Turquoise Energy Ltd. News #91
  covering August 2015 (posted  September 2nd)
Victoria BC
by Craig Carmichael

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

Highlights: TWO Improved Axial flux Switched Reluctance Motors: "AFSRM" or "ARM" and "Transverse Flux" Motor (Building ARM is in progress - see Month in Brief, Electric Transport topic)

Month In Brief (Project Summaries)
- Plan that got changed - ARM Motor - An even better(?) SR motor layout! - Mounting a Motor for a Car Wheel Drive - Solar Water Heater - Peltier Cooler Voltage Tests - Battery DES Chemicals arrive

In Passing (Miscellaneous topics, editorial comments & opinionated rants)
* Solar PV Power coming of age - Comet 67"P" for "Perspective" - Putin/leadership/spread of power - More Collapse Warnings and Signs Add to the Chorus - chem trails seriously interfere with solar collection - Fuchsia Berry Pie - Funnies(?)

- In Depth Project Reports -

Electric Transport - Electric Hubcap Motor Systems
* ARM motor project
* "Transverse flux" SR motor idea & project
* Making an offset motor with belt drive "Electric Hubcap" Plug-in Hybrid EV Installation: wheel end parts.
* Sprint: work stopped again owing to competing attractions. (ie, by very promising looking new motors, again!)

Other "Green" Electric Equipment Projects
* Solar Hot Water Heater
* Peltier Module/Thermoelectric Cooler Experiment: supply voltage versus attained cooling.

Electricity Generation (no reports [but see Solar Water Heater in Other Green Projects, and Solar PV coming of Age in In Passing])

Electricity Storage - Turquoise Battery Project (NiMn, NiNi), etc. (No reports)

No Project Reports on: Magnet motor, Lambda ray collector, evacuated tube heat radiators, CNC gardening/farming machine, Electric Weel, battery making, aquaponics.


August in Brief

The First Plan...

On the 2nd I prioritized what I thought were my main project objectives:

1. Make a 12 magnet placement jig for Electric Hubcap rotors, similar to the one I did for the 8 magnet Electric Caik rotors.
2. Mill the rotor (same one) slots for the new style magnet strapping.
3. Attach the magnets. While the epoxy is wet, patch up the gouged rotor wall in the motor.
4. Reassemble the motor and reinstall it in the Sprint.
5. Try it out with the Kelly controller to see if it works. (I'm really still not quite sure whether it will or whether all the propulsive forces cancel out.)

PATH A: Kelly controller works

6. Connect the torque converter control pulley linkage to the shift stick.
7. Try to get Sprint moving.

PATH B: It doesn't work, or after path A is done

6.(8.) Edit the unipolar controller circuit board and finish it up.
7.(9.) Print the artwork and Etch the board.
8.(10.) solder on the components.
9.(11.) put together the controller.
10.(13.) Test the controller on the Caik and Hubcap 4:3 motors.
11. Go to path A if it wasn't done before.

There was more of course, but I figured that would be at least a full plate for August. I did have a little holiday coming up!

   But the seemingly straightforward plan immediately got off track. I started (ignoring the task numbers) by trying to get the LED headlights working in the Sprint, since I already had half the dash apart. Something weird was happening and I soon wished I had left the headlights alone. Then I pulled the shifter cable out to see how I might set it up. Later I thought to cut one end off the cable so it fit the same inside at the shift stick but had a fresh cut end to work with under the hood. And it let me slip off the offending pipe that diverted the cable in the wrong direction at the firewall.
   On the 7th I wired the headlight switch and the right headlight back to the original wiring, and put a halogen headlight back in the right side. The only way it would come on was with "momentary flash" - not with regular low or high beam settings. Apparently they hadn't been working in the first place. (Didn't the dealer drive it to my place with the headlights on? Who can remember so far back.) There had to be a break between the light switch and the high-low beam switch - but there should be nothing but a wire and maybe a connector between them. Would I have had to deal with that somehow regardless? I decided to give that a rest and bought some copper pipe for a hot water solar collector.

   Then the motor plans started to morf.

The ARM Motor

   I rather suddenly determined on a good axial flux switched reluctance motor ("AFSRM") design on the 9th. It should be easier to make than my supermagnet motors, and if it's really better, how much farther do I want to go developing the BLDC type I'll probably never make any more of? Especially if the high RPM capability of SRM.s eliminates the whole need for a variable torque converter transmission for EV.s?
   The new design still uses my same toroidal coils, as "cup electromagnets" with a steel outer ring, to attract matching "overlapping ring" rotor poles. If the very tight tolerances and narrow flux gap can be reliably accommodated, it promises to be a winner, with much more torque than what I was thinking of previously and (I expect) less torque ripple than other 3-phase SR motors. With a 'solid' steel rotor having no magnets or wires, very high RPM.s can be easily and safely attained, and since power = torque * speed, small higher power motors become more feasible. Outboard motor conversions will work better with high RPM motors, and the external wheel drive to "hybridize" cars, with just a fixed gear ratio, gets much easier.
   With no supermagnets and minimal copper they'd be intrinsically the cheapest of motors, and with waterjet cut steel parts they'll be easy to make and to assemble. I can see motor kits being much more practical to put together, more 'professional' in form and parts, and easier to assemble than the BLDC/supermagnet type.
   The inherently more reliable unipolar motor controller I've been doing for some months now is (somewhat fortuitously) the right type for both the "BLDC4:3" motors and for the reluctance type.

   The new concept had been vaguely brewing in my mind. It was loosely based on an odd feature of a 'theoretical' design AFSRM in one of the research papers, a four phase motor with 16 coils and 20 rotor poles. So what made that seemingly "written in stone" 3 to 2 configuration necessary in three-phase motors? I had just successfully run the BLDC magnet motor having 4 magnet poles per 3 coils instead of two. For SRM.s were there not multiple possibilities? On the 9th I decided to try out some layouts for visualization.
   Since I had "donut" coils with a metal ring around them, forming something like "cup magnets" (here, "cup electromagnets"), then instead of straight lines, I could have rings on the rotor that matched the narrow flux gap rings of the coils. As the steel ring came near alignment with the coil flux ring, they would be strongly attracted to line up. The reluctance would change rapidly over a short angle of turning which creates strong torque, but it would be less abrupt than with straight lines. Good torque but with less torque ripple would doubtless also produce less noise, which is listed as a complaint against switched reluctance motors. If after a rotor ring(s) lined up with the phase A coil(s) another ring was in position to start lining up with coils B (or C), the process could be repeated, and then repeated again with coils C (or B). After that, the second rotor coil would be in position for phase A again.
   What was needed was not two rings per three coils, because each time, the next ring would be too far from the next coil for good torque. Instead a number of rings had to be found such that the next ring for each phase was positioned just about right for high torque just as the previous phase shut off.

Four rings per three phases
When ring 1 reached coil A as shown, A would shut off and C would turn on,
but ring 4 is too far from coil C for strong torque at first.
The same situation will apply for each change of phase.
(Coil A is shown with the outer steel ring. The rest are just the cores.)

   2, 4 or 5 rings per three coils weren't enough to get the next ring very close to the next coil. 3, 6 and 9 are even multiples of three so the rings all align with the coils at once instead of in a 3-phase sequence. 7, while it might have been close to a good distance, had two rings almost equi-distant on each side of the coil at turn-on time, which would cause nearly as much reverse torque as forward. 10 filled in too much space to make a rather solid rotor instead of distinct rings. But 8 seemed to work out. As seemed intuitive afterward, the ratio at the ring centers was about 1/3 metal ring to 2/3 air gap between... seemingly just right for three phase operation.

   The energization sequence (clockwise rotation of the rings/rotor) pulls ring 1 to coil A as shown, then 4 to B, then 7 to C. Then 2 pulls to A, 5 to B and 8 to C. Then 3, 6, 1 etc. By the time ring 1 is back at coil A, there have been 8 electrical cycles for one electrical rotation. And since there are six coils, two for each phase, there are 16 cycles for each physical rotation. So instead of 6 changes of active phase through broad, mostly low torque areas, it's 48 total coil activations in the sequence, each for a short distance through the highest torque zones.




ARM motor with overlapping rings rotor. Bearings including thrust bearing are installed.
Coils await outer rings and attachment to the plate, and then it needs optical interrupter
rotor position sensors mounted. Then wiring, assembly and testing. oh, and painting!
The outer flats are optical interrupters for sensing.


Test assembly of ARM motor case.

   Far from my original design of AFSRM motor that would probably run okay but with low torque and power, here was something to challenge the state of the art - and drive a car with! If it could provide start-up torque with a drive ratio for a single pair of toothed pulleys, like 3, 4 or 5 to 1 (making it 3000-5000 RPM at 100 Km/Hr for typical 13" wheels), that would be ideal. (If the ratio gets much higher, there'd probably have to be a double reduction configuration with two belts and extra pulleys.)

