Turquoise Energy Newsletter #141 - February 2020
Turquoise Energy News #141
covering February 2020 (Posted March 7th 2020)
Lawnhill BC Canada - by Craig Carmichael

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

Month In "Brief" (Project Summaries etc.)
 - Carrying Heat Uphill: A way to look at heat pumping - Open Loop Air Heat Pumping Experiments - Another Higher COP Device - Nickel-manganate/Zinc Battery Developments - High Efficiency Titanium Dioxide Solar Cells and "Pebble" Surface Cover Glass - NiMH D Cell Refill

In Passing (Miscellaneous topics, editorial comments & opinionated rants)
 - Permian Period Amphibian Evolution (continued) - Small Thots - ESD

- Detailed Project Reports -
Electric Transport - Electric Hubcap Motor Systems (No Reports)

Other "Green" Electric Equipment Projects
* End of the Solar Hot Water Tank (the mineral rich water stank) - Hot Water Tank Protective Anodes
* Very High COP Open Loop Air Heat Pumping - With Water Cooler Pump - Means for Improvement? - New Heating Unit - Compressors Today Are Not Made For Air Heat Pumping (& compressor design ideas) - Value of the outdoor heat exchanger - Install & test of that - A Better Way to join PVC Pipes to Copper Pipes (no pipe fittings) - The Grand Potential - Additional Heat Pumping/Air Compressing Info Gleaned From Web

Electricity Generation
* My Solar Power System: - Monthly Solar Production log et cetera - Notes. - Analysis of a Whole Year's Figures

Electricity Storage
* Turquoise Battery Project (NiMnOx-Zn in Mixed Alkaline Salt electrolyte)
 - Powder Electrodes Conductivity - Solid Porous Electrode Briquettes; Plaster - Some theory/ideas - 50x50mm Electrode Compactor - tests - Cylindrical Cells? - Okay, How About Flat With "Rebars"? - Cupro-Nickel Current Collector

February in Brief

   February is a short month, but somehow this is anything but a short newsletter, and it has taken me some time to complete it.

   This month I continued experimenting with the open loop air heat pumping (hoping to save myself electricity and money and have a warmer kitchen as well as it seeming to be an exciting and highly valuable project anyway).
   And I did some more work here and there on the new chemie batteries. They're getting close! (...How many times have I thought that and had to stop and try a new tack?)

   Aside from that I did just a couple of little things - some paint and plastic battery holders - for the ground effect vehicle model. Its two powerful ducted fan motors need about 2750 grams of lithium batteries to get full power, almost doubling the entire weight of the model.
   There was no pipe into which two rows of the lithium cells fit, or even one row fit nicely. I used pieces of the '1.25 inch sprinkler system' PVC pipes, slit them open and then shaped them by softening them in the oven.

   And of course I did a little garden raking/weeding/cultivating to prep the soil for spring. The 2019 crop failures were staggering all over the world, shipping is way down, and a few shortages have been reported in the USA with some large groceries starting to limit the amounts of some items like flour, sugar and rice, albeit to "5 bags" - more than most of us would buy anyway... so far. "Ice Age Farmer" on youtube said hardly anyone he talked to was taking the coronavirus threat seriously, so if just 3% of the population trying to stock up was causing shortages, imagine what will be left when everyone starts trying. So I want to grow more. At least I get good potatoes and most of my fresh vegetables. Maybe it's time to build that chicken enclosure?

The Fake and the Real McCoy:
The four prong implement on the right/top has been
my favorite gardening implement since I bought
it around 1980. I even put a new handle on it
when the old one rotted off. It's a rake, hoe and
cultivator, and one can pick up weeds with it,
shake off the dirt, and dump them in a bucket.
There's less bending down, and when there is, it's a
loose weed, not a tug and break the root off.
The tines have never bent or broken. Maybe
a good tool to wield on a CNC gardening machine?
The imposter on the left/bottom is less useful, so
it just sits in the garage and its original handle
is still pristine.

   Daylight started returning over the month and the solar panels came at least somewhat back to life as the monthly tallies show. But tops was under 6 KWH per day. March 6th and 7th made a dramatic rise to about 9 KWH - it was at last sunny (except for a couple of small - ugh - snow flurries), and although it's still quite cold, the sun is getting higher and the days are lengthening.
   The end of February marks a year of logging solar PV collection daily performance, and I made some notes on the annual figures obtained. Overall the sun made 22% of my power, but it was 35% in the seven best months. As expected the winter months were very low collection, with high electricity use for heating. Even with the woodstove doing most of it, 500 watts of heat in the trailer to keep it from going mouldy added a big chunk to the total, as did bedroom heat at night. And the electric car added some to the bill in all months.

Carrying Heat Uphill: A way to look at heat pumping?

   Here's a visualization if we need to take some of the seeming magic out of the workings of a "conventional" refrigerant based heat pump: There's a hill. It's not as high (as warm) as the occupants want it. Instead of getting new dirt from elsewhere to make the hill taller, they shovel dirt from the bottom of the hill and bring it to the top. The work done by a heat pump may be seen as similar to the work being done to bring the dirt up the hill. There's dirt (thermal energy) everywhere, but to make the top of the hill comfortably tall (warm), work must be done. But it's less work and expense to "concentrate" existing, nearby dirt into a hill than to get all-new dirt delivered direct from the distant quarry (power company) to make a hill.
   Now, obviously the higher the ground is around the hill (the less cold it is outside), the less work it is to bring up the additional dirt, and higher COP ('coefficient of performance') is obtained, and vise-versa: if the hill towers above the frozen countryside, much more work has to be done to carry it up and the equipment has more trouble doing it, too. The work increases, but the COP gets lower.
   The analogy is incomplete without noting that the sides of the hill are continuously slumping off into the surroundings - the heat gradually leaks out of the building, so more and more dirt has to be brought up to keep the hill from shrinking until it's level with the surroundings.

   The analogy rather breaks down with the open loop air heat pump. Here the compressor brings dirt from near the top of the hill and piles it up a little taller in one spot. From there it spreads around pretty much to where it came from. It doesn't take much work to do that, so the COP is (in principle) very high.
   But this misses the open loop and outdoor parts of the system: the 'spent dirt' (room temperature compressed air) from the radiator pipes exits to the outside, goes down the hill, but through the outdoor heat exchanger. The exchanger is like a ski lift circuit. As the dirt goes down, and ideally gets carried down into a valley below the base of the hill (is refrigerated below outdoor temperature as it is released and expands), its weight carries an equal weight of fresh dirt up from the base of the hill virtually for free, passively, effortlessly, regardless of the height of the hill. So the compressor gets its fresh dirt from the top of the ski lift instead of from the bottom of the hill and only has to lift it a little higher.
   A house still needs more heating as outdoor temperature drops, but here the COP in principle remains the same - potentially very high - or at least drops only a little.

Open Loop Air Heat Pumping Experiments

   First I tried out the small water cooler (fridge type) compressor I got February 1st in the same piping setup as last month's experiments. It seemed to work about as well as the lab vacuum/compressor pump - not a real improvement. But a coefficient of performance ("COP") of 4 or 5 is the maximum today's heap pumps can manage, and that in weather well above freezing, where this was only a little above 0°C. Obtaining roughly those levels of COP at that outdoor temperature would seem to make the technology better even with my jury-rigged experimental equipment. And I'm sure the potential is there to do much better yet.

   The next week I got another refrigerator pump from the refuse station, this time a bigger one from a large fridge. (I had to make a larger box.)
   On the same trip, Mike gave me two four-foot long finned copper pipes. These would save me from having to make more copper finned pipes before I could do more experiments.
   The compressor blew lots more air than the previous one and the power consumed went up with the output pressure, from 100 to 150 watts, instead of staying pretty much stationary. I had high hopes. I had to make a bigger box for it, and I made a new radiator piping loop across the dining area and out through a wall. Not wanting to make a big hole in the wall, I dispensed with the outdoor heat exchanger.
   I was disappointed that this one didn't seem to work any better than the previous, if as well. It was drawing more power (120 watts instead of 75) and still only heating as well as the 420 watt radiant heater - if that well. That suggested a COP of 3 to 4, tops.

   It seemed that perhaps the outdoor heat exchanger was more effective and more necessary than I had thought. I soon installed one on the side of the house.

Outdoor heat exchanger: Tall box on the outside wall (same box), and duct and pipe
through wall into house for incoming (warmed) and outgoing (still compressed) air.

   It seemed obvious that air compressors weren't being made with heat pumping in mind. I thought about what an ideal air compressor for heat pumping would be like.

   Then I cut the first (water cooler) compressor apart with an angle grinder and a hacksaw. It was immediately apparent that there was a major bottleneck to the air flow, in the form of a 16 inch long copper pipe of only 1/16" inside diameter. All my efforts to make the airflow as free as possible once the air was out of the unit were annulled by that.
   I improved it by cutting off that pipe and putting in a 1/8" one at the compressor body. That's 4 times the cross section -- and I only made that pipe 1/2 inch long before it "telescoped" out into a bigger one.

   Then I (finally) had a look at the big new Makita air compressor. If it wasn't made for the job, at least it seemed closer than anything else I had so far. It drew 550 watts with no air compression (drain valve open) which increased to 1100+ watts at 120 PSI. (Does that make it 50% efficient at compressing air? Anyway it indicates at least 550 watts of energy (and probably more) is going into actually pumping and compressing air, above and beyond simply spinning the motor and oscillating the piston.) The motor's fan also blew air under a shroud to cool the compressor, which didn't get too hot while I had it running. If I made a large box and put it inside, where a fan would blow its heat into the heating duct (as with the fridge compressors), and attached the radiator pipe directly to the bottom of the air tank, it should work better than any other other unit I had.

   But it had so much power that it (surely) needed much more radiator pipe, and surely I would have to install the outdoor heat exchanger. And it needed a large new box with a fan on the side. I wouldn't finish all that in February. So I tried out my "improved" water cooler compressor first to see how it worked and what lessons might be gained. It already had a box with a fan and the fittings. All it needed was some pipe fittings and lots of PVC ("electrical") tape to close up the housing. It should last long enough for an experiment of two. If eventually the oil leaked out and it seized up... oh well. It was free.

   I thought about the amount of power being used by these various compressors. Ideally, if there was no radiator pipe or if the end of the pipe was wide open, and no pressure would build up in the system, the compressor is just blowing air like a fan. The duct fans were drawing about 3-1/2 watts. Running with open ducts the compressors started at 75 watts and went up from there to 750. (750 - one horsepower - was for the Michelin. The larger Makita was a little better at 550 W.)
   This means that all the compressors are wasting huge amounts of energy. Of course it takes some energy to turn a motor, and to slide piston rings back and forth in a cylinder. But that should be small, almost trivial compared to the energy needed to compress air, where real work is being done.
   One can literally feel the truth of this when pumping up a bicycle tire with a hand pump. Disconnected or at no pressure, the lightweight pump handle, shaft and piston assembly plunges freely, almost effortlessly. It may slide down the cylinder just from gravity. (Of course the more pressure there is in the tire, the harder it is to push the handle to the end of the stroke to pump it up further because then it's doing actual work.) Why do electrically powered compressors use so much power - or even any significant amount of power - in free air?
   One can see this is at least partly owing to internal restrictions to the air flow: they are compressing air, but it is decompressing before it even comes out of the compressor. Nowhere was this more obvious than in the 1/16th inch, 16 inch long pipe in the water cooler compressor. It was heavily compressing air in the cylinder and making heat there regardless of whether the air outlet was wide open or almost completely closed off. And thus the power consumed (75 W) didn't change much with external load, since most of the load was internal. (After I put a larger pipe in, it used as little as 55 watts pumping 20 PSI; 50 watts at 0 PSI.)
   In a compressor designed for heat pumping, designed for optimum efficiency, I would hope that the watts consumed would go up with pressure by a factor of 5 to 10 when pumping to it operating pressure. and also that much of the heat would go out in the air being pumped. The heat starts in the compressed air. The "bicycle pump" air compressor with effective (copper) radiator fins and a one-way valve would also be the start of the radiator piping inside the opening of the duct, a continuous solid copper finned pipe with no obstructions. (I am presently thinking of how to make a "rotary" compressor using a screw mechanism of some sort to move a rather large piston with a long stroke in and out. That should have less vibration and noise than one with a short crank arm flying back and forth.)

   In a test at the end of the month the outdoor heat exchanger's performance was disappointing, raising the outdoor temperature air barely half way to room temperature. It may need to be made substantially longer to gain more heat exchange area between the outgoing warm air and the incoming cold air. Or it needs to exchange better over the same length. However, the performance was better than without a heat exchanger. It really does seem to be a key component of the system and I'm sure a better one will raise the COP. (I improved it. It helped.)
   On March 3rd I put together a configuration with the larger fridge pump. The warm air came out the end of the heating duct 4 to 5 degrees higher than the room temperature instead of 2 to 3. Drawing 115 watts it seemed to heat a little better than the 420 watts radiant heater, suggesting a COP of at least 4. So the COP wasn't higher but as the compressor was more powerful it was pumping more watts of heat. But the temperatures at night weren't much above freezing and it was quite inadequate for the amount of heat needed. Time for the bigger compressor!

As March started I had a pretty good setup with the large fridge compressor in a box with a fan,
finned pipes and a duct with air blowing through it, and the outdoor heat exchanger, improved with
some thermal breaks in the copper pipes and some reductions to the incoming air flow space.
(Here including a piece of foam partly blocking the inner opening in a strong wind.)

   When efforts are made to create really high efficiency equipment for air heat pumping, we will start to see the ultra-high COPs dangling in front of us.

Another Higher COP Device

   A reader sent me a link to another interesting device that uses similar principles to get more energy performance out of heating and cooling. This one is used to desalinate sea water. As well as making drinking water at sea, this process will doubtless someday be highly valuable for helping to turn deserts into fertile lands.
   As best I understand the operation, the initial heat comes from the sun to vaporize sea water. But instead of simply letting the evaporated water cool to re-condense, its heat in cooling and condensing is used to heat more water up to near vaporization temperature. Then (if I have this right), still pretty hot, it warms the next 'batch' of water part way. This is done, apparently in several steps, making it akin to my outdoor heat exchanger using the heat of the outgoing air to warm incoming air for the compressor. The water will be cool for drinking sooner, too.


