Turquoise Energy Ltd. News #30
Victoria BC
Copyright 2010 Craig Carmichael - August 3rd 2010


July in Brief (summary)
  * Torque Converter - "type 5: clock escapement"
  * Battery tab spot welder
  * Working Ni-Mn chemistry!
       Battery chemie promises highest energy densities, "green", indefinite life, maintenance free, low cost.

Chevy Volt: I'm suspicious
* Delays, luxury car price, no TV ads... and specs show abysmal efficiency: HALF the miles/KWH of the EV-1.
  * "Who Killed the Electric Car" is to be repeated with MiEV.
  * It's all part of a century old pattern.
  * Epilogue: A "typical example" of electric car suppression?

Mechanical or Magnetic Torque Converter Project
  * "Crown" of twisted sprockets
  * Escapement "anchors"
  * Principle seems to work
  * Punch for making improved twisted sprockets; then a new one for bigger ones.

Simple Capacitive Discharge/Resistance Spot Welder Project
  * Battery electrodes use spot welding for secure connections
  * Electrodes: heavy Cu wires sanded to cone-shaped ends
  * Trial 1: 10,000uF charged to 15 volts makes feeble welds on thin battery jumpers
  * Trial 2: 112,800uF @ 15 volts makes a bigger spark
  * With short, heavy leed wires it zaps holes, pits
  * With practice & technique, welds thin battery tabs to battery.
  * Shortest project yet, under 4 hours!
  * Improved Model July 22nd

Turquoise Battery Project
  * Short Summary
  * New Nickel-Manganese Battery Chemistry Works! - for the first time ever, 2 volt alkaline cells!
  * Higher EV/PHEV battery energy densities are now possible, eg 140 WH/Kg and up!
  * Made: Large 3" x 6" flat plate Ni-Mn Battery
  * Depressing press news: bolt-down electrode compactor is stronger than 12 ton hydraulic press
  * Too much pressure; lots of bubbling: changed from sealed dry cells to flooded, vented cells
  * It's alkaline, even with salt electrolyte. pH read 12-13.
  * Copper structural materials don't work in the positrodes - rapidly corrode away. Trying nickel-brass plates.
  * Copper negatrodes seem great!
  * VERY gradually starting to hold charge - the voltages are going up and lasting longer as the weeks pass.
  * Stacking electrodes for higher voltage cells: ...dip in paint or wax or something to seal edges?
  * New, more powerful electrode compactor; bigger electrodes?
  * Side thought: why are there no high energy density nickel-iron batteries, eg, 70-80 WH/Kg?
  * Exciting new Ni-Fe research - from India - is also applicable to Ni-Mn. (next two items)
  * Better electrode binder?
  * Catalytic H2 + O2 => H2O gas recombiner reduces gas pressures: allows maintenance-free sealed cells after all!

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.
* Simple Spot Welder for battery tabs, connections (in This newsletter)

all at: 

July in Brief

Latest Torque Converter:
- Motor rotor/axle with escapements (top),
- converter drum (bottom),
- Nylon escapement bushings "pipe" (inset)

   As I finally had the basic design details in my head, everything looked good for trying out an 'escapement' torque converter idea. A trial strip of aluminum with sprocket teeth was made.
   Three escapement anchors were made, a plastic 'template' and then two 'real' ones of aluminum, and each one's shape was better refined. When I started, I had only a meager idea of what a 'right' or 'wrong' shape might be. I drilled and tapped 6 holes around the rim of the motor rotor for their pivot axles.
   A hand turning test showed it seemed to do what it was supposed to. At low speed, the motor rotor and the converter drum turned freely against each other, but by around just 30 RPM, the escapements buzzed back and forth, rather noisily, and the force on the stationary rotor was up to around 2 foot-pounds, with two of six anchors installed.
   Some time was lavished on making a punch and die to stamp out 25 larger sprocket teeth around the rim of a new sprocket strip. And I realized the drum could be made with from the strip and a circle of flat aluminum plate, instead of finding a frying pan of exactly the right size and shape. This would be helpful to all future builders, assuming the design worked. Then I finally set to figuring out the best size and shape for the escapements and making them.
   After several hours of grinding, sanding, filing and burnishing I got three that didn't jam and tried it out. Performance was dismal! Heavier and more pieces, within reason, wouldn't be enough to get it to move the car.

   Another vague plan is forming in my head, this one for using the same sprocket gear and similar arrangement, but driving the escapements back and forth, which would directly turn the gear and wheel, rather than relying on oscillating masses.

Capacitive Discharge Resistance Spot Welder

   On the 15th I phoned around for a spot welder for battery plates and leeds. There seemed to be none for sale in town. For some comic relief I decided to see what went into one. A little investigation on the internet disclosed it's done in essence by suddenly discharging capacitors, using a foot switch to trigger the discharge. Electrodes were just heavy copper wire. Dispensing with all formalities, I ended up being able to weld thin battery tabs with a small power adapter feeding 0.11 farads - two dozen 4700uF, 25 volt capacitors from Queale Electronics (their largest size), which I later upped to 32: 0.15 F. These went via fat wires straight to the electrodes, which arced as soon as the contacts were made. I found out other welders use from 1 to 4.5 farads, but these operate through an SCR switch that limits the current. Without the switch my 'underpowered' unit seemed to be 'good enough'.

Battery Energy Densities Comparison Chart,
showing the promise of Ni-Mn catalytic dry cells
with stacked bipolar electrodes.
(...vertical is about right, but it should be way past the right hand edge of the graph!)

   I decided the Ni-Mn battery chemistry looked sound enough to warrant making a "full size" cell with the 3" x 6" electrodes. Loosely filling the compactor with 'Ni' mix and compacting it resulted in a 20 AH positrode about 4mm thick. A matching 25 AH negatrode was made, which turned out to be under 2mm, total 6mm per pair. This cell seemed at first to be a failure, and the copper framework of the positrode quickly corroded itself out of existance. I changed it from a sealed dry cell to a flooded wet cell and replaced the copper mesh with a nickel-brass plate. Gradually it started to charge to higher voltages and drop off more slowly - over a period of weeks. I almost gave up a couple of times. In spite of using neutral salt electrolyte the cell became quite alkaline (explaining the copper corrosion), about pH 12 or 13, but not 14 according to my pH test strips.
   Now after finding a bit more info I think I may 'form'/charge the electrodes in a flooded tank at higher currents than I've been able to use in a confined cell enclosure, but then use them in sealed dry cells. Manganese is element #25, beside iron #26, and as far as I can tell, Ni-Mn cells should last as long as Ni-Fe - perhaps a virtually unlimited cycle life as sealed dry cells according to the latest Ni-Fe battery research - 2004, from India. The voltage, about 1.9 "nominal" volt cells instead of 1.2 is the big difference. The Indian research found a catalytic device to combine H2 and O2 gasses created during charging into water, reducing pressure and making sealed Ni-Fe cells possible for the first time. (I have what I hope is a better, simpler, cheaper catalyst - antimony, which is mixed into the electrode material.)
   The Indian research also gave a figure for electrode compaction pressure, something I've been seeking for 2-1/2 years: 4-1/2 tons per square inch for their iron electrodes. My compactor of 1/4" steel plate bulges noticably in the middle when the bolts are done up, and I finally bought 1/2" plate steel to make a new compactor, which will have closely spaced bolts to exert greater pressure. It should be at least in the ballpark. I can judge by electrical resistance readings on the positrode, which are rather high with the present compactor. The "bolt-down compactor box" is a key tool for "DIY" battery making without a costly factory.

Chevy Volt: I'm suspicious

   What's with the Chevy Volt? Here's a chunk of my opinion:

   It's no surprise to me that it seems to be taking its time getting to market, or that the price is too high.
   The electric range is abysmal considering its very considerable lithium battery pack. In fact, the pack is 16 KWH, said to give an electric driving range of "up to" 40 miles. (Why then pricey lithium? - you can get that sort of range with cheap lead-acid batteries.)
   The EV-1 on the other hand had a 21 KWH Ni-MH battery pack that gave a range of 100 miles. This says the EV-1, though it weighed almost 3000 pounds, got about 100/21 = 4.75 miles per KWH. (This approximates other published figures.) The Volt works out to "up to" 40/16 = 2.5 miles per KWH - HALF the fuel economy of the EV-1! So it uses twice as much electricity from the grid as the EV-1 (or most anything else) to recharge. This means it needs a very large lithium battery pack, driving the retail price way up, to go its very modest distance on electricity.
   What is the problem? It would seem the power train and the motor itself must be awfully inefficient. It's a huge induction motor - 110 HP, the equivalent of at least a 265 HP gasoline engine. That sort of size should be more suitable for a large truck. Compare that with a typical 10-30 HP EV motor... and with the 5.5 HP Electric Hubcap (which admittedly has yet to prove its practicality). And lithium batteries are inherently not the best technology for electric vehicles, though some good workarounds seem to have been developed... that add even more to the cost. Notwithstanding what's being said, I'll bet that the Volt's lithiums will only last a few years (about the same as lead-acid properly treated with sodium sulfate), whereas it looked like the Ni-MH batteries in the EV-1 "would outlive the cars".
   The whole effort looks shoddy, even pathetic -- from the world's biggest automaker, who nevertheless was in deep financial trouble until bailed out by the public purse on the promise they'd make electric transportation.
   The same shareholders control the nickel-metal hydride EV battery company, Chevron-Cobasys (formerly inventor Ovshinsky's Ovonics, then GM-Ovonics) that made the batteries that made the EV-1 et al so successful over a decade ago. But since 2001 it never makes, sells, further develops, nor (via numerous acquired patents) permits to exist, any Ni-MH EV batteries. Of course they won't restart production and put them in the Volt. That's the last thing they want! Electric cars having been forced upon them by a disturbed public, they've managed to pretend the proven best batteries, cheaper than lithiums, are no good, and devolve the whole publicly funded program to this one long delayed model and rumors of another to come from Ford. Where are the PHEV vans, trucks and suvs? Why are the great, economical batteries left to rot -- and where is even basic good engineering?