Then, Another New Motor Design

   While I was on holiday, someone sent me a link to a new "transverse flux" BLDC motor layout. It made some strong claims, and it looked like it could be turned into a potentially superior reluctance motor layout. The most interesting feature is that each phase has just one coil, with the wire wrapped all the way around the entire stator. Without having to wrap a wire around each separate magnetic "tooth", the number and density of the teeth can be greatly increased. More points of magnetic force could greatly increase the potential torque per area and per volume.
   I would make a radial flux type of layout with the stator ouside the rotor, and with narrow "horseshoe magnet" teeth. With no coils overhanging the sides, even with the three phases having to be separate and side by side they are thin, and the motor can be thinner than my axial flux BLDC "pancake" motors. If the "horseshoes" were a left and a right leg each from a mild steel plate 3/16" thick, with another piece of 3/16" thick plate between for the 'center' of the horseshoe (and a space to run the wire), that's 9/16", times 3 phases is 27/16", for a motor that could be as little as 2" thick. If the torque per area is effectively increased and a smaller diameter rotor can have similar torque, a whole motor of a given torque and power becomes substantially more compact. Cheaper. Lighter. Simpler to make.
   When this splendid info came the waterjet parts were already cut. It seemed too late to simply cancel the "ARM" design, so I've continued building it. I'll have two distinct models to compare.

Mounting a Motor for a Car Wheel Drive

   The troubles with mounting a motor directly on or in a car wheel are (a) with no transmission it needs excessively high torque and (b) the unsprung weight of this heavier motor interferes with 'handling', the ability of the wheels to bounce up and down to drive over bumps and dips smoothly. A few months ago I thought up an idea for mounting a motor beside a rear wheel (behind it or ahead) with a belt drive, which solves both problems without adding much friction. (Flat belt drives are 99% efficient and toothed ones must be pretty close.) This month it dawned on me what actual pieces to use to implement such a system, and as I had them all I put them on the car to illustrate.

   From the picture, visualize a sturdy bar with "belt guard" edges, going straight back from the red hub to a pancake motor mounted on the bar behind the wheel, with its axle sticking through to a smaller toothed belt pulley. Since the bar is attached to the wheel (hub) and the motor to the bar, the belt will maintain its normal tension as the wheel bounces around with road bumps. The rear end of the bar is attached to the car body/frame to hold it horizontal, but with enough play to accomodate small movements as the wheel bounces.

   Now it needs a high RPM 'pancake' reluctance motor having the torque to turn the wheel with some fixed 'gear ratio' and the power to run the car on the road.

   See the detailed project writeup for more details.


Solar Hot Water Heater

   On the 8th I thought again about the solar hot water (TE News #88). In a house where people are living, solar hot water is the best way to reduce the electricity bill, heating much more water than solar PV electric for a given panel size. A 4' x 8' (32 sq.ft.) panel is typical for a minimal system. With all the plumbing involved it's not really practical to bother with anything much smaller. Two panels is better for a family.

   I had a copper tank (saved for many years now) for the preheated water. I decided to put a 'proper' copper collector panel together, similar to the one I made in 1979. Since copper pipe is rather costly I went to Ellice Recycling to see what they had. I wanted 2-1/2' pipes, but at 5$/pound, it really wasn't much cheaper than new pipes, and because they were odd lengths that had to be cut with various leftover waste, they were actually more. I spent a fair bit of money there before I realized this. I had a few thin aluminum fins that clipped onto 1/2" copper pipe and will buy more at Emco, to paint black and form the collection surface. They're 2' long by 4.75" wide. Since they're for 1/2" copper pipes, I'll still have to make something to extend the collection surface of the 1" top and bottom pipes.

  The pipes weren't cheap. I bought them in dribs and drabs, but it was probably over 200$ for the collector alone. Now I remember why alternatives to copper are ardently sought. But it's the best for a collector (and perhaps woodstove connected) system that might get quite hot.
   An opposite problem is cold - freezing of pipes. Copper is prone to splitting or joints pushing apart as the water freezes. I'll circulate house water through the collector and into the tank to keep the plumbing simple. There will of course be valves but it's a pain to drain the collector, high up on the roof with valves in the unfinished attic, for the winter. I keep thinking a small water heating element in the collector's lower header pipe, with a thermostat set just above freezing, would cause the water to circulate to the tank and prevent everything from freezing. But where does one find such a heater?

   On the 28th I drilled thirty 5/8" holes in one of the twelve foot long 1" pipes to silver solder three foot long 1/2" riser pipes to, then I cut it at 8'. The fins above and below will bring it out to fill the almost 4' x 8' frame. A 4' x 8' piece of clear plastic (lexan?) would cover the top with a bit of overhang. I tried to fit the pipes and they didn't go in the holes easily. But if they weren't a close fit, they couldn't be sliver soldered. I wished I had had something other than a twist drill, which doesn't make very exact holes in thinner material like pipe walls. Then I tried to solder in a pipe. While the silver solder melted off the wire and beaded up near the joint, I couldn't seem to get the copper hot enough for it to flow with the propane swirljet torch. Of course, there's a big copper pipe carrying off the heat, and the slower the heat, the more it spreads out without getting the heated point up to the required temperature. Next try will be with the scary old naphtha gas torch someone gave me. It really puts out heat.

Peltier Module Cooler Voltage Tests

   The Peltier cooler hadn't performed well when I went camping. When I bought it to "reverse engineer" I had never intended to use it. I had lost the aluminum heat transfer block and hadn't reassembled it so well. After all this time, I thought to hook it up to the the lab power supply and  test out my ideas of better performance at lower voltages. (I'd rather have Peltiers with 160-210 thermocouples to run at 12 volts, but except at very high price they unbiquitously have 127 themocouples, which puts them in an operating area that cools but has poor efficiency.)
   Also, my complaints about their short life expectancy (fine for occasional camping use, but not for continual home operation) might be alleviated if they weren't being run so hard. That would make a big difference in their practicality for appliances - heat pumps and refrigerators.

I found it worked best at 9 or 10 volts. The temperatures got coolest, and using much less power than at 12 to 14 volts. 9 volts got it coldest, with half the power of 12 volts. Just as I suspected! (See detailed project report for figures.) But I should try it all on the Superinsulated Thermoelectric 12 volt Fridge, which is working well to start with, and maybe try the large 14 amp Peltier module, run at lower voltage and power. It might just cool the whole fridge better with the same or less power.

Battery Making - DES ingredients arrive

   When I returned from camping trip, my order for choline chloride and ethylene glycol had been delivered, theoretically allowing me to try making some "Ethaline" DES electrolyte. Then again, I need to find time to check out the directions and do it. One can only tackle so many things at a time, and I definitely have less energy than when I was younger.


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

Solar Photovoltaic (Solar PV) Power Coming of Age

   The price for smaller solar panels is over 2$/watt, but the price for "full size" (about 1m x 1.6m) panels of 200-300 watts has dropped to an amazing 1$/watt even in shrinking Canadian dollars. If I went into a glass shop and asked for a piece of tempered glass that size with an aluminum frame, I'm sure it would cost more - without a solar collector and wiring glued to the glass!
   And the capacity of the panels (and so presumably the conversion efficiency) has been creeping up, from a typical 200 watts to 250 or better. This month I bought a 260 watt panel for 263$ (admittedly that's the wholesale price), to put on the east sloping roof for more morning power. I'll then have coverage pretty much from sunup until sundown.
   But I seem to have missed making a contribution with my nanocrystalline titanium dioxide borosilicate glaze frit, which was to be sprinkled on solar panel cover glass and melted in like little sidewalk pebbles. That would reduce reflections, difract the light inward, and panels would pick up more ambient and low angle light when full sun isn't coming straight on, for a higher daily total output.
   In 2010 new solar installations became cheaper than new nuclear plants. Solar has continued to cheapen and nuclear to rise in cost even without considering serious environmental costs such as the Chernobyl exclusion zone, and as the Fukushima nuclear plants continue to disgorge dangerously radioactive waste into the Pacific ocean. After this disaster in 2011, Germany decided to phase out its nuclear power plants, and offered a serious incentive program to homeowners to install panels on their roofs, and (as best I recall) as of maybe a year ago had an installed base equivalent to three nuclear plants. Power plant operators were complaining of falling demand for their output. I heard Ontario did the same thing briefly, but soon stopped. I had calculated in 2012 DIY solar PV paybacks (well, just for the wholesale panels, and on kilowatt hours produced or consumed at the then current prices) of 10 years where there's lots of sun (2000+ hours/year), and 20 on the cloudy-all-winter Pacific northwest coast (1000 hours sunlight per year) at about 2$/watt for panels. These should now have dropped to about 5 and 10 years on this account, and the cost of electricity in BC has risen from about 9¢/KWH to 12, which makes it well under 5 and 10 years. And there is no indication that electricity price increases are going to stop any time soon.
   Solar power isn't a 100% continuous power solution like river hydro, lambda ray collectors or perhaps magnet motors would be, or near to continuous like wave power on the west coast of Vancouver Island, but it's becoming a viable way to cut energy bills and reduce dependence on the power grid - power in an emergency. But if its use continues to grow, it may be practical at some point to consider an electric grid ringing the globe: the sun is always shining somewhere.

