Turquoise Energy Ltd. News #86
  March 2015 (posted  April 3rd)
Victoria BC
by Craig Carmichael

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

* A new design concept for an add-on wheel motor system
(see Month in Brief)
* Porous nickel sheet and powder for nickel battery electrode (see Month in Brief, Electricity Storage)

Month In Brief (Project Summaries)
 - Aquaponics & LED Lighting Progress - 'Ultra efficient' Electric Chevy Sprint project to be resumed - Axial Flux Switched Reluctance Motor (AFSRM, replacing unipolar motor) - Add On Wheel Motor: a New System Design - AFSRM Generator for Windplants et al - Electric Weel Generator - My/R & D finances? - The Usual Mazda Batteries Update - Turquoise Battery Project: Ni-Ni battery production ideas, great new nickel negode idea!

In Passing (Miscellaneous topics, editorial comments & opinionated rants)
 - Ceres, Vesta and Vegetation - Unsustainable World Population

Electric Transport - Electric Hubcap Motor Systems
* 'On hold' Chevy Sprint Electric Hubcap/Planetary gear torque converter & Centrifugal clutch is to be completed.
* Electric Caik Axial Flux Switched Reluctance Motor - AFSRM (was unipolar motor)
* Active Reluctance Generators: no cogging and output voltage regulation!
* A Switched Reluctance EV Motor by Ricardo.com (85KW!), AFSRM design papers
* Electric Weel - Huge low RPM generator [at last] nears completion

Other "Green" Electric Equipment Projects
* Aquaponics & LED Grow Lighting - tilapia pool - Beans?!?
* Cordless Lawnmower with NiMH dry cells: User Review

Electricity Generation (No reports)

Electricity Storage - Turquoise Battery Project (NiMn, NiNi), etc.
* Cylindrical cells with carbon rods?
* Ni-Zn or Ni-Ni acid battery with oxalic acid?
* Ni-Ni cell with nickel-brass sheet for negode?
* To make porous nickel sheets and porous nickel powder! (from nickel-brass sheet and monel powder)
* Packing peanuts for carbon film, carbon nano-particles?

No Project Reports on: Variable Torque Converter Transmission, Magnet motor, Lambda ray collector, evacuated tube heat radiators, CNC gardening/farming machine.

March in Brief

Aquaponics & LED Lighting Progress

    I spent the early days of the month digging and working on a 'trench' pond in the garden for the tilapia. it looked like it would be in the way of access wherever I dug it, then I got the idea to put a bridge across it. So far that's just a fat plank. There wasn't much extra pond liner, and I sandwiched the edges between boards to keep it from drooping down. The building supply ran out of 8x16" paving blocks, and I bought some cement and sand and made some larger, thicker ones of my own. I may get around to a second batch some time.

   The water was too cold for tilapia, and I went to a pet shop and got a dozen goldfish to keep any mosquito wrigglers out of the water. (But it's probably too cold for them too.) I got some duckweed and put it on top, and when there's surplus I can take it in and feed it to the tilapia - goldfish don't eat it.

   Inside, the beans in the aquaponics drain-down bed continued to grow, hitting the ceiling and blocking most of the light from the window. One flower appeared for a couple of days and then wilted, with no more appearing. The beans were just one of a number of things I planted in there last fall, but they overgrew everything else and prevented any other crop. I certainly won't try growing beans in indoor aquaponics again unless I find out why they're not producing beans.
   The tilapia fish also continued to grow. The large female was almost 12" long in a 12x12x12" space. They will certainly want that (~400 liter) trench pond once it warms up and I get the beds connected and a pump running. The little one was 2-1/2" long by month end.

   I didn't do anything with LED lighting except decide that a switching regulator would give a lot more flexibility and perform better than the linear one, except with certain combinations of voltages and emitters which weren't nearly as much "the norm" as I expected them to be before making and using a few lights with the varying solar panel/battery system voltages. (And I was so proud of my minimal components constant current linear regulator circuit!)

Axial Flux Switched Reluctance Motor (AFSRM)

   Out of sequence, I'll mention that it now looks like the AFSRM project now looks like it will stretch out a few months. The motor, while very doable, is also very different from the BLDC magnet motor, so there'll be a lot of new design. That suggests it might be a good idea to complete the current version car drive first. It's all been sitting nearly assembled and ready to install and try out in the car. It's worth completing it and seeing if I've finally found a workable path to a high-efficiency vehicle drive.

   The "switched reluctance" or "SR" type of motor seemed to have all the advantages of the bipolar (or unipolar) BLDC types, would use the same motor controller as I had just made for the unipolar, and had reduced weight, cost and rotor thickness, plus greatly increased RPM capability compared with my rotors with magnets.
   By about the end of February I made up my mind to convert the new Electric Caik motor from unipolar to the SR type. The only differences seemed to be the rotor and perhaps a few stator wiring changes. The apparent similarity proved to be deceptive.
   After thinking about it while working on the tilapia pond, on the 12th and 13th I designed a rotor that I hoped would have a relatively low level of torque ripple, that being said to be the worst feature of the reluctance type of motor. Later a correspondent pointed out that the main reason for the high torque ripple compared to BLDC magnet motors is that the SR coils can only pull steel, not repel it, so more poles are needed to get equivalent evenness of forces.
   The rotor had the aspect of an "iron cross" with a circular outside edge, and I made it by cutting out the four sections between the four "lobes" from a ready-made "brake rotor disk", cast from "soft magnetic" metal - that wouldn't magnetize when rubbed with a supermagnet. The cross section "sort of" followed the curves of the round coil cores.
   It felt quite different installing a thin, magnetless rotor that didn't try to grab tools off the bench and suck itself into the motor unbidden while your fingers were still under it. And once it was on, it didn't try to cog to different positions. Only the friction of the grease-stuffed bearings kept it from freely spinning.
   It also didn't turn, with 10 amps of current. The magnetic forces were too weak. The sort of huge flux gaps and plenty of flux for a relaxed BLDC design is replaced in the SR type by tiny gaps and a struggle to find enough flux for decent torque. It needed a wholely different design and layout.

   So I took two coils, the rotor and the lab power supply to deliver a constant 10 amps to the coils, and started doing some electromagnetic experiments on the bench to maximize the forces. These pointed to better ways to go, and I made modifications that should get things working. A steel plate behind the coils works to complete the flux circuit. Then someone suggested using two rotors, one on each side of the stator, to double the flux. This seemed like a good plan. I also thought of putting a ring of metal around the outside rim of each 'donut' coil to concentrate the flux like in a cup magnet, where a weak ceramic magnet has "supermagnet" levels of flux in the small space between the round magnet and the rim surrounding it.

A cup magnet

Although I got the metal, I hadn't found a heavy sheet metal shear to cut it into 1" strips by month's end. (Tin snips do an ugly job.)

   Then in the latter part of the month it occurred to me to do a web search, and I found two technical papers (PDF.s) describing axial flux switched reluctance motors (AFSRM.s).
   It turns out that the axial flux SRM layout is indeed much superior to radial flux, but for different reasons than with BLDC magnet motors. The flat rotors and coils with a flat profile can have substantially larger adjacent surface areas for flux interaction. Also, the usual two rotors with the stator between them virtually doubles the forces available with the same current. With the right designs, levels of torque can be attained that will move vehicles by direct drive, for in-wheel motors as small as 12" diameter. Such motors are however pretty heavy for good vehicle suspension handling.
   I plan to finish my small "AFSRM" as conceived and test it out, along with the motor controller, which luckily is the same one as for the unipolar magnet motor, that I designed and made earlier this year. But the next and larger design for a car motor should probably be more along the lines of the test motor built and tested in the Japanese AFSRM design paper.

   Along the way I made a balanced "arm" to measure static "locked rotor" torque, in conjunction with a weigh scale. (After shaping the pointer, I cut and ground metal off an "L" piece at the far end until it balanced.) It'll be far more sensitive than a torque wrench. The "pointer" that presses on the scale is exactly 6.0" from the center of the shaft, so it's easy to convert the "weight" measured into foot-pounds.

Add On Wheel Motor: a New System Design

   Also along the way, I've conceived what I think should be a good solution to large and heavy in-wheel motors, as proposed in the two papers and many other places. Placing an SRM directly on a wheel's axle where its RPM is limited to the low speed of the wheel, is a waste of one of its big advantages - the ability to run well and safely at high RPM.s. A motor running at 0-5000 RPM will be much smaller and lighter to get the same power as one running at 0-1000 RPM connected directly to a car wheel. With 1/5 the torque it is geared down with a fixed ratio, eg 7 to 1, to deliver sufficient car-starting torque, and because over-revving for this motor probably means (say) around 10000 RPM or higher, 7000 RPM on the highway is fine. The rotors spinning at such high speeds are (as I envision them) solid pieces of steel that won't fly apart, and fairly light as steel rotors go. That means no variable torque converter, clutch, or gear shifting is needed. And that leads to major simplifications.
   The first type of gear reduction that came to mind was a large planetary gear, perhaps with plastic 'planet' gears to eliminate lubrication requirements. The motor would be as originally conceived, in the "hubcap" position, bouncing up and down with the wheel, probably with some flexible coupling making for some level of "debounce".

   Then a solution that seems superior came to me. The "on wheel" or "in wheel" motor is conceptually elegant, but then one hits bumps in the road, which are more easily dealt with without a considerable extra weight on the wheel. One solution that's been used is to mount the "wheel" motor within the vehicle. Drive shafts with CV joints then connect it to the wheel.
   My plan, for rear wheel drive only, is (AFAIK) new as applied to cars: For an outside "add-on" electric pancake motor mounting, a telescoping 'arm' runs from the center of the wheel to somewhere on the car body, either in front of or behind the wheel, about level with it. (sketch below) The wheel end of the arm moves with the wheel as it bounces up and down with bumps. But the body end is almost stationary. The telescoping or flex mounting piece attached to the car body to allow the arm to move with the wheel is very short. (Probably with more than one attachment point for support against twisting.)
   The smaller, lighter motor is mounted on the arm, near the body end. It doesn't move around very much, thus it will have little effect on suspension handling. But being on the arm which is connected with a bearing et al to the hub of the wheel, its position relative to the hub of the wheel is fixed. It is then necessary only to connect the motor to the wheel with an efficient drive belt: toothed, flat or poly-V, with pulleys sized to give the desired reduction. And the pulleys can easily be changed to see what reduction ratio works best on the road.
   By putting the pulley on the axle inside from the wheel, and the motor in a "hump" (eg) behind  the axle in the luggage space, the same arrangement could even apply to a vehicle as manufactured, with the drive belt replacing the more costly, heavier drive shafts with CV joints.