   I believe they want the Volt to be the only, the worst and the highest priced piece of trash they can make it, as long delayed as they can hold it back, without attracting a lynch mob and without having president Obama seek elsewhere for a solution. Then, if it ever goes on sale, they'll "phase it out" as soon as they can get away with it "because there's no demand."

   I also read in the papers ('fraid this is from memory from a month or two ago) that the Mitsubishi MiEV electric car "is coming"... but that the few coming to the west (or was it only to Australia?) this year will only be leased, not sold, and that then they will all be recalled and sent back to Japan. It sounded like the same thing as in "Who Killed the Electric Car", happening all over again - brazenly announced in advance! They're supposed to be available in late 2011, but I suspect there will somehow be further obstacles to buying one next year. (BTW, the MiEV gets 80 miles from the very same size battery pack, 16 KWH, that takes the Volt "up to" 40 miles.)

   At the same time, we are shamelessly deluged with TV ads for all types of gasoline vehicles, or at best occasionally for non plug-in hybrids. Seen any of those glowing TV ads for the Volt or any other EV or plug-in hybrid?
   These people, who around 1900 seized control of our fledgling transportation economy, are not on the same side as the rest of us. This enemy within is now fighting desperately to keep electric vehicles out of the public's hands and by this repetitious TV and other propaganda to try to make us forget there ever was such a thing. The appearance of electric cars or plug-in hybrids for sale to the public at economic prices would be a death blow to their giant oil oligopoly, and they'll fight to the death to prevent it. They shut down, undermine and suppress electric car and EV battery development and commercial enterprises as they make their appearance, and let the public conclude from seeing nothing ever actually available for sale that they must not really be economically feasible. They have waged this war successfully decade after decade. How long will we continue to let them dictate to us that we shall burn fossil fuel, with wanton inefficiency, for all our transport needs?

   Hah! That was going to be the end of my tirade, but here's a great epilogue: a typical example.

   Someone has sent me a link to Evergreen Electric Vehicles - with a promising looking electric car "about to be introduced" at the 2008 Seattle Auto Show. This "medium speed vehicle" (up to 60 Km/H) must have been the product of much fine development work, and there are videos of it driving around, beautifully. Whether or not it ever made it to the show, it would seem it caught the attention of the enemy: the web site was last updated in 2008. It's full of great reports and good news that, as always, abruptly ends with no explanation.
   If president Obama had given any portion of a billion dollars to this Washington company (and provided it and its key personnel with protection against corporate, financial and physical violence), these cars could well be in production, on the roads, and probably in high demand with a big waiting list by now.
   Their web site: www.evergreenelectromotive.com/

The Evergreen 2008 EHC

   When people see one individual company, idea or person disappear, it's puzzling, but the obvious conclusion is that something individual happened to it - its business plan was poor, financing didn't come through, the product was uneconomic or bad circumstances forced it out of business, or perhaps the head of the company had some sort of personal or family trouble or an accident. But various such companies, some with "corpse" web sites, can be found -- and where are the successful ones, electrifying our transport system? For example, try looking up companies doing nickel-zinc batteries. There were two or three when I looked a while back, also a flat-plate Ni-MH battery company, all with nothing recent on their sites and no explanations. The ones I tried don't answer e-mails. Why would all these various people go to all the work of developing products, creating companies and web sites, and then just vanish on the eve of commercial production?
   And what happened to the people who invented engines that run on compressed air a couple of years ago - two independent designs in two separate countries? And the big, safe plastic compressed air tank whose side would just rip open and release the air if it was hit or punctured? "Now we don't need to burn fossil fuel to drive cars any more." they said on TV. But where are the cars running on air? What happened to the entire promising technology? The engines worked. Everything was going great - until they hit the TV news. The first news was also the last.
   And whatever became of all those electric car companies that started in the 1970's with the US "oil crisis"? Popular Mechanics magazine et al were raving about them and detailing their features, but I don't know of a single electric car having actually been sold in that era. I met a person who invested in one of them after an electric car ride, saying "This is the future!" and he said "Nothing happened. It just never went anywhere." He was certain there had been interference.

   History has recorded the "disappearance" (read murder) of Diesel from a ferry, the burning (read arson) of Edison's Ni-Fe battery factory and the nefarious dealings of GM with Constantinesco to ensure his fuel saving mechanical torque converter was never used, but it would sure be nice to know what happens to all these other disappeared companies, and to the inventors and product developers that created them! I bet one could look back and find the decades are littered with sabotaged attempts to introduce non-petroleum transportation and better electric car batteries, each naive startup wondering why no one else had already done it.
   We can eat all the best food we want, and try different diets, develop new multi-vitamins, be nature-friendly or whatever, but it's hard to make transportation healthy unless people start to realize there seems to be a huge tapeworm within, clothed with money and power but somehow swallowing everything good and... according to what I've heard... spitting out the evidence in mid ocean from tankers, in 45 gallon oil drums filled with cement.

Mechanical or Magnetic Torque Converter Project:
The Quest for Torque Leverage Without Gears

   Last month's magnetic design seemed to do all the right things... but with so little force it seemed impractical to try to up it to car moving magnitude. So I went back to the clock escapement idea, and I had a design in my head by the end of June. I sketched out a modified design July eighth, then made a modified version of what I'd sketched.
   For the rim of the output drum, I ended up cutting 37 slots between 37 solid "sprockets", each element 1/2" wide. I cut slits in from the edge, and simply bent out the pieces for the slots. I elected not to do this to the rim of the drum itself, which had proven so handy for so many of the TC designs. When I curved it around the drum rim, it looked like a crown of some sort. It also didn't quite fit and had to be trimmed down to 36 teeth and slightly spaced in from the rim.
   I had thought of putting 12 very short escapements around the rim, but finally opted for the idea of six somewhat larger ones.
   Next I made a plastic escapement piece. I had finally realized that its two hookes don't have to be at opposite sides of the drum, merely that one point be at a tooth while the other is 1/2 way between teeth, that they be angled appropriately for wherever they were, and that they should pivot as near to the rim as feasible.
   Then with the fit of the triangular "hookes" I realized that the sprockets should really be 'points' - vertical lines - rather than occupying 1/2 the space around the rim. Rather than cut the aluminum thin, I twisted the sprockets with a pair of pliers. This achieved the same effect against the triangles without leaving thin, brittle teeth. One broke off while I was twisting it anyway, and 3 or 4 more were much weakened.
   The whole thing seemed a bit hokey and it now looked a bit like a giant bottle cap. Nevertheless I persevered. I cut a metal escapement out of 1/8" aluminum using the plastic one as a template, and when I'd cut down the points until it didn't jam everything, it rattled back and forth as one rotor was rotated against the other.

Left: Torque Converter drum with "crown" of twisted sprockets.
Right: Motor rotor with flanged axle and the first three prototype "escapement anchors".
The washers hold the drum at the right height for the escapements to mesh with the sprockets.
Fish scale hanging down was to measure the torque.

   The torque should vary by the square of the speed difference between the two rotors. Either unit could be turned "freely" at very slow speed, but at a fairly low speed the driven rotor pulled the other around with a force of around 1 pound.
   With the two anchors (the second one being heavier), the force seemed to be around 4 pounds at 30 RPM. (Although I had a fish scale to measure the force, the unit repeatedly jammed and it wasn't possible to get a steady reading.) However, boldly extrapolating anyway, that would be 16 pounds at 60 RPM, 64 pounds at 120 RPM, 256 pounds at 240 RPM, 1024 pounds at 480 RPM, 4096 pounds at 960 RPM, and 16384 pounds at 1920 RPM. That, of course, would be assuming that nothing busted, wore out, or jammed! And the motor certainly won't supply thousands of pounds of force, ie, it surely won't have the power to attain those highest RPM differences. (Wasn't the Avro Arrow engine 20,000 pounds of thrust?)

   I looked at the various angles and shapes of the third prototype escapement, and started to see there was a "correct" shape, and that it was shaped "wrong". The square part of the edges was sometimes able to meet the sprockets instead of the triangle edges because there was nothing to specifically prevent it. This was the main source of the jams. The right shape wouldn't be able to pivot into a position where it could jam. A "wrong" shape could be made to work by attachments that would limit its travel, but with a "right" shape, that would be unnecessary. If I hadn't made the unit and the parts, I wouldn't have realized there was a "right" shape and any number of "wrong" shapes.

