

Motor rotor/torque converter input rotor with pivoting torque converter arms June 4th,
"upside down" in torque converter output drum.

   The arrangement has a side effect: the inner end of the outer arm pieces hits the center hub edge and limits the amount of swing of the arm (2 or 3" from one side to the other - can be increased by grinding the inner end of the piece). Whether this is detrimental, valuable or neutral remains to be determined.
   With the limited pivoting, the arms won't swing back far enough to fit in the blocks intended to hold them back until the center of the next magnet is reached. I'll try without them first. The arms are longer now and with more inertia should take longer to swing out, which could solve the problem 'automatically'.

   On the 5th I mounted the rotor and the drum on the axle and tried turning the rotor, holding the drum steady with a stick. The arms were about the right length, and the the force impulses to the drum seemed to be pretty much unidirectional, but they seemed much too light to move a car. As before, the greatest impulses were with the rotor going a very low speed. The rotor was easily turned; obviously the motor could take a lot more without being in danger of being stalled - this would be a benefit of the increased outer diameter.
   I could easily double or even triple the amount of steel attracted to the magnets, but I suspected that would still be insufficient. The way to get a much greater field and force would be to put supermagnets on the ends of the arms and have supermagnets attracting supermagnets.
   This brings the design back to being similar to one of my early ones (that I didn't finish putting together owing to lack of confidence in it) with a sliding aluminum "flat donut" plate that could pivot back and forth about 2-1/2" (from which the five slots in the drum rim remain), but with three key differences:
 1. the mechanical foundation - pivoting arms, is both simpler and more robust than the sliding plate. It should be much less subject to heat and to wearing out.
 2. there are no axial forces acting on the magnets: the arms pull towards the rim rather than to one side.
 3. The hinge of the pivot pins is only about 2/5 of the way out to the drum, giving the motor a leverage that allows it to start turning more easily in the presence of heavy magnetic cogging forces. The amount of the leverage depends on a number of factors and can't simply be ascribed as being 5:2.

   On the 6th, an idea about something someone said some months ago, which had been in the back of my mind, came forward. He said something along the lines of "It's really hard to pull magnets straight apart, but they can be slid apart sideways much more easily. Your design is sliding them apart; it needs to be pulling them apart to get the main force."
   I decided if I was going to put magnets on the arms, perhaps they should be edge-on instead of face-on to the drum rim magnets. Turning the magnetic lines of force would have a modified effect. Experimenting with a 1"x1"x.5" magnet held with pliers seemed promising, then I thought, what if the drum magnets were also edge on? I put another 1x1x.5 magnet in the side of a vice and put them both edge on.
   Sure enough! There was a slight repulsive force as the magnets approached each other from the left or right, then a very powerful attractive force once they were in line. They could easily be slid apart up and down, but to the left or right was very difficult: although they were "sliding past" each other, it was closer to the level of force of "pulling apart" rather than "sliding apart".
   Note: Using a piece of unmagnetized steel with the magnet edge on showed very low levels of force. If a magnet is to act on steel, a pole facing the steel is better. Forces between two edge-on magnets seemed to be a magnitude greater than any configuration of magnet on steel, so magnet to magnet was the obvious choice.

   I expected that this force could be multiplied to any desired extent by stacking magnets. If I stacked two rim magnets to get a 1"x1"x1" cube, oriented edge facing in, it should almost double it. Even better would be magnets that were magnetized "sideways" compared with the usual. With a similar double magnet on the arm a huge force would be needed to pull them past each other.
   Instead when I ran the experiment it didn't feel any stronger with the doubled magnets - perhaps not even as strong. To verify this I got a fish scale (no fish were harmed) and measured the amount of force needed. With single magnets, it took 12 pounds of force to slide them apart with a .08" sheet aluminum spacer between them (similar to the minimum gap they might actually have installed), and 16 pounds with the magnets touching. With double magnets making cubes it took only 9 and 12 pounds! It seems the "sharp edge" of magnetism was the vital factor. Obviously, with 2" tall magnets instead of 1" and with the same .5" edge, the force should be around double, but there wasn't enough height for them. (1.5" might fit, but I'd have to buy them that size - the heat from cutting the magnets - at least with the zip disk in the angle grinder, which seems the practical way - loses so much magnetism that the 1.5" would be little or no better than 1".) Later I realized that while a 2" magnet wouldn't fit inside on the arms, the ones on the outside could overhang a bit. I tested a 1" against a 2" in the vice and found the forces were about 15 and 20 pounds, a modest but worthwhile increase.
   Also noted in passing in the experiment was that unless the magnets were sticking well out from the vice and the pliers, the tools' steel shorted out much of the magnetic field.

   Later (10th) I tried a 1x2x.5" magnet with the 1" edge against the 2" edge. If the space could be found this was a modest improvement, giving about 17-18 pounds with the spacer. Finally (11th) I tried the obvious optimum: two 2x1x.5" magnets with their 2" faces together. The force, with spacer, went off the end of the 25 pound fish scale, probably 30-35 pounds. Here is the obvious configuration if there's room. With such forces, one might even be able to use fairly wide air gaps for loose fitting tolerances. That would make them easiest to make. For me at this point, it would also mean switching to a new motor rotor (ironically one with a taller center hub is ideal, to accommodate the magnet lengths) and a new torque converter drum - a lot of rebuilding.