   There are of course disadvantages. The belt is subject to wear, and will need occasional replacement. And this drive train, short as it is, should be more or less enclosed to protect it from road dirt. ...And it just doesn't seem as "cool", somehow, as a motor in line with the wheel.

First sketch of the new idea wheel motor mounting system.
The motor (presumably) is mounted on the front of the bar with the pulleys behind the bar.
The wheel pulley bolts to the wheel lug nuts, and has a center bearing to attach the bar to.

    For a while, improvements to this layout were occurring to me almost by the hour!
- Upper and lower body end bolts/pivots can give anti-twisting support, which it would doubtless need.
- The motor can be mounted right over the pivot point at the back to minimize effects on suspension/handling.
- Belt adjustments by motor mounting position, eg with slots for the mounting bolts. (many potential variations)
- Safety cover over the belt & pulleys - to keep out dirt and gravel as well as fingers.

   Suddenly the potential for DIY creation of electric car drives with small motors at home, including "hybridizing" a gas car rather than "converting" it, comes back into focus as an attainable objective!

AFSRM Generator

   The absence of cogging also made me think of SRM potentially as a generator. My magnet BLDC motors wouldn't be, eg, good wind power generators because the cogging would prevent them from starting to turn in a lighter breeze. With no cogging, the reluctance motors should make great windplant generators. A "problem" is that without a motor controller, they would just spin without making electricity. On the other hand, with an appropriate controller they could probably be coerced into putting out a constant voltage regardless of RPM. (Within attainable limits of course.) That would be a tremendous advantage. Where one might have a typical "passive" generator whose output voltage varies linearly with RPM, followed by a very flexible DC to DC converter to output a constant voltage, instead one could have an "active" reluctance generator putting out a constant voltage at a maximum power point regardless of wind speed. The DC to DC converter would be replaced by the motor's controller so the controller would be an alternative electronic component, not an additional one.

Electric Weel Generator

   With a SR generator still being undesigned, work continued on the original plan for the big generator for the floating hydro power unit. We got the stator coils wired together and put the magnets on the rotor. It's almost finished.

Magnet Rotor. (More lexan reinforcing pieces are to be added to this main piece.)

My/R & D finances?

   I got my first monthly payment of 1000$ from CHIP bank - owed with interest to be collected someday from my house or estate. I also applied for Canada pension plan (CPP, 341$). My finances took another turn for the worse when I finally got my Turquoise Energy Ltd. income tax refund for 2013. For every project they didn't like they knocked off a percentage (just in case I wasn't worth 22000$ for any one or two of them), and the government has apparently that decided rent or equivalent for facility space to do R & D doesn't constitute a legitimate R & D expense. It looked like that would knock it down to 12000$. But I finally got my 2013 refund on March 20th 2015, nearly a year late, and it was even worse: somehow it was beaten down to just under 8000$. I expect that pretty much covers all the time I spent doing all the paperwork I have to do to qualify for the SR & ED program - but not the actual inventive work. It certainly does little to compensate for what I've invested in the projects in time and money. That has been financed with a considerable mortgage on my house. I wonder how much the audit itself cost the taxpayer, complete with scientific appraisal by a university academic who seemed to sneer at the fact that I had no university degree and wasn't going out of my way to try to get university collaboration. The salaries of the two Canada Revenue employees are both doubtless far higher than what I've been living and doing R & D on. Now I've started in on all the same paperwork for 2014, all seemingly no less exacting than if I was claiming millions of dollars. And when will I receive some pittance for having done that? Another product developer I know, and so probably many others, have been similarly affected. It seems the government is cutting everything that won't cause a public uproar (and apparently even some things that do), while it involves Canada in shameful foreign wars and moves us stealthily toward becoming a lawless police state, in step with the USA. It doesn't look like the next advances in human society will come from the west.

The Usual Mazda Batteries Update

   The electric Mazda with 11 batteries and 144 volts has become somewhat more practical again, but still only for trips under 6 miles or so, or where I can charge a while at the far end. I've been to a friend's at 5 miles distance for a 10 mile round trip a couple of times, with a 2-1/2 hour [slow, float] charge while I'm there, without getting too low before getting home. OTOH, these were evening trips with light traffic and I picked a route with few stops and just a couple of steep hills (relatively short ones), and used as little as 238 watt-hours per mile (1.7AH/mile @ 140V) rather than the average of around 280 or more watt-hours. (Now I think I know how the EV-1 used just 225 WH/mile or so, according to various reports. It's not that the car was a whole lot better. It's that California is always warm (warm lowers consumption) and has fewer steep hills and generally longer runs between stops.)
   The older NiMH batteries made from "D" cells don't work as well as the newer ones - the voltages drop more when climbing hills or accelerating. Curiously, this doesn't seem to be a reduction in storage capacity, only in current drive, as the voltages soon come right back up fairly even with the new ones after stopping, which means the state of charge is the same. This suggests that the electrolyte gradually escapes. That wouldn't hurt a flooded cell (especially as it can usually be refilled), but the "dry" cell doesn't have much to start with.
   I started thinking about the possibility of refilling the dry cells, by immersing them in water for some period of time. It's an experiment I'd like to try on some older cells, but I would want to run a few load tests before and after, and ideally try varying lengths of time and depths of immersion, so it'd be a fairly involved set of tests in order to come up with general guidelines for 'restoring' NiMH dry cells, if indeed it can be done. And different size cells - and even different brands - would probably have different requirements.

   I'll want more "identical" batteries, eg, 3 x 300AH of NiMH.s, if I get the Chevy Sprint going with a 36 volt motor. I thought about investing another 3000$ in "D" cells to get there. Ugh! What about my own batteries? Surely there must be some way to make use of the new chemistries I've found instead of just paying through the nose for NiMH or lithium!

Turquoise Battery Project: Battery Production Ideas

   I have two fabulous new chemistries. But I don't seem to be able to make practical batteries. I even started thinking about the nickel-zinc oxalic acid cells again (might not care about air exposure?), and about sheet zinc as an electrode. I thought about the way standard dry cells are made, and the fact that they work well and their carbon rod is just what would work with my salt electrolyte posodes. What about simply using the rods from "D" or "F" dry cells and adopting the same construction? Nickel-nickel with salt electrolyte should be doable as a dry cell. (The high voltage of nickel-manganese, fantastic chemistry attainment that it is, would make for too much gassing and pressure - they pretty much would have to be flooded, vented or valve regulated cells.)
   Then I figured that I could use the nickel-brass sheets as the outer layer, with the nickel in them as part of the nickel electrode. I would assume the (18%) zinc would oxidize away leaving the nickel (17%), active wherever it was exposed to the electrolyte, and the copper (65%) as current collector. That would ensure some sort of high-current active nickel electrode, but with quite low amp-hour capacity.

   Then I thought such an electrode might be improved by dissolving away both the zinc and some or even all of the copper, to leave a microscopically porous nickel electrode. (17% nickel and 83% air space for electrolyte penetration.) On April 1st and 2nd I extended this great idea. I have the nickel-brass sheets, and also monel powder (Ni:Cu 66:33%). If some or all of the copper was leached out of both these alloys (and all or most of the zinc) with HCl and H2O2 (or maybe ferric chloride), microscopically porous nickel should remain, with mostly nickel on the surface, exposed to the electrolyte, and any remaining copper inside as a highly conductive backing. That should make a relatively solid electrode with very high current capacity and a lot of microscopically rough, exposed nickel surface and hence excellent amp-hours per amount of nickel. One might be able to apply a flux [water soluble?] and sinter the powder and the sheet together into a porous electrode with a torch, then dissolve out the zinc and [some of?] the copper to also get the microscopic porosity. It sounds like a real winner!

A nickel-brass (AKA "nickel-silver" or "German silver") sheet to turn into a rolled-up electrode,
with an "F" cell carbon rod and its original plastic cap, and the prospective PVC 3/4"
plumbing pipe as an outer casing. (Sigh - the '3/4" end cap' for the bottom was the wrong size.)

   On April 2nd I tried an initial experiment, tossing the piece of nickel-brass into a solution of HCl + H2O2 I had sitting in a jar for printed circuit board etching, for about an hour. I taped over one side so as to etch only the other. Sure enough, the exposed surface looked very different when it came out. It had a 'matte' appearance instead of glossy, and a rough texture with fine ridges instead of smooth under a 40 x magnifiying glass. It looked very promising and I'll have more info next month. (I wish again I had a microscope. Maybe I can get someone to put some samples under the fab UVic electron microscope.) By evening I realized that that solution doubtless dissolves nickel as well as the other metals, so I'll have to look up something else. Edison used sulfuric acid to dissolve copper from nickel, dissolving only a little nickel, so it's known to be doable. I think I'll try ferric chloride. It might dissolve no nickel at all.

   Another interesting find was that someone has discovered that parcel packing "peanuts" can be turned into thin carbon film or carbon nanoparticles (depending on formulation) useful for battery electrodes, by heating in a kiln in an inert atmosphere, and sometimes with the right salts added. (The salts used weren't listed). A friend sent me a link to the article. I may look for ways to try it out. The hard part will be keeping the air out, and I've wanted to be able to do that in the kiln for a number of things.

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

Ceres, Vesta & Vegetation

   I searched youtube for any more information about the Dawn spacecraft mission to Ceres. Specifically I was hoping for information about potential "polycyclic aromatic hydrocarbons" spectral findings or "fluffy" surface textures that might (or might not) support the visual appearance of Ceres as having the same strange vegetation (assuming that's what it is) evident on several airless worlds of the Jupiter and Saturn systems. There was a 50 minute NASA/JPL press briefing about Dawn entering Ceres orbit, but no one asked about those topics. There was much interest in the two "mysterious" bright spots.
   Since the bright spots appear to stick up above the surrounding surface, they're almost surely ice extrusions formed by expanding water, pushing up through holes or cracks and then freezing, following meteor impacts that melt the ice into a pool that quickly forms a crust. I have yet to hear this or any plausible explanation for such features, which are ubiquitous on many worlds, from the space science community. Unless that's what are being referred to as "cryovolcanoes". But they'd be very short lived as active phenomena - probably minutes to a very few hours, depending on the size and energy of the impact.

   In the search for info, I ran across several quite "out there" videos: The bright spots are alien lights. Jpeg images blown up into pixelated squares show an alien city with square buildings. Ceres has air and water. (!)
   One starts to see why it's hardly possible to speak rationally of evidence for life on space forums without having people instantly eject you from their membership without bothering to read what you have to say. And Ceres is, after all, just an "oid" about 2% the size of Earth's moon. (BTW: Evidently the equatorial diameter is substantially greater than the polar diameter: 950Km versus 910Km, slightly different than the figure of 915Km I used last month. Since the gravity is so slight and the rotational period is just 9 hours, this oblateness may perhaps be accounted for by centrifugal force. Hmm... Saturn, which rotates every 10 hours, is similarly oblate.)