February 10, 2020 Mark Hutchins

MIT scientists have developed a solar desalinator which transports heat from the sun through a ten-stage process of evaporation and condensation. The group estimates a $100 device employing their innovation could provide the daily drinking water needs of a family.

Scientists at the Massachusetts Institute of Technology (MIT) have developed a prototype solar-powered water desalinator which they say achieved solar-to-vapor efficiency of 385% through a multi-stage process where the heat released as water condensed was recycled, flowing into the next layer to power the next stage of evaporation.

Rather than using photovoltaics to power electrically-driven desalination - a method which has been used in large scale applications already - MIT's process uses solar absorbers to gather heat from the sun and evaporate the saltwater.

A prototype on an MIT rooftop delivered water which exceeded local drinking water standards at a rate of 5.78 liters per hour, per square meter of solar collecting area. The university said that was more than double the previous record for water produced by passive solar desalination. By optimizing and adding further stages to the desalination process, the group estimates devices based on the concept could reach efficiencies as high as 800% - meaning that eight times as much energy as is initially collected from the sun would be available for the conversion of water into vapor.

Nickel-manganate/Zinc Batteries

   An enthusiast with casting equipment in Oregon is making copper sheet electrodes with cast zinc on one face. This is likely better than electroplating on the zinc. With the degrading zinc tabs and after the last very thin zinc sheet got holes in it, I think the copper sheet is needed to make a solid, changeless base for the zinc which oxidizes and reforms with discharging and charging.
   He is sending me samples.

   I also found that in pH 13 electrolyte (even with salt in it) cupro-nickel doesn't seem to degrade in the positive electrode, so it can be used instead of graphite for the current collector. That will be better - yay! And monel or cupro-nickel powder can probably be used for conductivity additive in the electrode. That too may be better than graphite or 'acetylene black' powder. Or better in conjunction with it.

   I designed a 50x50 mm flat cell case that would hold the positive electrode compacted by having some glued support points in the middle of a "porous" plastic separator as well as a glued-down outside edge. (The nickel-manganate electrode is installed under the porous grid with its tab sticking out the slot; the zinc electrode goes above it. The sides will be a little taller, with an upper slot, an a cover will be glued on top.)

   I still worried that a thin ABS grille is too flexible and will let the electrode swell and lose contact with the current collector somewhat between the posts. But on line I found a new 3D printing filament: PVB. It sounds better - stronger and stiffer than either ABS or PLA. (And unlike PVC, its environmentally benign.) Of course I ordered a roll to try it.

   I also made a very exact fit 50x50 mm electrode compactor punch and die from 1/2 inch steel plate. After I put a base on it, it should press good electrodes for my planned 50x50 and (eg) 100x100, 150x150 and 200x200 mm (by ~8 mm) flat cells.

High Efficiency Titanium Dioxide Solar Cells and "Pebble" Surface Cover Glass

[This probably belongs in some "detailed project report", but it's not something I'm actively pursuing at this time.]

   Longtime readers of this newsletter may recall that I did some abortive experiments with dye sensitized titanium dioxide solar cells a decade ago. (eg, TE News #29) I was trying tartrazine (yellow food color) as a dye. It has just the right light absorption spectrum. And I developed a special glaze with a high refractive index (nanocrystalline titanium dioxide borosilicate glaze). It could be ground up into a "frit" (sandy grains) and melted into the front surface of solar panel glass as little "pebbly" grains, tiny lenses that would aim angled light more straight into the glass and hence increase the power gained from light coming from steep angles and scattered sky light. I didn't get that far.

   Really this was two separate solar cell projects both using titanium dioxide and I confused myself as to what I was doing and what my objectives were. At the time I didn't even have any solar panels. If I had known where to buy "raw" solar cells, I could have tried out the pebbly cover glass idea separately with them, but at the time I didn't, and my TiO2 cells weren't working. And it seemed more and more like a lengthy diversion - wasn't I doing electric transport?
   And as to titanium dioxide solar cells I thought, "So what, even if they work?" I read about the almost "problematic" tendency of titanium dioxide based solar cells with liquid electrolyte to leak. Could I solve that? And I had no shortage of other projects. I threw up my hands for the time being, but I never got back to either project.

   Now some researchers have come out with "24% efficient" titanium dioxide solar cells, a higher efficiency than today's best silicon based cells and much higher than those being made at the time I was experimenting. I hadn't realized the potential of TiO2 for actually being significantly better than silicon, or I might have continued. Probably a non-liquid electrolyte that would work could be (or has been) found, too.

First: Cover Glass (In)Efficiency:

   In an article in PV Magazine, researchers studied the effects of creating microfractures in glass to assess how they deteriorated panel performance. It shows that the surface of the glass is important.


By deliberately stressing glass to simulate years of wear and weather exposure, and with all other factors being the same:

"The Turkish team ascertained reduction in both exergy and energy efficiencies were attributable to a large number of micro-cracks and deformations on the glass surfaces, which were responsible for considerably affecting the absorbance, transmittance and reflectance properties of the materials." [my italics]

    Those little pits and cracks, causing reduced performance, appear to affect performance pretty much the opposite of what I hoped to achieve with the convex 'pebbly' surface. Those little lenses would aim angled light more directly onto the active collector surface, whereas pits would do the opposite and cracks would also scatter the light.

   Well, I still have the formula and method: TE News #29, June 2010, "Glaze Mix 9". I also still have no shortage of other projects. Solar panels are a big industry. It would be one more exciting thing that could be developed and commercialized if there were some more people in on the act.

Second: 24% Efficient Dye-Sensitized Titanium Dioxide Solar Cells


   In 2010 silicon solar panels were maybe 12% efficient. Now they have hit and passed 20% efficiency in the best new panels. But if a solid electrolyte is or has been found, dye sensitized TiO2 panels might still be cheaper and easier to make and are said to be more environmentally friendly.

   With higher efficiency, fewer solar panels need to be installed, reducing the materials and installation costs for putting in solar energy. A typical house roof that might today be covered in about as many panels as it can hold might be only 2/3 full instead, or if filled would supply more energy.

The Australian research team which developed the device said the higher efficiency was achieved through a nanowire design which eliminates the interface inside the titanium dioxide band. ['interface': usually a transparent, conductive tin-antimony oxide "current collector" layer behind the glass - but in front of the TiO2. In my trials I was having trouble getting a good layer.]

February 13, 2020 Emiliano Bellini

Titanium dioxide forms the basis of the cell, with efficiency lifted by a nanowire structure.

A delicious looking dish of titanium dioxide nanopowder accompanies the article

Scientists at Australia's Queensland University of Technology have developed a quantum dot, titanium dioxide (TiO2) solar cell they claim offers better efficiency more cheaply than traditional crystalline silicon cells, as well as being more eco-friendly.

The researchers claim the cell boasts 24% efficiency, more than double the 8-11% lab-level performance observed in standard TiO2 quantum dot devices.

Some photons become trapped in the interface between nanocrystals in typical TiO2 quantum dot cells but the Queensland team claim to have removed the problematic interfaces.

"Our nanowire design eliminates the interface inside the TiO2 band, as it's just a single layer of QD [quantum dot] -coated TiO2," said research coordinator Ziqi Sun. "If we can remove this disordered interface, we can improve efficiency."

Commercial production

The nanowire crystal used for the cell was assembled in China with an advanced transmission electron microscope.

The research group said their cell concept could be expanded to bring the technology to production.

Most dye-sensitized - Grätzel - solar cells are based on titanium dioxide thin film - a cheap and harmless, water-insoluble inorganic material which is commercially available and widely used in industrial applications.

Low efficiency has remained the chief obstacle preventing such materials competing with crystalline silicon cells to date.

Titanium has also been used recently to raise the efficiency of perovskite cells. Research projects of that nature have been carried out by Japan's Kanazawa University and scientists from Russia's National University of Science and Technology and Rome's Tor Vergata University.

   And maybe some solid electrolyte can be (or is being) employed to eliminate the leakage problem. There would be a pretty rapid switch to this type of panel if it was available, cheaper, reliable and anything like 24% efficient. (What efficiency might I have achieved if I had kept working on it? These people have a much better idea of what they're doing and some fancy equipment. But I still suspect tartrazine would be the best dye. Might it boost efficiency even farther?)

NiMH D Cell Refill

Finally, here is a picture of the long dead nickel-metal hydride battery I cut the top button off of, drilled through a rubber seal beneath it, and refilled with some distilled water (amount not measured). (as mentioned at the end of January in the previous issue) It then took a charge - but didn't seem to hold it overnight. Some cycling might or might not help, but I didn't get around to doing anything more with it.
   NiMH cells are known to corrode if their voltage sits at less than 1.0 volts. Since this one had sat for probably over two years completely dead and discharged, one suspects that refilling cells that aren't completely dead yet, but which have become reduced in capacity owing to drying out inside, would probably work better - perhaps even restore them completely.

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

Permian Period Amphibian Evolution (continued)

* It seems the earliest fossil eggs that have been found were from the Triassic period (and the earliest egg shells are from the Jurassic - so Triassic eggs must have been leathery covered?). This has puzzled paleontologists: "One of the fossil record's most puzzling features is the absence of preserved eggs or eggshell of the first third of the known 315 million year history of amniote evolution."

   This is in fact excellent further support for my theory that there were no reptiles before the end of the Permian period. I believe that the creatures on land until the Permian extinction were actually (adult) amphibians (not amniotes) or an intermediate stage, "pre-reptiles", which have long been misidentified from the meager fossil remains as reptiles (or even "mammal-like reptiles", their anatomy being atypical for reptiles). The present view that reptiles - advanced land life - crawled out of the sea almost contemporaneously with amphibians - fish that later breathe air in adulthood - in the mid Carboniferous, long before there were seed bearing land plants to support an advanced land ecosystem and even before all the carbon dioxide had been eliminated from the atmosphere (and deposited as coal), on the face of it actually seems pretty far fetched. It probably didn't in earlier times before the great anatomical differences between amphibians and reptiles were well appreciated, and in that ignorance the error was originated. (See TE News #100)

Small Thots

* I had my hair cut short, virtually shaved off, and I discovered perhaps what it is that may be the advantage to that for keeping or perhaps restoring thinning hair: when you rub your scalp, you're often pushing the short bristles into the follicles instead of merely pulling at them. Furthermore, you're moving them around in different directions than with your usual hair style.
   Anyway that's what I've come up with for differences very short hair makes. And I don't know what if anything it has to do with the demodex follicularum mites or any damage they may have caused to the scalp and hair follicles. (It's hard writing a coherent piece on this when the ideas or answers come in so much piecemeal, over months and indeed years.)

* It seems Joe Biden is having a lot of trouble just making coherent sentences these days. Senile dementia has well set in. In choosing this near zombie, under criminal investigation in two countries, as the next "Democratic" party presidential candidate, the corrupt body politic is clearly showing it wants only a figurehead in the highest office, someone who will obligingly rubber stamp whatever they've decided on out of public view. They've had far too much trouble getting that out of the current occupant, and they've had a lot of trouble getting some of the other candidates with their own programs and ideas off the stage. (Not with the reasonable, honest candidates with sound ideals and sane ideas. Those are easily disposed of. Tulsi who? She got the mic? Well, just cut to a political commentator while she's speaking.)

(Eccentric Silliness Department)

* When is Obsolete Obsolete? It's been 2000 years since people were using plumbium (lead) for water pipes. Why do we still call them "plumbers"? Enough is enough already! Shouldn't they be called "copperers" or "pexers" or something?

* Stragedy: A dyslexic's cunning plan that ends up very badly.

* People in China are spraying disinfectant on bank notes, or boiling them. As the Wuhan corona virus spreads, people all over the world will soon be laundering money.

* "I must be an athlete... I have an athlete's build... athlete's reflexes... athlete's foot."

* Lettuce not beet about the blueberry bush. If we really carrot to help our fellow human beans, we would have mustard up the curry to chervilly collard up the scallions in our legislatures and insist that they turnip their efforts for those with who have been uprooted, those sorrel cases who have leeked through the cracks, to help them ketchup with the rest of us. Then there can be peas for olive us. (Vegetable stew for dinner?)

* Biologists know that nothing in life is unavoidable except death and taxonomy.

* The Americans invaded Syria and Afghanistan and have long been threatening Iran. I believe I have finally figured out what they're really after: Pistachios! They are native to these countries, and especially Iran. If they can dominate them like they do Latin America, and turn them into "pistachio republics", then these delicious but costly nuts can be much cheaper in the USA. No wonder they have wanted Iran so badly for so long!

   "in depth reports" for each project are below. I hope they may be useful to anyone who wants to get into a similar project, to glean ideas for how something might be done, as well as things that might have been tried, or just thought of and not tried... and even of how not to do something - why it didn't work or proved impractical. Sometimes they set out inventive thoughts almost as they occur - and are the actual organization and elaboration in writing of those thoughts. They are thus partly a diary and are not extensively proof-read for literary perfection, consistency, completeness and elimination of duplications before publication. I hope they add to the body of wisdom for other researchers and developers to help them find more productive paths and avoid potential pitfalls and dead ends.

Electric Transport
no reports

Other "Green" Electric Equipment Projects

End of Kitchen Solar Hot Water Tank?

(8th) With my mineral-rich well water, for whatever reason, once the water had sat in the tank under the kitchen sink for a day or two, it started to stink. A sulfur smell permeated the air near the sink whenever hot water was run. Or maybe it was straight hydrogen. I am puzzled as to why this hot water tank would smell when the large tank has always been fine, however long the water sits in it.