   In a couple of days I had decided not to make the rest of the escapements and proceed to powered trials with the "crown" gear. It might well be that I would make them all and then have to almost immediately make slightly different new ones to fit a new sprocket gear. Once I had an idea for a better sprocket gear, it was: gear first!
   On the 15th after making the spot welder, I cut a new strip of #12 aluminum, and then made a punch and die to punch the aluminum gear teeth the way I wanted them... which was similar to the way I had in mind in the first rough drawing except for punching them with a 45º twist as I had done with the crown. The differences between this and the crown are:
   a) the flats between the angled parts are retained for strength. This will also prevent the tips of the escapements from swinging in farther than they're supposed to - a cause of jamming of the "crown" gear unit.
   b) the holes are punched from the center of the strip and are supported both top and bottom - also stronger.
   c) I intended to use a special punch and die jig with the strip clamped down while punching, ensuring uniform, straight sprocket teeth with even spacings between them.

In concept I wanted 45º triangle teeth: /\/\/\/\/\/\/\/\/\/\/\/\/\ . These would have been hard to do with sheet metal. ( - but probably not impossible. I was thinking that they'd have to be milled from a thick piece of aluminum, but some jig to form it from sheet metal could probably be created. Hmm... hmm...)

In practice: _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ with 45º angles and the same spacings between points should work pretty much the same, though it will probably wear out faster. How fast remains to be seen. Could be a minute or five years for all I know so far. The escapements, their triangles also 45º, will hit pretty squarely against the edges, so they might last much longer than one would suspect at first glance - and how long solid aluminum triangles would last is also unknown.

First, small sprocket punch (.5" x .5")

   The next day, the 16th, I decided it was all too small, and made a punch and die 1" x 1", twice each dimension. There would be just 25 larger teeth (1.5" / 37.5mm point-to-point) instead of 37 (1" / 25mm). The bigger size was surprisingly much harder to make. It took two days. But I did a fine job of it - it should be able to bang out lots of good toothed torque converter rims... assuming the design doesn't change again.

The main components of the second, larger sprocket punch & die (1" x 1")

The punch at work

Later two "rails" were added underneath so the punch could be used on a flat surface
without the teeth hitting it as they were 'extruded'.

   (The astute and long-time reader of these newsletters may recognize that the die is made from the cut-off ends of the original 3" x 12" battery electrode compactor, left over when it was reduced to 3" x 6". It was handy with four holes pre-drilled and threaded.)

L: 37" sprocket strip of .081" (#12) sheet aluminum with 25 teeth, ready to bend into a ring sprocket-gear.
R: Escapement piece shape trials. I finally realized they had to be curved to account for the pivotting.

   The large size teeth had their own problem: they stuck in too far and hit the 11" diameter motor rotor - and extended to the holes drilled and tapped for the escapement pivots. It was necessary to mount things so the teeth were slightly above the rotor, which fortunately wasn't too hard to do with some sanding and by inserting some washers. I also had to drill new pivot pin holes at a lesser radius from the center axle.
   However, for a 10"/300mm diameter rotor as I now recommend for the motor rotor (eg 6129R), it would be a very good fit. It makes me want to run off and change everything to match my vision of "as ideal as stock parts intended for something else can get"! But I decided to leave that for motor version four since I had other things to do, and instead (sigh) drilled another six holes in the 11" rotor.

Rotor on axle, with 3 pivoting escapements, output drum with sprocket strip, nylon "pipe" for bushings (inset).
It was hard trying to center the hole in the nylon rod... until finally
I remembered I have not only a drill press but a machine lathe!

   It would be nice to eliminate the 12" frying pan as being a hard to get "custom part", and it occurs to me this can easily be done.
   The torque converter drum could be made from two flat pieces by cutting a 12" circle of thick aluminum plate (3/8" or 1/2" thick?), making the necessary center and bolt holes, and then screw the toothed strip (as above) into its edges. Optionally, an untoothed strip could cover the toothed one to add stiffness and perhaps keep dirt out.
   Or, perhaps better, a thinner circle plate could be used, eg 1/4" or 3/16". One widened edge of the gear strip would be slotted with "V" slots and the tabs thus formed bent over inwards to 90º, level with each other. This edge would be screwed onto the flat surface of the plate all around the rim.
   To get the aluminum plate round and true without a big lathe - one able to hold a disk over 12" in diameter - it could be cut roughly to shape with the center parts cut and drilled, and then spun on the motor itself to mark where to file down the outside edge to true it... or perhaps even hold the file or some tool at the edge with the motor turning, though safety and technique I'd have to leave up to the user on that idea.

   On the 31st I finally had the drum and the first three escapement pieces about ready to try, after hours of grinding, filing, sanding and burnishing, trying to adjust the dimensions of the escapements so they would toggle as smoothly as possible without excess play that would have them "banging" on the teeth, but without jamming.
   But when I tried it, again the force developed was wholly insufficient. It would take more than adjustments and fine tuning to make it work.

   I'm now thoroughly sick of trying to employ the principle of forcing masses to go back and forth at high speeds. It worked for Constantinesco, but I'm beginning to think it takes a lot more mass, making much more oscillating motion, than there is room for in an "Electric Hubcap".
   There has to be some simple way to create some direct link, something like variable gears where if "A" moves, "B" is forcefully pushed, not indirectly coerced by an oscillating mass that will shove out of the way with less force than it takes to turn the car wheel.
   But, now, if those same escapement pieces were being directly shoved back and forth by the motor rotor instead of oscillating freely, their back and forth motion would forcefully, and smoothly, shove the same ring gear and hence the car wheel around... hmm - I see the glimmerings of a new plan! ...It'll need more pieces... Maybe it's time to go study more aspects of mechanical clock mechanisms!

Capacitive Discharge Spot Welder Project

Need to Weld

   One of the recurring problems I've had trying to do batteries is to get the electrodes connected to an external terminal, through a sealed case. Solder (probably hard solder as well as soft) corrodes away - and probably isn't a good contaminant inside the battery. Regular welding would destroy the pieces.
   On the 15th I phoned around to see if anyone in town was selling spot welders of the sort commonly used for battery tab spot welding purposes. Nope! So I decided to check into what went into the makeup of one. At first I had planned to visit a welding supply shop and start asking around, but I was dubious anyone would know much about this specialized type of welding - then it occurred to me to check the web.

   It seems the essence is a discharge of capacitors that creates a sudden very high current through the two welding electrodes and into the parts to be welded. The resistance to this massive current flow is highest at the point where the two pieces touch, and the sudden heat there simply melts the two parts together, creating a "nugget" of weld material. So it's also called resistance welding. Higher resistance metals like steel and nickel work better than the best conductors like silver, copper and aluminum, as the heat generated at the spot to be joined is higher and melts more metal. The electrodes, simply fat solid copper wire (eg, #4) with cone shaped ends, were placed, then a foot switch activated the discharge. A 100-300 amp SCR is used to switch the power. However, it seems the SCR limits the current considerably, and much more capacitance is needed. But the whole switch arrangement can be omitted. Then the instant the second electrode make contact the arc happens. Much skill is evidently required for this technique.
   There were commercial welders (2K$+), two-pulse welders, welders with sophisticated solid state controls, welders made with microwave oven transformers, electrodes which activated the pulse when they reached a certain pressure against the workpieces, and other refinements. One to 4.5 farads of capacitors(!) are typically used.

   But the key is the spark. Discarding all sophistication, I soldered together a rig with four 2500uF capacitors, two leeds with alligator clips to attach to a 12 volt power adapter, and two leeds to the electrodes, which were just solid #10 copper wires with the points sanded into a cone shape.
   This actually made a battery tab sort of stick to a battery. So I went out to get more capacitors. I bought two dozen 4700uF/25v ones at Queale Electronics, total 0.113 farads. Those were the biggest they had except at very low voltage ratings. (This was all before I discovered the massive "one farad plus" ratings others were using. They say "try ebay" to find the capacitors.) I soldered them to two bus wires of #12 gauge copper wire. This turned out to be rather thin - #10 or even #8 would have been better. My unit is powered by a dinky 12VDC power adapter, but it's the massive current pulse that welds.)
   (Be sure to get all the capacitors the same way around, and to hook them up the right way around every time. Polarized capacitors connected backwards can explode with quite a bang. You could put a diode in series with the power adapter line resistor to prevent backwards hookups.)
   I was using #16 wire about 3 feet long to the electrodes. Results seemed unsatisfactory, and I checked on the web again. Someone linked via youtube was boasting about how his (rather complex) welder was only 500 micro-ohms and so could out-weld commercial models. Another person mentioned switching from 3' of #10 to 3' of #8 and getting better results. (I also learned that the "nugget" tends to form towards the "+" side of the join. The positrode should go on the thicker or lower resistance piece.)
   I switched to very short #10 stranded wire electrode leeds. It made all the difference! Hmm... ultra-flexible #8 audio wire would be nice. The device can still only be used to tack very thin pieces together. I tried #26 nickel-brass and couldn't make it stick (my battery electrode plate thickness), but when I thinned a piece in the rolling mill (or use a hammer) to ~1.5mm (.005"), I could weld it.

   The technique that seemed to work best was to press the parts together with the positrode and then firmly tamp the negative down onto the work, preferably but not necessarily quite near the plus. The force with which the contact points are pressed together is important, with too much force leading to too little heat and no weld, and too little causing holes to be burned through the thin tab piece.
   Naturally a pit is made in the work where the negatrode touches down. (See the pitted practice battery in the picture.) I can see how the foot pedal would be better... but this simple unit basically works, and can be made cheaply and quickly.
   (Reading later, I see it's best if the negative is on the thin tab piece rather than the positive. This guy seems to have the good stuff and good info: http://frikkieg.blogspot.com/ . Seems his welder has 4.5 farads of capacitors(!), so I was on the right track using so many myself... but underpowered!)