   Dawn had already been to the second largest (but much smaller) asteroid, Vesta, before it went to Ceres. What might be so different that Ceres has a similar appearance to Jupiter/Saturn "fluffy" icy moons, while Vesta looks more like a simple rock, like Earth's moon? Apparently, it's just enough of a temperature difference that all the water ice has sublimated off Vesta (max. 253°K, -20°C), while most of it has remained on Ceres (max 235°K, -38°C) so far. Vesta is just a little closer to the sun. But both are much warmer than Ganymede (~140°K) where the apparent vegetation doubtless originated. So far the only real indication I've seen of vegetation is the low albedo and the high contrast with the bright spots. This is perhaps a rather superficial indication, and with such a temperature difference, certainly more evidence is required one way or the other.

   Hmm... I guess there's nothing stopping me from doing my own video to explain what I think is, and might be, going on!... except finding time to do it. I can see quickly wanting to make it into a half-decent production, bringing in images and other findings from other worlds to compare and illustrate the points. It would become yet another project - AWG!

   I was disappointed to hear in the press briefing that Dawn's orbit is only going down to something over 100 Km from the surface. (I don't remember the exact figure.) It should surely be possible to get down to 30 or even 10 Km above the tallest features to sample some of the finer details... like "fluffy" vegetation. Perhaps it's not time for humanity to indisputably see and recognize alien life.

Unsustainable World Population

   It has been said that in primitive 'caveman' times the world population was about 10 million. But this can hardly be true! It was also estimated that before 1550 (before smallpox) there were about 50 million native inhabitants of North America. This figure, for primitive hunter-gatherer societies often (if not usually) at war with each other, is quite at odds with such an extraordinarily low global estimate. If North America contains, say, around 10% of the world's productive land area (besides desert or arctic), we may estimate that the world population for much of our million year history was probably around 500 million primitive people.
   As herding and simple agriculture became common, the food productivity and hence the population per area ratio perhaps quadrupled. This would indicate that the world could, and probably did, support around 2 billion people in more recent millennia. (IIRC, the population before world war one was around 1.8 billion.)
   When agriculture started to become mechanized, more people moved to cities and the land was gradually converted to huge farms - "agribusiness", and an enormous population expansion began. (I remember in elementary school (~1964) my teacher mentioned to the class, and was shocked, that the population had reached 3 billion people!) It is now well over 7 billion, and there are numerous very serious food concerns. Not only can bad crop years result in shortages, and not only is farming presently dependent on fossil fuel*, but with the gradual loss of various trace minerals from farmland soil without replacement it's been said by a USDA study to be unsustainable (which is in fact obvious without a study). The rising rate of arthritis, which evidently stems mainly from boron deficiency, is just one health affect just now being linked to this - to boron poor soil. Besides food, various resources including land are in short supply and the quality of life is deteriorating rapidly. This is exacerbated by the greed and hoarding of a few, and their fears of "uprisings", and war. (There are today over 50 million displaced refugees - more than at any time since world war two.) As stock analyst Greg Mannarino puts it on youtube, "The population is in a bubble, and when that bubble bursts, there'll be suffering on a biblical scale. People won't have the resources to procure the basic necessities they need to sustain their existence." (Not an exact quote.)

   According to various internet sources, some who rule our world would like to see the population reduced to 500 million, and they scheme means to accomplish it without bothering to inform people of their intent, educate them on the need for population reduction, determine public opinion or explore peaceful, agreeable methods to accomplish it. Obviously 500 million is almost absurdly small - a cave days "hunter-gatherer" population and just 7% of the present population. It might put an end to human progress. Certainly much would be run down and abandoned.
   They needn't fret now about the growing population because in the economic collapse looming in front of us, it's going to drop sharply. The most shocking estimates of "9 out of 10 Americans will soon be dead" seem disproportionate, but "1 out of 2" would probably be optimistic and "2 out of 3" to "3 out of 4" more likely. It seems likely to come out around 2 to 2.5 billion people remaining by perhaps mid century - about where it was when the 20th century began. Again while spectacular events and disasters are inevitable, most of the deaths will probably finally result from plagues**. At some point of overcrowding and poverty, these become inevitable.
   Hopefully the survivors, with the global awakening and quickening of consciousness that is and will be taking place, and the spread of knowledge via the internet, will have the wherewithal to start a better civilization - and to voluntarily regulate their numbers so everyone has enough for prosperity.

- - - - -

* To get agriculture off oil, not only is there the CNC gardening/farming machine idea (which would be powered from the electrical grid), but there seem to be a number of battery-electric tractor conversions out there. An electric tractor may be more practical at the present time than an electric car because it's never far from home and from its charging station - or perhaps from a quick battery swap.

** An interesting youtube news show that follows an amazing number of major events and "top importance" topics including food and disease concerns is "The News in Two Minutes" (TheNITM") presented daily Monday to Friday by "FullSpectrumSurvival" channel. It's hard to keep up with the dizzying pace at which the news items (with relevant web pages pictured) are presented! It's quite an antidote for TV "nothing happening here" news where usually one doesn't get the impression that events of real import are happening daily often in rapid succession all around the globe.

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

Construction Manuals and information:

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

Products Catalog:
 - Electric Hubcap 7.2 KW BLDC Pancake Motor Kit
 - Electric Caik 4.8 KW BLDC Pancake Motor Kit
  - NiMH Handy Battery Sticks, 12v battery trays
& Dry Cells (cheapest NiMH prices in Victoria BC)
 - LED Light Fixtures

(Will accept BITCOIN digital currency)

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

Daily Log
(time accounting, mainly for CRA - SR & ED assessment purposes)

1-2nd: Finished February newsletter/report (#85)
3-12: Made insulated outdoor tilapia pond
12-13: Designed motor rotor to convert Electric caik motor from "unipolar" to "switched reluctance" ("SR") type.
14: Made rotor. Made "torque bar" for measuring motor torque.
15: Installed rotor and ran a test. (which showed stator wiring needs modification for SR operation.)
16: Worked on Electric Weel development - finished stator assembly, wired stator coils together.
18: inverted 3 of 6 coils in Caik SR motor. Still no apparent torque.
19: Ran various magnetic experiments with two types of motor coils, the SR rotor, a steel backing bar, and supermagnets and a supermagnet rotor. Electric Weel: started installing magnets on rotor.
20: Ran a couple more SR motor experiment variations (repeated some to verify impressions).
21, 22: Studied up on SR motor control "standard practices" & advancing ideas & projects.
23: Glued remaining magnets to Weel rotor.
24: Searched for pipe suitable for making rings around coils for AFSRM. (no luck.) E-mail conversation with a designer about AFSRM design.
25: Bought 2nd rotor for AFSRM.
26: Epoxied strapping to Weel magnet rotor.
27: Got sheet steel to make coil rings from.
28: Battery Production ideas. (I have two better chemistries... now, how to make batteries that work and are practical?)
31: Studied another theory paper of axial flux motors for EV.s, describing a motor very similar to what I want to build.

Technical Papers Studied:

A Novel Approach to the Design of Axial-Flux Switched-Reluctance Motors
Tim Lambert *, Mohammad Biglarbegian and Shohel Mahmud
School of Engineering, University of Guelph, Guelph, ON N1G 2T6, Canada; 
A Design of Axial-gap Switched Reluctance  Motor for In-Wheel Direct-Drive EV
Tohru Shibamoto, Kenji Nakamura, Hiroki Goto and Osamu Ichinokura  Elec. and Comm. Eng. Dept., Tohoku University

Electric Hubcap Motor Systems - Electric Transport

Electric Caik Switched Reluctance Motor

   The more I look at reluctance motors, the more I like them! The more one looks, the more their unique physical and operational features bring tremendous application advantages to the imagination.

A Revolution for Car Hybridization?

   The SR motor should prove compact and light, and if the rotor is balanced, capable of running safely at very high RPM.s. Where my BLDC magnet motors might be good for 2500-3000 RPM if the magnets are well attached to the rotor (per the latest Electric Caik & Electric Weel construction), the SR motor with a simple solid steel plate rotor might be good for perhaps 8000-10000 RPM or more. If it has enough torque, and if it's geared down enough, it might be mounted on a car wheel simply with a fixed gear (or belt) providing sufficient speed reduction/torque increase to reliably start the car moving rapidly or uphill. The RPM on the highway can be very high without being dangerous or wasting a lot of power. Thus the variable torque converter could be dispensed with! That would revolutionize all my plans and bring the wheel mounted car hybridizing system back into immediate focus.

   Furthermore, 5000-6000 RPM for the Electric Caik size (doubtless quite attainable) is plenty high enough to get a small motorboat up on a plane in outboard motor conversions, driving the main shaft directly and without changing the typical manufacturer's gear reduction down by the propeller, which is so frustrating for conversions using typical higher torque, lower RPM BLDC motors.

   Such potential benefits were completely unforeseen when I started the unipolar motor project, which has turned into the switched reluctance motor project!

   I had been thinking the gearing down could be accomplished by a planetary gear, with the motor directly in front of the wheel. But near the end of the month, I thought of what seems like a very elegant solution with a drive belt, as described in Month in Brief. This simple installation plan brings mounting an external add-on motor to "hybridize" a vehicle back into close-up focus, as an easy DIY project.

Meanwhile Back at the Ranch...

   Of course the layout of the motor, as per the usual BLDC case, is 3-phase and in fact for the Electric Caik is like the image below except for being axial flux instead of radial. The Fleadh Electronics Ltd. web site has this to say [My boldings]:

Switched Reluctance Motor Drives - http://www.fleadh.co.uk/srm.htm

Three-Phase Motor
Offers simplest solution to starting and torque ripple without resorting to high numbers of phases. Hence has been the most popular topology in its 6/4 form. Alternative 3-phase machines with doubled-up pole numbers can offer a better solution for lower speed applications. But again watch-out for torque ripple especially in the voltage control single-pulse operating mode.


   (Perhaps it's noteworthy that this company includes major electrical equipment and motor manufacturers such as Baldor in their client list.) Another document I found said (among other things):

SRM Summary (www.SRMDrives.com)

Due to the absence of rotor windings, SRM is very simple to construct, has a low inertia and allows an extremely high-speed operation. <snip> Designing a motor with high constant power range to base speed (e.g. at least 4:1), is not hard to achieve with SRM, and has a great effect in designing a lower power motor that can produce significant torque.

The SRM has many advantages, mostly resulting from its simple structure: low cost, extremely safe and particularly suitable for hazardous environments. The SRM drive produces zero or small open circuit voltage and short circuit current.