   My payback assessment based just on wholesale panels is doubtless overstated for most homeowners. I tend to make my own electronic parts and I do my own installing. The prices for the other needed equipment - charge controllers and batteries, or grid tie inverters - haven't come down as much, and installation costs are probably about the same.
   Then my 12V CAT standard plugs and sockets, which I should be doing much more to promote and to market, can make 12V DC house wiring far more practical than it is.

   I discount the drop in insolation and loss of solar power from the perpetual chem spraying, which I note in a brief clip below, simply because I don't believe they can keep doing it for many more years. Hasn't it caused enough ocean life die-offs and climate chaos already? I hope by next summer this insanity has been ended.

Comet 67"P" for "Perspective" (At the risk of excessively belaboring this subject.)

   Something that looks strange from one angle may look familiar at another. For instance, people thought they were seeing strange "glass tubes" sticking up on Mars in Mars orbiter images. I myself was puzzled by these bizarre images. But when turned upside down so the light is seen coming from the top instead of from the bottom, they are seen to be sand dunes filling hollows in rocky areas.
   In a similar vein, if the picture of the comet shown last month is rotated to apparently "level" the surface (while acknowledging that the micro-gravity is surely pointing different directions in different areas even within the picture), the landscape I thought looked like vegetation... looks even more like vegetation. Right at the top is an apparently treed or bushy hilltop silhouetted against the sky. Doubtless it must be very different from Earthly vegetation. Further down, the appearance becomes more like aerial or satellite images of scattered patches of trees as seen from above.
   Perhaps one might decide it looks more like some form of crystal growth. But dead crystals don't emit seeds - which are probably what the organic particles being detected by the orbiting Rosetta spacecraft are, which are probably similar to the 'seed from space' collected in 2013 by a high altitude balloon above the Earth.

   It's always dangerous to jump to conclusions about other unfamiliar worlds, especially based on a single image. The "glass domes" or the "Face on Mars" are but a couple of well known examples. But if a picture is worth a thousand words, I'd guess that color photos of this comet from this range would be worth a thousand black and whites. So would close-up color images of any of those airless, icy worlds having fluffy, polycyclic aromatic hydrocarbon surfaces.


   Another interesting aspect to this comet is its "dumbell" shape. Shoemaker-Levy 9 was pulled apart at Jupiter by tidal disruption. Might 67P be two parts of a larger body that was pulled apart - two parts that barely managed to stay together in a similar event?

   Here I'll update on Ceres: Dawn has lowered its orbit, but it's still over 1000Km out. There are closer pictures, but without spectral info there's nothing that leaps out and suggests the dark material is aromatic hydrocarbons, and it doesn't even look much like "fluffy", at least from the present distance.

Vladimir Putin - Leadership & the need to spread power more widely

   For those who believe the demonizing western media and think of Russia's president Putin as some sort of relic from the cold war, I say: go to youtube, type in his name, and listen to him speak - at most any function or time on any subject. His eloquence and his straight answers to any question show he's a deep, sincere and philosophical thinker.
   However, I point out again that the time of any leader in the limelight is short, and that until real safeguards are put into place and power is more spread out than today, with a good measure reserved by the people directly, the chance of Russians finding another Putin after he's gone is about the same as that of Americans finding another Kennedy. Furthermore, many things that should be done must be okayed if not actively sponsored by the leader, but in the aggregate they are beyond the compass of any one man, and so they generally get left undone. Then when a less effective leader is elected, there is likely to be a disappointing era, and when one ineffective leader follows another with no real input from an aware and educated public, a collapse.

   This link is to American Veterans Today. Is this why the US government is afraid of their own veterans? Is it why Russians are much better off today than when Putin first came into office? Presented without further comment from me.

http://www.veteranstoday.com/2015/08/08/vladimir-putin-battles-satan-worshipers-and-new-world-order-agents-again/

“The continuing attacks on Vladimir Putin and Russia by members of the western political, military and journalistic elite tell us one thing – the Russian President is doing a good job both for the people of his country and in the international arena.”—Neil Clark


More Collapse Warnings and Signs Add to the Chorus (At the risk of excessively belaboring this subject too.)

   I recently got an e-mail saying a retired British official, Damian McBride, adviser to former prime minister Gordon Brown, is saying that the banks may close and people shouldn't trust that their credit and debit cards will continue to work, and that consequent supply chain interruptions will probably mean grocery store shelves will be bare.

WHOA! British Official Warning Public: Stock up on food, water, canned goods & cash - enough to survive 1 month - Banks may CLOSE (Stock meltdown)

http://investmentwatchblog.com/whoa-british-official-warning-public-stock-up-on-food-water-canned-goods-cash-enough-to-survive-1-month-banks-may-closestock-meltdown/

This link points to a prior article link in the British Newspaper The Independent, which was based on a twitter 'tweet' by McBride.

http://www.independent.co.uk/news/uk/politics/stock-up-on-canned-food-for-stock-market-crash-warns-former-gordon-brown-advisor-10469509.html

   There are of course many who think this is all ridiculous, but many others would say his measures don't go far enough. They would be an excellent start for those who've done no preparing at all so far. One spends hundreds of dollars a year on house insurance against the remote chance of a fire. Look at having some good supplies on hand as personal insurance against various forms of wider disasters, natural or man-made.

   The first hint was the dropping export of Chinese goods to the west, where the bulk of the middle class is now essentially broke. This led to the Chinese stock market plunge and the recent devaluation of the Yuan. Related to but not really because of these events, the first "shot across the bows" in the west was fired at the US stock market on Monday August 24th as the highly overvalued Dow Jones plunged 1000 points at the opening bell, after a weird 'glitch' in the trading system in July and some weeks of nervousness and small drops. By mid morning the president's "plunge protection team" had brought it back up substantially, and later in the week it was further restored. But who else is buying? Will they simply print whatever money it takes to buy all the stocks to prevent the bubble from bursting? Is the New York Stock Exchange then a market, or just a pleasing façade on some ugly economic realities? They may soon have to give up and let it burst, as further drops on Monday, September first would hint, or they may sustain it to the end with money printing while everything crashes all around.

   And this brings up the fact that "liquidity" is scarce everywhere. China, also short of cash, has liquidated over 10% of its US treasury holdings and isn't buying more. Added to this are all the oil producing nations, now selling off accumulated wealth funds because of low oil revenues. Who then but the US "Federal Reserve" will be buying those back, and new US treasury bonds? Will they triple their own balance sheet while continuing to force all American banks to also buy bonds, to keep the US government afloat? How many trillions of dollars can the economy absorb? If hyperinflation sets in, no one trusts currency any more, and quickly its demise is all but ensured. Everyone rushes out to cash out and buy things while they can, the stores empty and close, and the distribution network also goes into turmoil when the truck drivers (etc) won't accept payments in devaluing cash.
   As exclamation points on the end, HSBC bank seemed to be having trouble meeting payroll near the end of the month. They claimed it was a technical glitch, but employees complained they weren't getting paid. And the state of Illinois isn't even paying out lottery winnings over 25000$! A family "won" 250000$ but evidently will get nothing. If your dream, or your last ditch financial plan, was to win the lottery, it looks like you can forget it!

   Remember: currency and credit are only claim checks on wealth, not wealth itself. If the claim checks, now mostly just blips on a computer screen, disappear, are dishonored or are made virtually inaccessible, if the insolvent global banking system holding most of them collapses, if rules are changed through government or financial institution dictate, or through rapid inflation, life savings can disappear overnight. When in such times of sudden crisis the claim checks vanish or lose their value and people with too many of them and too little actual property are impoverished, actual wealth doesn't disappear, but it gets transferred from those trusting the claim checks - most people - to those more knowledgeable, prescient or just well placed, who hold their wealth more "outside the system". While most were impoverished, more people became millionaires during the great depression of the 1930s than at any previous time.
   Real property is wealth. Precious metals, silver and gold, are only one form, but the most tradeable one, of real property. So formed as coins or bars of specific weights they have been considered "money" for thousands of years. They can't be hyperinflated to worthlessness, and if held physically by the owner they can't be transferred away or locked in by the stroke of a computer key or government decree. They can only be taken by physical robbery -- which is of course never a danger to be neglected. But in turbulent times and with good precautions, it's less risky than trusting in claim checks or having others hold your valuables. Actual cash is likewise much safer possessed than held as accounts in the custody of untrustworthy trustees. (Interestingly, money in a Paypal account may be safer than in a bank(?) - while they transact with banks continually, they aren't leveraged and the funds held seem to be pretty much held outside the banking system. Bitcoin in your own encrypted, password protected wallet, seems pretty safe, and is hyperinflationproof.)

   A means of producing something or providing a service people want is a means to 'generate' wealth. But new things made usually equal old things recycled, and land is pretty constant, so there's only a certain, somewhat constant amount of real wealth on the planet. The 'cosmic values of social sustainability' again apply: Quality of life, from which must follow Provision for Growth and Equality of opportunity and laws. The land-man ratio dictates how much wealth is available per person. All else being equal (technology, farming, social development), if there are half as many people, everyone will (or should) be notably better off, and quality of life will continue to deteriorate with an ever growing population. (I'm just now reading Social Sustainability HANDBOOK for Community Builders by Daniel Raphael, PhD, which goes into these recently recognized principles and their applications in more detail.)

Chem Spraying Ruining Solar Electric Car Charging!