   Other than that I had been very happy with the 15 liter, 36 volt water tank. I had hot water in the sink in 2 or 3 seconds, as hot as I cared to set the thermostat to. But the stench was too much, and I couldn't figure out why it was there or how to get rid of it. (Now I think some mineral was causing the water to react with the tank metal, to oxidize it and bubble hydrogen. There was a lot of black sludge in the tank when I removed and drained it. Or maybe it formed hydrogen sulfide. Apparently it's a very common problem where there is mineral rich water.) I just couldn't run a whole tankful of hot water every day to keep it fresh. Even keeping the tank cold didn't entirely solve the problem, tho it seemed to slow it down.

   Finally I ordered a 3KW "on demand" water heater. This huge draw just while the water is running - the total of all my solar panels at noon on a sunny summer day - is the opposite of what one needs with off-grid type solar power - something that can collect heat gradually all day so it's there to be suddenly drained when you want it. Well, I'm not off-grid as long as there's a grid.
   Now I finally got around to installing it. I picked 3 KW, the smallest power available, because there was an electric heater under the sink at the floor (which I had never used - I only finally turned it on to find out which circuit breaker it was on.) and I could (and did) take the 230 volt #14 AWG cable from that to use for the water heater instead of having to run new, heavier wire all the way from the breaker box.

   I am not enamored with this unit. One must run the water quite slowly to get it warm, it takes 20 or 30 seconds before hot water is coming out, and I would never really call it "dishwashing hot" except at the lowest dribble of flow rates. Well, it did say it was for a "lavatory sink".
   One major impediment was that the outlet to the hot water faucet was at the bottom left facing down, and I had to put in a 16 inch long "U" of copper pipe with several odd fittings I had (a frustrating piece of plumbing) to aim it up and bring it close enough to reach the already long (18"?) flex pipe to the faucet, which then had a further 18" spout on it (a tall upside down "J" spout). What was especially galling was that the hot water came out of the heating element at the top of the unit, and there was a foot-long pipe inside bringing it down to the fitting in the most inconvenient possible spot, the bottom left. Since the hot water lever is usually the one on the left, in order for the heater to be clear of the cold water pipes it will normally be mounted on the wall under the left side of the sink. Thus the shortest run in most situations is obviously from the top right, and just as obviously, with the very limited flow and "on demand" heating, the run length is very important. I suppose they had leaks and electrical safety in mind, but after installing and using it I can see that the setup is ridiculous. They should have simply brought the output pipe outside the case to the top right (even closer!) and had the join away from it.
   It all made a total of almost 6 feet of piping between the "on demand" heater right below the sink and the end of the spout. With the tiny flow available to get really hot water, it takes quite a while to actually get hot water. If you run it faster, it never gets very warm. If I had used some skinny tube instead of 1/2 inch pipe it would have helped. But it would have been far better with the hot water outlet at the top on the right to allow the shortest connection to the faucet instead of forcing one to do a lot of extra work to create a very long one. To get decently hot water, one only cracks the tap open a little - and waits, and waits - and then if it isn't on quite high enough, the unit doesn't turn on and the water stays cold and all the waiting is in vain.
   I'm tempted to disconnect everything, take it out, disconnect the fitting, cut most of the pipe off, and try to bend it up to join straight to the flex hose at the top. (I've been doing lots of that sort of thing for the air heat pumping!) That would cut out about 2 feet of piping and so reduce the absurd wait for "on demand" warm or hot water. In fact, then I could put in a shorter flex pipe, too, and make it 2-1/2 feet less.

   Six kilowatts would obviously not by any means be overkill for an "on demand" kitchen sink water heater - I would now judge it as minimal. I think a shower or bath would take more - 12 kilowatts?

   But the water doesn't stink.

Hot Water Tank Protective Anodes

(10th) Now that I've gone to all that work - removed my under-sink tank and installed the "on demand" hot water unit - I see on youtube that there's something called a sacrificial anode in hot water tanks! A guy showed four hot water tanks cut wide open with a zip disk/angle grinder, and there they are: rods sticking down from the top. Apparently they're magnesium or aluminum and by corroding first, they protect the metal tank from rusting! How is it possible I have lived this long, and even having replaced several water tanks, I've never heard of them before? How is it possible people say a tank lasts 8 years (or whatever) and never mentioned that there's an anode inside which eventually wears out and then stops protecting the tank? Or is putting an anode in water tanks a "new" idea that I just never heard about -- like newer than the 1980s when I was (too frequently) replacing hot water tanks in my old house? I'm going to guess that that's the case, since my last water heater lasted 19 years 11 months. (I remember it because of something else at the time I replaced it, December 1996. It sprung a leak in November 2016 before I moved and I was cursing that it hadn't lasted just a few more months until it was somebody else's problem! Probably it had needed a new anode and with one could have lasted more decades yet! No doubt I should check my present main water tank.)

   I'm going to guess that in my mineral rich water the anode wore out quickly and that's when the water started to stink, because I don't remember it smelling until it had been installed for a while. When I removed it I unscrewed a nut attaching a bolt through the shell (whose purpose I had wondered about), and something fell down inside. No doubt it was the anode. In this small tank it wasn't made to unscrew from the shell and pull out the top. I would have to unscrew the heater element to access the inside and really see what's up. Sometime I will get to that...

Simple Air Compression Heat Pumping
- or -
How to Heat a Building Almost for Free

   I read on a "Q & A" discussion list ("Quora") that air heat pumping does work, but "at low efficiency" and needing to pump "quite a high volume of air at high pressure". I understood it was lower efficiency, but with potential Coefficients of Performance (COPs) being so high in the open loop air system one should be able to tolerate quite poor efficiency and still get excellent results.
   And having now looked at and tried some compressors, I think the low efficiency is in the way compressors are being made and not inherent in the whole process. Today's air compressors are simply not designed for heat pumping and they care nothing about design as related to that purpose. Just thinking of the simplicity and efficiency of a hand bicycle pump one can see that it's easily possible to do much better.

(3rd) With Water Cooler (Refrigerator Type) Pump

The water cooler compressor blew up a balloon very slowly

   I unsoldered the inner 'telescoping' narrow pipe from the larger output pipe on the 'new' refrigerator compressor to leave a maximum size opening with least resistance to air flow. It was a hard go, heating virtually to red hot - must have been silver solder. I tried to do the same with the input pipe, but the whole thinner pipe softened and fell off without melting the solder. I cut it off shorter. A last bit of the intake's thinner pipe was then loose inside and came out.

   I note that when I first plugged it in, it sprayed out some oil. Later I turned it sideways and more oil started dripping out the input pipe. So it still had some oil in it. (several fluid ounces as it turned out.) I suppose it will need occasional or even frequent lubrication now that the oil isn't sealed inside with the refrigerant.
   When covered and turned on it also pinned my 0-60 PSI air pressure meter in about one second!

   I went to put it into the heat pump box box and found it was 3/4" too tall. I had to rebuild the box. I filed a pipe with 1/8" NPT threads on one end for a friction fit over the compressor's output pipe. I threaded another fitting I had, to screw onto it and bring it up to 3/8" NPT to fit another adapter to the 1/2" copper water pipe size. (In 35 years I think that's the first time I've ever used that threading tap! But there it was in the set. The trick with NPT threads is that they're tapered. One must taper the end of the pipe slightly with a file, grinder or reamer before trying to thread it. Even lots of pro metal workers don't know that and can't figure out why it won't thread. But it's why pipe thread joints don't leak.)

(4th) I bought a 0-160 PSI air pressure meter. I worried about setting a 4 foot long heavy finned pipe on top of the soft, scrawny copper output pipe on the fridge compressor. It would just bend it and fall over, and probably kink it. I thought the least I could do was put a steady of some sort on the pipe. It would have been easy if the air duct didn't have to fit over the pipe and also into the hole in the box. I cut a chunk of aluminum and rigged something up. Afterward I thought of a better arrangement. I'll probably want to make that later, so there's no force at all pushing on the fridge compressor's little output pipe.

(5th) I changed the outdoor heat exchanger from 2 legs to 4. (Then I had to unscrew 2 to get the height just right for the other pipes.) I put everything together and turned it on at 21:00 PM. Now it wouldn't go above 30 PSI. Why did I buy a new air pressure meter again?

I closed the doors and turned off the LED garden lights (~200W). It started off with the following conditions:

Outdoor air: 2.5° (°C), no wind.
Room air (at kitchen table) 17.6°
Cold air in at floor: 16.0°
Air pressure in pipes: 20 PSI

   This time the compressor was so much quieter that the air fan was most of the noise. It makes quite a whir, but with the lab air pump I could hardy hear it, and not at all with the Michelin compressor. A very slight waft of cool air could be felt coming in the outdoor heat exchanger. Also unlike other tests, the compressor and its output pipe seemed hardly warm. Then I put my hand near the pressure meter and felt a leak at a soldered joint. I wrapped some electrical tape around it. But the threaded joint next to it was also leaking. I turned it off and 're-threaded' the pipe with some new teflon tape, and while I was at it I put the old lower pressure meter back on. I turned it back on. It seemed no matter how much tape I wound around the leak I couldn't stop it. But it was now next to nothing.
   Soon the warm air out was 19.3° (instead of 18.3°), almost 3 degrees warmer than the 16.5° air being drawn in by the fan at floor level. That had to be doing some heating. Later the temperatures were down a bit (disappointing but not surprising given it was 2° outside.) Warm air 19.1, Floor air back down to 16.0, and the room air 17.1. I went in and did some dishes and they came up marginally - about .2 degrees - a warm human, hot dish water and vapor.
   At 23:00 PM I finished and I turned off the pump and fan. Within 15 minutes the temperature was down 1/2 a degree, and .7 or .8° in half an hour. The heat pump had definitely been doing some good pumping to keep it from dropping much for two hours. Then I went to run the 425W electric heater for 1/2 an hour, but the temperature rose .4° in a very few minutes - in fact, when I turned the heater on and looked again it was already up by .1°. Perhaps I had left the livingroom door open for a few minutes? Or maybe it was because the fridge had come on? Then after I left the room the loose "tip over" switch decided to cut the heater off and the temperature dropped. I decided to ignore that test entirely and run it another 1/2 hour. The thermometer read 16.9° the whole time until just before 00:30 it went up to 17.0. At that time I turned it off and turned the heat pump back on.
   I put my hand on the tape and managed to squeeze off the leak. It didn't seem like much, but the air pressure rose from 35 PSI to 46. That was surely a substantial loss, having that air decompressing into the room instead of outside. It was too late at night to stop and fix it. Manñana! The temperature dropped a bit over 40 minutes by 01:10 AM, but not as much as with no heat. or with just 75 watts of it.

(6th) In the morning I returned to the charge. I re-soldered the pipe fitting and put it together again. An hour later, thinking I still heard hissing air, I found another leak, inside the box. It was hard to get into and I had only tightened it by hand. (There was oil from inside the compressor all over the fitting.) I took the wall with the fan off the box and got two wrenches in and tightened it up. The compressor motor outer case was just a little warm. It seemed the heat transfer was very good, especially compared to the previous compressors. I was now running it at about 45 PSI. It seemed to be getting the room slowly warmer. Then I realized a window was open 1-1/2 inches for fresh air and I closed it.
   It was quite something, because nothing much seemed to be happening. What heat? There was noise, and a fan was blowing cold air out the end of a dryer hose duct. How could that heat a room? Yet that air was a couple of degrees warmer than the air coming in at floor level and just a bit warmer than the temperature of the room, and would continue to be a bit warmer than the room regardless of room temperature. A low temperature rise is just what is wanted for high heating COP. I went away for an hour and the room temperature had risen by about 1/4 of a degree. That would make it about the equal of the 425 watt radiant heater... except that the outdoor temperature had risen to 5.5° and the LED lights (200w) were on. (Too many variables!) At 13:45 I turned the unit off, but at 14:00 (for once) the sun came out, heating the room through the patio doors and making further comparisons of temperature changes useless.

   Here are some typical readings, obtained in the evening after running the unit for a couple of hours:

Outdoor air: +2° (dropped from +5° earlier)

Room air: 17.3°  (at table height - livingroom door was open allowing woodstove heated air in, so this temperature is meaningful only in comparison with the two below.)

Cold air into duct: 16.1°  (room air at floor level)

Warm air out of duct: 19.1°  (so the air was being raised by 3.0°, to 1.8° above room temperature, as it blew through the box and duct past the compressor and stegosaur pipe. With a faster fan it would have been even less rise - but at least as much heat.)

Top of Pipes°: not measured in this instance. (It seemed to be typically only 1/4 of a degree above the warm air temperature, so it would seem the "stegosaur" radiator pipe (now with about 30 inches of fins) was quite effective, and or that the pipe wasn't being warmed very fast by the mere 75 watt compressor.

Compressor Body°: Not measured, but felt like 35+°. (Here is where there may be some important losses and COP drop.)

Pressure in pipes: 44 PSI

Pump power: 72 watts (Varied from ~78W cold to ~72W once well warmed)

Duct Fan power: 3.2 watts

Means for Improvement?

   This time I didn't see (feel) any major "hot spots" on the compressor where the pumped heat is being wasted. But it was definitely warm, and there may still be one or more bottlenecks inside the outer body of the compressor. The real compressor flops around inside when you tip it, "clunk clunk", so it's obviously just a refrigerant-leak-proof outer shell that's seen. (It probably cuts down the noise quite a lot.) Does it restrict air flow? I suspect something in there is reducing efficiency. (It sure was!) If I get up the nerve, I may cut open the compressor body and see what further gain might be had by using just the actual bare bones compressor from inside the housing. Actually I think I'll get another one to do that to.

   Also I'm not sure how effective my outdoor heat exchanger is. With the small compressor there's almost no air to be felt coming in through it, and what it does need may be coming in through other leaks in the room. One of the doors, a sliding door to a very cool room with lots of free air flow, has big gaps under it and on the side. It could draw air through that perhaps more easily than through the heat exchanger. If this is in fact the case, an effective COP rise - and some air definitely coming in through the exchanger - might be observed in a tighter space.