Furthermore most SRM converters are simple because the current is unipolar. The SRM drive is immune from shoot through faults, unlike the inverters of induction and brushless dc motors.


   SRM Drives speaks of high speed motors up to 100,000 RPM, and while that must describe quite a small diameter motor, indeed a solid chunk of steel is highly resistant to centrifugal forces. (But I still remember they guy in the 1970's(?) who was promoting high speed flywheels for energy storage. His experiments and grand ideas were featured in Popular Mechanics and Popular Science magazines month after month (remember when there was no internet?), until they just stopped appearing. It turned out he had been killed, and his entire lab destroyed, by the "explosion" of a huge flywheel at 50,000 RPM. Even the best steel has limits!)

   The Electric Hubcap and Electric Weel would be "Alternative 3-phase machines" with more poles and lobes. (Come to think of it, the Hubcap version with 3 coils per phase instead of 2 or 4 might not work out - instead of NS it would have to be NNS or NSS. This concern was negated by later magnetic configuration ideas, below.)

   In further study of the switched reluctance motor, I was surprised to see the seeming complexity of the control circuitry to drive it as shown on Wikipedia. (In fact, I think the first time I saw the name "reluctance motor" was in the description of an "Asymetric Bridge Converter", and the weird circuit, with six wires to the motor coils instead of three, is what made me reluctant to take any interest in reluctance motors - plus the name just sounded weird. Unfortunately I didn't even look them up.) But on closer inspection, other than needing 6 power wires to the motor, it's the same six transistors and six diodes as in a 'regular' 3-phase bridge driver. But again, as noted in the SRM Drives summary above, the current is unipolar, and "glitches" transiently turning on transistors can't cause a catastrophic short ("shoot through fault") from B+ to B-. In fact, both transistors of a phase are turned on to energize the coil - the fault condition in a regular half bridge.

Power Circuitry (Wikipedia)

The most common approach to the powering of a switched reluctance motor is to use an asymmetric bridge converter.

There are 3 phases in an asymmetric bridge converter corresponding to the phases of the switched reluctance motor. If both of the power switches on either side of the phase are turned on, then that corresponding phase shall be actuated. Once the current has risen above the set value, the switch shall turn off. The energy now stored within the motor winding shall now maintain the current in the same direction until that energy is depleted.

This basic circuitry may be altered so that fewer components are required although the circuit shall perform the same action. This efficient circuit is known as the (n+1) switch and diode configuration.

A capacitor, in either configuration, is used to suppress electrical and acoustic noise by limiting fluctuations in the supply voltage.

   The rising current shutting the coil power off is of course what I call CRM - current ramp modulation. So my controllers, ever since the A3938 version in 2011, already have it. (In the MC33035 based controller so far, it's in an approximate form.) No complex DSP or processor control is required to achieve it.

The yellow lines are current to the active phase coils. The red lines show inductance with rotation, maximum
inductance being attained when the rotor pole lines up with the coil. (The coil should be 'off' by that point of rotation.)

Left: Current ramps up quickly at low speeds and attains destructive values unless limited.
The rate depends on the inductance of the coil, which increases as the rotor pole metal goes by.
When the current reaches a maximum value, the coil is briefly shut off to allow it to drop back down.

Right: Owing to back EMF, current ramps up slowly at higher speeds, and either the current limit
isn't reached, or it's reached more slowly and less often and the CRM is less active or inactive.

If the control current is set lower for less drive to the motor, all the yellow lines are
reduced in amplitude.

    The second "n+1 Switch and Diode" circuit seemed more intriguing at first. Where I put in a coil to dissociate the supply voltage from the actual supply to the coil switches during voltage spikes, this circuit uses another mosfet. This is probably superior - more predictable and controllable.
    But the high-side transistor is turned on whenever any of the others is turned on. It'll get hot 3 times as fast. It would seem then that at least for larger motors such as mine, the asymetric bridge type is used so as to spread the load between three sets of drivers.
   On the other hand, my controller with the in-line coil has only three driver transistor points instead of six, with no troublesome floating high-side gate drives. If the coil can adequately isolate the supply from the battery for energy pulse switching purposes, it may be the optimum solution. It was derived from not knowing what a "standard" solution might be, since I was initially trying to drive a unipolar magnet motor - a new type of motor for which I didn't expect to find an existing type of controller. Occasionally not knowing "standard practices" in advance can be an advantage!

   A problem noted with switched reluctance motors is the generation of acoustic noise owing to flexing of the stator with ever-changing magnetic loading. (No doubt this is to an extent true also of BLDC with magnets on the rotor, and especially it would be of unipolar.) The polypropylene-epoxy composite bodies of my motors should damp the sound and make less noise than 'ringing' metal bodies.

Concept Drawing to define rotor shape

   Another consideration is electromagnetic inductance into the rotor steel. Soft magnetic material (eg, high silicon steel) is used, and some have used laminations or other induced current restricting materials.
   I considered using 1/4" x 2" bar stock and welding two pieces together in an "X" (or something along those lines), drilling out a 1" center hole for the axle. That would give a "disk" with four good lobes, but not quite what I thought would be the best shape. I finally decided that the most practical thing was to take a pre-made 7.8" "disk brake rotor" disk and cut out four shapes to leave the desired four lobes in what I hoped might be a design with minimized torque-ripple. My biggest concern was that this rotor was only 3/16" thick. But at worst, the maximum torque would be lower. It would also be easier to cut than thicker metal. I drew this up on the evening of the 12th, drew it on the rotor with a felt pen on the 13th, and cut it the day after that, using a jigsaw with a metal blade.
   It came out looking somewhat like an "iron cross" medal with the outside edges being the original circular rim segments. The extra metal around the outside edges should give the coil more to attract when it first comes on as the lobe approaches it, while having sufficient clearance between the coil core and the lobe behind, plus it would spread the center of the lobe and reduce the higher torque as the lobe approached the center of the coil. Or so went my reasoning. I thought I cut it pretty well and I didn't try to balance it. First get it to run! Then the details.

   On the 15th I installed this four-lobed rotor into the motor. I seemed very strange having no annoying magnetic forces at work trying to suck the rotor toward the stator. And when it was assembled, there was no cogging - no force except bearing friction. (...which was considerable. Oh well, some grease will ooze out of them as it runs.)

   And I had started to consider that a very useful device would be a torque wrench or other torque measuring device that measured in "inch-ounces" or some such units instead of "tens of foot-pounds", that I could put on the motor shaft. That could give an idea not just of torque, but of torque ripple, by measuring the static torque at various angles of rotation. A Texas Instruments document suggested that torque ripple in SR motors can be minimized by careful mechanical design, which mirrored my own thoughts.
   I wasn't seeing what I needed at local auto and tool stores. Then I thought of making a balanced bar to put on the motor shaft, with a 'pointer' on one end. This would press on a small mechanical weigh scale when the motor was energized, to give a reading in grams or ounces. Knowing the length of the bar, foot-pounds or newton-meters could be derived, to a fine scale. To find the torque at different angles of rotation, the whole motor would be rotated to keep the bar level. Since the first objective would be to measure differences in torque at different angles, only a small force where the motor could be held by hand would be applied. (Maybe even just from the DC lab power supply to one pair of coils at a time, to get exactly the same current each time.) I made this also on the 15th. I found a nice bar on which it was simple to shape a 'pointer' to press on the scale, at exactly 6.0" from the center of the shaft.
   Static torque per amp - including maximum and minimum if there's much ripple - can be derived by knowing the current which gave the figures. (That's another reason for using the power supply, in "constant current" mode.) To measure torque with the motor rotating will still need something like a dynamometer.

   I foresee a problem with determination of back EMF per RPM ("Kv"): With no magnets to induce electricity into the coils, the rotor can spin freely all it wants and essentially no voltage will be produced. Apparently if it's being pulsed by the controller this changes, but it sounds like it'll start to get complex -- couldn't any desired back EMF be generated at any RPM?

   On the 16th I hooked up the lab power supply to one set of coils and turned it on to 10 amps. I turned the rotor to find where the torque would occur and how strong it would be. (Funny I never thought of doing this with any previous motor!) To my surprise there was no apparent force or torque at any point. My best guess was that with a north electromagnet pole at both ends of the engaged rotor lobes, the repulsion of  the two norths canceled the attraction to the metal. I'd have to rewire half the coils - or flip them over - to create north-south magnetic circuits.
   I manged to flip 3 coils over on the 18th. But the slotted optical interrupter drum seemed to rub on the wires, which I had done my best to get out of the way. I might need to start banging on them with a hammer to get them aside! Not my idea of a simple adjustment. For the moment I removed the drum. I again hooked up the power supply to one pair of coils, and again there was no perceptible force. If I put a supermagnet by the coil, I could feel a small amount of repulsion or attraction, but not with the rotor. And not with a fat chunk of steel that would stick fast to the supermagnet. My motors with the supermagnet rotors work great with these coils. Plain steel doesn't seem to work at all. Apparently with the reluctance motor I was moving into unexpectedly unfamiliar territory.
   Of course, usually motor coils have metal behind them, connecting them all together magnetically. The coil cores are usually integral to the die-cut laminates that comprise the stator. Connecting metal didn't seem to be necessary with the supermagnet motors. Going from steel backing disks to plastic composites had improved performance a lot. The magnets, which were connected together by the steel rotor plate, seemed to make for a complete enough magnetic circuit. Here the steel in the rotor just doesn't seem to help. Perhaps I needed to go back to a metal coil backing plate in the stator? Since there are no supermagnets to generate current into it, it shouldn't cause the losses that it did with the supermagnet rotor.

Configuration Experiments

   The next morning (19th) I took two older coils with 63 turns of #14 AWG wire (instead of 21 turns of #11 - normally 3 were wired in parallel instead of 3 in series to get to 63 turns total for Electric Hubcap motors) and set them on a steel bar, about the same distance apart as in the motor, opposite magnetic polarities up. These would of course give 3 times the force with the same current. I put the rotor over them and applied the same 10 amps. This time the lobes pulled fairly strongly down onto the coils. Pulling up on the rotor, it could just lift the heavy coils and bar. With the coils on the bench instead of on the steel bar, the force was much weaker and it would by no means pick up the coils, or either one alone.
   I tried the same thing to repel a single small supermagnet. It was stronger with the steel backing the coils, but less notably, and the forces were stronger overall.
   Next I tried again with a spare Electric Hubcap magnet rotor. It was again somewhat stronger with the plate than without, but even with an inch flux gap, it had good force either way.
   Finally I tried putting both coils the same magnetic polarity and tried the reluctance rotor again. This still worked, and worked much better with the steel bar than without, but the force was somewhat weaker than with opposite polarities.