The primitive spot welder:
- 24 capacitors totaling 112,800uf (@ 25V).
- two electrodes - #10 wire.
- two clip leeds to 12 - 18 volt power adapter (observe polarity!)
- resistor (I used 3.3Ω/5W) to limit charge rate from power adapter, and limits current after initial spark.
- the fatter the wires in the discharge circuit the better.
Note electrodes attach at center, not at an end: 1/2 the current flows through each side of the buss wires.

Welding practice: welded rag-tag end of a battery tab to 2nd battery - note all the pits in the battery case  where the negatrode touched after placing the positrode on the tab at the weld point. Though it has acquired a couple of extra holes and pits, the tab is well stuck on! Sunglasses might have been better than just "UV blocking" glasses for the sparks. But I soon started simply looking away before connecting.

   This was quite a satisfying little project with other projects dragging on and on - a successful prototype in about 4-1/2 hours including this writeup. (...writeup later expanded.) I may of course try to make further improvements later to weld thicker pieces.

First Improvements: By evening I realized I could cut the capacitor assembly in half and put it in two rows, and attach the electrodes to both rows in the new middle. This would double up the "skinny" #12 wire busses to give the effect of #9 wire instead. In fact, I would only cut one wire, and fold the other one back, to make it two rows. Say, maybe 3 rows of 8 would be even better!
   I'm not sure whether adding 8 more capacitors (4 rows of 8, 0.15 F?) or getting the effect of thicker wire would have greater effect. Certainly going from 36" of #16 wire to 10" of #10 for the electrodes made the difference between failure and success at least with small, thin pieces. Of course, I could do both. And I could go up to about 20 volts, but 16 volts from the "12 volt" power adapter seems a more comfortable margin for 25 V rated capacitors that are being asked to do a very hard job.

   In lieu of remote activation of the pulse, a means to prevent pits in the workpiece might be to clamp a piece of copper to the work. Bring the second electrode down on the copper and pit that instead. Being a low resistance, it should pit less. (Just setting the copper on the work doesn't make a good enough connection - I tried it.)

Before month's end, I put all the above improvements into effect,
using short lengths of "#8" (really #9 ...if!) very flexible audio wire.
A 12 volt, 1 amp power adapter, hooked to the alligator clip leeds, takes a few seconds to recharge
the capacitors for the next zap. (In theory, the series resistor should be 12 ohms or more.)

Copper isn't the best thing to try to spot weld - I didn't manage to weld the copper
electrode mesh shown to its leed wire - only to burn off some pieces of the mesh.
It did weld a couple of nickel or nickel-brass battery tabs quite nicely, though if I pulled hard enough,
all the welds broke off before the tab metal ripped. (With commercial battery tab welds,
it's about 50-50 in my experience.)

Turquoise Battery Project


   As happened a couple of months ago, I've done so much work and writing on this project it's become a long, tedious and somewhat disorganized report. This time, I'll leave the text as there are many details of potential interest to anyone working on batteries, but I think this shorter summary just touching all the key points is in order.

Chart shows the promise of Ni-Mn EV battery chemistry with proposed constructions:
sealed dry cells of bipolar electrodes stacked to any desired voltage
with catalytic recombination of H2 & O2 into H2O.
Later calculation (after only a guess for the graph) showed the horizontal placement
should have been entirely off the right hand side of the graph: 500-750 WH/L...
and so are Ni-MH or Ni-Fe dry cells (350-560 WH/L).

The rather 'hacked' test battery under charge.

   Both electrodes contain a mix of Ni and Mn for different reasons. In fact, if one makes a positive electrode of 60% Ni(OH)2 and 40% MnO2 'additive', it has about equal amp-hours whether used as a positive or a negative electrode.
   AFAICT, it seems the biggest difference between Ni-Mn (2.5 open circuit volts) and Ni-Fe (1.35 volts) is the need to add a bit of egg albumin to the manganese negatrode in order to raise the hydrogen overvoltage to handle the higher reaction voltage.
   And this month, I was pointed to a fabulous 2004 research report from India on new sealed Ni-Fe batteries [www.nickel-iron-battery.com], which mostly looked equally applicable to Ni-Mn. A few original Edison flooded Ni-Fe cells are still in use today almost 100 years after they were made.

   I decided to make my "full size" Ni-Mn battery with 3" x 6" flat electrodes, which turned out to contain material for about 20 amp-hours in 4mm (Ni+) and 1.5mm (Mn-) thick electrodes. I used copper mesh collectors and leed wires. I started with a "dry cell" with limited electrolyte, but the pressure rose rapidly. (The sealed case worked great!) But I soon flooded it and vented it, and found it would handle much more charging current - hundreds of milliamps instead of tens. Later, after reading the report from India, it looked like the best thing was to initially charge the negatrodes separately in a flooded tank, and then use them in sealed dry cells.
   A new catalytic technique to recombine O2 and H2 into water was employed that kept pressures remarkably low, which would allow much larger sealed cells than are currently practical. Their catalyst, however, uses platinum, which is pricey stuff (and cerium). I hope to find one that uses economical ingredients, like antimony oxide.

   The report also mentioned 675 Kg/sq.cm. (4.8 tons/sq.inch) as an "optimum" electrode compacting pressure for their iron electrodes - finally!, the first actual figure I've ever seen on the subject. I probably get around 1/2 that in the bolt-down compactor.
   It further mentioned use of PTFE (teflon) suspension as binder paste to "glue" the electrode together. This is probably better than "CMC gum" and I'm trying to buy some. Seems the only place to get it is China.
   I am also wondering if it's possible to find some inert but electrically conductive or semiconductive organic 'binder' substance that might be both 'glue' and increase the conductivity of the electrode, giving high current potential. Maybe that's a pipe dream, but it would be great if a small battery would, for example, start a car engine.

   Although I used neutral pH KCl salt electrolyte, hydroxide liberated by the charging reactions turned the cell alkaline, about pH 12 - 13. (That's still somewhat less caustic, and using edible salt saves you from having to work with "caustic potash".) In the alkaline environment, the copper collector plate and leed of the manganese negatrode were fine, but the copper in the nickel positrode soon dissolved into green goop. Even the stainless steel terminal bolt had some nasty corrosion. I put in a piece of nickel-brass for a collector plate, which is faring better but not well. By month's end there was green goop everywhere in and around the battery. (proving it's a green energy battery!) I eventually took some copper mesh and nickel-brass wire to Victoria Plating and had them nickel plated for $70 - a bargain after the prices I'd been quoted for any sort of actual nickel metal product. I'd rather have taken stainless steel mesh as being naturally more resistant, but they told me that can't be nickel plated. It then occurred to me that painting an alkaline earth element, hmm... barium carbonate?... preferably the hydroxide, but it's hard to make, on a positrode metal might help protect it from alkali, so that pure nickel might not be necessary. Until now, only pure nickel has been made to work in alkaline positrodes, but there may be another way. I'll be experimenting with that.
   The cell took the whole month to initially charge, and in fact it's still charging. I almost gave up more than once. It would go for a couple of days sometimes without there seeming to be any increase at all, then the third day the voltage would be a bit higher and it would hold it a bit longer. By the end of the month it would hold over 1.8 volts for a few minutes, and the figures were still slowly rising. I'm wondering how much hydroxide is being converted to carbonate via CO2 in the air with an open top for so long -- and when that and the corrosion of the positrode plate will overtake the charging and improvement will stop. I think all the "green goop" corroding off the positrode, and perhaps impurities from the CMC gum or impure pottery supply chemicals, is making for (as usual) very high self discharge, and that when everything is pure and right with nothing corroding away, it will hold charge fine. Initially charging the Mn's in a tank should be considerably faster and soluble impurities would dissolve out.

   Again, along with the chemistry and tests is the idea that bipolar electrodes with no collector plates could be stacked inside to create dry cells of any desired voltage in a single cell of almost "ultimate" energy density. It would also be cheapest. The edges could be sealed by dipping the complete assemblies in (take your choice) paint, glue, epoxy, wax... hmm, preferably something rubbery and flexible for when the nickel 'trodes swell a bit.

Cross Battery Checkup

   I cross checked the graph figures by weighing and measuring an actual Ni-MH battery. That should have similar amp-hours to Ni-Mn but at 1.2 nominal volts instead of 1.9. The nickel side is the same, and Mn gives similar amp-hours by weight to very good metal hydrides, and the figure is in the same range as the theoretical calculations.
   The 30.5 gram Ni-MH "AA" cell contains almost "nothing extra" and so has almost "ultimate" energy density: 2.6 AH * 1.2 V / .0306 Kg = 102 WH/Kg. Now account for the voltage difference: 102 WH/Kg * 1.9/1.2 V = 162 WH/Kg. The vertical placement of Ni-Mn on the chart is thus approximately borne out.
   An older 1.6 AH Ni-MH cell weighs only 25 grams instead of 30, which actually accounts for almost 1/2 the difference in capacity. 1.6 AH * 1.2 V / .025 Kg = 77 WH/Kg. (not 63 WH/Kg)
   For energy by volume, which I haven't tried previously to work out, the "AA" cell measures 14.4mm (diam) * 49mm ("+" button not included), or 5.54cc. 2.6 AH * 1.2 V / .00554 L = 563 WH/L. That's already off the chart. Multiply that by voltage: 563 * 1.9/1.2 = 892 WH/L. Since the case will be thick plastic rather than thin metal, it won't do quite this well, but it certainly seems I should have tried working it out before I made that chart - these are not going to use up a lot of cargo space! The 1.6 AH Ni-MH fares worse here: 1.6 / 2.6 x 563 = 346 WH/L. That's the only figure that's on the chart's horizontal scale. An equivalent Ni-Mn is hardly likely to be under 500 WH/Kg, thick plastic case and all.