   I can visualize the 'optimum' embodiment of the design more easily than I can take apart and re-assemble everything (and more easily than I can find the ideal drum): an ideal 12" diameter drum of aluminum with 2.1" straight rim walls (no curve to the base), .5 x 1" steel arms with 3/8" pivot bolts, .5 x 1 x 2" outer arms, sideways, next to the magnets outside of them, and 1/8" stainless steel plates to attach the magnets to these.
   The present question is: do I have to build the optimum to get the car moving properly, or is the design with lighter parts and lesser magnetisms that I can throw together fairly easily now good enough? Or, the other nagging question, is it possible that even the optimum isn't good enough?

   If the arms swung freely, the slight repulsive force would hold back a slow moving arm until its inner end was past center. But as the motor sped up, the centrifugal force flinging the arms outwards would overcome the weak magnetism and put the magnets directly across from each other hopefully just after the pin passed the center point, so the force should mainly or all be one way and should increase with motor speed.
   I found there was actually a considerable backwards force as the magnets latched onto each other: the arms jumping forward pulled the drum backwards. It occurred to me that if the arm magnet was able to twist sideways, or to move inwards away from the drum rim magnet, the arm perhaps wouldn't have to pivot, and the arm magnet would simply twist or retract as it entered the zone of light repulsion approaching the rim magnet, then straighten or spring forward again as it entered the strong attraction zone, to give the drum rim the sharp tug that should turn the wheel.
   The one "sticking point" would be if the motor came to rest with all the magnets lined up, would it have the force to break loose and start spinning? The swinging arms get around the problem, giving the motor a lot of leverage to pull the magnets apart.

   Well, there's still a lot of "hopefullys" in here and it's shy on descriptive math weighing the balance of the forces. But formulae characterize the various parameters of a given design, and a given design is exactly what we didn't have yet.

   On the 7th, I mounted a 1x2x.5" magnet edge-on outside of the drum rim, held by aluminum angle pieces, having of course removed the flat mounted magnets. The setup at this point allowed experimentation by hand turning the rotors on the axle along with easy removal of the drum with no other disassembly. Ideas could now be tested at a far greater rate than was possible at any earlier time with any other design.

   Approaching the (vertical) drum magnet with a magnet held horizontal disclosed that the forces attracting them were virtually annulled. (In fact, twisting the loose 'arm' magnet horizontal proved to be the easy way to separate the two supermagnets.) However, when they were lined up, there was a very strong force wishing to twist the arm magnet back to vertical orientation. The forces on the approach and exit tended to want to spin it the other way up.
   If the arm magnet could be mounted "hinged" so it could twist 90º, from horizontal to vertical with one face (eg, north) up or to the right only, it would turn horizontal on the approach to the drum magnet - from either direction - then would suddenly go vertical just about as the magnets passed each other. Thus there would be little or no "backwards" force on the approach but the most powerful forward tug just as they passed. The tug comes from the sudden slowing of the outer end of the five arms, which kinetic energy is imparted to the rim via the magnets. The faster the motor is spinning, the greater the velocity or the arms and the greater the centrifugal forces holding them straight out, and the greater the impetus as they pass the rim magnets - provided the magnetic forces are sufficient.
   (It would seem necessary to stop the twist a few degrees short of vertical to ensure that it twists back, since at exactly vertical, it would equally want to flip both ways and might stay upright.)

The Test

   When I tested the finished assembly on the car wheel, the forces seemed to be all in the right direction, there was no vibration, and the turning force increased with motor RPM. Unfortunately, that turning force was much too light by at least an order of magnitude - I could rather easily stop the jacked-up car wheel by hand.
    If it had seemed like even 1/4 as much force as needed, I'd have started working out ways to add enough arms and magnets to make it work. But to add ten times as many supermagnets or get them ten times bigger seems problematic. At some point, it's easier to just do a greatly overpowered motor with ten times the needed horsepower in order to get the needed torque without gears. But then it wouldn't fit on the outside of a car wheel, so the main purpose is defeated.

What's Next?