   I didn't actually measure the forces with a scale. That would have taken much longer to set up. When I speak of "somewhat" stronger or weaker I mean somewhere around double or half. "Much" weaker might be 1/4 to 1/8 as much force. The flux gap with the reluctance rotor was much smaller than for the magnet and magnet rotor, eg 1/8" to 1/4" versus 3/4" to 1".

   Evidently the lack of steel backing for the coils with the supermagnet rotors simply means using a smaller flux gap to get the same force. The  With the lobed steel rotor, it appears to make the difference between "works" and "doesn't work". The small flux gap might preclude having the "wall" making separate stator and rotor compartments.

   As the day wore on I thought of more things to try. The toroidal iron powder coil cores worked fine in the supermagnet motors, but their magnetic permeability is much lower than solid iron. How would iron laminate coil cores work, like the ones I was originally making with nail gun nail strips? Always wanting to keep a few "artifacts" of earlier work (and packrat that I am), I went out to the garage and found two sets of the old original Electric Hubcap coils. I took one in, snipped a wire so I could activate two coils alone, and put those on the steel bar. That had more somewhat more strength than the toroidal core coils - not earth shattering but it was a notable improvement. With the powder cores I could just pick up the coils and bar with attraction to the rotor lobes. With the nail strip cores it was a fairly solid pick-up that took a little shaking to break loose. I repeated the test a few days later with similar results. Then I remembered that the 2" (O.D.) x 1" laminate type actually has more iron. The 2" x 1" toroid cores with the 1.25" hole in the center have only 60% as much. Doubtless that explains the strength difference, rather than any qualitative difference between materials.

   I put the shaft through the rotor so it sat just above the coils on the end of the shaft. With either set, the rotor would turn until two lobes were in line with the coils. That didn't necessarily mean there'd be a lot of torque, but a motor made this way would definitely run.

   But I thought of yet another trick. I had purchased a couple of years ago some ceramic "cup magnets" that I saw on line when ordering supermagnets. Although ceramic magnets aren't very strong, these boasted an impressive pulling force. The magnetic metal "cup" surrounding the magnet concentrates the flux into a narrow gap between the magnet and the rim, and just in front of that gap is its powerful pull. Why couldn't this technique be used with an electromagnet to greatly increase its strength in a narrow flux gap between coil and rotor lobe?
   The cup magnet technique should also make the steel backing redundant, since it completes the magnetic circuit locally at each coil. Plus, since that's the case, it wouldn't matter which polarity the coils of each phase were: the three coils per phase in the "Electric Hubcap" motor size should be fine.
   But testing the "cup electromagnet" idea out was going to take more than a quick setup: parts would have to be made. (and from what?) And cups, if used, would have to fit in the cramped Caik motor stator area.

   I found a pipe (chain link fence type) that seemed to be about the right size for the outer rim of the 'cups'. It barely fit over the coil wires - and might need a few protruding "next layer" turns removed. But it was rather thick walled. Not only might it be hard to fit, it was going to take some cutting to get even-length 1" cup sections. (at least, if I cut them with an angle grinder.) I decided to wait until Monday and see if there was an electrical conduit pipe size that was better. (And if it could be cut - evenly - with a pipe cutter, so much the better!) Those were about the only types of pipe I could think of that just might be suitable. The place I went to then (24th) didn't have any large enough.

   While all the rest was in progress, I had been on the motor controller e-mail list. A couple of people, one in particular, were very interested in the SR motor concept. When I explained I was doing axial flux, he replied on the 24th and suggested having two rotors, one on each side of the coils. This idea grew on me. I had thought this impractical and seemingly unnecessary for magnet motors, but here, struggling to find more flux and magnetic force - torque - it seemed like a good idea. Later I continued reading the axial flux SRM paper and found "everyone" is using two rotors.

   I could still put a pipe/ring around the coil (or bend 1" wide sheet steel into circles), but instead of being like the cup magnet, both ends would have open gaps and would interface with a steel rotor. It would even out axial forces, too, so there'd be no pull towards one end of the motor at any time, quite unlike the continuous pull of magnet rotors toward the stator side. There should be, if perhaps not double, at least substantially more torque than with a single rotor.

   Here I realized I'd hit a point of almost complete redesign of the motors. They'd have to be re-done more or less from scratch. My experience in designing and building molds and motors should shorten the process substantially... but with something new, you never know for sure. Somehow the coils will have to be held in the center, inside metal rings, with very thin flux gaps to adjacent rotors on each side, too small to permit separator walls. God only knows how thick the outer perimeter would have to be to protect against a steel rotor flying apart at very high RPM. On the other hand, the chance of that happening is pretty remote, and the mass of the rotor is relatively small. Perhaps the biggest bonus for SR is that the solid steel rotor is safe up to very high RPM.s.
   The optical position sensor unit could perhaps use the petals/lobes of one of the rotors itself as the optical interrupters. It might have to be installed last, inserted through the outside perimeter wall. In fact, it might best be done as three separate inserted sensors. Or there could be a slotted drum with the electronics mounted in one end "bell".
   How "fat" will the pancake be? (Stator coils 1.0") + (2 rotors 1/4" = .5") + (two flux gaps of 1/16" = .125") + (clearance, rotors to end walls [with bearings] .375" = .75") + (two end walls 3/8" = .75") = 3.125". If the actual build pans out, that's over 3/4" thinner than the single magnet rotor type. (Not counting the protrusion of needle or "trailer wheel" bearing hubs. A single rotor unit would work out to just 2.675" thick!)

   On the 25th I bought the second rotor. These are (I believe) cast metal, with a built-in 1" machined and keyed center hub. The next day I wrote to the chief author of the AFSRM paper hoping for comments on my design. Then I tried to magnetize a cut-off piece of the first rotor by rubbing it with a supermagnet. It wouldn't hold magnetism! The soft magnetic rotor material is the best, and I had it. A rotor that holds magnetism would cause losses. It would be good for functional "gosh it runs" motor tests, but not for a finished motor. I guess you wouldn't want a disk brake that picks up metal particles, which would scratch it and wear it out, and it was made soft magnetic to prevent that. Evidently I lucked out.
   Next I decided that the only way to get an outer ring just the right diameter would be to roll it from flat stock. On the 27th I went to a sheet metal fabrication shop and got a piece of sheet steel, about 8" x 48" x .032" (20 gauge) ... from their scrap bin. That was a little thinner than I thought best (.04-.05?) but beggars can't be choosers. Would two winds of it around the coil work, or would that just diffuse the magnetism? I decided one layer would be best. I wanted soft magnetic material, but thought it might be hard to find, and I forgot to bring a magnet to test with. Anyway, I wanted to get the motor running before worrying too much about efficiency. To my surprise, the sheet wouldn't magnetize. Had I just 'lucked out' again? Or was it made that way (a) because it was cheaper, or (b) deliberately, because you don't want your furnace ducts picking up shards and filings of steel either?
   At this point I tried several bits of steel, including a mild steel bar, and found most of them wouldn't magnetize. A threaded rod and a drill bit did and they would then pick up steel washers, so I wasn't doing something wrong. Apparently "soft magnetic" steel is much more common than I thought, and the annoying magnetization and pickup of filings and bits of steel by drill bits, files and screwdrivers isn't "the norm" with most steel stock types.

   The design of the coils, it seemed, needed only a little change. The biggest constraint was that nothing could stick out either end, even a little bit, since there would be rotors flying by in very close proximity. (.5 to 1mm?) The second one was that each layer of wire would make the gap between the inner toroid and the outer ring wider. Where I had been using 21 turns of fat #11 wire per 12 volts, turn #21 started a third layer since just 10 turns fit on each layer. Then, in order that the inner end of the wire could come out without protruding past an end, the second layer couldn't overlap the first wire. At first I thought that meant eliminating turn #20 as well as #21, leaving just 19 turns - a 10% reduction. Then I realized turn #20 could end just before the protruding inner wire, putting the inner and outer wire right next to each other - a lesser 5% drop. Okay! But if the outsides of the coils were to hang over, past the outer edge of the coils, then the protruding wire would be okay. At first glance, that would seem to be just lost flux. OTOH, it's mainly the leading edge of the coil where the flux is most needed. The loss is in the middle of the outside edge.

   I took a coil, which had the old cotton insulation, and unwrapped two turns to get it down to two layers of wire. (They unwind disturbingly easily considering they're epoxied on.) Then I cut a piece of the sheet metal ~1.05" x 9.5" with tinsnips. I wrapped it around and found it only needed 8" length. The O.D. was just ~2.55", or less than .3" radius of wire and outer ring around the 2.0" core. With modern insulated magnet wire, it would be slightly thinner, with slightly less than 8" of 1" strip sheet metal required.
   If the core to ring gap was just, say, .25", but considering the core itself had .375" thick walls, I estimate the best flux for the rotor occurs within, say, .25" to .35" of the coil. With a gap of (<)1/16" and a rotor 3/16" thick, the entire rotor is just within .25". The actual rotor metal may expand that flux distance, too, which would mean substantially thicker rotors could be accommodated. The outer ring of soft metallic steel would replace the ilmenite paramagnetic coating of my coils for magnet rotors.
   IMHO I think the best form for a coil for axial flux SR would be two parallel "slot" lines in the radial direction, at the front of the coil and the back, rather than the toroid. That would also modify the best shape of the rotor to pretty much straight radial front-back edges. However, the toroid coils are what I have at the moment. At least they enclose the most core per length of copper wire.

   But if each coil is but little changed, the stator as a whole needs to be radically different, first because of the narrow flux gaps, and second because of the double rotor layout. I can't think of any reasonable way to put it together except to epoxy it all into a blob - a very flat, precision width blob. That certainly doesn't help with heat dissipation. Oh well, that's the way the axial flux windplant makers do it.

Switched Reluctance "Active Generators"?