   I plan to try out many ideas and materials in the coming months - starting with a more powerful electrode compactor, hoping to get better conductivity within electrodes.

   (This month's full gory details follow.)

Ni-Mn Battery with "Full Size" Electrodes

   The positrode I so casually grabbed to mate with the Mn negative for a test battery in May turned out to be the last lanthanum perchlorate electrode when I checked into it, rather than nickel hydroxide. Oops! I was fooled by the greenish colour, forgetting that not only nickel hydroxide is green, but the perchlorate is almost the same green - slightly brighter. The battery as such didn't hold charge, but I did find out that the manganese negative didn't seem to bubble a lot of hydrogen.

   The chemistry looked promising enough that I decided to make a Ni-Mn battery with my planned "full size" 3" x 6" electrodes, the size of the compactor.
   I would find out how much material makes how thick an electrode, and hence obtain a much better approximation of how many amp-hours to expect from what size and weight of battery.

Nickel Positrode

   I decided to try for 50 amp-hours. Nickel hydroxide is 289 AH/Kg if the valence change is 1.0. With the manganese additive raising the oxygen overvoltage, the nickel valence is said to go from about 2.25 to as high as 3.8, a change of 1.55, evidently with a considerable proportion of NiO2 present in the charged product. (Below valence 2.25, the conductivity becomes poor. My batteries are starting with straight Ni(OH)2 at valence 2.0, no doubt explaining some of the high internal resistances.)
   There is a fair level of uncertainty of what valence the nickel will actually attain with salt in the electrolyte. Having no special equipment, I can only measure the amp-hours attained in a successful battery and adjust formulations based on the result, a process so laborious I'm unlikely to define an "optimum" mix as part of the project.

I decided on these ingredients:

   Using 1.5 valence change as a working basis, 50 AH / 433 AH/Kg = 115 grams (of Ni(OH)2).
   Using half as much MnO2 as Ni(OH)2 (33%:67%), 57.5 grams
   Using more monel powder (by weight) than Ni(OH)2 for good conductivity, 125 grams.
   Adding 1 wt% (of the Ni & Mn) Co2O3 for even more conductivity, 1.75 grams.
   Using 1 wt% (of the total) CMC gum for 'glue' to hold the electrode together, 3 grams.
   About 70 cc of water.

Some Notes:
* I forgot to put a thin layer of calcium hydroxide on the collector screen as I've done with a couple of previous positrodes - a possible way to protect it from corrosion besides making it of nickel? I'd like to try barium hydroxide, but it seems harder to make. It might even help to protect the metal collector plate/leed wire structure.
* I think I'll also add the 1% antimony oxide to the positive electrode next time, as another possible way to reduce gas pressure by causing H2 and O2 to recombine into water.
* It's amazing to start with mostly turquoise green nickel hydroxide, mix in a couple of black things (MnO2 & Co2O3), and end up with black - no hint of green.
* I suspected that 125g would be more monel than needed, which would needlessly lower the AH/Kg of the battery, but since poor conductivity seems to have been my biggest problem so far, I just hoped that would be 'lots' and it would work well.
* But it turned out that the dry electrode initially had resistance of tens of kilohms. That was completely uncharged. Perhaps when the Ni(OH)2 reaches a valence of 2.25, it will be considerably improved.
* Also the readings are suspect. When the ohm meter automatically changes ranges, the apparent resistance 'jumps' by an order of magnitude. Still, it's nothing like ohms or tens of ohms.
* I've mentioned discharging NiOOH to beta Ni(OH)2 with H2O2. The opposite is to charge it to NiOOH, using NaClO (bleach). I sprinkled a few drops of bleach to absorb in, to see if that would result in lower resistance readings. It did seem to help, though it wasn't a big improvement.
* I used a copper mesh and a #10 copper leed wire. The leed connected to a stainless steel terminal bolt. The same arrangement was okay on the negative side, but the wire corroded right through within a day or two in the positrode. Furthermore, where copper touched the bolt, the stainless steel too was corroded.

   When I made the electrode from the mix, only about 40% of the 50 AH worth of loose mix fit into the compactor with 1/2" sides. It compacted into a 20 amp-hour electrode 4mm thick, 145g. (140 AH/Kg. If it wasn't for needing the rest of the battery, that would be great! The monel drags the figure down considerably.) I could probably up that to 30 AH, 6mm thick next time if the conductivities are good once it's charged. Apparently a battery of 60 amp-hours equivalent to a 100 amp-hour lead-acid battery will be larger than I visualized last month.
   Since I had more than 1/2 the mixture left over, I made a second 4mm electrode briquette, with no collector screen (134g - remarkably within a gram of the same weight as the first without its 11 gram collector screen), to be part of a central bipolar electrode for a higher voltage battery. My case was fat enough for two sets of electrodes of this thickness - would I dare try a 4 volt battery instead of 2? (I didn't.)

Depressing News: the Hydraulic Press gets Bad Press
...the bolt-down compacting system is Better!

   When I compacted the electrode with the new 12 ton hydraulic press, I pumped the handle until it took an awful lot of force to push it any farther. I pressed in the middle of the electrode, somewhat towards each end, and then the middle again. The powders compacted from 1/2" thick as poured in down to about 4.5mm. The electrode seemed pretty solid, but for comparison I dug out the bolts for the bolt-down compacting system and torqued them down. The electrode went down about another 1/2mm to 4mm thickness.
   It seems my bolt-down electrode compacting system is better than a 12 ton hydraulic press! Evidently the only point to buying a hydraulic press was to demonstrate this point. A 45 ton press might achieve about the same results as the bolt tightening system... for $2000 more!

   To speed up the labourious bolt-compacting process, perhaps all I needed was an electric nutdriver, and a holder for the compactor so I could work the driver without having to hang onto that at the same time?
   A 4.5 amp drill might hold a nut driver bit quite nicely, and it has a lot of torque. I tried that next.

   One thing I'll do differently if I make another bolt-down compactor will be to use a 7/16" or 1/2" steel plate for the bottom, as the 1/4" thick bottom plate bulges a bit in the middle. I may change this regardless. The 1/4" top plate also bulges, but it pushes down the 3/8" thick x 3" x 6" die piece onto the electrode, which piece (I trust) doesn't flex.
   Another thing would be to have even more bolts around the edges - at least one per inch.
   I'm also wondering now why I decided to reduce "full size" electrodes from 3" x 12" to 3" x 6". Although they'll each be somewhat easier, it means making and installing twice as many batteries to get the same result.

Manganese Negatrode

   I decided to go with a larger amount of almost the same mixture as last month for the negatrode, but with just a little of the finer nickel particles as derived from the hydroxide. Since I'd used 40 AH worth of positive material and since the battery should be positive limited (so it creates O2 when overcharged rather than H2), perhaps making about 50 AH for two negatrodes would be about right:

Expanded copper mesh collector sheet/backing with Cu leed wire, 6" x 3" (11.8g)
Monel alloy powder - 55g (conductivity, ??)
Ni(OH)2 - 5.0g (conductivity - reduces to Ni)
MnOOH (?) - 55g (active ingredient - overdischarged state)
Mn metal powder - 20g (active ingredient - charged state - add this last)
Sb4O6 - 1.25g (increases hydrogen overvoltage)
CMC gum - 1.35g (glue)
Albumin (egg white) - a smear (significantly increases hydrogen overvoltage even in PPM quantities)
HOH distilled - 15cc

   Assuming the Mn has in fact been successfully reduced from MnO2 to MnOOH, the balance is 30g of Mn overdischarged to valence [III] and 20g charged to [0], with [II] being the expected discharged state. It should thus take 15g of Mn to bring the 30g of MnOOH up to "normal discharged" state, leaving 5g of 50 charged. Since the nickel side was 40 AH, that should also leave 5g of uncharged Mn(OH)2 when the positive is fully charged. Assuming gas permeability, the migrating O2 from the "+" side during overcharging will spontaneously discharge as much Mn as is needed to prevent the "-" from overcharging and generating hydrogen.
   One might question whether there's any value in having undischarged manganese in the negatrode when the positrode is fully discharged, and also whether 5g, 10% extra, is enough uncharged material during overcharging to prevent hydrogen generation. I'm not going to try to answer these questions at this point.

* I found out this month that the best way to charge the negatrodes is probably immersed in a tank, with nickel plate positives bubbling oxygen gas. They're put into the battery later. That means the initial state of charge isn't important. Thus pottery supply NiO and MnO2 (assuming purity is okay) are probably as good as powdered metals, and they're cheaper and more readily available.
* Quantities should be adjusted to account for the oxygens. (I do wonder if the electrodes should be re-compacted after charging, especially if the compaction pressure was low to start with. Of course, the electrode needs space to hold the manganese hydroxide when discharged, too - but one wouldn't discharge the Ni again of course... or the copper mesh!)
* In the past, it has been considered that a sealed alkaline cell should be positive limited on charge so that if overcharged, oxygen is formed first and is converted back to water by discharging the metal [manganese] to hydroxide at the negatrode. Generation of hydrogen, which there was no way to eliminate, was thus prevented regardless of overcharging. With the catalytic recombiner, it would seem more desirable that both O2 and H2 start to be generated at about the same time for catalytic conversion, rather than to have one or the other form first and build up pressure. The research paper had negative limited cells. I'd guess that equal capacities would be ideal. The Ni side valence only goes down to 2.25, about 5/6 discharged, before the user will decide it's not putting out very well, and it would be the same for the Mn side, 5/6, if it had the same capacity. Thus the usual desire to have some reserve negatrode capacity would be in practice fulfilled.