   The best idea now looks like the clock escapement type of mechanism. The hard part is to figure out one that looks robust enough to transfer the torque forces from the motor rotor to the slower wheel rotor without the serious vibrations of the shifting plate design, and without quickly falling apart.
   I now have the glimmerings of a design that might work. It's going to take at least another month to put it together.
   The idea is little "many" short two-point nylon "clock escapement anchors" sized so that at least a dozen will fit around the outside of the motor rotor, with the points sticking into slots (rectangular holes) in a strip (aluminum?) around the inside of the rim of the drum. Each one will provide only a small force (like all my other designs have done so far anyway), but they'll all add up. (Ack! It has to be made so it can be assembled without dismounting the anchors! The slots will have to be open to the open edge so motor and drum can be fitted together. But I digress.)
   The pivot holes for the anchors will be drilled through the motor rotor between magnets, and threaded. (Tentatively, 1/4" bolts for axles.) 2, 3, 4, 6, 8, 9, 10 or 12 anchors can be installed with balanced forces, as seems appropriate when the forces are measured. I can start with just two, or even one, and check the results without making the whole works, in case changes are needed to the design. (...why do I have the feeling it'll need more like 24, 36 or 48, and how will I fit them all? At least 24 and possibly 36 could be doable by stacking them two or three per axle, if the slots in the rim are wide enough.)
   These anchors will swivel back and forth very fast. Since E = .5 M V^2, faster gives more energy than heavier. But I can add steel weights to nylon anchors too, if it looks like that'll help, or make them of steel entirely.
   In principle I think I'll also keep the magnets mechanism in some similar form, with the external drum magnets, as a means to lock up the converter to 1 to 1 when the torque required is light.


   Perhaps it would be interesting to run through the several designs of torque converter since the start of the project a over year ago.
   My first idea was that one could use a typical induction motor design adapted to axial flux, but instead of electric coils creating a rotating magnetic field, have supermagnets on the motor rotor. That way, the output rotor would be driven by a supermagnet strength rotating field. The fundamental problem with that was that the motor itself was still driven by electromagnets, and the unit still couldn't have more torque than they would provide. The motor wouldn't run fast enough to provide any more torque than without the new mechanism. And all the slip was just losses: there was no way to convert or transform the torque, only to couple the original, insufficient torque to the output.
   The second idea seemed good in principle: magnets would flip up and down on hinges. They'd flip down next to a magnet on the output rotor just as it was going by, giving it a tug forward without a preceding tug backwards. The third design was a variation on the same theme. A plate on the output rotor holding all its magnets was to slide backwards an inch or more as the motor rotor magnets reached them, move forward to the end stops, and then the inertia and the attraction between magnets would give the output rotor a 'hit' of torque as the motor moved on. There were two main problems: the first was that if the magnets lined up with the motor stopped, the motor probably wouldn't be able to break free and start spinning. The second was that the mechanical construction didn't seem up to the level of the forces involved. I built some of the parts of these two designs and tried turning them by hand, then gave up without completing them.
   A couple of times I studied clock escapement mechanisms. They seemed to have the basic idea: when you try to run it too fast, it gets very hard to push. If you turned it around, running the motor fast should create a lot of torque. But it looked like the 'escapement' pieces would be getting bashed around at very high speed and would probably break or wear out quickly. But perhaps something that had those beveled hits, with a higher mass and hence lower speed...
   The fourth design was purely mechanical: a plate would spin with the motor, and would hit ramps on the output rotor at intervals that would cause it to shift sideways, in and out. The hits on the ramps would be 'torque hits' to the car wheel. I went through about 3 variations but found they all had serious vibration and not enough turning force, or that the turning forces were so short the tire absorbed them. In retrospect, the main cause of the vibration was that since it wasn't one solidly fixed assembly, it was actually much easier for the whole motor to bounce in and out than for it to cause the car to move - so it did.
   In between a couple of these variations, I thought (again) about Constantinesco's 1920's designs, and looked up some more recent patents as well. The big trouble was that his design took linear space between the driving motor and the driven output. Then I thought, what if the motor was mounted on the car body above the wheel? It all started to look quite straightforward when I found a ratchet wrench that was strong enough to move the car without a problem. (I put it on the car wheel and stepped on it - the car rolled.) But early in construction, I suddenly realized that since the car and wheel moved separately at bumps, the torque applying rod would simply make the car bounce up and down instead of turning the wheel. I was also having serious doubts about the efficiency of forcing a heavy wrench and steel bar to shift back and forth at high speeds.
   So I finished the last version of the shifting plate and tried it out. The vibration was still terrible, for the reason given. Somehow, I had to shift the axis of action from axial to transverse with a workable design. Transverse forces would be in the wheel turning plane, and furthermore a number of them spaced around the rim would cancel out their vibrations.
   The first idea for this was swinging arms on the motor that almost hooked, sort of clung, to pins around the inside of the rim of an output drum, yanking them along before they let go. It might have worked, but I didn't get it made.
   On April 30th I doodled out a diagram of a mechanism with the same swinging arms but with magnets on them, going past more magnets inside the rim of the output rotor drum. I figured the arms would freely swing forward magnetically when they reached the magnets, then un-freely pull away from them. Then I also figured that since only attracting forces were involved, the arm magnets could simply be replaced by steel. But when briefly tested (in May) the forces seemed too light, and also somewhat 'balanced', pulling both directions: as the arms swung forward magnetically, they also pulled the drum backwards.
   The unique feature was that the motor would pry the arms and rim magnets away from each other with leverage, so the motor would need far less torque to start up with the magnets 'latched' in the attracting position, the frustrating problem of all the other magnetic designs.
   To get more radius of action and more prying leverage, I went back to the larger 12" (aluminum frying pan) drum of the earlier magnetic trials, and then to putting the magnets on the outside of the non-magnetic rim. This had the benefit of allowing me to see what the arms were doing and run more experiments faster.
   Finally some experiments disclosed that two magnets edge-on to each other have much more 'ideal' characteristics for the job, especially if they could pivot as well as the arms swing. This led to this month's design, which had the right rotational forces, but they were much lighter than I expected, and the amount of supermagnet mass needed to up them sufficiently would seem impractical.
   Now I'm thinking again about clock escapements. This time, with the forces radial and balancing, and as many small "anchors" or whatever as possible or as needed, spaced all around the outer rim, to transfer motor forces to wheel, increasingly with motor speed.