   With a motor controller it's possible to use the SR machine as a generator. As with the bipolar BLDC, it's accomplished by attempting electrically to slow it down - to drive the motor in reverse of the direction it's turning. I am interested in this, as it must be possible to change the generated voltage by varying the control, at any given RPM. Thus the large Electric Weel might be done as a SR machine that could equally be a motor or a generator, and as a generator (its intended use) the output could be programmed or set to a desired voltage, or to the point where it presents the 'maximum power point' ("MPPT") output at whatever RPM it's turning, as opposed to simply putting out a fixed "volts per RPM" that may not be a convenient value.
   As a magnet rotor, magnet-driven generator, the maximum output occurs with alternating north-south magnets, four rotor magnet poles per three phases. To run it as a motor it has to have two magnet poles per three phases. So configured as an optimal 'passive' generator, it won't run as a motor. For the SR version, which must have 'active' control, the motor and generator would be the same machine.
   With magnets, there is usually cogging with the iron in the coil cores. Hence, the best permanent magnet generators where low or zero cogging is required have used two large diameter rotors with a lot of supermagnets (a couple of dozen), and thin air core coils between the rotors. The area needing this most is wind power, where even slight cogging means no rotation of the propeller and hence no power in light breezes. The active reluctance generator solves this problem too as it has no cogging. And with appropriate control (microcontroller?) it should be able to set an output voltage and maintain it at any wind speed, instead of having the voltage vary with wind speed. This should eliminate any need for a separate DC to DC converter since it can regulate its own voltage. So in many installations there may be no more circuitry overall.
   This situation might also apply to the Electric Weel being used for hydro power. I have been somewhat concerned that the cogging of the huge Weel, combined with the strong gearing up of the speed of the water blades, might in the worst case prevent the mechanism from starting itself. (The flux gap can of course be adjusted... still...) The magnets hadn't been installed on the rotor yet. If we switched it to reluctance now, that would solve potential cogging problems, and obtaining a desired output voltage -- all in advance.

   But testing of the motor on the 19th showed there would be a bit more to the motor configuration/construction than seemed to meet the eye. And there was the little tested motor controller with the hot transistors. The time to get a properly working unit might stretch out unduly, while the rest of the floating hydro power unit was ready to go. The developer came over that afternoon and we started installing the magnets onto the rotor per the original plan.

Other Switched Reluctance Motors For EV.s

   Early in the month someone gave a link to a press release for a larger switched reluctance motor, under development. They think SR is "the next generation" of EV motors, and I agree. Why they aren't the present generation is beyond me. Apparently most others have been no faster off the mark than me in this regard. Of course, in general all the advantages and features Ricardo is touting for their SR motor apply equally well to any I might make. And I think the axial flux will be better -- certainly it'll make motors ideally shaped for wheel mounting. An interesting feature of their motor is the "distributed stator winding". That can't be done with individual coils, and also would likely require more elaborate control electronics than I presently intend to make.

   Later in the month I found more technical papers specifically on axial flux switched reluctance motors (AFSRM.s) that have been helpful in modifying if not redefining what I want to build.


The new electric vehicle (EV) motor has been designed and built in prototype form by Ricardo as part of a collaborative research and development project, RapidSR (Rapid Design and Development of a Switched Reluctance Traction Motor). Using a conventional distributed stator winding, the Ricardo synchronous reluctance electric machine is a highly innovative design that makes use of low-cost materials, simple manufacturing processes and uncomplicated construction. It has a rotor made from cut steel laminations, which are used to direct and focus the flux across the air gap. By maximising this flux linkage between the stator and rotor, performance can be optimized within a tightly packaged, low weight and rare earth element free design.

“As the market for electric vehicles grows globally, there is an imperative to explore alternatives to permanent magnet traction motors which require the use of expensive and increasingly difficult to source rare earth elements,” commented Paul Rivera, MD of the Ricardo hybrid and electric vehicle systems business. “The Ricardo prototype that we have announced today demonstrates what can be achieved by using the latest electric machine design processes in the creation of a high performing, compact, lightweight, and rare earth element free concept.”

Since its launch in 2012, the RapidSR project has been researching the design of next-generation economic electric motors that avoid expensive and potentially difficult to source rare earth elements typically used in permanent magnets. By developing effective CAE led design processes as well as prototype designs, the team has created a framework for the future design and manufacture of electric vehicle motors that offer the performance, compact packaging and light weight required for EV applications, but at a significantly reduced cost compared to permanent magnet machines. Ricardo’s partners in this research include project leader Cobham Technical Services – which is developing its multi-physics CAE design software, Opera, as a part of the project – and Jaguar Land Rover. The research is being co-funded by the UK’s innovation agency, Innovate UK.

“By bringing together state-of-the-art simulation technology with advanced electric machine design we have created a highly credible next generation EV motor concept that shows considerable promise,” added Dr Will Drury, Ricardo team leader for electric machines and power electronics. “The Ricardo prototype is now built and will be rigorously tested over the coming weeks in order to validate the extremely positive results that it has shown in simulation, as a concept that provides an exceptional balance of performance, compact package, light weight and low cost.”

   Mid-month I found a paper discussing designing of a two-rotor axial flux reluctance motor for car wheels:
AF-SRMs-03-00027.pdf from a team at the University of Guelph in Ontario.

A Novel Approach to the Design of Axial-Flux Switched-Reluctance Motors
Tim Lambert *, Mohammad Biglarbegian and Shohel Mahmud
School of Engineering, University of Guelph, Guelph, ON N1G 2T6, Canada; 
Published: 3 March 2015

   It was 28 pages and it took a while to read it, even without trying to 'get into' most of the electromagnetic equations. I think it was just theoretical and nothing had actually been built - there were no photos. Their ideas seemed to parallel mine except they were trying to make a larger motor that would turn the car wheel directly without gearing down. It seemed rather heavy at 41 pounds. Addresses were given and I e-mailed the chief author for possible comment about my own design.
   One important feature I noted was that instead of being typical 3-phase it had 8 coils and 6 lobes on the rotor. The 8 coils were to be driven by 8 half bridges. Since 8-6 is symetrical at 180°, four half bridges could each drive 2 coils, making it 4-phase. I assume this arrangement must give more torque. In the 3-phase version, generally only one phase is powered at a time. For the 4-phase to have more torque, two phases out of four will have to be energized at least part of the time, or the current in the active coil will have to be higher per copper area, in order to achieve that. That implies more complex control than my simple plans. In fact, it will probably mandate using a microcontroller in place of the simple motor controller chip. But maybe that's not a bad thing since it can accomplish other control functions like regenerative braking without adding more control components later.

Energy Conversion Efficiency (ECE) of Axial Flux SRM Compared to Other SRM.s,
showing the desirability of using axial flux for SR motors.
(Note that ECE doesn't mean overall efficiency, which will be quite high. Most
of the energy in a coil is returned to the supply when the coil is switched off.)

   This paper had a reference to another axial flux SR motor design paper, which I looked up on the 31st:  yazaki_5_1_ronbun.pdf . This paper is only 6 pages, but it was packed with good info, and it wasn't just theory. They actually built and tested a test model.

A Design of Axial-gap Switched Reluctance
Motor for In-Wheel Direct-Drive EV

 Tohru Shibamoto, Kenji Nakamura, Hiroki Goto and Osamu Ichinokura
Elec. and Comm. Eng. Dept., Tohoku University

   On looking over some of this paper such as the graphs shown, it becomes evident that the two rotor axial flux design is surely the "way to go" for a switched reluctance over-a-horsepower power motor.

   The Japanese paper described two motors. The second one is the 3-phase AFSR motor of 12" diameter of which a model was actually built and tested. It was the most interesting to me because it's near my size and parameters. It has twice as many coils as I usually use, 18, with the sort of "co-linear" shape of leading and trailing edges that I had identified as probably having the most flux and hence the most torque per coil and per amp for this type of motor. The 12 narrow rotor poles are substantially different from my "lobes", partly in keeping with the narrow coil cores.
   An interesting feature is the "Stator support link", which apparently splits the coils in half, or at least the copper windings (it's not really clear), but which must provide a means of fastening the stator in the center of the motor, between the two rotors. The text mentions it being related to the two rotors shrinking the tiny gap owing to magnetic attractive force to the stator. (Once again this model had no external shell - the rotors were the outside and the stator connected to the stationary axle.)

   Theoretically it would deliver up to 302N-m of torque, which is just over 200 foot-pounds and should be adequate to run the Chevy Sprint with no gear or belt reduction!
   The motor said the motor had a speed of "up to 330RPM", allowing a vehicle to achieve a speed of 30 Km/hr. This seems like an absurdly low top speed for a motor with solid steel rotors. A test graph showed up to 800 RPM, and with loss of torque with speed at very low speeds. Of course, this depends also on the applied voltage and the back EMF of the design, so it may be that the applied voltage could easily be higher (or the number of turns in the coils reduced to lower the back EMF), resulting in higher available speeds.

   Since I already have the Chevy Sprint configured with a transmission under the hood to which the motor could be attached, a 2 or 3 to 1 speed reduction could be obtained with a chain or belt drive (to the original differential), giving more torque, with a higher speed - up to about 2300 or 3500 RPM at 100Km/hr on the highway. Even if the torque of the motor and controller I build don't attain those in the paper, it would still have plenty of force.

   I can see how all the little short segment, intense flux transitions could deliver much more torque than the wide lobes of my present design - over very narrow areas with rapid coil switchings per RPM.
   Maybe I'll concentrate on getting the first motor done and working, and the controller working, and all installed on the outboard using the present design, while I consider all these probably desirable designs for larger car motors. I do have one or two reservations. Foremost, their design seems to have a lot of torque ripple. Second, the low RPM.s indicated, including rapid torque reduction with RPM, must be the result of the rapid flux transitions from the narrow coils and rotor poles. The high torque may come with the high torque ripple 'built in', and it may limit the available RPM, with high back EMF at low speeds. I may still find the 9 coil, 6 rotor poles version has more desirable characteristics. ...or maybe 12 coils and 8 rotor poles? ...or?... And maybe some rotor pole edge shaping to reduce torque ripple, even at the expense of maximum torque. Doesn't the motor have to start the car rolling even from its minimum torque positions? Better to raise those. Or, I might change the pie slice coil shapes to something more like diamonds or rectangles, which would be the same as edge shaping and also would allow some air flow gaps across the coil windings for cooling. (And rectangular pieces of steel would be easiest to cut - assuming I can use solid soft magnetic steel instead of laminates.)

Electric Weel Motor (Generator)

   Rick Linden of Pacific Coastal Geoscience, the designer and builder of a floating river hydro power generator, came over on several occasions and we worked on finishing the Weel generator for the unit.

   First we finished wiring up the stator coils. We wired it as two sets of 4 coils in series per phase, which left 12 wires to hook up, and it can be configured as 8 coils in series (higher voltage) or two sets of 4 in series (higher current). The actual voltages and currents will depend on RPM, and on the adjustment of the flux gap.

   We epoxied the 32 magnets to the rotor's outer steel ring, 16 at a time in two separate sessions. Since the rotor unit was too big to put in the oven, a day or more had to be allowed after each operation for the epoxy to set. Before the second session I made a jig to allow safely inserting the second set magnets (north up) between the magnets of the first set (south up). Once in place, each magnet was clamped down to ensure it wouldn't snap over against one of its previously installed neighbors.

   In a third session, we put the main lexan ring in and wrapped the polypropylene strapping around the magnets and through the slots in the rotor. This should ensure they're well attached and able to take considerable centrifugal force. A 26" diameter magnet rotor is an impressive looking rotor!