   The Mn negatrode reaction, one with the elusive solid-solid (non-dissolving) reaction products on both charge and discharge:
  Mn(OH)2(s) + 2e- <=> Mn(s) + 2OH- [-1.55 V; 976 AH/Kg or 1513 WH/Kg of Mn]
  I neglected to weigh the electrode (with screen and leed wire) before installing it, but it should have weighed about 81 grams. The thickness was about 1.5mm. Add the 145g for the positive and it's 226 grams for 20 amp-hours, or 88 amp-hours per kilogram. Since it's two volts, that's 176 watt-hours per kilogram, considering only the electrodes. Add to that the case and the electrolyte.
   The center electrodes with no collector screens or leed wires are about 25 grams lighter, 201 grams or 99 AH/Kg, 198 WH/Kg.

 For a 12 volt battery:
 201 * 5 + 226 = 1231g
 20 AH * 12 V = 240 W-H
 240 WH / 1.231 Kg = 195 WH/Kg

Again this is for the electrodes themselves. The case and electrolyte will reduce this figure. However, it is perhaps instructive to note that current lithium batteries are said to produce around 80 to 120 WH/Kg (there may be some somewhat better), and that by adding less inert and heavy conductive metal filler powders, lower current but higher energy per weight Ni-Mn can be achieved.

Testing the Battery

   When the Battery was assembled, I tried charging it... with the usual disappointing results. At least a couple of times I almost gave up. But this time, fixing obvious problems that soon arose, trying a few things and, mostly, sheer persistence paid off, because in fact it took weeks for it to come up to its expected voltage and stay there when the charge was removed. Later I learned of a faster way of 'formation' of the electrodes to get the initial charge.

   First I decided to fill it with electrolyte, a flooded cell instead of a dry cell. It worked way better. I could put in hundreds of milliamps of charging current instead of tens, and yet the voltage stayed much lower. I decided it would have to be a flooded cell instead of a dry cell. (Later I reconsidered this.)
   Also the pressure crept up visibly to 5 PSI as I watched for a few minutes. In 2-1/2 hours it was up to 18 PSI. Through the clear plastic lid I could see the water bubbling. I unscrewed the pressure gauge to release the pressure. I decided it would also have to be a vented cell. Anyway my sealed case design seems vindicated! With a glued-on bottom, a rubber gasket at the top, and the bolts done up, there was no sign of leaks.
   There was quite a lot of bubbling for the first day, which didn't smell like hydrogen AFAICT. This subsided, but after about a day the currents also dropped a lot and the charge voltage went way up, to over 3 volts. It turned out to be because the copper connections of the positrode had corroded entirely away - and a piece of the stainless steel terminal bolt in contact with the copper was eaten away. It looks like copper is good for the negatrode, but not on the plus side, and also that stainless steel wasn't going to work either.

Corroded bits of copper mesh and wire from the nickel positrode.

The positrode was pieced back together, on a sheet of nickel-brass with a stainless steel screen.
This was wrapped in cellophane and re-used.

   So the question is what to use for connections? I suspected the nickel-brass (AKA "German silver", "nickel-silver") wouldn't last either, but I stuck a sliver of it in behind the now unconnected electrode to try it out. The battery went back to accepting high currents without the voltage running away. I suspected that only pure nickel was going to work for long, just like in alkaline cells - if it worked.
   Later I went down to Custom Plating and found they'd nickel plate a good size chunk of my extruded copper mesh and a piece of heavy nickel-brass wire for about $70. That should work except it may corrode in from the cut ends eventually, and anywhere it gets scratched during handling or when compacting the electrode. (Maybe they'd plate finished assemblies for a reasonable price, too.)

   Next I suspected there was a short circuit through the separator sheets (the Arches 90# watercolour paper and the cellophane). I took the battery apart. I found 3 holes in the paper. I probably made the holes sticking in the sharp sliver of sheet metal.
   The negatrode appeared to be in fine condition. The positrode fell apart, its copper mesh skeleton dissolved away. I reassembled the pieces on a sheet of nickel-brass, wrapped it in cellophane, cut a new paper sheet, and re-assembled the battery.

   Then I tested the pH. The pH was 12 or 13! Evidently in spite of adding neutral KCl salt, I had created an alkaline battery. That would explain the corrosion.
   In retrospect it doesn't seem too surprising, since the manganese (MnOOH - ?) gives off OH-'es as it charges, which will mingle with the K+'es in solution to form KOH. And the Ni(OH)2 in the negatrode as well gives up its OH-'es one-time, afterwards being fine metallic Ni. (I wonder what happens to the Cl-'es? Probably they remain active in the solution, especially seeing it isn't pH 14.) Here we see a situation analogous to the lead battery renewal process, where starting with neutral sodium sulfate salt in distilled water, the liberation of more sulfate from the battery plates yields an acid electrolyte of sulfuric acid and sodium bisulfate.
   Evidently then nickel or nickel plated metal will be required for the "+" side. I wonder how practical it is to make up my own nickel plating solution, seeing that nickel mesh and wire seem, absurdly, to be 400 $/pound and up from anywhere I can find them?
   This also means that the alkaline voltages from the table should apply: +.48 or +.51 for the nickel (depending where you read it) and -1.55 or -1.56 for the manganese, about 2.03 to 2.07 volts. Still much higher than Ni-Fe, -Cd, -MH or even -Zn!

   After a few days the cell still didn't seem to want to charge further and come up to its proper voltage, like most of my batteries. I decided to take it apart, rinse it out, and replace the grungy looking electrolyte.
   The manganese negatrode looked much the same as when I made it and only a little colour was evident in the rinse bath. But as I was rinsing it, I realized I hadn't heated it to cook the bit of egg white that was in it. Could it have washed out or degraded, and that be the reason it wasn't charging? If I couldn't make this battery work, only a new negatrode would tell for sure.
   The nickel hydroxide positrode, wrapped in cellophane, oozed greenish and yellowish stuff from start to finish. What was all that? Answering that question might also answer why the battery wasn't performing as expected - after all, I know I had a working nickel positrode earlier, that I  had coupled with a commercial battery cadmium negative.
   Next thing to check in the replaced KCl electrolyte was the pH. After a bit of time and charging, it was about 12, still very alkaline.
   The negatrode was bubbling strongly under charge, even worse than initially. The obvious suspect was the uncooked egg white.
   A simple thing to try before making a new negatrode suggested itself: dry the electrode, mix some egg white with water, pour it on and let it soak in, then heat the electrode at about 110ºc in the oven for a while to cook it. I did this. Then I reassembled the battery cell and put in on charge. Soon it was up to 2-1/2 volts, and something like a white merangue formed above the negatrode, which continued to bubble. It was yellow in places, where it looked a bit like what was leaking out of the positrode when I was rinsing it.

   It still wasn't holding much more charge. I added a couple of drops of methyl-ethyl keytone to the electrolyte. I don't know if it helped. A day later, it was holding just slightly more charge.
   The cell was, over the weeks, increasing in voltage. On the 17th, I read a piece on "forming" iron electrodes in a flooded tank with nickel metal "counter-electrodes" on both sides (doubtless bubbling oxygen when charging the negatrode), even though the electrodes were for use in a dry cell. That's probably the way it's supposed to be done, and the way I should be doing the manganese ones.

Hydrogen Overvoltage & Manganese

   I should perhaps clarify in detail the reason for the antimony and the egg white. A small amount of a transition metal in the negatrode raises its hydrogen overvoltage and this has been helpful with zinc, although zinc does charge without it under normal conditions. The transition element may be added as its oxide or in pure metal form.
   Antimony looks like the best to me and is evidently non-toxic, and the oxide is cheap at pottery supplies. Furthermore, there are indications that antimony (specifically) might cause a catalytic reaction to recombine hydrogen and oxygen into water, reducing gas pressure inside the battery. This has also been done using cerium and platinum, but platinum is costly.

   The manganese metal/hydroxide reaction is 300mV higher voltage than zinc, which has previously been considered to be the highest energy negatrode substance usable in aqueous solution. Instead of the manganese charging, hydrogen bubbles off the electrode: the water charges first. What I've never seen mentioned in battery literature is this (Science Magazine (23 October 1964, article: Effect of Traces of Large Molecules Containing Nitrogen on Hydrogen Overvoltage):
"Organic amines, present in very small concentrations (below 10-6 M) in 0.1 N H2SO4, cause a significant increase in hydrogen overvoltage, the effect being stronger the higher the molecular weight. The increase could be accounted for by the usual site-blockage concept. In the case of egg albumin, a drastic increase of over 300 mv was observed at 12.5 ma/cm2 for a concentration of only 0.01 part per million. A new mechanism is proposed in which the dielectric constant and hydrogen-ion activity are believed to be depressed in a region twice as thick as the usual transition region."