   I've heard it said that plain air heat pumping is less efficient than with refrigerant. No doubt that's why everyone has gone with refrigerant until now. But reducing the pumping 'lift' required from 20 or 30 or 40° (depending on weather) to 10° and maybe even well under that (independent of weather), makes allowances for even very low efficiency of equipment and technique to nevertheless attain high real COPs.
   If the 75 watt air compressor only does about as well as the 425 watt electric heater, that indicates an effective COP of about 5.5. Or it may be a bit less. Even if COP 20 or 10 can't really be attained in practice, at least not without a specially made air pumping unit, 5 or higher (no matter how cold it is outside) is still better than anything else out there. I can replace the 250/500 watts radiant heater in the travel trailer with the 75 watt heat pump - 1.8 KWH per day to keep it from getting cold and mouldy* instead of 6 or 12 (and I won't care about the noise). But I'm sure that if I can't improve them myself at home, once this goes commercial great efforts will be made to attain those juicy potential COPs.

   At one point I thought to check the temperature of the decompressing air at the nozzle outside. It was 4.0°, while the outdoor temperature was 2.2°. The fact that it was about the outdoor temperature instead of below indicates the outdoor heat exchanger wasn't doing well. I'm pretty sure the compressor had an easier time drawing in air from elsewhere, as one couldn't feel much air movement coming in from the exchanger. However, 4° is way colder than the radiator pipes it was coming from. The unit was definitely compressing and heating the indoor air, and decompressing and discharging it outside much cooler.
   That the leaks in the house were drawing in a bit of cold air only added a bit to the heating load. They didn't negate the overall heat pumping effect. (hmm, a couple of windows at the far end of the house were cracked open for fresh air.)

*On the 5th someone living 'off grid' told me a horror story of how his unheated travel trailer quickly got damp and went mouldy inside, and became unlivable and useless.

New & Improved Compressor

(8th) I went back to the refuse transfer station and got another fridge compressor. This one was from a large fridge instead of an office water cooler.
(9th) It seemed to put out far more air. It was around 100 watts running open, but if one covered the air outlet, it quickly ran up to 150 watts and even 185, and plenty of air really wanted out! My immediate impression was that it was far superior to the other one. Twice the power and more, and seemingly at least 10 times as much air. And it even was quieter! The power of the other one stayed pretty much constant at 75 watts regardless, while this one changed markedly if it had to compress the air more, indicating that its power was being much more fully harnessed to the work, not just into turning itself.
   This one was taller yet than the first and second ones and bigger around, so I would have to rebuild the box yet again with more changes. I decided to leave the present box and make a new one. Making and assembling 6 plywood sides into a box took the afternoon.

   I went to dinner at Mike & Heather's on board their boat. Mike was replacing the heating system on a neighboring boat and gave me two four foot radiator pipes of 3/4" copper with aluminum fins from the old system being discarded. What a great score! That would save me from having to make a bunch more stegosaur radiator pipes before trying out the new compressor - I was sure it would need more radiators than the old one.

   A parameter I thought of later was that the pipes might be more effective per foot if there were heat exchange fins inside the pipe so the compressed air would yield its heat more readily to the copper pipe. How much improvement that would make I have little idea, but it sounds hard to do. Perhaps here is a place for aluminum: short lengths of 'fin assembly' can be inserted and pushed into the pipe with a plunger of some sort until it's as full as desired. OTOH might it be noisy with compressed air running by little pieces inside? It seems to have enough hiss going around a sharp corner that I ended up making the corners rounded.

New Heating Unit

 12 foot finned radiator pipe
(16th) With these new components I got a couple more pipe fittings and another piece of dryer duct and started putting together a new unit. Since the 9th I had unsoldered the two new finned pipes and reassembled them in line so I could get the dryer duct over them. And I put in a pipe union and added the stegosaur. I had enough confidence to drill an 11/16 inch hole through the house wall so the compressed air radiator pipe would end outside.
   It seemed to me at the time that the outdoor heat exchanger didn't really do a lot, and I didn't bother with it for this build. The compressor just didn't draw that much air, and it seemed to get it from other gaps anyway - windows cracked open or whatever.

   Now I fitted out the plywood box and mounted the compressor and another fan in it. After I mounted the fan it occurred to me to wonder whether it should blow air from the compressor end, or from the far end and then through the compressor. If it blew from the compressor end the unheated air would hit the compressor and the warmest part of the radiator. But then it still had to blow it out the far end. That meant that warmed air would hit the coolest end of the radiator pipe, and it might not cool it any further.
   If on the other hand the air blew the other way, the coldest air would hit the coolest end of the pipes, and the pipe would be coolest when it when outdoors and hit the end. The somewhat warmed air would flow past the warmest end of the pipes and the compressor. And the warmer air would come out at floor level where the fan was blowing it out.
   I decided the second way was better. That seemed convenient, because I had accidentally mounted the fan on the wrong face of the plywood - on the outside. Now I wouldn't have to turn it around. (I may put it inside it later.)

The larger fridge compressor in its
box and all connected, fan pulling out

(17th, 18th) I got everything together and fired it up. It didn't seem to perform very well. The warm air out wasn't even as warm as the smaller compressor. I thought that since warm air rises it might circulate more slowly because of being pulled down instead of pushed up. Later I remembered that the uninsulated aluminized duct was now 15 feet long instead of 5, and that warmth would be radiating from it all the way along. So the air at the end should be cooler.
   Still, it didn't seem to be warming the room very much. Again it seemed less effective than the 420 watt electric heater. Why?

(19th) I turned the fan around. Then I turned up the air pressure from 60 to 80. (actually from 63 to 77, by closing the faucet on the outside end the the pipe a bit more. It's hard to set exactly.) One or both of these measures seemed to help. Temperature readings including room temperature slowly started rising.
   There was a join between the two sections of dryer duct, and I had closed it with alligator clips. I could remove one and put the temperature sensor in there. It was at least a degree warmer than at the warm air outlet, proving that a lot of the heat radiated out before it reached the far end.
   I ran the unit for about 6 hours. Initially it warmed the room a little, but it failed to get it above about 16.5 degrees at any time. This was however the warmest part of the house as I had tried to do something with the woodstove and then it smoked horribly. (I had to wait for it to pretty much go out so I could take the stovepipe off.)

   After that I ran the 420 watt radiant heater for a couple of hours. Once again it seemed to do about the same as the heatpump or maybe a little better. The good part was that the heatpump was 120 watts where the radiant heater was 420, suggesting a COP of at least 3 to 1. This at least seems to indicate that open loop air heat pumping works and is worth pursuing - it can save energy and money over simple electric radiators, and the performance is on a par with refrigerant based heat pumps but it's simple to install and cheaper.

   Along the way I noted that fridge compressors seem to burp oil out the compressed air outlet. I thought it would drain back inside since the outlet air pipe goes up, but even when it's turned off, the oil just sits and builds up in the pipe. It burbles more and more the longer it has run. One might drain it, but of course the compressor eventually needs the oil.
   So apparently, to accommodate in an open air system this type of compressor, which is made to run continuously and last a long time but in a closed loop of refrigerant, it needs a very tiny connecting pipe to bleed the oil back into the air inlet and hence back into the pump. Ug! That should pretty much solve the problem, at the cost of wasting a bit of the compressed air. I hope.

   But it would be nice to know where the bottleneck(s) is and be able to do something about it. It is said that air heat pumping is less efficient than using refrigerant. But why is that?
   Then, if piston compressors in general are inefficient perhaps something else could pump air? Where is the inefficiency in a piston and cylinder? The piston rings against the cylinder walls must surely be a significant source of friction. What about some sort of bellows type of device instead? The walls of the bellows themselves move, so nothing is sliding against anything else. Or what about the "balloon" type of power "cylinder" found in some Stirling engines -- might that have lower losses than piston and cylinder? (The internet is out again. Otherwise I would be looking this up and seeing for myself.)
   In the other end, how can the transfer of the heat to the room air be optimized? Obviously the bigger the radiators and the harder the fan or fans blow on them, the better the heat exchange from the radiator pipe to the duct air.
   What about the transfer of heat from the compressed air to the radiator pipes?

Compressors Today Are Not Made For Air Heat Pumping

The Ideal Case

(24th) I thought about the amount of power being used by these various compressors. Ideally, if there was no radiator pipe or if the end of the pipe was side open, and no pressure would build up in the system, the compressor is just blowing air like a fan. The duct fans were drawing about 3-1/2 watts. Running with open ducts the compressors started at 75 watts and went up from there to 750. (750 was for the Michelin - the larger Makita was better at 550 W.)
   What is wrong with this picture? It means that all the compressors are wasting huge amounts of energy. Of course it takes some energy to turn a motor, and to slide piston rings back and forth in a cylinder. But that should be small, almost trivial compared to the energy needed to compress air, where real work is being done. One can literally feel this when pumping up a bicycle tire with a hand pump. Disconnected, the pump handle plunges freely, almost effortlessly. If it's upright, it may descend by itself. The more pressure there is in the tire being inflated, the harder it is to push the handle to the end of the stroke to pump it up further. Why don't motorized compressors use so little power in free air?
   One can see this is at least partly (if not mainly) owing to internal restrictions to the air flow: they are compressing air, but it is decompressing before it even comes out of the compressor. Nowhere is this more obvious than in the 1/16th inch, 16 inch long pipe in the water cooler compressor. It was heavily compressing air in the cylinder and making heat there regardless of whether the air outlet was wide open or almost completely closed off - because it was almost completely closed off anyway. And thus the power consumed (75 W) didn't change much with external load, since most of the load was internal. When I changed that to a larger pipe, it dropped as low as 55 watts pumping 20 PSI.

   In a compressor designed for heat pumping, designed for optimum efficiency, I would hope that the watts consumed would go up by a factor of 5 to 10 when pumping no pressure compared to (say) pushing against 90 PSI. or maybe even 30 PSI -- all depending on the sizes and proportions --. and that much of the heat would go out with the air being pumped, into the radiator pipes. The heat starts in the compressed air. (That's the whole idea!) The "bicycle pump" air compressor with effective (copper) radiator fins and a one-way valve straight into the radiator piping inside the opening of the duct would form a continuous solid copper finned pipe with no constrictions.
   Then we might start seeing really high efficiency equipment, leading to the ultra-high COPs dangling in front of us waiting to be plucked.

And... (Hmm... This section seems repetitious...)

   If the compressor itself can be kept cool it's a big help. The Michelin compressor ran very hot and virtually didn't work - most of the heat never made it out of the compressor to the radiator pipes. The compressor walls are the first and warmest part of the radiator system. It would be best to have the heating duct fan blowing directly on the compressor, which should be made as a very good heatsink with fins. Instead, the Michelin compressor was inside its own housing with its air going to an internal tank through a small rubber(?) hose - not even metal. It was all outside the heat radiator system. How could it possibly not get hot? And refrigerator compressors are enclosed inside a steel case. How can the heat possibly be extracted so the compressor itself doesn't get warm? It's in the low "lift" that the high COP is gained, and if the compressor itself runs 30 degrees warmer than the surroundings, regardless of all other temperatures, is that not a 30 degree lift from ambient air? Yes, there is probably the prime cause of the poor efficiency. (At least my fridge compressors were small and enclosed inside the radiator air duct circuit, with cold air blowing on the outside. Probably that's why they worked even as well as they did.)
   I visualize now a long stroke compressor ("small bicycle pump") with copper heatsink fins along its length, astride the entrance to (or exit from) the the radiator warm air duct, and connected by solid copper to the first of the radiator pipes. The compressor must run as cool as possible, transferring its heat to the warm air duct and to the pipes. (We don't care too much about the motor running the compressor, but the compressor itself.) Obviously nobody now making air compressors is very concerned about that - they only care that it not get too hot to operate. Maybe the cylinder of the compressor/pump could even be a short section of copper pipe, with the "stegosaur" type copper radiator fins along its length (at least toward the end where the air is compressed into), and a one-way flap at the end? Maybe it should be something harder - cupro-nickel or monel? - or copper with a nickel plating on the inside - but something that would conduct heat well and that one could still solder copper fins onto.

Rotary Screw Compressor

(25th) Here is a new thought: It seems air compressors need to have a powerful motor to push the piston in against higher air pressures. A more powerful motor has more losses than a smaller one. How can some leverage be gained so that a smaller motor can push in a big piston with a long stroke that gives quiet operation, against heavy air pressure? There are obviously many ways, but this might be found most practical: screw it in. I can think of more than one way to do that. Three that might be valuable are:

1) Piston is attached to the end of a threaded rod. The motor turns the rod, which screws in and out via a nut at the base end of the cylinder. The piston may have a small bearing so it doesn't turn with the screw, but it has to move in and out with the rod.

2) There is no piston. The threaded rod that screws in is almost as big as the cylinder along its whole length, and it itself displaces the air in the cylinder as it screws in.

   Obviously for these the shaft/rod has to move in and out with the piston. That will make it tricky to connect to the motor.

3) The cylinder and piston are both threaded. Think of a big set screw turning inside of a long coupling nut. The motor turns a thinner hex shaft which turns the 'set screw' piston. In this case the threads are the "piston rings". The hexagonal hole in the 'set screw' goes right through it so a long hex shaft can attach to the motor to spin the screw without having to go in and out with the screw.

4) The piston is a hexagonal 'nut' running up and down a matching hexagonal cylinder. (Any matching shapes except round should work.) The center of the nut is driven by a threaded rod to the motor. Again the advantage is that the threaded rod is simply attached to the motor - it doesn't have to move in and out. A disadvantage is in trying to find any matching shapes for piston and cylinder other than round.

   There's the concept. I'm not ready to consider the smaller details yet - the relative value of the types above and different kinds of threads and pitches, or various seals and bearings, to have it all run at optimum efficiency with minimized friction.
   In all cases some sort of optical interrupter going up and down the shaft with the piston but outside the cylinder (or something) can tell the motor when to stop and reverse. A microcontroller controlling the motor can make the transitions soft and smooth for (hopefully) very quiet operation.

   If whatever it is is oil lubricated (as I assume for longevity), pointing the cylinder upward with the compression at the top end should result in the least oil going through the cylinder into the compressed air end even if the piston/cylinder gaps leak - hopefully virtually none as the compressing air pushes it back down through the same cracks it came through. (Some kids in the younger grades in school always said I was a screw-up. Apparently there's value in screwing up!)