Nanocrystalline Ceramic Motor Coil Cores Project

   I seem to be out of inspiration for this at the moment. (I may try using FeO again, this time well mashed with the mortar and pestle, or maybe steel grindings. Finding fine metal powders or lower oxides instead of fully oxidized metal oxide powders might be of considerable assistance. and costly.)
   Ultra-high permeability soft magnetic materials may or may not exist, and may or may not be found if they do, but good "FINEMET" nanocrystalline alloy cores are made, and good results have been obtained with composite materials. This much (my original goal) is known to be possible.

Nanocrystalline Reflective Rear Electrode for Dye Sensitized Solar Cells

Some Design Considerations - Rationale for using certain materials

   I chose tartrazine (yellow food dye from the grocery store) as a good looking photosensitive organic dye. Its wavelength of maximum absorption, 427 nanometers, is at about the wavelength of the strongest levels of solar irradiation (below, the peak of the yellow area, also the peak of the red on some charts, is at around 400nm).

   Totally off topic, this particular chart was especially interesting to me as it notes the absorption of specific wavelengths by ozone, oxygen, CO2 and water vapor in the atmosphere. That's interesting in itself, but also I thought from previous limited information I'd found that the 'cyclic' absorption lines on Titan from methane vapor were quite unlike from the spectrum of water vapor. This graph reveals instead a similar sort of spectral pattern.
   Titan is an atmospheric planet, between Mercury and Mars in size, that orbits Saturn. It's the only other world in the solar system with an air pressure similar to Earth's, and composed mainly of the same gas, nitrogen. It is also the only other world with land and rivers, lakes and seas - albeit of liquid methane.
   I see from a release June 4th, along with an earlier title announcing a lecture about it at a conference, that space scientists are finally starting to talk about the possibility there's life on Titan -- not because of the aquatic leaves and stems ("rocks"!?!?) in the Huygens lander images and evidenced in the Cassini T14 RSS pass, and not because of the forests of gigantic trees visually revealed and otherwise evidenced by the Cassini SAR radar (Cassini SAR image swaths of Titan's temperate and polar regions are similar in nature only to satellite/Google Earth images of forested areas.) Instead they're saying it for the equally evident chemical/spectral reasons: covering over of expected inorganics such as water ice by complex and unknown organic spectra with abundant benzene, and a seemingly unexpected absorption of expected energetic gasses (in particular H2 and HC2H, acetylene) at the surface. (The shortage of hydrogen at the surface seems to be what's finally caused the awakening, perhaps because it fit with a prior theory that hypothetical Titan life might use it for energy as a breathing gas.) So far they are still looking for "microbes" on a world evidently verdant with huge forests and aquatic vegetation of a scale unknown on Earth.
   On January 15th, 2005, the Huygens probe parachuted down and as far as I can tell splatted down on Titan's swampy tropical tidal methane seas right on top of a dune, just a couple of inches below sea level. The images obtained were immediately released to the public on the Huygens DISR website, in a form which was said (most unprofessionally) to be the original raw images: but which were actually grotesquely contrast magnified, and by completely inconsistent amounts. Luckily, René Pascal pointed me to two archives of the actual raw images. It took me until mid July in uncounted hours of study to figure out the true nature of the scenes: submerged wetlands (formed into dunes by Titan's powerful tides), covered in aquatic vegetation of stupendous dimensions. After that everything started to fit into place. Apparently scientists haven't got that far yet. Perhaps they were misled by their own unmentioned and bizarre image processing.
   My studies of Titan from 2005 to 2007 are on the web in my 'book', Living Titan. Some later thoughts (mainly 2008) on a number of worlds and the strange things and activities on them are at Space Update Notes Web Page. The main index to my pages on that site is: www.saers.com/~craig .
   But I seriously digress!

   The choice of borosilicate glass ("pyrex") for the glaze layer and for the front glass is because it is the most transparent glass at the strongest wavelengths. In fact its transparency is exceeded only by pure silica - quartz, which can't readily be purchased nor made in a typical kiln.