   Note that the "drastic increase of over 300 mV" is the very voltage by which manganese exceeds zinc in alkaline solution, and zinc charges. And that increase was given for an acidic solution and a very small concentration of albumin. It is probably even higher in alkaline solution, perhaps 0.5 volts or more, and with a greater concentration. Although the hydrogen overvoltage is different for different electrode metals, the antimony plus the egg albumin should surely provide enough increase to enable manganese to charge. Enabling manganese high voltage negatives in batteries using the 46 year old albumin finding is a fine new piece of "open technology"!

  (BTW the reason I don't use zinc is because of its problem of migrating, and growing "tentacles" (dendrites) during cycling, via the temporarily dissolved zincate ion. The migration of zinc to the positive electrode gradually reduces the capacity, and often the tentacles short out the battery. Otherwise Ni-Zn would make great higher energy batteries! Cadmium, right under zinc in the periodic table, has the same problem and Ni-Cd batteries rarely last their nominal rated 200 charges.)

Reducing Manganese Dioxide

   I added plenty of H2O2 to 10.00 grams of MnO2, to try to determine what form it became by the weight of the product. There was lots of bubbling. Then it had to dry out. I poured some water/H2O2 off once it settled.

MnO2 [IV] - molecular weight 87 - 10.00 g

As I saw it, the chief possibilities and their weights were:

MnO [II] - 71 - 8.16 g
Mn2O3 [III] - 9.08 g
MnOOH [III] - 88 - 10.11 g
Mn(OH)2 [II] - 89 - 10.23 g

   The measured weight once it had dried was 10.05 grams. (The resolution of the scale is only .05g.) I think it lost a tiny amount of Mn material in the water I poured out. I think I'll assume it's MnOOH. That's closest by weight and a likely reaction result. The best result would have been valence two - MnO or Mn(OH)2. A less likely possibility is that it acted simply as a catalyst that caused the H2O2 to decompose and is still MnO2.
   Whether it's MnOOH, Mn(OH)2 or MnO2 it's about the same weight to attain the same amount of manganese. Only the balance of charge shifts. This can be compensated initially by the proportion of Mn powder added, or later by charging or discharging the nickel side - if I recognize that it's out of balance.

   Note: On the 17th I found a means of independently charging/forming the negatrode, which means it could perhaps be brought to the desired state of charge independently of the positrode before the battery is assembled.

Useful Miscellaneous Notes/Conclusions

Clear Plastic Covers

   I found an advantage to having a clear bottom on the battery: most of my covers have leaked under pressure. With the clear bottom, I can see how well the glue is covering and if there are air bubbles or gaps. I can also see that the glue seal is much better if the whole thing is well clamped while it's wet (not a surprise!), which I did by assembling the case and screwing the top cover on.
   I can see that for a production situation, it would be much the best to make molded cases with only the top to be fitted afterwards -- as is the case with most liquid filled batteries. Pockets molded in to put in screws to hold the top on would be superior to a glued-on top IMHO. Then defective batteries could be repaired. Of course, I'm still thinking of small production. If virtually every battery comes off the line perfect and lasts for 10 or more years in EV use, it would matter little.
   But it looks like it's all academic... Ni-Mn cases will need to be vented and flooded rather than sealed and dry cells.

Vented Flooded Cell Cases and Battery Sizes

   After trying the battery and seeing that they were going to have to be vented flooded cells, suddenly the whole case problem becomes much easier - it doesn't have to hold pressure.
   Making the cases from flat pieces of plastic and gluing them together (well clamped till the glue sets!) as I had tried earlier on is probably simplest the way to go. If I or anyone gets the chelated lanthanum perchlorate chemistry going for higher energy density and lower pressures, dry cells in sealed cases will likely again become a topic of interest.

   And the terminals don't have to go through and hold against pressure without leaking. Indeed, for homemade batteries I see no reason not to just stick the collector wires through tight holes in the lids, or maybe loose holes and use RTV rubber cement or silicone to make a seal. It solves how to connect internal terminals in a way that won't corrode - by eliminating those connections entirely.

   A second aspect of vented cases is the elimination of size constraints. Where I would have used about 3.25" x 2.5" x 6.5" (internal) size for a 12 volt, 30 amp-hour battery that would take the pressure, I could now go for any size, even as big as typical lead-acid batteries, but with very high amp-hours. The one constraint is the size of individual electrodes, which still have to be strongly compacted.
   One thought would be to make a battery of the same 3" x 6" electrodes, stacked in parallel and in series to make a single 36 volt battery for the Electric Hubcap system. For example, a 36 volt, 120 amp-hour battery would be four electrodes wide, 18 pairs in series. That would be (internally) about 6.25" tall x 12.25" wide x 7.25" across, roughly the size and weight of one of the six lead-acid batteries that it would replace, about 50 pounds instead of 300.

   A final note is that thinner plastic might be used if there's no pressure, slightly decreasing battery weight and size. Then again, that would apply to batteries the size I was planning for sealed cells. If quite large batteries are to be made, they'll probably need thicker walls.

   An less final note is that after reading the Indian experiments into sealed Ni-Fe cells, I think the catalyst idea to almost eliminate gas pressures is admirable and that sealed Ni-Mn cells deserve further experimentation.

Electric Electrode Compacting

   Using a 4.5 amps variable speed electric drill as a nutdriver reduced the labour of tighening and loosening all those bolts on the electrode compactor. After torqueing them down with the drill, I tried torqueing them further with the wrench I'd been using. Some of them went a little farther, some wouldn't budge. If there were just a few more bolts around the edges, eg one per inch or per 3/4 inch, it would have been better compacted than with the wrench, with less effort and less stress on each bolt. But it seemed good enough. I stopped using the hand wrench.
   The chief problem was holding down the compactor. I ended up with it between two C-clamps, but once it got away and the spinning compactor sent tools on the workbench flying around.
   Another problem with the compactor (I may have said this elsewhere) is that the 1/4" steel bulges when the bolts are done up. The next compactor will use a 1/2" steel base.

   Later I added 6 more bolt holes to the compactor, bringing it from 14 to 20 bolts. The spacing is rather irregular, and on this compactor it would be hard to fit any more.

   Reading a Ni-Fe battery experiment paper, I see they compacted the Fe electrodes with 675 Kg/sq.cm, or 4-1/2 tons per square inch. If my compactor is roughly equivalent to a 50 ton press, 50/(3"*6"=18sq.") = 2.7 tons/sq.inch. That's just over 1/2. But it's the right ballpark, anyway, the right order of magnitude.
   They left the pressure on the electrodes for 5 minutes. Maybe if I just leave the box screwed together for 10 minutes... (Warning! Faulty logic alarm!!!)
   I sensed the desirability of making the new box with a 1/2" steel plate base (forget 3/8"!) and more bolts as closely spaced as practical around the edges, and I bought the metal to make it, with a bolt every 5/8".
   On the other hand, one must consider that nickel electrodes swell inside the battery - alpha nickel oxyhydroxide theoretically takes up to 44% more space than beta nickel hydroxide. (This is counterintuitive since fluffy nickel hydroxide actually has more hydrogen atoms in it than oxyhydroxide. The oxyhydroxides must be really fluffy!) The electrode then, compacted to satisfaction, somewhat uncompacts itself. (10-20%?) How much compaction then, is really useful?

  Naturally I want to make electrodes as large as is practical - I'd do 12" x 12" if I could, and maybe have one battery for the whole car. The gas pressure in sealed cells precludes such sizes. But if the compacting pressure is insufficient for 3" wide electrodes with good conductivity, one could make them 2" wide and get 1.5 times the pressure with the same force, 1.5" wide and get twice the pressure, or even 1" wide and get 3 times the pressure. If that sort of solution becomes necessary, I might end up doing foot long, skinny electrodes in narrow squared-up pipes. The idea could be tested using smaller inserts inside larger compacting boxes.
   Or perhaps there are other simple means of compacting electrodes. For example, the force could perhaps be doubled by compacting 1/2 the width at a time - if that didn't break the electrode in half. (A gradual bend at the "middle" edge of the die?) Or perhaps some sort of pressure roller arrangement, though I haven't come up with a design I'd have confidence in myself.


   The copper leed wire on the negatrode looks fine, and the expanded copper mesh collector is still holding the negatrode together after a month. Given that it seems to work, copper should be the best material given its high conductivity.

   The one from the plus side soon had such a coating of copper oxide it looked like it had green plastic insulation, then it and the mesh inside the electrode corroded away altogether. I'll have to try other materials.
   There's the stainless steel mesh for the positrodes next time, and nickel-brass wire. If the wire corrodes but not the mesh I can doubtless find stainless wire. If stainless corrodes I guess nickel plated (or __?) mesh will be necessary - with nickel plated wire if the nickel-brass also corrodes.
   I put the present 'positrode' back together with a nickel-brass plate wrapped in stainless steel mesh. We should know soon enough.

   I've been setting the copper or stainless mesh in the bottom of the compactor so it's at the rear of the electrode, but it occurs to me the best internal connectivity, and the best strength, will result if the mesh is in the center. It's hard to resist the mental temptation to think the wire should be at the back, but in fact the mesh permits electrolyte ion flow penetration, and the ion flow in the electrolyte and the electron flow in the wires are unaffected by each other.
   If the mesh is in the middle of a 6mm thick electrode, no part of the electrode is more than 3mm from it, instead of up to 6mm. The ions still have to penetrate 6mm either way.