(29th) I was concerned about the amount of friction this sort of system might have, especially one with large threads. I got out a stainless steel 3/8 inch nut and bolt, and they seemed to turn quite freely against each other, even with axial pressure between them and no lubrication. Even a long zinc plated coupling nut turned pretty freely. The chief concern for friction then was the piston rings against the cylinder walls. (Again I rather like the bicycle pump. Instead of piston rings, sealing material - leather? - was attached to the face of the piston, set up to expand and seal as the piston is pushed in, and then pull away from the cylinder wall as it goes back. Thus it also acts as the air intake mechanism. Perhaps I'll try something like that.)
   It looked like the chief reason to pick a diameter, at least up to about 1/2", was how many turns per inch one wanted. That sets the speed the piston moves in and out with "n" motor RPM and hence the amount of air pumped. A higher RPM motor would use a smaller threaded shaft. Without picking a motor yet, I thought a 3/8" threaded rod and coupling nut (or a piece of a coupling nut) looked about right. Low friction; too thick to flex when pushing against pressure. The seal, whatever it was, could be on the front face with a washer backing it, going out almost to the cylinder walls, and a rod or two could be silver soldered to the back to stick out the motor end. These would slide through something to prevent the nut from turning with the rod (without needing an un-round cylinder and piston shape), and provide a means for an optical interrupter to "see" when the piston reached the end of the cylinder.
  I also found in my stash some cupro-nickel (monel?) pipes, 1 inch and 1.5 inch. They both looked very smooth inside to use as cylinders. (I would not, I expect, be using the full lengths of these. Then again, why not? Fewer back and forth transitions is less stress, less noise.)

  With the whole mechanism being on-axis, the side-to-side vibration of the crank rod would be eliminated. However, that would be replaced by the continually reversing radial force. Would it be quieter? I hope so! If it was controlled by a computer the transitions could be made pretty smooth.

   Well! I had certainly hoped not to get into this air heat pumping idea so far as to be thinking about air compressor design! But it seems key to "ultra-high" COPs. Am I actually going to build one? Am I not spread thin enough already? Another one to put in abeyance, I think, and just see what can be eked out of existing compressors.

Fridge Compressor Disassembly

(21st) But first I wanted to take a apart a fridge compressor - the first one, that didn't pump air very fast. I sawed the thick metal shell apart by the seam with an angle grinder and zip disk, plus a hacksaw.

   It was quickly apparent where the air bottleneck was: The output from the compressor went through an extra long, itty bitty copper tube/pipe to the far side, where it came through to the larger pipe seen on the outside. All my efforts to get free air flow outside were wasted with this severe restriction on the inside before it got there. (In the fridge, the long tube would serve to isolate the vibration of the compressor from the outside shell - which is itself on rubber feet. Noisy as refrigerators are, the makers obviously go to a lot of trouble to make them as quiet as possible. And with the long length also the copper of the tube itself would be little stressed by the compressor vibrations.)
   Obviously the amount of heat pumped is determined by volume of air moving times the pressure of that air over outside air pressure. Obviously this pump couldn't push much volume through that tiny pipe without losing a lot of pressure, so the pressure had to be high with low air flow volume. That probably reduced the efficiency of the compressor and made it run warmer or hotter.

   High COP is dependent on a low temperature lift. Aside from that was there also a heat bottleneck from the compressor to the outside air?
   First, the compressor cylinder appeared to be made of cast steel instead of copper. Steel is doubtless good for durability, but it's not a good heat conductor. And cast metal is full of tiny voids (created as the molten metal cools and solidifies) - little heat insulators. Thus the inside of the cylinder would surely run much warmer than the outside (making a higher temperature lift), so even putting heatsink fins on the outside of the compressor wouldn't really solve the problem. I could see using a thin steel (or cupro-nickel?) sleeve inside a copper tube (or even just some nickel plating on the inside?), with heatsink fins on the outside.)
   Then there was the fact that the whole compressor was inside the outer heavy steel casing. No doubt the oil was slushing around, hitting the outside of the compressor and transferring its heat to the outer casing, but how effectively? I would have a hard time judging that, but it's probably a significant resistance to heat transfer.
   Then once the warmed oil sloshed against the outer casing, how effectively would it be dissipated to the outside air? My fan blew air around it the whole compressor, but there were no heat dissipation fins, and once again the case was steel rather than a good heat dissipator like copper.
   So it would seem there is actually a series of poor heat transfer steps, all together constituting a considerable or perhaps even severe bottleneck. This plus the restricted air flow may be sufficient in themselves to explain COPs of 3-5 instead of 10-15.

   They say air heat pumping is less efficient than with refrigerant. Maybe it's because no one has tried to make an air compressor to be efficient for air heat pumping?

and Reassembly

   It seemed to me that the worst feature was that 1/16 inch inner diameter tube, about 16 inches long, that the compressed air had to somehow make its way through before it got to a reasonable size pipe to flow in.

   But I couldn't see any way to attach a bigger pipe except to drill out the tiny hole on the side of the "washer" fitting larger, and it could only be a bit larger because then the hole would hit the end of the fitting. I got it up to 1/8" and found a tiny bit of pipe to fit into it. It was just 1/4" long and the inside diameter was just 3/32". Still, that was 9/4 of the cross section and 1/64th as long. I found another pipe to 'telescope' that out to a somewhat larger size in which to bring the air out of the case.
   Now it was coming out of the case direct with the inner compressor's vibration transferring to the radiator pipes. I hoped it wouldn't be too noisy. But with the refrigerator compressors so far, the air fan had been the bulk of the noise. Perhaps I could put a resistor in the fan power supply - or just run it off 24 volts instead of 36? - so it would run slower? But the first test would be with the fan the same to help show the improvement, if any, in the compressor's performance.
   But having done this much, I was nervous about trying to put the compressor back together without leaks. If it leaked it could only be run for short tests and a demos. Then the larger fridge compressor wouldn't start and I decided to try this one again after all.
   Later I removed even the bit of smallest pipe and expanded the hole to take what is the middle pipe in the photo. It worked much better, delivering more air and was down as low as 55 watts pumping 20 PSI instead of over 70 pumping nothing.
   But, the case seal being breached, even wrapped with PVC tape I was leery of leaving it running unattended in case it should run out of oil, and being such low wattage (65 W at 40 PSI) it would only make 260 to 325 watts of heat at COP 4 or 5. And it was much noisier with the inner compressor linked through the case straight to the radiator pipes.

Makita Compressor

(22nd) Odd as it may seem, I hadn't even opened the box of the big Makita compressor that I had ordered and received almost a month ago. I had wanted to try the refrigerator compressors as being smaller scale experiments and tolerably quiet. I figured the Makita would want about 25 feet of radiator pipe. That was going to take some making. But with the gift of the aluminum finned pipes, I now had at least had 11 or 12 feet.
   And I had been learning. The compressor has to go inside the ductwork in a box so that its heat, and the motor's heat, all contribute to the radiator air's heat.

   I was pleasantly surprised to see that the cylinder with some heatsink fins fed a fat 3/8" copper pipe that had aluminum fins on it, and fed into the tank in a 16 inch length. Here was a setup that would let lots of air pass through freely! The motor of course had a fan, and the whole had a plastic shroud so that that fan blew cool air over the whole works.
   Furthermore the compressor was oil filled. It was made to run continuously, not just occasionally. And there was an air filter, with an inlet fitting. That meant the air could be connected to the incoming air side of the outdoor heat exchanger. Then the air both to and from the compressor would connect outside the house, and would be separate from the radiator fan air.

   The compressor and the air pipe to the tank run pretty warm, not to say hot, even with the drain valve open (ie, no pressure in the tank). Or perhaps because it was open? With the the valve open, the unit drew about 550 watts. Once it was closed (after its 20 minute break-in period), that went up to over 1100 watts with the pressure nearing 130 PSI. To me that would seem to indicate 550 "wasted" watts (probably somewhat less, since it was nevertheless moving air) and 550+ of watts actually compressing air... over 50% efficiency. (That beat the large fridge pump: 100 W with the air output open and around 150 W max when compressing heavily -- 33% efficiency.)

   I decided I had been beating about the bush and that I should set it up. With just 12 feet of radiator pipe it would run warmer than it should, but with say 700 watts even a COP of 4 would be equivalent to 2800 watts of baseboard heaters.
   Now... yet another, and much larger, box to hold the compressor. I went into town to buy some plywood, and I also borrowed the rest of the radiator pipes that Mike had obtained. That make about 21 feet of radiator pipe. For the time being, that would save me from having to make a jig and produce more 'stegosaur' copper finned radiator pipes before I can try things out decently.
   With this compressor moving a much higher volume of air, I decided that the outdoor heat exchanger had to be done. All this was adding up to a bigger job than just swapping compressors. Then I opted to learn more first by experimenting more with the fridge compressors and less radiator pipe.

Outdoor Heat Exchanger Install

   Without an outdoor heat exchanger both fridge compressors, even after improving the smaller one, performed poorly compared to the first tests with one. I started to think the outdoor heat exchanger is a much more important component than I had realized, not to say a vital one. After all, it does the "heavy lifting" of bringing outdoor air up toward room temperature as it enters the house. Otherwise the compressor draws in cold air through window gaps and cracks.

(27th) For the outdoor heat exchanger I swapped my older pipe with the homemade aluminum fins for one of the commercially made ones with far more but a little smaller fins. It seemed to be about the right length for my insulated box "duct", and its smaller fins would let air go by more easily. To mount it against a house wall, I capped off the top end, and soldered in a pipe with a gentle 90° bend (two 45° elbows) and then a reducer from its 3/4" pipe to 1/2", to go through the rear face and the house wall. Inside the house another gentle bend (bent soft copper pipe) turned it parallel to and close to the wall to attach the radiator piping and dryer air ducts.

   Owing to having to make a hole in the wall of the house, I wanted this to be "the" real install, not just a jury rigged test setup. At first I thought to make a square hole in the house wall, about 3x3 inches, the same as the duct. Then I realized that as there would be no fins on the pipe going through the wall, it could be much smaller. In fact, it didn't need to be a whole lot bigger than the 1/2 inch pipe of compressed air going out. So I drilled a circular hole with a hole saw and put a piece of the PVC 1-1/4 inch (actually about 1.5" I.D.) "irrigation system" pipe in it as a duct through the wall. With a similar hole in the exchanger inner face, the PVC pipe brought the air directly from the exchanger box to the inside of the house.

   I pointed the whole exchanger vertically down the wall, with that hole and pipe into the house at the top end, and used four "shelf brackets" to mount it against the wall. (Weatherproofing later, but I did staple a window screen over the bottom to keep flies and bugs out.) Warm air rises, so the warming air should naturally circulate better as it goes up the duct, passing by the pipe fins and picking up the heat from the outgoing compressed air, and then into the house through the PVC pipe.
   Before putting on the screen I stuck in a couple of pieces of foam insulation at the bottom to narrow the outer air passage to reduce wind blowing unwanted air in. I could still feel a draft coming through the plastic pipe.
   The compressed air pipe ends at a restriction valve far enough below the bottom of the box that the cold outgoing air shouldn't be drawn back in. (And a piece of garden hose can be attached to end it even farther away if desired.) The vertical orientation opening at the bottom will expel any water in the outdoor exchanger part of the piping.

   On the inside, the PVC pipe could just end at by the wall as at present, or more piping and a hose could route the incoming warmed air straight to the compressor. I rather like the idea of bringing fresh but not cold air into the room - without opening a window. The compressor *should* still draw it through the exchanger. If the air is routed straight to the compressor then just exactly what the compressor needs is warmed and enters, and then it exits as compressed air through the pipes, then is discharged/decompressed outside at perhaps below outside temperature.

   If the outdoor exchanger's intake is connected directly to the compressor, its location is almost immaterial. But if it is to draw (warmed) fresh air through the house on its way to the compressor, it or its location could probably be improved. With it being on the middle of a wall, if the wind puts it on the lee (low pressure) side, the house would tend to draw more air through cracks on the windward side and less through the exchanger. And of course when it's on the windward (high pressure) side, instead it excess air would blow into it, to go out the now leeward cracks. (At least its opening points down, not straight into the wind.) Having the entrance to the exchanger away from the house wall in a neutral wind pressure position might be optimum. Or at a corner of the house might be better than the middle of a wall. Well, wind always makes a house harder to heat. The tighter it is, the less air the wind will push in and out.
   This mounting should suffice. At least, having mounted it I don't plan to change it. If it seems worth it, a possible improvement besides moving the whole exchanger might be to have a duct it draws air from, placed out in a more wind
pressure neutral location.

(March 4th) On measuring temperatures in tests, I found other more important factors than placement.
   I had already reduced the large air gaps around the fins (per image) by inserting some thin slices of styrene foam, because too much air was coming in without being much warmed. (And maybe doing a little more yet might further improve it!)
   Next, the pipes themselves were transmitting coldness from the outdoors and from colder sections below directly through the copper to warmer areas. The pipe just indoors from the exchanger hole was way below room temperature. The 9 inches or so of pipe sticking out below the unit to expel the compressed air, transmitted the outdoor temperature into the exchanger. The piping in the exchanger needed as many thermal breaks as practical to isolate the differing temperatures from each other. I unsoldered and put in three joins of PVC pipe to make gaps in the copper: bottom, halfway, and just inside the house.
   Even then, the incoming air was only being raised a little over half way from outside to room temperature - maybe 60%. With all due respect for the law of diminishing returns since the incoming air will never quite get up to the temperature of the outgoing, I hope to do better than that. It's probably a notable factor in the low COPs. How much would more heat breaks help? Or perhaps the thing to do is to double the length of the exchanger to 8 feet?

A couple of pieces of foam to reduce the outdoor opening,
and a screen to keep bugs out. (I could just see a bumblebee
hive or thousands of hibernating flies taking up residence.)

Outdoor heat exchanger "in context" on the wall.
(If I have to make it longer, it'll have to be oriented diagonally.)