"Pyrex glass (borosilicate type) is opaque to radiation in the UV-B band and attains a maximum transmission level at 340 nm and beyond (Acra et al. 1984). The coefficient of transparency for borosilicate glass, 1.0 cm in thickness, is 0.08 at 310 nm, rises sharply to 0.65 at 330 nm, and attains a peak level of 0.95-0.99 from 360 to 500 nm. (Weast 1972)"

    Although I purchased a piece of borosilicate ("wood stove door") glass to divide into solar cell test pieces, and stannous chloride and tin dioxide, having now tried to put the conductive tin oxide layer on some test pieces I must conclude that unless its horribly expensive I'd be better off to order some ready made. (Oh, wait, I forgot... it's become almost impossible to find goods on the web any more, with all the "parasite sites" that each appear a dozen times in the same search and claim they link to thousands of companies selling just what you want, dominating the results of any search for anything and knocking out the bona-fide sellers. Maybe local glass companies might know something...) The resistances I've heard of for commercially made pieces are ohms or tens of ohms, whereas the best I've got - on porcelain which process I can't duplicate on glass or glaze - is hundreds of ohms, with thousands and a cloudy appearance being my best on glass.
   Perhaps I can do better on glaze at high temperature, but it would mean a lot of manipulating glowing hot ceramics straight out of the kiln, and then putting them back in. It will certainly complicate the melt-quench. Unless I can think of a better technique than what I'm envisioning now.
   As an alternative, perhaps it would be easier to use an "open" metal grill that lets most of the light pass by as the rear electrode collector, and forget about making the glaze conductive. Of course, its thickness would greatly thicken the electrolyte layer, even with quite fine mesh.

   The 'glaze mix 9' seems, by some fluke of luck given that up to about mix 6 it was supposed to be for another purpose entirely, to be pretty much ideal for the reflective rear reflectors, or certainly 'close enough' for working purposes.
   I painted on a thin white zircon (zirconium silicate) underglaze and then a very thin glaze layer on two pieces of porcelain, and one with no zircon for comparison, and fired and quenched them. One with zircon was clearer than the other two, but all had colour, and showed clumps of coloured microcrystals under strong magnification. I painted on a second thin glaze layer and fired and quenched again. This time, the glaze looked uniform and clear except for a couple of streaks on one piece, the microcrystals evidently having given way to transparently thin nanocrystals in both layers.
   With thick underglazes (or two layers), the glaze wouldn't stick on the porcelain, and with thick glazes, there were a lot of dark, coloured microcrystals rather than transparent nanocrystals. It would seem that two or more thin layers is the way to go rather than trying to do everything with one coat of glaze.
   The zircon white underglaze seems a little brighter than the titania (titanium oxide) white underglaze - at least to my eyes. (Must try tin oxide, too! ...would that be called "tinia"? Of the three common glaze whiteners, Ti and Zr are adjacent in the same column of the periodic table.) In the comparison piece with no underglaze, the whole tone was less bright, and the glaze seemed 'rougher' at the microlevel, especially with only one layer. I expect the glaze was pulling 'impurities' out of the grainy porcelain, and conforming to its grain, instead of having fine white zircon smoothing its base.
   The reason for the titanium oxide nanocrystals in the glaze is because it has a very high refractive index (highest of any substance) and a very high dispersion - the prismatic separation of colours. It is likely to most modify and disperse the light to give good prospects of making it more susceptible to being absorbed by the dye on the way back out.

Zircon underglaze and one thin coat of glaze (no underglaze on rightmost piece).
The melt-quenched glazes were clouded with microcrystals.

After a second thin coat of glaze, fired and melt-quenched, the microcrytals were evidently refined down to transparent nanocrystals except in the piece with no underglaze. A new piece with only one glaze layer was fired and quenched in the same batch.

   I'd like to try expanding on the glaze layers idea a bit by painting 'tinia' on the last glaze layer, and then doing another layer of tinia on top of that. At some point, it should become conductive. (though it may lose transparency - all in trying it out and seeing if some way works.)
   An alternative configuration for the solar cells, if making the glaze conductive works, and IF the electrolyte is quite transparent, might be to have the nanocrystalline titania sintered onto the rear electrode instead of the front one. The rear electrode would then be the negative and the electrolyte ions would flow the other way around.
   Now if that titania could be formed into vertical nanotubes or nanorods impregnated with the dye, that "light guide" could possibly be still another improvement.

Turquoise Battery Project

June Experiments

    When the new version torque converter didn't seem to have enough 'kick' on the 23rd, I decided it was high time to get back to the battery project, sitting since April in promising condition. Perhaps the best thing to do next was to make a sealed Ni-Mn/KCl battery that wouldn't need watering all the time, and wondering what other things might be getting in or escaping though the open top.