   I've had the RX7 EV running for 3 summers now, and the solar panels on the roof for 4. Now charging the RX7 with solar power has become difficult because even the lightest of chem trails blocks the sun and spoils the effectiveness of the solar PV panels. The 12V fridge and the indoor plant light, together using around 80 watts, are okay, but if I've driven the car and it needs another, eg, 200 or 300 watts (from the 892 rated watts from the south facing panels plus 500 from the west facing ones), then whenever a chem trail, however light, drifts past the sun, the inverter goes into low voltage alarm and the power, which there should be plenty of, goes on and off because it's insufficient. It's not practical to charge the car batteries in banks, so I have to plug the car into grid powr, on an otherwise sunny day. Solar charging is unreliable to impossible in other seasons when overcast or frequent clouds come by, and of course towards evening (now almost at sunset with the west facing panels), but it has never happened in the summer repeatedly, day after day, until this summer. I'm also getting tired of collecting brown rainwater that foams when it's poured, on the rare occasions it has rained.

   I note that this month (about the 23rd) brings news of a big die-off of whales in Alaska. Their bodies are washing up on beaches. The suspected culprit: algal blooms (again), which would of course be caused in the open ocean mainly by the unusual level of nutrients for plants/algae in the water from the chem spraying.

Fuchsia Berry Pie!

   Someone told me fuchsia flower berries are edible. I kind of rolled my eyes inwardly. Like there would ever be enough to use! But I have a large fuchsia bush, and this summer I noticed there were really quite a lot of berries hanging down. Enough for a pie? I picked them. Well, they made a thin pie with no top crust. (And there were more later!) It was quite tasty, with mild flavor and much less gritty than pies from the prolific blackberries people insist on picking.

The fuchsia bush


The berries look rather like beans, but the soft texture and mild taste is of course entirely different



Purportedly Funnies (Originals unless someone else already thought of the same ones)

Veteran: A veteranarian in training.
----
"Dan, I know the transmogrifier has radically altered my appearance, but you know me, I AM Franz Jeckyl. To prove it, ask me anything that you know only I would know."

"Let me get a pen.   Now, what is the online password to your Royal Life Savings bank account?"

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)

 
August1: Disassembled Sprint dash to fix headlight wiring (no worked with LED lights!) Worked on July report.
2: Finished & posted July TE News (#90). Changed headlight wiring in Sprint because polarity was backward for LED lights - Ugh! And they're still not working.
3: Cut shift cable.
4: Started in on the design for the 12 magnet jig.
5:
6:
7: Worked on Sprint headlight problem. (Turning into a big waste of time!)
8: Started on a solar hot water collector. Bought some copper pipe (but not enough)
9: Started newly inspired design of AFSRM rotor
10: Continued design of AFSR motor.
11: Figured out how to export to .DXF - Victoria Waterjet said the AFSRM motor rotor and plate files weren't suitable... needed defined curves instead of curves broken into short line segments.
12: Checking out other CAD programs that might put out .DXF files with suitable curve definitions.
13: Learned more or less how to coerce LibreCad into producing a .DXF file for the rotor. Took it to Victoria Waterjet (where he modified it since it still wasn't right, but had the essential curves and lines.) Did up bottom plate.
14: Fixed a problem with the plate, then did up the top plate as well. E-mailed them to Victoria Waterjet.
15-21: Holiday!
22: Read about new "transverse flux" BLDC motor - has advantages. Bought more copper pipe for DHW solar collector.
23: Worked out many of the design concepts for "transverse flux" SR motor. Started Peltier module voltage versus cooling experiment.
24: Went shopping, found/bought possible 6 inch "drum rotor" for said motor.
25: Picked up AFSRM (ARM) motor parts at Victoria Waterjet, got a hub at Princess Auto. Picked up a 6" disk with 1" hub there as well, for 'transverse' flux... "Horseshoe?"... Motor.
26: Turned hub lower end OD to diameter of thrust bearing ID. Started fitting motor together.
27: Carefully fitted motor rotor to lower plate. Finished first Peltier versus voltage experiment.
28: Wound 4 coils for "ARM" motor. Drilled holes in copper pipe for DHW (domestic hot water) solar collector.
29: Wound last 2 coils (needs 6), did some adjusting and fitting on the motor.
30: Touched up wires/epoxy on coils - and figured out how to get them to work despite slight overhangs on at least half of them.
31: Worked on newsletter

Electric Hubcap Motor Systems - Electric Transport

AFSR Motors for Hybridizing Cars

   For some reason, I seemed to be having a hard time getting started on the motor and controller work. Somewhere in the back of my mind was the nagging thought that the axial flux switched reluctance motor (AFSRM or AFSR motor) might be the best of all motors once developed (and probably cheapest to produce) and would supplant all my other types, and also that I now had a superb design for putting one on a car wheel. (TE News #87 & below) Furthermore, there was a good chance that if I drew out the design nicely, the mechanic at AGO might have time to put together the critical mechanical parts set.
   This came back to the fore in my consciousness on the 9th, as I thought about potential rotor designs with more "lobes", that might work better than what I had made so far. The overall value of converting a car from gas to electric at this time, however superior the expected performance, seems to me to be less than that of making a regular gas car into a plug-in hybrid (with similar great performance except for hauling the gas propulsion system's 'extra' weight around), using an add-on external motor.
   But having looked at the other designs, I was convinced that my AFSRM design, while it would run, would be rather low torque. It wouldn't be a really effective design. The six donut coils with rings could work. The rotor was the problem. Unless the steel torque elements on the rotor matched the curved lines of flux from the coils, the torque would be low.

   Let's see... I have the 2" O.D. donut coils, with the wire coils and then sheet metal rings around them, making a ring of very high magnetic flux with about 2" I.D. and 2.5" O.D. What if I welded overlapping steel rings of that size, maybe 3/16" tall, to a solid steel rotor, in a pattern such that just as one is pulled into alignment with coil A, another ring is set to pull onto coil B with high torque, and then the same for C, then A again? One might thus attain a large number of very high-force pulls per rotation. And with just .25" wide walls that don't align with anything except right where they're being pulled, the rings wouldn't exert much magnetic drag. It sounds ideal. Now... how far apart and how many rings? I figured that would be best drawn on paper to get a sense of the geometric proportions. Even better, how about cutting rings out of paper and arranging them in different patterns? First the 'ring draw' on a piece of cardboard. The cutting was much slower than the drawing. I eventually cut 11, just the number really required for the last layout.



I used six real coil cores roughly layed out on a real rotor plate per the Electric Caik motor. (The thrust bearing for visual effect would have been better if I'd straightened it out.)

First arrangement is the unipolar motor controller's "normal"(?) 4 rotor elements per 3 coils. The fifth (dark) ring starts the second set of three coils (A), (B) and (C), which is merely a duplicate of the first set. This is 'theoretically' good, but (for clockwise rotation) as rotor ring 1 lines up with coil A as shown, having produced its thrust, rotor ring 4 would be too far from coil C to have good torque at first. Each ring in sequence would be the same, making poor and lumpy torque.


Next I increased it by one ring to 5. This time, ring 3 would next align with coil B. It's getting better, but the metal of the ring is still pretty distant from the flux ring at first, which would make for bad torque ripple with low torque areas.

As with 3 rings, six rings per three coils would have all three phases line up at the same time, then there'd be no torque at all. No coil could be turned on without producing reverse torque as the motor went further. Depending on the starting rotation, the motor wouldn't be able to start, negating the whole point to having 3 phases 120 [electrical] degrees apart!


So the next choice would be 7 rings per 3 coils. At this point, there's a lot of metal, and we need to ensure there's no undesired backward force seriously counteracting the forward force.

In this case, ring 6 is quite close to coil C, and should jump into alignment with it when C is turned on. Since only coil C is on (one coil ON at a time), alignments of other rings with other coils won't cause backward forces. But looking more closely, I would hardly trust ring 5 not to jump backward to C instead of 6 to jump forward. They're too equidistant for my liking, and the torque will be reduced until 6 gets closer and 5 gets farther from the coil.

[Ignoring that...] Once 6 and C were aligned, 4 would be ready to attract to B, and so on, and the energization sequence would be CBACBA... Note that this is opposite to the 5 ring arrangement and the same as the 4 rings. But that merely reverses the sense of the forward-reverse switch.

If counterclockwise motion was desired instead, coil B would go on next and attract ring 3 equally but opposite to C and 6, then ring 5 to coil C for the sequence ABCABC...


Finally(?) I tried laying out 8 rings per 3 coils. At this point there's more metal in the rings than space in between. But is that bad, or good?

Ring 4 [note: should be placed a little farther to the outside] is almost aligned with coil B, and it should have strong torque to pull into alignment with it. Close together as the rings are, ring 3 is farther away so it won't do much backwards force. Ring 5 will be minutely helpful, almost canceling 3.

Then 7 will be equally close to C (as 4 was to B), and following that 2 will be ready for coil A to turn on. then 5 at B, 8 at C, and then 3 at A. Two rings are skipped in each sequence, but they are different ones in each successive sequence. This may be the highest torque arrangement because the rings are each energized in turn just approaching their highest torque area, in rather rapid succession, and nowhere else. As it happens, the geometry just works out much better than with 7 rings per 3 coils -- for these rings and these coils, at the spacings shown.