Thermal Breaks & A Better Way to join PVC Pipes to Copper Pipes (no pipe fittings)

   Perhaps of interest, I made the thermal break PVC pipe pieces soft and flexible like rubber hose by putting them in the kitchen oven at 250°F. That lets them fit over the copper pipes or soldered copper pipe fittings. Then they harden again as they cool. (Wear gloves!) They're soft for just enough seconds to put them on both pipes. Any delay and they have to go back in the oven. Then I put pipe clamps around them over each copper pipe. One I had to take off with vise grips and then put it back on, and then it leaked. I could hear the air hissing out. I took a hot air gun to it and managed to get the clamp to seal. (The joins held 120 PSI.)

Inner thermal break.
Pipe on left going into heat exchanger is definitely colder.
(Aluminum sheet was to protect wall unsoldering the copper join.)

Outer thermal break at bottom of outdoor heat exchanger box.
This one went over a 3/4 inch pipe and a slightly larger 3/4 inch pipe fitting -
the advantage of soft and flexible!

While I was at it, I had slit this larger PVC pipe, and I softened it in the oven and turned it into a flat PVC sheet. (~5 x 12 inches)


I took a fair lot of readings in doing all these tests

(28th) I hooked up some radiator pipe and the "improved" water cooler compressor. I had a hard time with leaks. The little compressor's connections weren't very robust to start with, and now as modified they were pretty fragile and came apart a couple of times. Each time I had to untape the cut-open housing, fix it, and and tape it all up again so the oil wouldn't spew out.

   I was rather disappointed in the outdoor heat exchanger. The compressed air pipe going out was only 13° where it entered the exchanger (the room was 17°, so the incoming air was cooling the pipe!), and what felt like a cold draft was blowing through it from outdoors at 8°. Outside it was 5°. 13-5=8; 3/8=37.5% of the lift available from the pipe. It was working and doing a fair job of cooling the radiator pipe, even to a foot or so into the room, but it wasn't warming up the incoming air anything like as much as I wanted - only 37% of the way to the pipe temperature, and also the pipe was being cooled by 4°. 17-5=12; 3/12=25% of the desired lift. That's nowhere near the "law of diminishing returns" where more length of exchanger would be helpful. Still, the principle is there. Even 3 degrees passive 'lift' is that much less 'lift' needed from the compressor.
   But it has to work better to get really high COPs. My thoughts are that the smaller fins on its pipe leave too much gap around the outsides and the outdoor air flows through too freely and perhaps with poor heat exchange. Perhaps I'll try putting the other pipe with the larger fins back in. Other than that, I'm now thinking that routing the exchanged air into a hose that goes straight to the compressor intake would slow the incoming air down and let it heat up more on the way in, while also not letting any extra cool to cold air into the room. Less air will get warmer if the exchanger isn't large enough to warm a greater volume at a fairly high indoor-outdoor temperature differential. Otherwise, an 8 foot exchanger length would have much more warming capacity than 4... and of course 12 feet more than 8, et cetera. -- provided of course that there's enough heat in the outgoing air pipe, and of course until the incoming air is near room temperature and won't pick up more heat no matter how much longer it is.

   With the room typically at 17.3 degrees, the fan took in floor level air at 16.6 to 17.0 degrees and the warm air came out the duct (which only had the one 'stegosaur' finned pipe) at 18.5 to 18.8 degrees. So the heat pump was heating air by almost 2° and that air was being blown straight into the room. Another 4 feet of my aluminum finned radiator pipe in open air (no duct) doubtless contributed a bit more. (Of course the compressor itself, hidden inside its steel shell, was running much warmer than any temperature to be felt outside, in spite of the expanded outlet air pipe.)

   The cord for the fan, which was more than long enough, started mostly in the livingroom, and it seemed to slip farther and farther through the tight fitting livingroom door every time in and out. I didn't try very hard to push it back, and finally in the kitchen area it was taut at ankle height and I tripped on it, which shifted the box and made a pressure leak. I couldn't seem to close the leak while it was running. Pressure went from 45 PSI to 35, and then I turned the air pressure down to 30 at the outside valve so it would be letting out more air there than at the leak. But it got worse and I turned it off when it was down to 20. Thus ended the tests for February.
   But until near the end of the test the room temperature had been staying a pretty constant temperature despite only ~65 watts of power and 5 degrees outdoors. Then I turned it off and turned on the 420 watt radiant heater. This seemed to keep the room at a similar temperature, maybe a little better. Not bad to do similar with 65 watts! If it was even 325 watt equivalent, that's a COP of 5. Present day heat pumps don't get that at 5° outdoor temperature.
   The noise would be unbearable for regular use. With the compressor's long, thin internal pipe removed and the output connected straight to the radiator piping, everything vibrated. The larger fridge compressor is fairly quiet and I won't modify it. For manufacturing, something like my rotary screw compressor idea might be a lot quieter.
   And a large diameter fan for the air is also quieter. Or maybe a "squirrel cage" fan like in typical house furnaces?

   I'm getting rather tired of soldering copper pipes and fittings. (29th) Instead of changing the pipe in the outdoor exchanger again, I have the idea to add more foam insulation on the inside to narrow the duct. March 1st: did it.
   Then I realized that the copper pipe, as well as delivering the air to and from the fins for heat exchange, is conducting cold from outside along its length to inside. Otherwise, there was no reason the pipe should be colder than room temperature inside the house near but not at the exchanger. I think I should break it into three pieces with rubber pressure hose beatween them as thermal breaks:
* the outside end below the exchanger where the compressed air is discharged.
* lower and upper lengths inside the body of the exchanger.
* another thermal break where the pipe enters the house.

I can change the one long length of finned pipe inside the exchanger for two short ones. Rats, more soldering to do!


   I decided the next try would be with the larger fridge compressor. If one could only get around 5 to 1 COP with an existing compressor and my apparently inadequate outdoor heat exchanger, I should use one that drew more than 65 watts to get some useful heat from it. It would draw about double that, making 500 to 650 watts instead of 260 to 325, which ought to heat up the kitchen area nicely except on the coldest (or windiest) days.
   Compressors make water in the pipes from air humidity. I had noticed "burbling" sounds the time I had used it previously, and when I plugged it in (after having to fiddle with the connections to get it to start - ug!) it spewed out a couple of tablespoons of orange colored water. So I decided to elevate the box and have the pipe come out horizontally instead of going upward. If the pipe started at the highest point and all sloped down, all liquid should travel along it and come out at the outdoor spigot.

The large fridge pump setup. Right to left:
*The box with the fridge pump and fan.
* 'Stegosaur' copper finned pipe in flexible dryer duct. (with air pressure meter)
* Some aluminum finned radiator pipe mostly not in duct.
* Out through wall into outdoor heat exchanger. (2/3 plugged with foam owing to high wind blowing in way too much air)

The compressor & fan in the box end.

   To intrude slightly into March, on March 3rd I put together this configuration with the larger fridge pump. The warm air was coming out 4.5 to 5.5 degrees higher than the room temperature instead of 2+ to 3+.
   The room temperature (once again) didn't rise with the door to the livingroom (and the woodstove heat) and all other doors and windows to the kitchen area closed, and in fact it sank very gradually. Was it performing no better? Then I noticed frost and discovered it was -1°C outside. The thermometer that read '+5' had been dropped and was reading wrong. So actually it was doing pretty well.
   (I have been struggling all along with a digital thermometer that reads to .1°C but can take upward of 5 minutes to get to and stabilize at a substantially different temperature. I ordered 3 more digital thermometers on line in late January from China... will they arrive?)
   I adjusted the air pressure up and down, which changed the warm air temperature slightly and the compressor watts considerably. I got the highest temperatures with around 60 PSI. As the watts changed substantially with pressure, it looks like lower pressure is higher efficiency, but up to 60 or 70 PSI, higher pressure makes slightly higher radiator temperatures.
   I started leaving this running to help warm the kitchen area. But only when I wasn't there - the fan was too noisy for when I was. (Next job: cut down fan speed. After that: put in the powerful Makita compressor with all the radiator piping I can get or make. It's noisy, but should heat up the place so fast it can be turned off much of the time.)

The Grand Potential

   Assuming high COPs can be attained, what might open loop air heat pumping replace? Here is a picture of the house circuit breaker panel.

There are 11 double (230 volt) breakers:

* Workshop Sub Panel (The well water pump is the only major and frequent load)
* Clothes Dryer
* Hot Water Tank
* Kitchen Stove

* SEVEN double breakers - over half - are for electric baseboard heaters, in a perhaps typical electric baseboard house:
  - one   - dual 30 amps (~5400 W)
  - two   - dual 20 amps (2 * ~3600 W
  - four - dual 15 amps (4 * ~2700 W)

   Without bothering to check out the rating of each heater, at 75% of breaker capacity that's 85 amps, 230 volts or about 23,000 watts of electric baseboard heaters. Now just imagine instead a heat pump with a COP of 15. That would need about 1500 watts of electricity - one single 15 amp circuit breaker - to do the full capacity of all those 23,000 watts of baseboard heaters. And it would have to be awfully cold to cause it to run full time or most of the time. So it could heat the house for anywhere from zero to say 80 $/month, depending on weather. If all those breakers were removed the breaker panel would be just over 1/2 full and it is apparent that the 200 amp panel has become "overkill".

   That kind of performance could for the first time keep me from lighting the woodstove most of the time in cold weather, keeping it mainly as a backup for power failures. Might it even be practical to heat a greenhouse modestly warm, and light it with LED lights, for extended season or even year-round gardening?
   And of course homes all across the north the world over could benefit the same way in heating season. In some ways, this seems almost as good as harnessing free energy. Winter loads on northern power grids will be dramatically reduced when this catches on.
   Another configuration would do cooling this way as well, again with an outdoor heat exchanger doing most of the 'lifting' ('dropping?'). It's not needed in the very moderate coastal climate here, but would surely be just as valuable a product for warmer zones.

Additional Heat Pumping/Air Compressing Info Gleaned From Web

Here are some general heat pump stats for closed loop refrigerant based heat pumps, in a New Zealand report:

Energy efficiency of [Refrigerant Based] heat pumps for space heating
(Source: http://www.level.org.nz/energy/space-heating/heat-pumps/ )

As heat pumps only move heat and do not actually generate it, they have a very high ratio of heat output to energy input. This heating energy efficiency is expressed as a coefficient of performance (COP), while cooling energy efficiency is expressed as an energy efficiency ratio (EER).

Typical domestic heat pumps have a COP of 2-4.5, which means the heat pump produces about 2-4.5 times as much heat as the electricity it uses (under optimum conditions). Some heat pumps have COP's as high as 5.7. The efficiency of an air-to-air [via refrigerant loop] heat pump decreases as the temperature difference between source and supply increases - as outside temperatures drop, the heat pump's energy efficiency [actually, its COP] reduces.

Heat pumps are currently the only form of heating, aside from solar, where the COP is (usually) greater than 1, which makes them the most efficient [they mean "effective" or "economical"] form of purchased space heating commonly available. [Solar heating COP can be greater than 1? How? Am I a nit-picker?]

The cooling energy efficiency ratio (EER) is typically about 2.5-4.0, which means the heat pump produces about 2.5-4 times as much cooling power as the electricity it uses. Some heat pumps have EER's as high as 5.8.

   Yep! There are the low COPs that drop with outdoor temperature, pumping the least heat when it's coldest and most needed. ...with the advantage in warmer climates that they can cool as well as heat. But they continue:

"In a BRANZ study of 160 households with heat pumps, more people said their energy costs had increased since acquiring a heat pump than those who said their energy costs had fallen. Only 15% described running costs as excellent. Occupants keep their homes warmer than previously."

(23rd) Then, here is a link to an article about recovering and using waste heat from air compressors. They don't expect to get as much heat out as the electricity put in. They consider high temperature waste heat to be a free asset for heating instead of a liability and make no mention of COP anywhere. It would seem their thoughts, like those of everyone I've been able to find so far, are far from considering the potential of air compressors as heat pumps.


"Compressors are better at producing heat than compressed air" says Tom Taranto of Data Power Services.

   And here is a more in depth look at heating and cooling by compressing a gas (air) than I have given or even thought of. But after all, real gasses aren't "ideal" gasses:


   Luckily this more involved bit of theory and math doesn't change the air heat pump idea. It was never dependent on a specific volume of gas reaching a specific pressure or temperature for a specific compression. Rather I realized the heat - however much from however much pumping - will be carried off into the cylinder walls and into the compressed air flow into the radiator pipes. And I also assumed the actual pumping efficiency would be under 50% because everybody said so - but assuming at least 25% could be attained.

   I note (again?) that on a 'quora' web discussion page someone said heat pumping with air takes a pretty high volume of air at a high pressure. But just how much volume at exactly what pressure works best for any given real compressor and set of radiator pipes is probably easier to determine by doing it and trying out different pressures than by academic formulae when complex parameters are not all known (there are a lot of variables), and where gross mistakes might go undetected for too long into a design.

Electricity Generation

My Solar Power System

Month of February Log of Solar Power Generated [and grid power consumed]

(All times are in PST: clock 48 minutes ahead of sun, not PDT which is an hour and 48 minutes ahead. DC power output readings - mostly the kitchen hot water heater for some months, then just lights - are reset to zero daily (for just lights, occasionally), while the others are cumulative.)

Solar: House+DC, Trailer  => total KWH [grid power meter reading(s)@time] Sky conditions

January 31th 03.76, - ,  937.32 => 0.17 [85Km,charging; 71512@17:00] dull. dreary. dim. rain. 3°.