   This time I considered thoughts that had come to me about the manganese while I was making the previous electrodes in April.
   For manganese, I had purchased MnO2 (local pottery supply) and Mn metal powder (Micron Metals). The form in which these are added to the battery probably deserves consideration, since starting with one electrode charged and the other discharged is bad. It seemed easy to add some of each to get the desired starting charge state, but each form has its own small issue:

   The MnO2 (Mn valence IV) seems to be the common 'pottery glaze' form of manganese, but if put into a battery it makes a pre-charged 'positrode' (MnO2 being the usual positive chemical of non-rechargeable dry cells). For a negative, it's 'overdischarged'. Expected discharged form in salt solution would be Mn(OH)2 (II) or MnO (II). Starting from a higher 'overdischarged' oxide, less conductive and probably bulkier, seems counterproductive. One solution is to add a higher percentage of metallic Mn. But MnO2 is a cheap oxide and Mn powder (typical of pure metal powders) is expensive. To reduce the valence of the Mn oxide from four to two (or even to three) would be quite helpful by itself, and would reduce the metallic Mn required.
   NiOOH (III) can be discharged to Ni(OH)2 (II) with hydrogen peroxide, H2O2. Oddly enough, this oxidizer seems in effect to steal even more oxygen, probably from the water in the presence of the nickel compound, which bubbles off as (presumably) O2.
   The first experiment was to see if it would do the same for MnO2. It did seem to work the same way. (The gas didn't smell, so it wasn't H2, and that's pretty much the only other choice besides O2.) A week later, I realized the bubbling had stopped because the hydrogen peroxide was exhausted, not because the MnO2 was all converted. I added some more H2O2 and it went back to frothing away. I converted all my MnO2.
   I bought some more to weigh the product before and after next time. I should be able to tell by the weight change if it's MnO, MnOOH, or Mn(OH)2... and also make sure it's not actually oxidizing to some even higher valence form!

   The pure metal, Mn (0), is the negatively charged form in the battery. Yet I started to think: when the Mn powder was mixed with liquid, it might well discharge spontaneously before the hydrogen overvoltage ingredients were added. It all looks pretty black - I can't really tell MnO2 from Mn powder in the damp mix, let alone discern MnO or Mn(OH)2. If it was discharging, I wouldn't know.
   The obvious thing to do here is to mix the hydrogen overvoltage and all other dry ingredients first, then add the Mn powder last.

   Another consideration that may have been a factor in the difficult charging of the previous Ni/Mn mix is that I don't have (and can't seem to get) nickel powder. I added the nickel as discharged Ni(OH)2 (II) as well. Now that I think about it, this presumably has to charge up to Ni (0) before the manganese can start to charge, since it's a lower voltage.
   So perhaps I should be using monel powder instead? (It's around 2/3 Ni & 1/3 Cu).

   I looked at both of them at 40x magnification. The "blackish" monel looks like shiny little rounded chunks of silvery metal - which of course it is. The nickel hydroxide looks like green laundry soap flakes. While the fluffy Ni(OH)2 'snowflakes' tend to be of similar size to monel grains, some are much smaller, and when compacted they'll break down much finer.
   It seems to take some doing to charge the Ni(OH)2 to Ni metal one-time, whereas the monel is already "charged". Depending how intimately the Ni needs to be in contact with the Mn to effectively raise the hydrogen overvoltage (or whatever it does that seems to help), the fine Ni(OH)2 may do a much better job. The best way to find out is probably to try them both, and to try a mix with half of each. (At the rate I've been running battery experiments, that could take months!)
   If the compaction is good, the electrode should certainly be highly conductive with all the monel or nickel, and the manganese metal and lower oxide being quite conductive as well. If the "+" can be made half as good, a pretty small battery might start a car engine.
   Another thought is to buy NiO at the pottery supply if monel powder proves too coarse. Why run down my Ni(OH)2 supply when it's what's needed for the positive electrodes and is harder to get?

   While I was at it, I took a magnified look at MnO2 and Mn powder. The metal chunks in the "grayish" powder had a bronzy sheen and were much more irregular than the monel particles. The particle size was similar (well, I did order them both as -320 mesh! Perhaps I should be ordering micron sizes.)
   The MnO2, quite black at any scale, was clumped into quite large particles, but these easily broke down with compaction. They didn't seem to get as fine as the Ni(OH)2.
   When some was dry, I inspected the Mn(OH)2 (or is it MnO?). (That's the trouble with pouring 3% H2O2 into some powder... it got soaked into a wet slurry. Oh well, pretty much dried overnight when spread around a bit. But I doubt I added enough H2O2 to reduce all the MnO2 powder. I really should calculate it out and then measure the ingredients - that would be easier if I knew exactly what the reduction reaction was.)

   I decided to use a previous nickel "+" electrode that was already in the battery test case. (A slightly dangerous thing to do since it hasn't proved itself and I don't remember exactly what's in it - but easiest.) The final mixture for the first "-" was:

Expanded copper mesh collector sheet/backing, approx. 1.3" x 3" - 1.75g
Monel alloy powder - 10g (conductivity, ??)
Mn(OH)2 (?) - 7g (active ingredient - discharged state)
Mn metal powder - 5g (active ingredient - charged state)
Sb4O6 - .2g (increases hydrogen overvoltage)
CMC gum - .2g (glue)
Albumin (egg white) - a smear (significantly increases hydrogen overvoltage even in PPM quantities)
HOH distilled - 2cc

   In the electrolyte is Na2B4O7:10H2O (borax) and KCl (potassium chloride), not to mention water.
   The KCl is of course the neutral pH salt electrolyte. The idea of the borax is that if hydrogen is created by overcharging the negatrode, instead of making H2 gas, it'll change it to a borohydride, which, being like the borax, soluble, will diffuse over to the positive electrode to give up its 'H's and discharge some NiOOH to Ni(OH)2. This will create heat but eliminate buildup of hydrogen gas pressure.