   Tentative layers of the battery would be:

-- nickel-brass plate
-- stainless steel mesh grid (with calcium hydroxide?) If nickel mesh was available that would be better.
-- nickel 'positrode' material as detailed above
-- all of above wrapped in cellophane with one layer facing 'negatrode'
-- sheet of Arches 90# watercolour paper or equivalent
'optional', for multi-cell batteries of any higher voltage, multiples of two layers compacted together:
-- manganese 'negatrode' material: no connection leed, no collector mesh - just the plain briquette
-- (Thin calcium hydroxide layer?)
-- nickel 'posode' material - just the pure briquette.
-- this "bipolar briquette" wrapped in cellophane
-- sheet of watercolour paper.
-- manganese 'negatrode' material as previously described, compacted around expanded copper mesh, with #10 copper leed wire
   (leed wire flattened and wrapped into a folded edge of the mesh, left round at terminal end)

   I'm considering whether edge sealing might be done by something like wrapping electrodes in a "U" shape with cellophane and heat gluing the edges to form bags with open tops. Or perhaps make cellophane bags separately and slip the 'trodes into them.
   A way to get surer separation with the paper might be to fan-fold it left and right between electrode pairs.
   And, with a little paper overhanging at the bottom (of course!), perhaps dip it in tar or paint or epoxy or wax or something rubbery to seal the whole bottom edge, and perhaps the sides.

   July 29th: The nickel-brass is still getting corroded in the positrode. I've replaced another n-b sheet 'leed' after its remains had fallen off. This time, I used a nickel plated wire (of nickel-brass), though I had flattened it and tried to weld it since having it plated, so it could hardly be said to be truly protected.

Real Watt-Hours

   The nickel-manganese test battery - with one cell in it but room for two - was 822 grams: thick plastic case, electrodes, excess water, internal plastic space fillers and all:
20 AH * 2.2 V / 822 g = 53 WH/Kg.   (Final batteries will perhaps triple this figure!)

   Even with all the waste weight, it already compares favourably with (properly derated) lead-acid at: (40 * 0.6 =) 24 WH/Kg, or with the Changhong 10 AH size nickel-iron battery:
10 AH * 1.2 V / 625 g = 19 WH/Kg.

Show me the Money!

   John McCain wanted to offer $45,000,000 to anyone who could come up with a better battery for EVs if he became president of the USA.
   But he lost the election, and perhaps it wouldn't have been available to a Canadian anyway. In the usual course of events, the inventor is left to starve as his invention spreads over the world... or for various poor reasons goes to waste.

Electrolyte Reaction Comparisons

   With Ni-Mn, Ni-Fe and Ni-Cd, the net result during charging is that hydrogen comes out of the nickel hydroxide and and hydroxide comes out of the "-" side, making H-OH, water. The Changhong Ni-Fe cells are visibly fuller (about 1/8" in the 10 AH size) when charged than when discharged.
   With Ni-Zn, the ZnO (which evidently forms instead of Zn(OH)2) "-" gives off its oxygen, also making water with the nickel's "H"es, though half as much.
   With Ni-MH, the H's from the nickel "+" are simply packed into the hydride negatrode alloy's 'pores' and the net water content remains the same.
   With the chelated lanthanum chloride/perchlorate chemistry, the "+" side oxidizes by absorbing oxygen during charging instead of by releasing hydrogen. If that's coupled with (eg) manganese, the two OH'es from the Mn(OH)2 yield the O it needs and an H2O, but half as much water as with a nickel positrode supplying more hydrogen. As well as replacing the bulky, heavy nickel side of the battery with much less material for the same amp-hours, this is the chemistry I hope will have lower gas pressures as sealed dry cells. Alas, I must leave its development for later or for someone else. I've created the concept and many techniques which may be employed in its implementation and published it all in these newsletters, with increasing clarity of concept over time from the initial muddy ideas. I hope that's enough that its development can be completed, and that that will be done in the same "open technology" spirit with which I started it.

   I expect the perchlorate chemistry would have tolerably low pressures and could be sealed as dry cells, but Ni-Mn is still an improvement over anything else out there. For this unpaid researcher who needs to move on, Ni-Mn will do.

Separator Sheets

   Real cellophane ("Pacon creative products, Cellophane" - from Island Blue Print) seems to work well as a microporous membrane to wrap up an electrode. Beware of plasticized "cellophane", which is non-porous! The Arches 90# watercolour paper (from Opus Picture Framing) makes a great "non-woven" separator sheet, a uniformly thick and dense mat paper with wet strength including in salt and in alkali. Both of these were found at art supply stores.

Side Thought: High Energy Density Nickel-Iron

   As I think about it, it seems a puzzle that no one has ever made high energy density Ni-Fe batteries for automotive use. They would be cheaper than Ni-MH, long lasting, and as a century old technology no patent can be used to kill them. The reason they're only 20 to 40 watt-hours per kilogram as made today is largely because of the construction, and also because the "Ni" side chemistry hasn't been updated: they're made only as "pocket batteries" almost the same way Jungner and Edison made them 100 years ago. The perforated metal plates weigh perhaps as much as the electrode material they contain, and the open spaces between plates are filled with quite a bit of water. And then because it's all so much bigger than it needs to be, the cases are heavier. This construction is also quite costly compared with modern types.

   It seems to me that Ni-Fe batteries with an energy density of somewhere between 70 and 100 watt-hours per kilogram could be made simply by making them with the same sort of electrode briquettes as for Ni-Cd et al, compacted around nickel foil or mesh to replace the heavy exoskeletons. The higher end of these figures, not out of line with the best EV batteries currently available (at a much higher cost!), would apply if my electrode stacking construction (topic above) was applied and with the best modern "Ni" formulations.

   Of course, all this is suddenly redundant just as I write it, since Ni-Mn should provide a very similar low cost battery with almost twice the energy - judging by the chemistry and characteristics of manganese as far as I can tell and the stated efficacies of the techniques I've hunted down and employed to raise the hydrogen overvoltage.

   I think the fact that high-energy nickel-iron was never done shows just how little creative thought has gone into battery development, in spite of intensive battery research efforts (mostly geared at pricey lithium types). Everyone seems to take the development side of "R & D" almost entirely for granted, assuming that once the principles for inventing something are discovered through research, there's no need to support it - development to the "invented" state will happen "by itself". Here is a case! While people have long struggled to find decent, economical batteries for electric & PHEV vehicles, a good, practical working solution has been sitting right under our noses for almost half a century, simply for want of being developed!

New Nickel Iron Sealed Dry-Cell Battery Research - Equally Applicable to Nickel-Manganese!

   After writing the above, I found recent research (2004 report) that has been done to make a sealed, maintenance-free nickel-iron dry cell. No figures were given in the paper for the energy density, but it appears it was essentially the sort of battery I wrote of above - even better. It also says the authors didn't know of any similar research ever having been done.
   It is exciting and promising. The only down side is that a new catalyst that keeps the gas pressure amazingly low (hence allowing the sealed cells) contains platinum, which can't help the economics.
   The fact that this valuable research was done in India shows not only that there is creative genius in that land, but, considering how badly we need such batteries, the unfocused disarray which is the state of R & D in the west. A critical eye overseeing battery technology and allocating research funding would have seen, or at least been receptive to the idea, that Ni-Fe with modern constructions was an undeveloped area worthy of research, and probably of development to commercial stage even without sealed cells. If now the Indian government has followed up by helping to get the sealed cells commercially produced, the west could soon be importing its best EV batteries from India.

   This study was located via a new information page from Victoria BC, www.nickel-iron-battery.com . It looks like I can probably do maintenance-free Ni-Mn sealed dry cells after all! A catalyst without pricey platinum would be an advantage, and that alone should bear more research.


   The above experiment also mentioned using "6 wt% PTFE Fluon Suspension" as (presumably) the binder to hold the negatrode together. PTFE has also been stretched to make microporous separator sheets used in batteries (not to mention in gore-tex). Previously I'd seen "not more than 1 wt% CMC" as a figure and material to use. I presumed this to be the same as "CMC gum" from the pottery supply, and presumed them both to be sodium carboxy methyl cellulose. I had also purchased a bag of "Veegum", a bentonite clay mix, to try out.
   So I I thought I should try to track down a source for PTFE suspension - it's probably better, and it would seem you can use more of it, presumably making physically stronger electrodes for handling until they're in the battery, without causing other problems.
   Also, they mention rolling out the 'putty' electrode material into a sheet (and then folding it over the mesh, which would get the mesh in the middle as discussed above) before compacting it. Rolling the 'dough' form would help get a uniform thickness and hence uniform compaction, which I didn't get on the first large size manganese negatrodes - there were 'thin' spots that weren't well compacted.

   Next it occurred to me that both those binders are insulators. If one could find a stable but electrically conductive, or even semiconductive, binder material, the electrode would handle far more current, even with less compaction. "Small" batteries with the 3"x6" electrodes might start your car instead of providing only tens of amps. Perhaps now that I have a working chemistry, I'll go back over my earlier experiments with acetal ester and the like, and see if any such binder can readily be made.

   It's doubtless worth finding the best possible binder. An electrically conductive one would be another 'coup' for this project. And of course I'm going to try doing those bipolar electrodes with no 'backbone' whatsoever, and they have to hold themselves together, which the CMC isn't going to accomplish.

Victoria BC