February 1th 04.02,-, 937.33=>0.27 [55Km,charging; 71555@17:00] Mostly dull, snow 0°. Collectors covered with snow except the two on the lawn, which are set at a steep angle.
2th 05.06, - , 937.55 => 1.26 [71590@16:30] Sunny. 10 panels still covered in snow. 1 was clear by late afternoon.
3th 05.39, - , 937.78 => 0.56 [71620@17:00] overcast & rain, 3°. Solar panels were clear of snow (viewed at end of day).
4th 05.74, - , 938.00 => 0.57 [71653@17:30;75Km,CarChj.] same as 3rd.
5th 07.28, - , 938.98 => 2.52 [71689@17:30] a bit of sun in the morning, then cloudy.
6th 09.67, - , 940.16 => 3.57 [71715@17:00] Sun for the afternoon! 5°. Wow!
7th 10.91, - , 941.00 => 2.08 [85Km,charging; 71760@19:00] mostly cloudy. 3°
8th (2.6*), - , 942.48 =>4.08 [55Km,chj; 71797@20:30] Mix of sun and clouds *Estimate: Power bump in eve lost reading.
9th 01.27, - , 943.15 => 1.94 [50Km,chj; 71830(est)] mostly cloudy. 8°! (forgot to read - all are estimates from AM-10th.)
10th 2.56, - , 943.85 => 1.99 [71854@18:00] mostly cloudy.
11th 5.30, - , 945.48 => 4.37 [60Km,charging; 71885@18:30] I thought it was light clouds. Did the sun come out?
12th 6.16, - , 945.98 => 1.36 [71914@17:30] Uggy. Wind & rain in eve.
13th 8.45, - , 947.69 => 4.00 [90Km,chj; 71951@18:00] Some sun, some clouds, some drizzle, some jet trails.
14th 9.63, - , 948.39 => 1.88 [85Km,chj; 71994@17:30] Clear almost all night, then cloudy all day.
15th 12.23,-, 950.35 => 4.56 [60Km,chj; 72035(est)] Some sun. 5°. Sleet in evening.
16th 14.57,-, 951.91 => 3.80 [72064@18:30] Some sun. 2°.
17th 16.64,-, 953.31 => 3.48 [55Km,chj; 72104@19:30] Much more cloud than sun.
18th 19.46,-, 955.58 => 5.09 [72125@18:00] More sun than cloud? Cold wind, frost AM, 3°.
19th 20.52,-, 956.14 => 1.82 [55Km,chj; 72163@18:00] overcast until sunset. 4°.
20th 22.35,-, 957.21 => 2.90 [72195@17:00] pretty much overcast - lighter spots. 6°.
21th 25.17,-, 958.73 => 4.34 [85Km,chj; 72228@18:00] Light overcast, hazy sun later PM.
22th 27.64,-, 960.50 => 4.24 [55Km Chj; 72264@19:00] Frost AM. More light overcast.
23th 30.69,-, 962.14 => 4.69 [72289@18:00] Some sunshine! Frost AM. Hit about 5°.
24th 32.15,-, 962.97 => 2.30 [72325@18:00] Frost. Rain. Clouds. Wind.
25th 34.75,-, 964.53 => 4.16 [72350@18:00] Frost. Clouds. Rain. Sun later PM.
26th (3.33*),-,966.53 =>5.33 [72369@18:00; 55Km car charging] Clouds. Rain. Bit of Sun. *est - Power failed for an hour ~16:30, lost reading. 13°.
27th 03.11,-, 968.71 => 5.29 [72406@23:00] Rain. Sun. Rain. Sun. Rain & sun. 10°. Frost & stars after dark.
28th 06.34,-, 970.82 => 5.34 [85Km,charging; 72437@18:00] A fair bit of sun but some clouds.
29th 09.00,-, 972.49 => 4.33 [55Km,chj; 72475@19:00] At one point, the sun came out.

1st  10.36, 973.25 => 2.12 [40Km,charging; 72500@19:00] Heavy overcast, rain: dull.
2nd 14.06, 975.32 => 5.77 [72528@18:30] Mostly SUNNY!
3rd -  Power failure and huge windstorm brewing. I decided to unplug the grid tie inverters and leave them off for now. I don't want them damaged by power bumps. I turned them on mid-late PM on the 5th. I also put a "watch" battery in the house power logger so it wouldn't lose readings in power failures (which did again start at 0). Probably about 1 KWH attributed to the 6th was actually collected on the 5th. Still, the increase from 5.xx to 8.xx KWH is considerable.

6th 05.88, 979.21 => 9.77 KHW [35Km,Chj; 72665@18:30] Sunny (with snow flurries). (Includes ~3/4 KWH from 5th PM.)
7th 11.13, 983.09 => 9.13 [55Km,chj; 72695@18:30] Sunny.

KWH-  # of Days (February)
0.xx  - 3 (near start of month)
1.xx  - 6
2.xx  - 4
3.xx  - 3
4.xx  - 9
5.xx  - 4 (more 4.x and 5.x as month went on.)
6.xx  -
7.xx  -
8.xx  -
9.xx  -

Monthly Tallies: Solar Generated KWH [Power used from grid KWH]

March 1-31: 116.19 + ------ + 105.93 = 222.12 KWH - solar [786 KWH - used from grid]
April - 1-30: 136.87 + ------ + 121.97 = 258.84 KWH [608 KWH]
May  - 1-31: 156.23 + ------ + 147.47 = 303.70 KWH [543 KWH] (11th solar panel connected on lawn on 26th)
June - 1-30: 146.63 + 15.65 + 115.26 = 277.54 KWH [374 KWH] (36V, 250W Hot Water Heater installed on 7th)
July  - 1-31: 134.06 + 19.06 + 120.86 = 273.98 KWH [342 KWH]
August 1-31:127.47 + 11.44+91.82+(8/10)*96.29 = 307.76 KWH [334 KWH] (12th panel connected on lawn Aug. 1)
Sept.- 1-30: 110.72 + 15.30 + 84.91 = 210.93 KWH   [408 KWH] (solar includes 2/10 of 96.29)
Oct.  - 1-31:  55.67 + 13.03 + 51.82 = 120.52 KWH, solar [635 KWH - from grid]
Nov. - 1-30:  36.51 +   6.31 + 26.29 =   69.11 KWH, solar [653 KWH - from grid]
Dec.  - 1-23: 18.98 +   .84* + 11.70 =   31.52 KWH, solar + wind [711 KWH + 414 (while away) = 1125 from grid]
Jan.  - 6-31: 17.52 + ------* + 10.61  =  28.13 KWH, solar+ wind [1111 KWH from grid]
Feb.  - 1-29: 56.83 + ------* + 35.17  =  92.00KWH, solar + wind [963 KWH from grid]

* Solar hot water tank removed because the mineral-rich water stank. (Why doesn't it stink in the main house tank?) Now the solar DC system is only running a couple of lights - not worth reporting. So there's just the 2 grid tie systems, house and "roof over travel trailer".

Analysis of a Whole Year's Figures

12  month total March 1 2019 to February 29th 2020: 2196.15 KWH made; [7927 KWH consumed from grid]

   In the usage figures it must be considered that I have an electric car, which is always charged at home. Owing to my country location I drive around 900 Km a month even tho I only drive every second day or so, getting around 7 Km/KWH (10%(?) less in winter weather). Assuming 85% charging efficiency, that's 151 KWH/month as a rough estimate. (For comparison, the 500 watt heater in the trailer to keep it from going mouldy uses 360 KWH/month! And it doesn't keep the trailer warm enough to live in.)
   The wind power generated and used was too little to bother about. If I get a grid tie inverter for the windplant, or if I once again make more use of the battery system, it will contribute some small increase to the totals, but while we occasionally get very strong winds here, there are way more calm days than windy ones, and winds providing over 100 watts from this windplant are even more rare.

   Apparently all the renewable energy made over the year was only 22% of my total electricity usage. Well, I knew there wouldn't be much solar energy in the winter around here with the short, cloudy days and the sun just 15° above the horizon at noon - which is of course just when a lot of extra juice is used for heating, even with the woodstove supplying the bulk of the needs. (That may change! Open loop air heat pumping dangles the promise of very cheap electric heat.) Using the above estimates 1812 KWH were for the car. One might say the solar electricity if nothing else at least paid for all my energy costs for driving it, with 400 KWH to spare.

    Most of the solar energy was harvested in the 7 best months from March to September 2019: 1854.87 KWH. In that same period 3395 KWH was used from the grid so the ratio was 35% solar to 65% grid power. The car would have been around 1057 KWH of that, so home use would have been 2338 KWH, of which 800 came from solar.

   It might almost be said that in the worst four months put together (November, December, January, February), less electricity was made than in any one of those seven best months (220.76 KWH).

   So if I was serious about, on average, supplying all my needs with solar, I would need about 55 solar panels. That many would make up in summer for the deficit in winter. (Effective heat pumping would reduce that. Arguably it could save 25% of the total power consumed, which would bring it down to 41 solar panels.)

Things Noted - February

* As February went on, the days lengthened, there was (a bit) more sun and it was higher in the sky, and solar output increased substantially, from "very little" to around 5 KWH/day - which is still only about 1/3 as much as on sunny days in the summer.

* The total for the year from March 2019 to February 2020 shows that with just 10 to 12 panels, the solar output was only 22% of the energy used - not insignificant, but not a major replacement, especially in the about five months with the least sunlight.

* There was only one sunny and nice day in January, and one in February until the 23rd. Even well into March it forze every night, and there was snow. Egads! (No wonder I haven't been milling the last of my log into lumber!)

* The windplant spun for some hours on several days, but I didn't make any use of it except to run a little heater in the garage - just to keep it from revving up too high. The heat could be felt by the heater, but was trivial for actual heating - tens of watts. (Now if that could be multiplied by COP of 10!...)

Electricity Storage (Batteries)

Turquoise Battery Project: Long lasting, low cost, high energy batteries

Dip-Plated Zinc Electrodes

An impressive stack of zinc ingots cast
from discarded boat/ship protective anodes

   A reader in Oregon, Peter Carlich, who has been collecting zinc said he wanted to make batteries. He offered to cast zinc electrodes. I said it just needed thin zinc with lots of surface area per volume, which didn't lend itself to casting. I said I was going to try electroplating thin copper sheets with zinc. He then offered to dip plate copper sheets in melted zinc. That sounded promising and I said "Sure!" He has proceeded with that and is sending me some samples in the mail. He also took some pictures.

Zinc melting setup.

The backs of the sheets were protected from plating with a clay mixture,
and the fronts were covered with flux so the zinc would adhere well.

Zinc plated copper sheet.

Powder Electrodes Conductivity

(3rd) I took the nickel manganate mix from the previous electrode, dried it, and torched it. I compacted it to 5 tons and replaced the other electrode with it. I still didn't get good conductivity from it. Well, it was just graphite powder for conductivity.

   In reassembly I noted that the zinc sheet now had some holes in it. Perhaps it was just too thin a piece, or perhaps in pressing so much of my weight on it I had squeezed out some of the gel and left it vulnerable.

(4th) I then took the one with the graphite felt (which was still in three main pieces rather than crumbled into little bits), dried it and torched it. I decided to use a new sheet of zinc for the other side - and simply a sheet, no processing. I wouldn't worry about cycle life, amp-hours capacity and dendrites for the moment.

Solid Porous Electrode Briquettes - Plaster

(8th) I had been racking my brain trying to figure out what I might add to an electrode to keep it from swelling when liquid was added to the assembled battery. On this morning suddenly the words "Plaster of Paris" entered my head. I thought, maybe that's what the electrode needs. It sets solid when it dries, but it's still porous.

   I had also decided to try mixing some short strands of carbon/graphite fiber into an electrode. I took my bag with just a few loose scraps and cut all the fibers into lengths of 1/4 to 1/2 inch or so. They didn't need to be very long to touch each other and to touch the current collector to make random connections all through a 3mm thick electrode. It took a few minutes, and it all added up to a whopping .35 grams. I added that to the powder from the previous electrode that had had only graphite powder conductivity additive. The total was 6.05 grams.

To this I added 1.95 grams of powder from a bag labeled "casting plaster", total 8.0 grams. (This I had purchased some years back with a like quantity of high temperature silica powder potentially for aluminum casting.) Then a couple of drops of the sulfonates-rich dishsoap and a little water.

   This mixed into a nice paste, which I put in the compactor. It seemed too wet. From experience I thought it would ooze out through every tiny crack if I put it in the hydraulic press, but I did anyway since that might mix things even better, at the micro level, and then I would redo it. Instead only a little liquid oozed out. The powders stayed and made a ~2mm thick electrode.

   Then I set it by the fan from the woodstove to dry and set the plaster. Four hours later it had developed a mottled appearance with white areas - perhaps "not so well mixed in" plaster patches?

   I pulled it out of the die. I turned it over and found out why it hadn't oozed out: the graphite fibers had oozed down and filled the gaps. That I turned it over without it crumbling or breaking was in itself a good sign. The bottom face was uniformly dark.
(9th) The dried electrode measured in the hundreds of ohms range with the meter leeds across any two points on the surface on either side, although it was quite variable across different points. This was encouraging since they are often far higher once the electrode is dry. (When it's wet, one is measuring ionic conduction in the water, not the electron conduction in the substance. This always measures as a low, but in an electrode meaningless, resistance.) I measured a piece of dry Ni(OH)2 electrode (doubtless well discharged) from a commercial NiMH "D" cell. It was upper tens to lower hundreds of thousands of ohms - a couple of hundred times less conductive. All that graphite plus the graphite fiber strands certainly must have worked, and the plaster didn't seem to have hurt it. And nickel manganates has higher conductivity than nickel [oxy]hydroxide to start with. The potential is there for very high currents.
   And in these manipulations it still hadn't crumbled or cracked. Yay! The chief question now was: would it swell and lose conductivity when wetted in the cell, or hold its form? On finding out that plaster could take 72 hours to completely set, I decided I'd better leave it for another day or two before trying.

Some theory/ideas

   What was "Plaster of Paris", anyway? I was only sure it must contain calcium. It turned out to be calcium sulfate hydrated with just one H2O molecule per two CaSO4's. When wetted, it "sets" by turning into the dihydrate form, CaSO4 . 2 H2O, which is solid.  - also known as gypsum or plaster. It takes some hours to set, and apparently "up to 72 hours" to be completely set.
   (This form of calcium is as opposed to calcium oxide/CaO/lime/quicklime. When wetted this turns into Ca(OH)2/slaked lime/calcium hydroxide, which in the open sets to limestone/calcium carbonate/CaCO3 by reacting (if wet) to CO2 in the air. It is the base active material for cement, mortar and concrete. One reverses the reaction and turns CaCO3 back into CaO + CO2 by heating it to 800°C - which I have done in the mini kiln to make CaO.)