Epoxying on the PP strapping to secure the magnets and the lexan rotor.

Aligning the lexan center height with the steel ring using spacers, and
weighing down the strapping to keep it in place while the epoxy sets.
The magnet side is down to keep the straps even against the plywood with no protrusions.

The 'magnetized' rotor.
Lexan rotor reinforcing pieces are to be installed after the shaft is keyed to ensure alignment.

   The pull from the magnets to the stator will of course be immense, probably getting up into the "hazardous" range for fingers even with the safety wall between the rotor and stator ensuring a considerable flux gap. (I look forward to getting all the details figured out for producing switched reluctance machines, which won't have any permanent magnets to mount or to fight with.)

   At month's end the main things left were to machine the dual key slot into the main shaft, and then to place the lexan reinforcement pieces on the rotor and glue them with methylene chloride. The shaft was to have been keyed elsewhere, but the machinist was too busy so it didn't get done. Neither did I get to it. Thus it sits, nearly finished. (We got it done on April 3rd.)

"Green" Electric Equipment Projects

Aquaponics & LED Grow Lighting Project

LED Grow Lights

   Early in the month one emitter in the white light made with the LED emitters from dx.com started flickering. A couple of days later it went out, and another in the same row started to flicker. The next day it didn't work at all. This is a theoretical weak point of my lights, that if one emitter of four or five in parallel quits, the same current is pushed through the remaining emitters and so they work harder. Still, the currents were under 1/2 the specified maximum, so the four remaining emitters should have had no problem. I already noted that the dx.com emitters had higher voltage drop and also weren't as bright as other seemingly identical emitters I ordered from a Chinese store on aliexpress.com. And the 10W dx.com LED "light bulbs" probably made with the same emitters keep burning out and needing an emitter replaced. While they run warmer than ideally in both lights, I finally conclude they are just crappy emitters, to be avoided.

   On the 17th I tried swapping the two LED grow lights, with different wavelength blue (450nm) and violet (420nm) emitters. I keep saying how my linear current current regulators are really efficient at 12 volts. But the violet one was now getting 13.65 volts from the solar system instead of 12.0 volts from a power adapter, and soon the transistor heatsink bar was quite hot. I measured the voltage across the lights and it was just 9.75 volts - apparently the violet emitters had somewhat less forward voltage drop (3.02v) than the blue ones (3.23v). At 12 volts the supply efficiency would thus be just 81%. At almost 14 volts the power transistor had about 4 volts across it and was thus dissipating 6 watts extra in a 15 watt light, total 21 watts. That's only 70% efficiency. That's probably tolerable since 14 volts means the solar panels are working... but it runs too hot! Evidently I could either have put an extra row of red LED.s in it for 11.65 volts, or used 3 rows of violet emitters and just one of reds (10.85v).

   But I'm starting to think that the real solution, for best power supply and LED emitter voltages flexibility, is, after all, to use a switching regulator. It might probably be only about 85% efficient where I - ideally - get up to 90% with the linear, but the "ideally" is hard to get. The switching regulator would be good at all supply voltages and with any combination of emitters I might put in a light. And it would make 1/2 the heat of 70% efficient.

Tilapia Pool

   On the 3rd I finally decided that the tilapia pond would be a trench in the garden, made with the rubber(?) pond liner. Weeds from the garden could be pulled and thrown in, where the tilapia would hopefully eat them, reducing the fish food bill. And it was close enough to the greenhouse to pump water to and from to flow through greenhouse plant beds. And more secure - easiest to fence off from raccoons. I couldn't figure out where to put without blocking a path until I thought of simply putting a bridge across the trench. Tilapia like cover anyway. I dug out the outline of the trench, not to full depth, and looked again the next morning to see if I actually wanted it there. Wherever it was in the garden, it'd be somewhat in the way.
The trench ~18x22x76".
People were wondering who I was going to bury.
I don't think I've ever dug the garden down to the subsoil before!

   With it being outside, I decided that the bottom and sides should be insulated so it could be heated to extend the season, or perhaps, with an insulated top cover, even be used all year. Fiberglass was out, and the cost for styrene foam would add up. But not as much as electricity cost to heat an uninsulated pond!

   It took about 3 days to dig out the pond. Then I filled it from drums of rain water, to about 95 imperial gallons, 115 US gallons or 430 liters. This left a lip of about 6" to keep fish from jumping out. Then somehow it took another few days to put a concrete surround around it. The store ran out of concrete paving blocks, and I ended up making some as well as buying some.

   I saw small insects flying near the water and alighting on it. I thought about breeding mosquitos. The water was too cold for tilapia, barely 10 degrees C. I could either get a heater, allow mosquitos to breed... or get some goldfish. I got 11 small ones, "feeders" according to the pet shop. Then I got some more duckweed from my brother's ponds. The goldfish won't eat it. It grows rapidly and I can feed the surplus to the tilapia without having them make it extinct. By the end of the month, the water was up to about 14°. Theoretically tilpia wouldn't die, but it surely wouldn't be good for them. Trout come to mind... So does hooking up the whole aquaponics system in the greenhouse, including a small solar water heater Jim Harrington gave me to help warm the water in the pond and system.
   I'm not really clear just where I'm going here. Having dug the pond, I worry about raccoons catching the fish if they're in it. But the largest tilapia is almost 12" long, in a 12x12x12" aquarium space, and will have to be moved somewhere soon. (There's always a 200 liter plastic drum with the top cut off, I suppose.) The one little tilapia is now about 2-1/2".

   I decided to use the big swimming pool as a reservoir for watering the garden in the summer, city water having become rather costly. Someone said he had one and a seam split and all the water was lost. So I don't want to invest too much in it that would be lost if that happened, ie, valuable fish. I channeled water from my house roof downspout and from my neighbor's garage roof. (It spills out of his gutters and turns that end of my yard into a swamp. I put an open top "barrel" with a hose tap at the bottom just across the fence to catch it.) Just after it was set up to fill there was a great 36 hour deluge, and it was 11" deep. With further March rains it got to 20" - probably 1500 imperial gallons. I suppose when I see inevitable mosquito wrigglers I'll throw in some goldfish. Come to think of it, I could surely put in some duckweed! (I did.)


   I was about ready to rip out the bean plants from the aquaponics grow bed. They had grown as well as any pole beans in the garden - better - but with no flowers it all seemed pretty useless. I had added a bit of potassium chloride for potassium and trisodium phosphate for phosphorus, but nothing was changing. On the morning of the 16th there was one orange flower high up. This would have been most welcome 2 months ago! Now it merely dampened my resolve to be done with them and try other plants, that I had finally arrived at. The flower shriveled in a couple of days, and no new ones appeared.
   One thing the beans did appear to have done quite well was their job of removing fish waste - ammonia that gets converted to nitrates by bacterial action - and converting it to plant growth. The water in the system stayed quite clear with little algae growth. (On the other hand, the tilapia seem to eat at least some types of algae.)
   Other than that, they're a bit of an odd ornamental house plant!

Cordless Lawnmower with NiMH "D" cell batteries

   Quite a while back Jim Lawrence, the friend who helped me with my web site, had bought a cordless lawnmower. These silly things use lead-acid batteries, which of course don't last long. After he had bought replacement batteries that only seemed to last a year for not much less money than the original price of the mower, I volunteered to replace them with 40 NiMH "D" cells. (24V, 20AH.) The only part a bit tricky was the charging. On the charger cord I put in a diode and a resistor to drop the voltage a little from what the PbPb types took for charging.
   Later as I learned how the fat copper straps I was using would work loose with vibration, and how such 'battery packs' could be dangerous with only the low-temperature plastic insulating the cells from each other, I "recalled" the mower and re-soldered it all with #16 stranded wire and tarpaper sheaths around the cells. Since then, the only problem has been that a wire in the charger assembly came loose once.
   Now Jim has sent me a review on the mower with the NiMh battery. (Yes, we were already mowing lawns here in Victoria BC in February. We had no snow this winter - only a few frosts.)

Hi Craig:

Find attached a couple of photos of my electric lawnmower powered with your metal-hydride batteries.

The pros are as follows:
1. Mower is super light in weight and easy to push around.
2. Batteries have been running around five years and seem as good as new.
3. Needs no oil or gas.
4. No fumes or bad smells.
5. Quiet as a vacuum cleaner.
6. When turned over for cleaning, no fuels or oils drains out.
7. Always starts instantly.
8. Easily does my lawn and even the next door neighbours lawn on a single charge.
9. Requires virtually no maintenance.
10. Unlike regular batteries, they are not toxic.
11. Inexpensive to power.

The cons are as follows:
1. The cutting blade still needs regular sharpening.
2. Needs a steady trickle charge to keep the batteries full.

Pro or Con?
1. Initial costs of batteries are high but after a few years costs average out much better than any alternatives.

Hope this helps for your next review. Maybe you should start selling an upgrade kit...$500 for potentially 10 to 15 years of use...good investment.


Electricity Storage

Turquoise Battery Making Project
Cylindrical Batteries?

   Seeing I couldn't seem to stop the flat plate cell leaks, apparently through the graphite foil, and that this was preventing me from making batteries, I thought again of the carbon rods used in standard dry cells. I started to wonder how it would work out if I made batteries with "F" or "D" cell carbon rods, with similar construction to simple standard dry cells?
   Ideas started forming. I didn't see how I could reliably jam things together under pressure without breaking the carbon rod and ripping the separator sheets before it was together.
   But there's the fact that the nickel oxides in the positive side always expand. Perhaps if I made use of that I could make things so they'd start with just enough play to set the layers in place to put together the cell, and when water was added, it would swell and everything would be a tight fit. This seems to be worth a try, given that nothing else seems to have worked, and since it seems to work somehow for billions of dry cells. Whether it would be a "dry cell" (AKA starved or limited electrolyte cell) or wet/flooded seems to be of little concern at the moment.

   So the "layers" of the cell would start with the carbon rod in the middle.
   Surrounding that would be the fat posode cylinder of nickel manganate with carbon black. I'm not sure how the graphite felt could be worked in since it would be compacted from one end, but perhaps I could cut some strips or bits of it to throw in. This electrode would be compacted inside a pipe, closed at the bottom and with a metal rod just a trace fatter than the carbon rod. A die would push into the pipe, filling the space between the center rod and the pipe. All these would be much longer than the intended electrode in order that the powder could be poured in and then compacted in one press.
   Around that would be the separator, probably paper plus a couple of layers of PP non-woven fabric, then more paper. But I'd try different things to see how many layers were really necessary.
   Surrounding that would be the negode of nickel and monel, and perhaps more carbon black. This would be done the same way as the posode but of course with larger diameters so it would fit around the inner parts, just loosely enough to slip over the paper without trouble.
   Finally would be a metal pipe, tube or can to hold it all and form the negative terminal. (This got changed to a PVC pipe with end cap, with a sealed bolt for a terminal.)
   A plastic disk piece would fit the top and if necessary the bottom. Perhaps the tube could be crimped on the ends over this/these pieces. (Perhaps I could save the crimped metal terminal on the carbon rod from the original cell.)