   On the first attempt to charge the battery, the voltage went up to only ~~ 1.6 volts while charging and rapidly dropped off to under a volt with the charge removed. The pressure went up to 12 PSI in a few hours without the battery being charged. Overnight I reduced the charge. In the morning the voltage was only marginally higher, but the pressure was up to 18 PSI.
   I unscrewed the pressure gauge and sniffed the air coming out as I did. I didn't smell much, and it didn't seem to be hydrogen. That seemed to say the negative wasn't bubbling hydrogen gas. So I concluded that the positive electrode was already charged and giving off oxygen on further charge, thus the negative wasn't reacting with it to charge properly.
   I took the battery apart and put the positrode in 3% H2O2. There were lots of tiny bubbles for over two hours, and even after four, a little shake would start it mildly fizzing again. This evidently means it was charged to NiOOH and was now being discharged to Ni(OH)2. After that I put it in water to rinse it out.
   As I left it in a considerable puddle of H2O2 and the bubbling continued for such a long time (even at that I can't say it had really ended), I think my previous attempts to discharge nickel positrodes used much too little H2O2 for much too short a period of time to discharge them more than a fraction of their full charge. This may well account for a few disappointing experiment results.

(FWIW The reaction might be: 4 NiOOH [III] + 4 H2O2 => 4 NiOHOH [II] + 2 H2O + 3 O2.)

   On further testing the cell, the positive connection (nickel-brass) corroded away, and in the restoration I put the meter probe at various points in the electrode briquette itself. This disclosed that while charging there was a great voltage gradient through the thickness of both electrodes. The closer the probe to the interface between the electrodes, the greater the voltage drop from the collector screen. Evidently both electrodes have very high internal resistances.
   This could explain many poor experiment results! And I've started to realize that if a battery has 10 AH of material and it'll only charge at (eg) 20 mA, it'll take over 500 hours to charge it! A few hours, or even days, may not bring it up to the expected voltage, explaining that many "failures" weren't due to the electrochemistry at all. When the area of best conductivity to the collector becomes fully charged, the voltage goes up to some ridiculous level, and I finally conclude it isn't working right at all. But much of the rest of the electrode has yet to be charged, so as soon as the charge is removed, the voltage drops below the expected voltage. And if the current that can be stuffed into the poorly conductive battery isn't above the level of self discharge, it'll never get charged.
   Perhaps my electrode compactor, though forming the powders into a seemingly solid 'briquette', isn't pressing hard enough? My briquettes crumble more easily than the ones from the Ni-Cd dry cell, though none are immune to losing material when mechanically stressed. It would seem I probably need to find a still better compacting technique. Maybe double the screws? -- screws going in from both sides of the compactor to increase the available pressure? Maybe pound the compactor with a big hammer when it's already fully tightened, and then see if the bolts will turn in a bit farther?
   I decided on the 29th to look into hydraulic presses. A big 45 ton press is $2000, but a 6 ton one is just $100, and a 12 ton press is $270: nothing like the high capital costs I'd expected. I even got a used 12-ton at Barclay's Exchange for $150. (KMS Tools had just sold their last 6 ton one: I saw it - it waved at me on its way out the door. But I'm guessing that 6 tons would be a bit underpowered anyway.) And I had found doing up a bunch of bolts to compact an electrode would have been a serious production bottleneck even for enough homemade batteries for a car. The press is just a common 12 ton hydraulic jack in a big heavy steel frame.

   BTW I wish I had some much finer copper mesh. I got a coarse mesh roll at an art store (Opus) and I was probably lucky to find that. It doesn't help conductivity when much of the active material is somewhat distant from the collector. Also BTW this was the first time I had the mortar and pestle to give the mixture a good grinding and mixing before compacting it.

Higher Voltage, Cheaper, Higher Energy Density Batteries: the easy way?

   In considering making higher voltage batteries with multiple cells, I had thought of putting a "+" and a "-" electrode on opposite sides of one collector sheet, when I was using nickel-brass brass collector sheets. Then when I started using metal grilles, it occurred to me one could more simply use a metal grill wrapped around a thin piece of hard plastic.
   Now I'm thinking: why even use the plastic? The voltage of both electrodes is necessarily the same, so why does the electrolyte need to be blocked, especially if it's a dry cell with no extraneous liquid that might bypass the electrodes? Why not just compact the two electrodes together into one bipolar briquette? And then, with all the current simply flowing across the thickness, why use any grill or collector sheet at all? Why not have nothing but the active electrode briquettes? This would be the simplest and cheapest, with the highest possible energy density.
   Here we hark back to Volta's original 1798 "electric pile" in a modern form. Except for a collector grille or plate in each end electrode, the whole battery is full of nothing but active electrodes, with the requisite separators between them.