9 rings per 3 coils would again have rings aligned at all three coils at once and wouldn't have any torque to start at some points.


I was pretty sure 10 rings would have so much ring with so little space between them it would start losing the torque by the coil attracting more than one ring. Later I tried it anyway. Not only did that look like a problem, but like the 7 ring setup, it looked like two rings would be almost equidistant, pulling in opposite directions.

Having so casually placed the coils, I later checked to see that they were 7.75" at the outside diameter. [Electric Caik rotor size]. They were about 8", so I adjusted them and tried again. The results were the same. 3, 6 and 9 rings are out because they line up with all three coils at once. 2, 4 and 5 left some large gaps - very low torque areas. 7 and 10 made for rings trying to pull almost equally in both directions.


   8 rings - doubling up the 'typical' 4 - looks like the arrangement for good torque. This could make for some tricky fabrication - cutting and welding or machining. Even the optical rotor position sensing device will need 16 slots!

I was hoping it would want fewer rings. But that's why I cut all those cardboard rings and layed them out; to find out. And now that I have, it makes sense that if the ring width is .25" and each coil is on 1/3 of the time, the spaces between rings should be 1/2", which is about what it is (widest part). But I only realized that after having done the layout and perhaps because of it.

There will be 8 activations of each coil per electrical rotation (one per ring), or 16 cycles per physical rotation since there are 6 coils, two per phase, and so 16 rings. That compares with just 4 for the BLDC4:3 motors and two for regular 2:3 BLDC.

   The next morning (10th) I was casting around for materials and ran into a washer about the right outer diameter but with a little smaller inside and so thicker walls. It occurred to me then that instead of having so many rings, one might simply make them like that - inside smaller or outside larger. That way, with the 5 ring system the edges of the rings would be closer to the flux areas when they activated and there'd be more torque. In fact, adjusting the ring size would be a way of fine tuning the torque ripple versus peak torque. And the 1/3 metal to 2/3 space ratio could be attained with the 5 or even 4 ring layouts. But fewer rings would surely be lower torque. Or the rings could be thinned a bit for the 8 ring layout if that seemed desirable.
   One can of course try out new rotor designs for real in one motor, by creating and mounting the desired rotor and an optical sensor ring with the matching number of slots. The one that seemed most advantageous would be kept.
   And to skip ahead a bit, when I set the pieces together after the rotor was made, the rings seemed a little more distant from the coil flux rings for the starting rotation of each phase than I expected, which will reduce the torque until they get closer. Thus the torque may have a fair amount of ripple. Maybe I'll do a second rotor with bigger outsides on the rings, which should solve that. Or maybe check out 10 or 11 thinner rings, which just might make higher torque than 8 after all (with of course higher coil sequencing speeds).

   The motors in the research papers had two rotors, one on each side of the stator. But absolutely flat, rigid and level construction on all points is vital. I decided to try out a single rotor design - it would be way simpler and more robust. It might be good to aim for about a .025" (.6mm) flux gap. The needle bearings hold more precisely than the automotive ("trailer wheel") bearings, so they're the pick. The center holes will be perfectly sized for them to hold them in alignment. The thrust bearing will help with precision leveling between rotor and stator. A flat plate of mild steel would seem to be vital as the base for the stator both to hold the coils exactly and to complete the magnetic circuits at the bottom.

   As I thought about doing a drawing for the mechanic, I started thinking more about having a circular bottom plate with a center hole sized for the bearing cut by waterjet. And then, why not waterjet cut the bolt holes for the coil holding bolts?... and then, for everything else? And then, why not do a top plate, too? And why not have the whole rotor, with optical interrupter holes and all those overlapping rings cut out of a plate by waterjet, too? That would take all the precision grunt work out of it. Any number of rings becomes as simple as any other (except to the waterjet machine), and it makes it easy to try out the most promising ring layouts and variations of them!
   Suddenly making an AFSR motor started looking, if not trivial, at least simple enough. I called Victoria Waterjet and they said they can use .DXF files. OpenSCad, the program I design 3D objects in for the 3D printer, can export to .DXF. So I could create the designs in software I already knew, and have them precision made: BINGO! (And there I was, about to go down to Western Equipment to buy a bunch of those washers, to cut them up and weld the pieces onto the bottom of a steel disk!)
   Next feature: Princess auto sells hubs that fit on a keyed shaft to weld chain sprockets onto. I could make the rotor center to fit one of those, and weld it. And I could machine out a bit of the bottom end of the hub to fit it to the thrust bearing. Finally, a 3/8" aluminum plate could be rolled into an outer shell (only about 3" tall) with the metal roller at Victoria Makerspace, and then perhaps be welded into a "perfect" tube. Or, given that SRM.s are said to tend to be noisy, perhaps a PP-epoxy outer shell would be appropriate as a sound dampener -- if the top bearing position could be centered accurately.

   Knowing how cramped it was inside the Electric Caik motors with 7.5" diameter, I decided to use the 8" diameter to the outside of the coil cores; 3" radius to the centers of the coils and rings. That would leave more space for the optics and wiring in the middle - and enough room for running wires between coils. By evening of the same day, the 10th, I had the bottom plate and the rotor with 16 overlapping rings designed in the CAD program. Suddenly, the design was coming together with blazing speed, with most of the nebulous loose ends pegged!
   I went through some SRM literature on the web late in the evening. Before I went to bed, I had the thought to use the outer curves of the rotor rings as the optical interrupter slots and solids. The optics could be mounted on the outside wall - and probably would be wired from the outside. This too seemed like a simplification, and I had just the right opto-interrupter units for it. It looked like it would be faster and easier to make this motor - including the prototype - than my BLDC motors.

   What a long way this project came in two days! However, on the 11th came a check. After figuring out how to export the files to .DXF format, which to my surprise took the computer a couple of hours, I drove out to Victoria Waterjet. As with the .STL files for 3D printing, the .DXF curves were broken down into a series of short straight line segments. I was told those were very hard on the waterjet machine, which would repeatedly stop at the end of each line and then restart in the next direction, 'jittering'. The 3D printer doesn't stop, but I know my drill-router also breaks down G-code curves into line segments, and it too stops at the end of each segment, so I was familiar with that, and if it's a problem for the waterjet machine, I have to accommodate that.
   That meant I'd have to find and learn to use another CAD program that would put out .DXF files suitable for the waterjet machine, and redo the designs in it. This would take it from "done in a day" to days or even weeks, considering the learning curve for new software. I had downloaded the 2D CAD program LibreCAD a couple of years ago, hoping to use it with the CNC drill-router, but it hadn't worked out. It was pretty unintuitive. I had to ask the author how to do some pretty basic things. It did them well if you knew how. But by now I had essentially forgotten how to use it.

   I went into AGO on the 12th and found they had an assortment of CAD programs that did .DXF. I was given a CD with an old MS Windows version of Autocad. But when I got home I thought an old version of software might not be the best, and I decided to try LbreCAD again ("better the devil you know" - tho I never really got the hang of it before). The program was again unintuitive and the on-line manual was sketchy at best. It was enough to get you to hurl your computer out a window, but eventually I got the hang of some basic things and suddenly most of the rotor layout magically appeared.
   When I took that out to Victoria Waterjet the next day, it was just lines and circles on his screen with all the solid and area "fills" missing, but he managed to cut, paste and manipulate it (using a CAD program called Rhino3D, 700 or 800$), which he manipulated with great ease and familiarity) until it was how I wanted it. I did up the bottom and top plates on the morning of the 14th, and simply e-mailed them. Then it was just a matter of waiting for them getting around to it. I didn't ask how many weeks that might be. I was leaving for a week holiday anyway.
   Here are the .DXF ("Drawing eXchange Format") files as done in LibreCad, which as I found may need some tweaking when applied:
AFSRM-Top-Plate-LCad.dxf
AFSRM-Rotor-LCad.dxf
AFSRM-Bottom-Plate-LCad.dxf

Note: I moved the three bearing holder holes outward a bit on the bottom plate to miss the thrust bearing. I threaded the plate holes, filed the square bearing holder holes outward a bit, and used regular 5/16" hex head bolts. This was unnecessary on the top plate. They should be moved in to (IIRC) 1.5" radius from the axle and cut to 5/16" diameter. The center bearing holes are sized to form one side of the bearing holder, so only one pressed needle bearing holder is used instead of two.

   Aside from the top and bottom plates, the other body component is the outer rim. As long as it's stiff and the right diameter (10"), I think it could be anything from aluminum to PVC. A plastic shell might help dampen vibration and noise. The coils are 1.0"... the rotor is .025" above the coils and is .25" thick, and the hub sticks out maybe .125" above that... clearance from hub to upper plate .05"... The thrust bearing might add .1" or less. That's a whopping 1.55" height needed on the inside, for a total of just over 2". (That doesn't count how far the bearings and shaft stick out.) Remembering having to raise the Weel's sides later I'll give it some extra height, but  even at 2.5" x 10" it looks like this is going to be one very skinny pancake motor! It will however weigh about 25 pounds, 10 pounds heavier than the present Electric Caik. I could change the top to aluminum and save a couple. The bottom needs to be very stiff, and also to carry the magnetic circuit. But perhaps that too could be replaced by a thicker slab of aluminum and some steel rings under the coils to lighten it a bit more, and at least get it under 20 pounds or so.