   Before crimping the top on (or before screwing in the bolt?), electrolyte would be added. This should cause the posode to swell and take up the slack spaces, tightening around the carbon rod and pressing against the separator layers.

Some thoughts... A Nickel-Zinc or Nickel-Nickel Acid Battery?

   As I thought of configurations with a carbon rod electrode in the center, I kept thinking of the rolled-up dry cell. Nickel-nickel would be the obvious choice. The high voltage of nickel-manganese with its bubbling during charging would probably preclude such a tightly wrapped up configuration. What would the current conductor layers be? The graphite foil would crack if rolled, so that leaves the graphite foam and flexible sheet graphite gasket material.

   For the negative, the zinc sheet came to mind. But the reaction voltage of zinc would have it oxidize before the nickel... it would then be a nickel-zinc battery. Iron might be good; maybe stainless steel mesh. Or perhaps the copper mesh. There'd be nothing to stop the negative terminal from being metal - only the plus side has to be carbon. Pieces for the negative could be tack welded together and to the mesh. The carbon rod and a sealed metal rod for terminals would allow for very minimal leakage of electrolyte. It could work.

   But what about that zinc? A sheet of zinc works quite well as an electrode. The trouble with zinc is that intermediate-state soluble zincate ion that causes it to form dendrites, branching "roots" of zinc during charge and discharge. This both corrodes away electrode substance and tends to short out the cell.

   This brought me back to a theoretical idea I had quite a while back, and I made a demo cell just to show it could be done. The reason lead is used for lead acid batteries, despite its high atomic weight and consequent low energy per kilogram, is that lighter metals would dissolve in the acid. I looked for an acid that wouldn't dissolve zinc and nickel compounds. And I found one: oxalic acid. Neither nickel, nickel oxalate, zinc nor zinc oxalate was soluble according to a table. (The state of charged nickel, oxide, and its potential solubility might be a bit of a wild card, as well as the exact reaction voltages - AFAIK this is pretty uncharted territory.)
   Back then, I made a little test cell with a piece of nickel hydroxide electrode from a Ni-Cd dry cell and a zinc rod to illustrate the point. It worked but had little current and seemed to deteriorate. The voltage dropped and charging didn't entirely restore it. But it was open to the air in a beaker and the posode current collector was made for pH 14, not pH 1, and would obviously deteriorate. It was hardly a definitive test except for showing an initial voltage of [IIRC] about 1.7 volts.

   The likeliest nickel reaction (I think) would be something like: Ni(COO)2 + 2e- <=> Ni2(COO)2 + (COO)22− .

   Now the likely zinc reaction is: Zn + (COO)22− <=> Zn(COO)2 + 2e- . The environment is acidic and the compound isn't ZnO or Zn(OH)2 . Will the zinc still form the intermediate soluble zincate ion, or will the reaction go direct as indicated above? The zincate ion is more usually associated with an alkaline environment. There have been a lot of people who've tried a lot of different things to prevent zinc from growing dendrites in battery reactions, but I don't think anyone has thought of this before.
   To be honest, the thing I like about the zinc is simply the ability to roll up a solid sheet of it in with some paper and a nickel posode side, for a very simple and pretty high-current negode. I'd just as happily roll up a thin sheet of nickel... if such a thing could even be found for sale, much less afforded! Maybe I could use nickel-brass, and the zinc would oxidize away. I actually have some of that! (nickel-brass, AKA "nickel-silver", Cu:Zn:Ni , ~65:18:17%)

   If the zinc doesn't work out and yet the nickel side does, one could do a nickel-acid battery with with oxalic acid and nickel compounds in both electrodes. In fact, that would be pretty much identical to the nickel-nickel salt cell except for the electrolyte. That brings me right back around to that cell. Except for the new idea of using the nickel-brass sheet.

Nickel-nickel (Salt) Cell with Nickel-Brass Sheet Negode

   So a plan for a cell (wet or dry) is (a) the carbon rod from a standard "F" (lantern battery) cell, (b) my "usual" nickel manganate electrode as a cylinder packed around the carbon rod in a manner similar to standard dry cells, (c) separator sheet(s) ringing this posode, (d) the thin nickel-brass sheet as a ring around the outside of the paper, (e) and finally a common 3/4" (or 1") x 3.5" long PVC water pipe around the whole thing, with a PVC end cap glued to the bottom. (or a flat insert that doesn't stick out and increase the diameter.)

   For the top, I may be able to use the original top from the "F" cell if it glues okay to the PVC. It'll need an extra hole near one side for the negative terminal unless that terminal can be an exposed piece of the outer ring, which could be soldered to directly. That would have to have some glue all around it between it and the outer pipe so nothing could leak. I have some misgivings about that, but I don't like trying to tack weld on a terminal to the sheet, either.
   The inner electrode would be as discussed at the top of the 'storage' section. I'll elaborate on that when I go to make it.

   To use a sheet containing nickel for an electrode as well as the current collector would seem like folly except that I know it works with zinc sheets. Standard dry cells are made that way. OTOH, the amp-hours capacity from a simple sheet of zinc in the dry cell come from the fact that it's gradually eaten away, and it's not just the initial surface that reacts. So I may make a cell or two this way, but some powder material - monel - will have to be added to the sheet to give real capacity. The layout might be a good one, tho. I wonder if there's some way to "glue" the powder to the sheet metal so it holds together during assembly?

To make porous nickel sheets and porous nickel powder!

  Another thought was that hopefully the zinc would migrate and make some void spaces exposed to the electrolyte in the remaining cupro-nickel. I wonder if one could leach out all the zinc to have a porous electrode sheet somehow? Wikipedia verifies that HCl will leach out zinc, forming ZnCl2, which is soluble. It might be preferable to do this in advance of using the sheet. (Hmm, it looks like HCl might well dissolve the nickel, too! Something else?)

   That led to another idea, expanding on that... Edison's technique for making thin nickel sheets was by electroplating nickel to copper and then dissolving the copper with sulfuric acid... circuit board etching solutions also dissolve copper... and the way people separate gold plating from copper is by dissolving the copper with H2O2 and HCl solution. (That would leach out copper, but again would probably also dissolve the nickel. Ferric chloride might take only the copper, or the copper and zinc... Must read up on this!)
   Then, what about leaching out some of the copper to make an atomically "rough" surface of nickel particles? That would increase the exposed nickel surface and hence the amp-hours of the sheet. This could be the subject of some experimentation. Ultimately one could perhaps dissolve all the zinc and copper and be left with relatively pure "porous nickel", which might make an excellent electrode - or might disintegrate, or disintegrate as it oxidizes on discharge. If pure porous nickel isn't good, there's probably some optimum level of dissolving the copper that would make a conductive substrate with a maximum of nickel exposed to the electrolyte.
   Then, even a thin layer of monel powder somehow bonded to that would make still more nickel surface area. Perhaps the monel powder too should have some of its copper leached out to create more nickel on the surface? Or even all of it to make porous nickel powder?
   I wonder if there's some chemical means for bonding or alloying the monel powder to the nickel-brass sheet? Or maybe just torching it [with a flux?] to "sinter" them together? (With both(?) pre-treated to reduce copper and maximize nickel at the surface. That should make an amazingly conductive electrode, with high exposure of the microscopically porous nickel to the electrolyte yielding excellent amp-hours per quantity of nickel. And it surely won't fall apart during or after insertion into the cell. That would be a real winner!

   All these things should be experimented with! I'll definitely make time for battery development work again.

   I've never wanted to roll up sheet metal before. Now I have two jobs wanting smoothly rolled sheet in a month. I suppose even if I had a commercial sheet metal roller, it wouldn't roll into a 3/4" diameter, and if it did, I wouldn't be able to get the rolled 'tube' off the roller. I think I'll have to make something to do a smooth rolling job.
   Ideally I wish I could make nickel-brass pieces into one-piece cans, having sealed sides and bottom, with the top rim to be crimped around a plastic lid (with the carbon rod in the center) to seal it all. That's the real 'production' method. It ought to make everything a lot faster. But stuffing a rolled up piece into a PVC pipe "can" might be easier for home production.

Carbon/Graphite from Packaging 'Popcorn'/'Peanuts'

   A friend sent me a link to an article about turning packaging 'peanuts' or 'popcorn' into nano graphite forms with heat in an oxygen-free atmosphere, 500-900°C. I could probably do it in the mini kiln if I can keep the air out, ie, make a sealed container of some sort that can take the heat. Maybe it could replace carbon black, or graphite foam, depending how it comes out.

Packing peanuts could be reused in better batteries


When a new lab was recently being set up at Purdue University in Indiana, a lot of the equipment arrived in boxes full of protective packing "peanuts." Unfortunately, few facilities exist for recycling the little pieces of foam, so they typically end up sitting in (or getting blown around) landfills for several decades. A team of Purdue researchers, however, discovered that they could find use in better-performing lithium-ion batteries.

When a lithium-ion battery is charging, the lithium ions are stored in one of its two electrodes, known as the anode. Ordinarily, anodes are made out of graphite. Led by Prof. Vilas Pol, the Purdue scientists instead set out to create a new type of anode made from carbon.

They started by heating packing peanuts (which were made from polystyrene or were starch-based) to a temperature of between 500 and 900 ºC (932 to 1,652 ºF). They did so in an inert atmosphere, in either the presence or absence of a transition metal salt catalyst. Depending on the peanut material and the approach taken, the result was either carbon nanoparticles or carbon microsheets. In both cases, they made excellent anodes.

This was partially because they were about one tenth the thickness of graphite anodes, allowing for quicker charging times. Additionally, they exhibited less electrical resistance than graphite. In more precise terms, they demonstrated a maximum specific capacity of 420 mAh/g (milliamp hours per gram), as opposed to the theoretical 372 mAh/g maximum for graphite anodes.

What's more, the carbon anodes stood up to 300 charging cycles without a significant loss in that capacity.

The microsheets were particularly effective, as their porous structure allowed for more contact area between the anode and the battery's ion-carrying liquid electrolyte. That said, the researchers are now working on making them even more porous, to further enhance their electrochemical performance.

According to Purdue, the packing peanut conversion process is inexpensive, environmentally-friendly, and should be practical for large-scale battery production.

Source: <http://www.purdue.edu/newsroom/releases/2015/Q1/new-processing-technology-converts-packing-peanuts-to-battery-components.html>Purdue University

Victoria BC Canada