   It's getting tempting to try making a "full size" 12 volt Ni-Mn battery with 6 cells of 3" x 6" electrodes, maybe 50 or more amp-hours. But I'd better work on improving the conductivity first - it would be of little use for transport if it only supplies (eg) a half amp for 100 hours instead of 50 amps for 45 minutes.

A Battery Comparisons Thought

* Left: "D" cell size battery. (Ni-MH "D" cells are up to 10 AH, 1.2 V, "1000 charges", 100 WH/Kg.)
* Center, Right: Ni-Fe 10 AH, 1.2 V pocket cell battery, 20 WH/Kg. (Larger ones are said to be up to 50 WH/Kg.) These last for decades. The active chemicals are inside rectangular "pockets" which are combined into flat plates. Changhong uses battery production line equipment purchased from Varta in Sweden, so their batteries are about the same as Varta's.
* The "hacked" electrodes at the back-left was a Ni-Ni experiment in the spring.
* I imagine the same batteries as Ni-Mn and think: 2V cells and almost double the energy, for all types!

   Knowing a few more details now about batteries, I have an observation for comparing various chemistries with a little more savvy, especially alkaline types versus lead-acid. The most important ratings for EV batteries are WH/Kg and WH/$.
   Lead-acid battery ratings are deceptively high. They're rated at a 20 hour discharge rate, and the amp-hours are much reduced at high rates. And even with sodium sulfate keeping them from decaying, they can only be discharged to about 60% depth of discharge for electric drive because the voltage seems to start seriously dropping off when much current is asked for beyond around 50%. With Ni-Cd or Ni-MH, one can tap around 90% of the charge out before the voltage plummets, and the amp-hour ratings are for a shorter discharge time and they degrade less at high rates. Therefore, most other batteries store perhaps 60% more usable energy than a lead-acid battery with the same rating. A 100 AH lead-acid battery should therefore be compared with a 60-65 AH battery of most other types. The same derating scale could be used in reverse for the cost, ie, multiply the lead-acid cost by 1.6 to compare it to other chemistries. And the weight, already heavy, must also be multiplied by 1.6 to compare it fairly with other types.
   Thus if I can make a 60 AH 12 V Ni-Mn battery (a rough estimate, perhaps a little on the high side) by incorporating everything I've found) in a 4" x 4" x 8" case (128 cubic inches) it would compare with a 100 AH size 27 deep cycle lead-acid battery of 12" x 7" x 8" size (672 in^3) - five times the size.

The Ultimate?
for the front battery to be usable energy equivalent to the rear one,
cheap, safe, and superior in every other way.

Lead-Acid/Sodium Sulfate Battery Renewal Project

Sodium Sulfate for Battery Renewal (Click for Prices, Info.)

   I finished restoring the batteries I got last month and a couple of others. There was tremendous variation in initial performance after draining the acid and adding water and sodium sulfate, one battery seeming to be pretty fair after one or two charges. Two others maintained only 8 volts at 10 amps for just a minute or so on the first try, but one rapidly improved. The other seemed to get just marginally better each cycle, slightly perking up in voltage and time, but never really seeming to work properly.

Shorting out Shorts?

   I tried for a while to burn out "shorts"(?) in some cells by shorting the battery out with a light jumper cable (often the battery still read a volt or two while "shorted"), or giving them a very heavy load - with the caps loosened or removed. Some of them didn't respond. One was loaded with car headlights, 50 or 60 amps, which started flickering and suddenly went brighter, then again blinked brighter still. Voltmeter readings jumped up as well. I thought, "Aha! That's done it!" but then it flickered some more and went dimmer again. It went brighter and darker several times, but ended up not fixed. That was the best I attained.
   I could hear the water boiling in the battery, presumably where the current was heating the shorts up. There was a strong smell of hydrogen gas after a while and I moved the operation outdoors.

   There are two main problems: first is that the liquid cools the short, making it much harder to burn out than if it was in air. Second, charging the battery seems to strengthen and build up the short. But in order to try to burn the short out with a heavy load or short circuit, the battery must be charged. It's working against itself.

   With one battery, I dumped out the liquid then refilled all the cells except the problem one. Then I charged it for a bit, then shorted it. The theory here is that with little or no liquid in the cell, the only path for current is through the short, and the short is no longer protected from heat by being submerged.
   After a minute or two, the voltage suddenly dropped from about 1.5 to 0.5 volts and stayed there. (I "shorted" it through a thin jumper cable, not a total short.) Then I refilled the battery and tried to charge it. It charges up to 12+ volts, but gradually drops again. I think there's still some little shorting connection left over that discharges a cell.
   The battery was pretty flat when I started.  Perhaps I should have charged it longer before shorting it, so it would have supplied more current to burn out the whatever it is. I'm going to use the battery a while and see if it clears up by itself.

   I won't say curing battery shorts can't be done, but my success has been pretty limited so far, and for considerable trouble. Better I should concentrate on getting the Ni-Mn batteries working!

Victoria BC