   When I returned from vacation, I found that someone had e-mailed me a link to a "Transverse Flux" BLDC motor. After thinking about this for a day or two, making a reluctance motor with a similar layout seemed so promising that I e-mailed Victoria Waterjet and asked them to cancel cutting the axial flux design. But it was too late. So I decided I would make this design first while I thought about the new one. At least I'd be able to compare two quite different designs, one done with iron powder in the coils and the other with mild steel, and see what happens as the RPM goes up.

   On the 25th I picked up the cut parts. I had to file out the center hole of the "lacy" rotor a bit to fit the center hub. (Better that than too loose!) Then I turned down the center hub on the bottom to fit inside of the thrust bearing. The bottom ring of the bearing was to go in an indent on the bottom plate, which at 10" diameter with an odd size center was too big to turn on my lathe. I took it down the street to AGO Environmental Electronics for the machinist there to do, and it was done the next morning.


  
After getting the waterjet cut steel parts I made an aluminum body ring, threaded the case bolt holes,
and put together the case to see what it looked like. Everything seemed to fit together nicely.


Inside: bearing with rounded edges only goes part way through.


Outside: regular pressed bearing holder holds bearing centered in hole.

   With these thick steel plates for stiffness, it'll weigh almost 25 pounds. One might try, say, 3/8" aluminum for the top plate, 1/4" steel for the bottom, and 3/16" for the rotor, to lighten it up. ...if the bottom plate and rotor don't flex too much and close the flux gap during operation.

   The next question mark was details about making and mounting the coils. I may "fill in" the centers with epoxy, or PP epoxy, somewhat recessed. Presuming that will stick to the cores, it'll give something solid for the coil bolts to go through without anything sticking up above the 1.00" tall cores. The cores will only take 9 wires (11 AWG) when they can't stick out the slightest bit top or bottom, so 18 wires in 2 layers instead of 21. (It can't extend to 3 layers because the steel ring has to fit to the outside.) Unless I put the wires into the rolling mill and flatten them just a little bit. But then, could they be kept straight "on edge" while winding anyway? I had my doubts, and decided to just use 18 turns. At worst, a little lower torque, power, and voltage than it ought to have. Since it's so hard to estimate what it'll have to start with, I can just add 15% to the figures if I figure out a way to wind them better. Maybe micrometals.com has some slightly taller toroids?

   On the 27th I fiddled with the rotor/bottom plate assembly and 6 toroidal cores with no wires. First the rotor plate wouldn't quite fit down onto the hub. I figured the inner lip of the hub must be slightly rounded, and I filed out the lower edge of the hole in the rotor at a 45° angle to match. This time it went on with no gap showing anywhere between rotor and hub.
   The fronts of the rings seemed a little far from the highest torque areas at the point where each coil would turn on, which may make for weaker torque points - torque ripple. If measurements indicate, I may want to try a rotor with rings of slightly larger outside diameter later.

   I wanted the rotor to sit and turn about .025" above the coils. To get this, the rotor hub and bottom plate had to be carefully turned on the lathe to set the hub the right height on the thrust bearing. After measuring things with a feeler gauge for a while, I decided to take the hub down a few thousandths of an inch. After that it seemed about right. It didn't turn with zero wobble, and different coils had different clearances, meaning the base wasn't totally level, flat and even. The lower ring didn't sit flat, but on further examination, it was the thrust bearing's race ring that was slightly bent, not the bottom plate. But it was close enough, and I couldn't flex the rotor enough to make it hit the coil cores.


   It became a slightly different story when I wound the coils on the 28th and 29th. It was easier to say "absolutely no protrusions above or below the cores" than to achieve that when confronted with wires that weren't perfectly straight, gooey, slick wet epoxy, side guards of plastic that annoyingly insisted on curling in under the windings, and a winding setup that wasn't made to prevent said protrusions. Some made it, a couple didn't.
   I finally decided that instead of winding more coils and trying to achieve ultimate perfection in the prototype I would match each coil to a specific position and grind a dip into the plate where the offending wires would touch it. (I had got used to there being no chance of coils causing shorts with the PP-epoxy plastic cases in the BLDC motors!) And if necessary I'd add enough of a shim under the thrust bearing lower ring so that everything had clearance, even if it would result in more than .025" gap most places. IF I stick with this model I can improve the winder pieces, and make 6 or 9 winders so all the coils of a motor can be done at once instead of two at a time, to make it considerably easier and faster. (The epoxy must set in the oven for an hour after winding.)
   I finally decided to epoxy the outer rings onto the coils first, then sand them to exact 1.00" evenness. Then I'd epoxy the finished coils onto the plate, making a thicker patch of polypropylene-epoxy in the center of each coil. The patch, covering the mounting holes, could be drilled out and nylon(?) bolts threaded in to better secure everything. ...or maybe just use lots of epoxy all over to glue them down, maybe in 2 or 3 coats? Being on the stator, they wouldn't have to endure centrifugal force.
Rotor and coils on bottom plate. (Rim and top plate behind.)            




I found a "hidden" item 'invisibly' increasing the top and bottom gaps: the paint on the coil cores is at least several thou thick.
Ideally it should be sanded off. (Note the sanded spot on the wire insulation - it protrudes past the end where it shouldn't.)

Also note the apparent solidity of the "iron powder" core. An ohmmeter shows that the iron particles are in fact not
insulated from each other, although the resistance is high enough to prevent significant iron conduction losses.
I conclude that they are probably sintered iron powder cores, or at least that the particles are pressed together with many tons
of force. I didn't see that on the manufacturer's web site. (micrometals.com) This is merely an observation.
But also the sanding reduces the grainy appearance of the iron.

If I was making iron powder components myself I expect I would simply use epoxy saturated with iron powder
and pour it into a plastic mold because it would be easiest for me to do. But it might not work as well.

Another Layout: "Transverse Flux" Motor

   Just when you think your design must be pretty much the best thing going, somebody comes up with some great new idea. After my exciting new "ARM" motor was all in train and the designs for the steel parts were at the waterjet company awaiting cutting, someone sent me a link to a website with an interesting new "transverse flux" BLDC motor layout, which looks like it should adapt well to the SR motor type. [ http://etmpower.com/ ] It took me a while to figure out from the diagram just what was happening. The on-edge magnets with the "offset horseshoe" electromagnets should make impressive torque. It seems to show only one phase, presumably of a 3-phase motor if one wants torque at all points of rotation. I don't see any way the phases could be interleaved in one stator, so I guess they must go side by side, making the rotor also 3 times as wide, which seems like a waste of magnets and rotor material. But perhaps it isn't once it's all calculated out -- and the SR motor won't be using any magnets.

   The coil winding should be very simple since each phase is just a single large coil wrapped right around the entire stator. (Probably the stator should be made of iron powder, but I'll do a prototype from waterjet cut steel pieces to make a lower RPM version just to try it out and check the construction layout, and test the torque et al. With really good luck, I just might find solid mild steel is fine for the desired RPM range anyway.)
   The flux goes around the coil wire, transverse to the stator in any direction, so it could equally have the rotor inside the stator, outside, or axially, by aiming the stator "teeth" and rotor poles/lobes in the appropriate directions.
   For axial flux and flat rotor for a "pancake" shape, one could put one phase on each side of the rotor, but I don't see where the third phase could fit without adding a second rotor. On the other hand there's nothing intrinsically evil about a radial design. Since the coil wires aren't run around the teeth, one could make the parts with quite thin stator 'teeth' and rotor lobes. It could be stacked with lots of them like gear teeth, making lots of magnetic interaction points. Where my layout has two fairly large rings active at any one time, this could have maybe a couple of dozen finer lines. So the issues mentioned for radial designs in the AFSRM research papers could disappear - perhaps plenty of torque, even very high torque, can be developed. And the radial direction (with magnetic points all around the circumference) also puts no axial bending stress on the rotor - a distinct advantage with the tiny flux gaps. The thrust bearing, handling of such loads, and precision axial placement become unnecessary. I feel this is moving beyond the dual rotor AFSRM.s in the research papers - they are no longer my "design models" to follow after.
   But the rotor needs a wide, "toothed" outside rim, hard to cut in one piece with the waterjet, and if it's solid it'll be very heavy. It might be made as some sort of cast part, a bit like a large-rim gear or a brake drum with outer "teeth".

   I visualize, perhaps, a stator with 3 side by side rings of inward facing horseshoe magnets, offset for the three phases, and a rotor with wide teeth across its outer rim it to match up with all three in sequence. Or perhaps the 3 horseshoes could be in line and the rotor lobes skewed instead (made from three pieces)... perhaps less width would be required that way. Hmm, maybe not. Either way, it's like putting a bar across the horseshoe magnet. The force pulling it to center is strong. As I figured this all out, I sensed that here was an even better, even easier to make, new motor design! I'm starting to wish I just wasn't seeing them! I can hardly settle on one design when I keep seeing something markedly better can be made! At least - so far - it's the same new motor controller for any of these new motor designs.

   Let's see... use 1/4" mild steel for all the stator pieces. One with the teeth facing inward to go on one end, one full diameter but jut a ring around the outside in the middle. The wound coil wire goes inside from this. Then another piece the same as the first, the three forming the ring of inward facing "horseshoe" electromagnets. And they'll all have bolt holes at the same points of rotation so they all line up when bolted together.
    One piece will need a slot for the coil wires to come out through. Hmm... or they could come out at one end, through the gaps between the stator horseshoes. Since the pieces are bolted together the finished (epoxied?) coil can simply be placed in the right place during assembly.
   The teeth will be 1/4" wide. If they were ~1/2" long there would be 17 teeth in an 8" rotor, with ~1" gaps between them. The rotor would need teeth 1/2" long and 3/4" wide, the width of the three stator pieces. For the 3 phases, it would need to be 9/4" or 2-1/4" wide. This will probably be the hardest part to make. Waterjet cutting it from multiple pieces of mild steel would make it very heavy. (But it should be possible to cut them from the pieces left over from the stator rings - they should be pretty much just the right size.)

   The second stator phase would be identical to the first, but the bolt holes would line up so the horseshoes are 1/2" (1/3 of a tooth spacing) rotated from the first phase. The third phase would be another 1/2", and the next 1/2" is phase A again. A 1/8" gap could perhaps be left between phases, making it all 2-1/2" wide instead of 2-1/4". The gaps would isolate the phases magnetically (hmm... is that useful?) and assist with cooling.
   Or instead of 1/4 x 1/2" teeth with an inch space between them, one could put in twice as many 1/4 x 1/4" teeth, with 1/2" between them. (34 teeth for an 8" rotor.) The coil sequence would switch twice as fast. Would that give more flux than 1/2 as many of the larger size? Probably. For once, one of those computer flux simulation programs just might be useful. Maybe.

   I'm not sure of the best approach to the optical rotor position sensing. There are a couple of ways I can think of offhand but nothing very convenient. The LED.s and phototransistors certainly can't go across a 2+ inch wide interrupter tooth gap.

   Considering the three stator phases side by side, and that they have no wire overhangs sticking out the sides: if they are only 3/4" wide each, (or even 9/16" using 3/16" thick plates), this radial layout motor is actually thinner than my axial flux BLDC pancake motors. Flat covers could perhaps be bolted right onto the outside edges of the stator, and depending on the rotor, the bearings could be pretty much internal. This might make it not much more than 2.5" across. (And here I thought having the 3 phases side by side would be a waste of space and material!)
   And while the outside stator theoretically adds diameter, the flux gap is tiny, and the stator itself forms the outer casing. It's probably only a little diameter increase over the axial flux machines with no active parts outside the rotor, but with a ventilation air space and a 1/2" thick outer protective rim. (The contours of the radial design will provide air cooling spaces despite the tiny flux gap.) So it's definitely smaller overall.
   And if the many magnetic flux points should provide the same torque with a smaller diameter rotor, the whole motor will be downsized -- or will have more torque in the same diameter.

Toward the end of the month I found a 6" drum - a circle cutter rather than a brake drum - and a 6" rotor
with a 1" I.D. shaft hole. I would turn the rotor down a bit so it fits, tightly, inside the drum,
adding the center axle interface and better stiffness.

The cross bars can be bolted to the outside through holes in the rim (different widths could be tried),
or the voids between the bars could be cut out of the drum rim so the rim forms the bars.
Neither of those sounds as precise and well balanced as waterjet cutting or machining.


Making an "Electric Hubcap" Plug-in Hybrid EV Installation by the New Plan


Concept for wheel drive motor assembly
[TE News #86]

   Having determined what seems like a better way to mount a motor to drive a wheel, and with the tire off owing to some brake trouble, I removed the "Electric Hubcap" mounting tubes that have extended from behind the wheel to the front for 5 or 6 years. (One was pressing a bit on the parking brake cable, and it had eventually perhaps become slightly kinked, jamming and holding the brake on while driving. That was the apparent problem, but I remember that same brake cable sticking over a decade ago, too. I managed to free it up at that time with some oil. No such luck this time!)

   Then, having given it some thought, on the morning of the 15th it came to me that a 1" or 1-1/16" trailer stub axle with four bolt holes, just such as I had tried out in 2009-11, was the right piece. It attaches to the car wheel against four extended lug nuts by short 12mm bolts. Its holes are actually a 4" (101.6mm) spread, but a little filing on the inside brings them down to the 100mm spacing of many 4-bolt small car wheel lug bolts.
   With this sticking out from the wheel (about 6-7" long), first a spacer is used to place the drive pulley out past the body of the car. (It looks like it should be about an inch wide - or maybe I could cut off the mud flap and have little or no spacer.) Then the pulley is bolted on to the spacer to turn with the wheel - to turn the wheel. (The details remain to be worked out. If toothed belt pulleys and parts prove too hard to get, a heavy chain and sprocket gears could be substituted.) Outside of that, a cut-down trailer wheel hub is placed on bearings on the axle. Then a spacer (if necessary) and then the axle nut and cotter pin on the outside.
   The hub makes a perfect, stable place to mount the wheel end of the motor bar. Instead of being a flat piece, the 'bar' will have sides going inward all around as a guard over the belt. Thus it will look like a belt guard, but  it will be made stiff enough to carry the motor and the tension of the drive belt.
   Since the hub on the wheel holds the motor bar and prevents it from twisting, all that's needed at the rear end is a single hinge or pin to the car body or frame to hold the bar horizontal. This has to be somewhat flexible so the bar can move back and forth a bit as the wheel bounces up and down on road bumps, and the single attachment point makes this easiest to do. Centering the car attachment point on the motor axle is ideal, but anywhere near the motor will do.

   I actually had most of the required pieces including a sample toothed belt pulley and a cut-down trailer wheel hub, so I took them out to the car to demonstrate how they would fit and take pictures. It'll probably want a larger pulley (or chain sprocket), like maybe to 10" to 12" in diameter, and the trailer wheel hub that the 'bar' attaches to could be cut down just so that it fits the shaft and the regular cap can be put over the end of the axle.
   The motor mounting bar should go pretty much straight back (horizontal) from the trailer hub, with its sideways guards covering the belt and pulleys. The motor would bolt onto the outside, with the shaft and pulley sticking through to the 'guarded' inside. Again the motor would be a "pancake" variety, not one that would stick out a long way to the side, which could also put strong twisting forces on the mounting bar.

   The horizontally flexible bar-to-car linkage/attachment pin remains to be worked out. It might go through another plate with a slot, welded to the main 'bar' but behind the motor pulley. There'd be a bolt with a large washer on the end of the pin so the assembly can't slide off the end of the pin. (and another washer inside so it can't slide in and the pin touch the pulley.)

   Obviously it'll be harder to change a tire, but not as bad as previous arrangements. One can (I expect) simply undo the four bolts from the wheel and drop the wheel end of the bar down any available amount - and it's not very heavy, the motor end being still supported. Maybe swing it outward a little if some good slack was left at the pin end mounting. Then the wheel can be removed and replaced as normal.

     
Two more views: from rear, 'exploded'


Variable planetary gear transmission with Centrifugal clutch in Chevy Sprint

   I started by trying to get the LED headlights working, since I already had half the dash apart. I grounded the "always hot" pins at the lights, and changed the headlight switch to supply +12V instead of ground... but no +12V was getting to the headlights. What was between the switches and the headlights? Nothing, according to the Chilton's manual. The Haynes manual showed something about "daytime running lights" from 1990 on, but although I have a 1990, it has 1989 construction including a carburetor instead of fuel injection, the engine plate indicates 1989, and separate buttons for the headlights instead of a combo switch. I wish I had just left the headlights alone. But it turned out that even re-wired back to the original the halogen headlight would only activate on "flash". Apparently there would be no way around dealing with it.

   Next I removed the entire shifter cable, which I found could be pulled right out of the cab into the hood compartment through its little hole, fat sections and all. There was no way to get all the stuff off it, and it came to me that I'd probably be better off making a whole new cable, the way I wanted it. The next morning, the 3rd, a better idea came to me: cut the outer end off and keep the whole cable with the inner end. That way the inner end was already done. I cut it with a zip disk, which went through in 2-3 seconds. I found I could salvage the crimped-on end bits after all - the rod (if I could get the old cable out and crimp in the new end) and the rubber boots and outer clamp pieces. One part of the puzzle was coming together, except that I couldn't run it through the firewall where I wanted because there were obstructions there. Well then, perhaps I could make it pull a pivoting lever that pulled the tensioning rope? That should work.

   That was as far as I got when the first new idea for an improved reluctance motor lured me away.


Other "Green" Electric Equipment Projects
Solar Hot Water Heater

   On the 8th I thought again about the solar hot water (TE News #88). In a house where people are living, it's the best way to reduce the electricity bill. Solar electric is great, but the panels are only (say) 15% efficient at converting sunlight into electricity. So to gain the heating power of a 32 square foot solar direct water heater - call it 60% efficient(?), 128 square feet of solar electric PV panel is required -- 7 or 8 solar PV panels. give or take a couple. (or it might be less as they seem to be getting more efficient.) A 2-2/3' x 12' solar hot water collector can go just above the eaves, below where four of the existing PV collectors are mounted, and without producing electricity itself, do more to cut the regular electricity bill than all of them.

   There were the plastic swimming pool collectors I bought a couple of years ago, thinking to make a quick and easy pumped system. But I really prefer the simplicity and automatic operation of thermosyphoning - water flow by convection. In spite of needing large pipes that can take city pressure at hot temperatures, and having to be drained down (or the collector to be heated) in winter, a thermosyphoning system has its own savings in parts and headaches. And the tank can easily be connected to the woodstove for winter water heating, which can save even more than the solar, and in the winter when the water coming in is coldest and there's no sun.

   So I decided I'd try to put a 'proper' copper collector panel together, similar in size and shape to the one I made in 1979. Since copper pipe is rather costly I went to Ellice Recycling to see what they had. I wanted 2-1/2' pipes, but they had a lot that were just 2' long. I had a few thin aluminum fins that clipped onto 1/2" copper pipe and could buy more, to paint black and form the collection surface. They are also 2' long. And they are 4.75" wide, so I'll need about 30 of these 2' pipes soldered between two 3/4" diameter, 12' long header pipes.
   Then I figured that I could and should have sun collection fins on the top and bottom header pipes as well as on the riser pipes, and that the fins above the top pipe and below the bottom could add about 3+3=6" to the height. That would give the desired 2'8" collector panel height. (2-2/3' x 12' = 32 square feet.) I might even have to make the fins 2" or 2.5" rather than 3". And I was pretty sure I'd have to make them, that they are only available (at least at the one store I know of that has any at all) for 1/2" pipe.
   I've come to hate fiberglass with a passion. I decided to look for some alternative insulation that would withstand the somewhat high temperatures in the collector under the heat fins and copper piping. And what for the solar preheated hot water tank? These days, hot water tanks usually have expanded foam insulation except right around the electrical connections, where you're left to deal with... ugh, fiberglass! Perhaps I could do it like the Peltier module fridge: use 2" extruded expanded polystyrene foam construction sheets, cut to fit around the tank, and 'glue' them together and more or less seal them to the tank with a can or two of spray foam. And some packaging tape to hold it all together while the stuff hardens. I don't think it needs an outside shell unless it's to keep rats from chewing on it in the attic. Hmm... maybe (again like the fridge) some naugahyde wrapped around it. Foam pipe insulation is of course readily available for the pipes.

  On the 22nd I bought some more copper pipe. I weighed some, and discovered that brand new pipe wasn't so much more than salvaged, when the salvage price was 5$/pound. (I can't believe they get anything like that much when they ship it out.) But I decided to use 1" header pipes. These were about 55$ each before taxes for two, and the bill (with a pipe cutter and 4 fittings) came to over 300$ at the plumbing supply. Now I remember why alternatives to copper are ardently sought. But it's the best for a collector (and probably woodstove connected) system that might get quite hot.
   An opposite problem is cold. I'll circulate house water through the collector and into the tank to keep the plumbing simple. There will of course be valves but it's a pain to drain the collector, high up on the roof with valves in the unfinished attic, for the winter. I keep thinking a small heating element in the collector's lower header pipe, with a thermostat set just above freezing, would cause the water to circulate to the tank and prevent everything from freezing.

   On the 28th I drilled thirty 5/8" holes in one of the 1" pipes to silver solder the 1/2" riser pipes to. Was I really going to do 60 silver solder joints, each with the chance of a leak, for just two foot long risers? That seemed like an awful lot of heating and soldering for such a small bits of pipe. Maybe I should make it 6' x 6' - cut the soldering in half and have more collector area? But that would interfere with the thermosyphoning unless the tank was mounted horizontally near the peak of the roof. Or 4' x 8'? I decided not to decide and left it sitting.
   On the 31st I decided the riser pipes would be 3' and the collector 8' wide. A 4' x 8' piece of clear plastic (lexan?) would cover the top with a bit of overhang. I tried to fit the pipes and they didn't go in the holes easily. But if they weren't a close fit, they couldn't be sliver soldered. I wished I had had something other than a twist drill, which doesn't make very exact holes in thinner material like pipe walls. Then I tried to solder in a pipe. While the silver solder fell off the wire and beaded up, I couldn't seem to get the copper hot enough for it to flow with the propane swirljet torch. Of course, there's a big copper pipe carrying off the heat, and the slower the heat, the more it spreads out without getting the heated point up to the required temperature. Next try will be with a scary old naphtha gas torch someone gave me. It really puts out heat.
Peltier Module / Thermoelectric Cooler Experiment

   When I returned from camping I thought to test the cooler. It obviously wasn't working as well as when it was new. Originally it could get down to 5 or 6°C. This was doubtless because I had taken it apart to "reverse engineer" it and I had tried various things with it. I hadn't then planned on ever using it, and while I kept it and threw the parts back inside, I wasn't as careful as I might have been. I couldn't find the original aluminum cold transmitting block, and I probably hadn't put it back together as well as the manufacturer. After all this time (and being unable to find its plug-in power supply), it occurred to me I could run the cooler from the lab power supply, and also that in doing that I could try out different voltages to see what results they gave, and see if they were in line with my complaint that peltier modules should be run in their most efficient range, like 8 to 10 volts, instead of 12 to 14 volts, for a 15 volt rated module.
   The converse of that of course is that for the ubiquitous power supply voltage of 12 to 14 volts, one should use a peltier module rated between about 18 to 24 volts (around 152 to 203 thermocouples instead of the ubiquitous 127) to get higher efficiency operation, at least for camping coolers cooling by around 20 to 30°C or less. Why is it such modules hardly seem to exist, while 127 element modules are "a dime a dozen"? This has been a source of frustration almost from the start of experimenting with them.
   A related observation is that with the solar collector putting out nearly 14 volts all day, the 15 or 16 volt rated peltier modules in the peltier fridge are being driven pretty near their upper limit, which probably has a lot to do with the short service life I've been finding. That's just one more reason not to drive them so hard.

   From the 23rd on, I left the cooler running at each voltage for at least several hours, usually a day or more, for the temperature (taken at the floor of the cooler) to stabilize. I got the following results.
Date/Time
 Volts
 Amps
(Approx)
 Watts
Consumed
(Approx)
 Cooler Bottom
Temperature
 Room
Temperature

August 24th
 7.0
 2.5
 14
 12.3
 23.5

August 25th
 8.0
 2.5 (?)
 20
 11.2
 23.4

August 26th
 9.0
 3.3
 30
 9.5
 ~23

August 26th (from 9AM, readings 3PM)
 10.0
 3.7
 37
 9.8
 ~23

August 26th (from 3PM, readings 9PM)
 11.1
 4.2
 47
 9.9
 ~23

August 27th
 12.0
 4.5
 54
 10.7
 ~23

August 28th
 9.0
  - (not read)
  - (30?)
 10.9
 ~23.5

August 30th
 10.1
 3.7
 37
 9.6
 ~22 (cooler)


   Everything seemed great up to the 27th. It seemed to be proving my point quite well. Then I set it back to 9 volts from 12. But the temperature didn't go back down to 9.5°. Instead it went up to 11°. The reason for the change was likely moisture build-up inside the cooling unit, and probably frost buildup on the peltier since the heatsinking didn't seem very effective. The next day I opened the cooler, and sure enough a puddle of water drained out of the cooling unit. So the tests probably were showing as much the moisture conditions as the intended voltage versus cooling relationship. I turned it off. I should probably turn the cooler off for a day to warm up and dry out before trying each new voltage, and then take the readings after a specific number of hours, say 6 or 8 - as if I could remember to return at the right times - and just when it had been turned on. On the 29th I tried 10 volts in the aired-out cooler. But it didn't seem to improve it.
   It then occurred to me I could also try out the thermoelectric fridge with the power supply at different voltages. That will be a better test than the poorly functioning cooler. And then I could try out the large, 14 amp (but still 15 volt max - arg!) modules at reduced voltages. Somewhere in there there should be more cooling capacity at the same or lower power. Assuming I get such results, the obvious next step (unless some better configurations of peltier modules at a reasonable price make an appearance) is to make a DC to DC power supply that takes the 12-14 volts from the solar collector and reduces it to the optimum voltage, probably between 8 and 10 volts. Preferably a switching supply of over 80% efficiency. (Let's see... getting 9V from 13.5V with a linear regulator would be 67%. 10V from 12V would be 83%... the linear regulator is probably good enough since there's lots of power when the sun is out.) But it's all unnecessary losses - a peltier module optimized for efficient operation at 12-14 volts and for fridge/cooler temperatures (ie with ~160-210 thermocouples instead of 127) would be better.
   I can make a DC to DC switching power supply for dropping voltages if I run out of other things to do. Including a model for LED lighting. Meanwhile... 3 or 4 ten watt resistors could drop the voltage to the solar fridge?

http://www.TurquoiseEnergy.com
Victoria BC Canada