   It seems likely or at least possible that the layer of CaO I've been putting in my positive electrodes could react with the organic sulfates in the dishsoap to form gypsum and organic oxides. (Eg, Ca(OH)2 + ? H2O + sodium lauryl sulfate => CaSO4 . 2 H2O + sodium lauryl oxide.) But I don't see how that would disperse to solidify the electrode. Maybe I'll try mixing a bit of CaO (lime powder) into the electrode next time instead of quite a lot of CaSO4 . 1/2 H2O (plaster powder). Or maybe some of each. Well, let's see what happens with this one first!
   Of course the above substances are all electrical insulators AFAIK. But they are porous and should allow ions from the electrolyte to pass. At the ultimate, one might encase the compacted electrode and current collector in a "box" of cement or plaster all around the outside. Theoretically that should keep the electrode from swelling, yet allow electrolyte through.

(10th) I wonder how a layer of cement or gypsum would perform as a separator between electrodes? In fact, I think that would be a good way to do it. Paint the layer of CaO / Ca(OH)2 onto the face of the electrode. Then instead of putting the cell together immediately, leave it a couple of days to set, to react with the CO2 in the air. Maybe a couple of coats. Then there's a thin, insulating and insoluble but porous surface layer of calcium carbonate/CaCO3/limestone - on the electrode. That should completely replace the need for a separator sheet or grille. The jell layer on on the zinc would also help insulated the common surface. And the limestone would be quite a thin layer compared to a grille, making for fast response to sudden current draws.
   And maybe a bit of limestone mixed in would be better than plaster for holding the electrode together, too? Limestone has the appeal of being completely insoluble in water, whereas plaster is slightly soluble and might gradually migrate.

   A small aggravation is these things taking days to set. I tried to put on a second coat 3 or 4 hours later, but it just lifted and mixed with the first one. Well.... on to something else then!

50x50mm Electrode Compactor

   The high conductivity measurements seemed to portend near readiness to make production size cells, something a little bigger and heftier than the test cells. I rather liked the 100x100mm size idea. (4"x4" - hopefully good for a few amps.) But that would need a lot of pressure to compact, and even a 1/2 inch steel compactor might flex a bit if not bend. I thought that instead one could make 50x50mm electrode bricks and just put 4 of them together in a square on one current collector to make 100x100. I could also do 50x50mm and 50x100mm test or small size cells and (eg) 150x150mm (6"x6") or even larger large ones (up to what the 3D printer could do - 150 or maybe just barely 200x200 mm).
   Thus the idea to make a 50x50mm compactor punch and die. I had a big piece of 1/2" thick stainless steel. Someone happened to be discarding it at the refuse trasfer station just when I was there looking for something else. Yay! (And just after I had bought a piece of ordinary 1/2" steel! Stainless will be better.)
   There was lots of the stainless to make as many compactors as I wanted. It occurred to me that one could stack four compactors in the press and compact them all at once. They would all need the same pressure - just 1/4 of what a single 100x100 compactor would need. Another idea to save for later. One would be enough work to make.

   I had thrown the "stainless steel" plate on the grass outside the storage room months back. When I picked it up... the bottom was rusty! But it was still shiny on top. What was it really? Two layers? I decided to use the other steel after all. I spent some hours cutting out the pieces for the die and the punch. I would file and grind them to exact dimensions another time.

(22nd) That time came over several rather grueling sessions of filing and grinding, several hours in total. On this day I had the die to the right size. I think.
(25th) I finished it this day. I took considerable pains to get it perfect. I must have done well because on the last edge filing, the punch that didn't quite fit in any direction, suddenly fit through, equally in all four directions.


(11th) The new graphite foil tab had cracked and folded the same as the previous one and became virtually disconnected. Obviously I needed a new cell with a layout that would work. However, I stuck in another piece of graphite foil, and while the connection to the electrode was never satisfactory or consistent, it was there. (Wiggling it around or otherwise adjusting always made the voltage under load go up or down.)

* Short circuit current was just .25 amps, but there was almost no fade after 10 seconds.

I did a load test with 60 ohms, and without much charging beforehand. It started out over 1.4 volts, but in under 10 minutes fell below 1.3 volts.

* The voltage stayed over 1.2 volts for almost an hour. That was a big improvement over previous electrodes.
* I kept trying to get better connections, which made at least tens of millivolts difference (up or down), including opening the cell a couple of times. In doing this I noted that the electrode brick seemed to have not swelled, and it hadn't broken. It remained a one-piece brick, which could be shoved around. Yay! (I'm not sure why it was smaller than the space, that it had room to move!)
* I took out the separator grille and tried just the two electrodes together with the calcium oxide/carbonate layer for separation. It didn't work. If it wasn't a short circuit, there was at least some significant conduction between them.
* I replaced the separator grille with a piece of parchment paper. The voltage went up by about 25mV. (We'll keep the parchment paper, then?)
* The conductivity still wasn't very good. A pound or two of weight on top helped considerably. (No doubt it would with a gold plated current collector, too.) OTOH gold plated copper proved less than reliable in its first test. I clamped another stiffer alligator clip over the one on the graphite and got a few more millivolts.
* After an hour and a half (~60 mA-H), it hit 1.0 volts, and it didn't go much longer. I could visibly watch the drop accelerate. It plunged from .95 to under .90 volts in seconds - just like a real battery that was right out of juice.

   Voltage recovered in 15 minutes to 1.271. Then I put it on charge (@ 1.943 V). This time the voltage didn't jump up to near a full charge voltage, only to 1.6 volts. It started charging at 100mA and gradually over minutes the voltage rose and the current dropped. In 5 minutes it hit 1.750 volts and was still drawing 65mA. This certainly seemed like a huge improvement over any previous. It was taking a charge the way an electrode - and a cell - should.
   After 15 minutes I realized that I had the current meter on 0-200 mA, and that that meter range itself imposed a consderable resistance in the circuit. I moved it to the 0-10 A range, and I tried briefly shorting the cell. Short circuit current went up to half an amp, albeit with considerable fade over 10 seconds. Then voltage and charging current went up and charging was going much faster, initially back up to 80mA. (or should I say, on this scale, .08 amps?) The cell hit 1.9 volts (@ .03 amps) in 11 more minutes.

(12th) After charging some again the next morning (I left it off for the night), the voltage quickly dropped into the 1.7 volt range. Anyway I ran a load test. It wasn't as good, running not much more than an hour.
   Then I charged it again. This time the charge voltage was up to 1.93 and off charge it stayed over 1.8 volts for a bit. But short circuit current was down under 1/2 an amp, so I wasn't optimistic. But I ran another 60 ohm load test, and this time the voltages stayed higher longer and it lasted exactly 2.5 hours before dropping below 1.0 volts - almost 50 mA-Hours. Now that's at least starting to sound promising! Would it get better yet, or was that it? Or would it start dropping? At least it hadn't started dropping like MnO2 did by the third test.
   A six minute test then stayed above 1.5 volts but I shut it off. Then 1.5+ volts for 8 minutes and over 1.4 for 45 minutes. It ran for an hour late that evening with the highest discharge voltages yet. It looked like it would go at least 3 hours, and maybe longer. (That would have made it a very late night.) So apparently it improves with cycling - as I see it, turning the Ni(OH)2 and KMnO4 into nickel manganates (NiMn2Ox, Ni2MnOx) - and probably some extra KOH and O2 gas. Whatever is actually happening the chemistry, and the plaster and cement making it a good brick, now look clearly like winners.

   I also 3D printed a 50x50mm case in nylon. I filled it with water to check for leaks. Later there was a puddle under it. But had I filled it too full, and the water crept over and went out a screw press? I wiped off under it and filled it again, not quite so full.

(13th) The nylon case had swelled up with the moisture and the flat bottom was now dish shaped. So much for nylon for cases! And PLA had disintegrated in an alkaline environment. That took me back to ABS and needing a 'heated oven' (or a noisy heat gun blowing at the work) for the lengthy printing times - which are only longer with the Cura slicer than with the old Skeinforge.

   I also did another 60 ohm load test. I had taken it off charge some hours previously, and while I "topped it up" for 15 minutes, It looked like it was only going to be a mid-range for performance, losing voltage more quickly than the last two great tests. But I ran it down to 1.0 volts, and along the way the voltage drop slowed to a crawl - about one millivolt drop per minute, then even less - taking 3 hours and 33 minutes to drop under 1.000 volts, having delivered around 63m mA/hours of current.
   At 1.051v I put a couple of pounds of weight on top of the cell, and the voltage rose to 1.073. (It obviously gave some more time because of that.) Once again, the better the internal connections, the higher the voltages under load, and I'm sure the more milliamp-hours that will be supplied. I don't think I've done enough yet, even with the plaster and cement, to make them stay properly conductive. Perhaps very solid bottoms and tops of the cell cases, glued together under some pressure, would be better. After all, that's why so many batteries come as cylinders that can keep it all tightly compacted. Perhaps that's what it needs. In fact, one suspects what drove manufacturers to cylindrical cells in the first place.

(14th) Cylindrical Cells?

   Making flat electrodes and cells seems so nice and logical, and so simple - until confronted with the fact that they definitely seem to work best if well pressed together. In particular, the positrode brick needs to be well pressed against its current collector. The best thing to hold pressure is a sphere. That's impractical. Next best is a cylinder. But how would I make cylindrical cells utilizing my techniques?

1. Simplicity would suggest doing it way of typical standard cylinder cells: a layer of zinc plated sheet electrode around the outside, a single separator sheet, and a fat nickel manganate plug with a carbon rod in the middle. Some single-piece metal can would have to be found. It shouldn't be hard to paint it on the inside with the osmium dope layer - just pour some in, swish it around, and pour it back. Then put in a round piece of cardboard to cover the bottom. The electroplating would be similar: fill the cell to the desired level with the plating solution, put a zinc anode in the center, and plate for a desired number of amp-hours. For the agar jell, again fill up the cell, but heat it so the jell doesn't set until after it's been poured back out, to get a fairly thin coating. Next roll up and insert the separator sheet.
   There would have to be some jig to insert a compacted cylinder of NiMnOx mix. Perhaps a thin walled plunger with a piston of some sort to compact it and then press it out. (It seems Sanyo did it as three short cylinders, even in an 'AA' cell.)
   But this construction doesn't make for a very high current cell.

2. The best cylindrical battery would probably be made with rolled-up electrodes like Ni-MH and Ni-Cd cells. This makes for far more interface area between electrodes. The electrodes in the Ni-MH 'D' cells were about 50mm x 500mm, for a total of 250 square centimeters of surface. And both surfaces of the interior layers are in contact with each other, adding a lot more and making it perhaps 425 sq.cm. That's where they get the "50 amps" momentary current capacity, where the Mn-Zn alkaline cells 'only' put out about 10 amps shorted. Both figures probably represent 100-200 mA/sq.cm. of interface surface. (Mn-Zn in salt only gives around 5 amps in spite of the whole circumference of the cell inside the zinc being the interface area - 80 mA/sq.cm.)
   It would be no problem rolling up a thin sheet of zinc plate. For the plus side, perhaps I should be thinking of compactors that make semi-circular sheets, in diameters - or even spirals - that approximate the shape they will be? Then they can be pieced together in assembly, and the whole stuffed into a cell case?
   Of course, there needs to be electrode on both sides of a piece of flexible graphite gasket, or of gold plated metal foil, or else the strands of carbon fiber must to trusted to bring the current to the top edge where the graphite sheet or gold plated metal will connect and connect to the top button. Of course, gold plate would make for fabulous current current capacity.

   The thing that I keep thinking is that it all has to be done so gently. The jell and zinc plating must not be scratched. The separator sheet can't be breached or torn even a little bit, anywhere. I almost feel it's beyond my capacity. I can hardly imagine making and setting up the specialized equipment that would be required to do it reliably. Could I get them done by some battery company that already makes them like that? They would need some additional processes too, so it would be a very substantial investment on their part, too.

Okay, How About Flat With "Rebars"?

   So... the problem with flat is that the top and bottom will bulge and the electrodes, especially the plus internally and to its current collector, lose connectivity even if they don't swell. How can the top and bottom be kept from bulging? One way of course is to make them very strong. But that would make the whole cell fat and heavy. Another way is to have connections, "posts" connecting the top to the bottom at many points along the interior.
   This is sort of what "pocket electrodes" did in old flooded alkaline cells. There was a big plate, but it was divided into narrow sections with walls where the two faces were connected together, so the plate was effectively a rather narrow set of tubes that each couldn't bulge much, keeping the mix pressed together and against the perforated metal pocket which was also the current collector.
   I was trying this some years back but with ABS plastic pockets instead of metal. Hmm... maybe that wasn't such a bad idea. The zinc plated foil electrode doesn't need to be pressed - just the powder mix one. The problems at the time were (a) chemistry problems and (b) I couldn't 3D print "porous" plastic faces, that is, with zillions of tiny holes. Those problems are solved now. If I make a single porous plastic electrode shell with lots of 3D printed pins going through it between faces, it can be single or double sided to interface with one zinc 'trode or two - one on each face for a "double" cell that isn't much thicker.
   Rather than "walls" separating narrow pockets, I would do the posts, which would glue the faces together at many and various mid points. Then instead of a narrow current collector in each pocket (all needing to be connected together) there's just one big one, with a hole punched in it at the position of each post. And likewise for each 50x50 mm (or whatever) electrode piece. If there was a post every centimeter in both directions, the 50x50 would have 16 posts and of course the outside walls. Every square centimeter of both faces would be supported at all four corners. When I went to make it, that seemed like a lot of posts and I only put in four, 16.7mm between posts and posts-to-edges.
   And the cell construction could be pretty simple: just straight side walls, no special strength needed for the top and bottom. Drop in the bottom zinc 'trode (if double cell), then the porous plastic shell (single or double faced), the the top zinc 'trode. The shell is the electrode separator.

(18th) Looking at the electrode shell I had made, I got the idea that if I extended the back plate out, it could have the exterior cell walls and be the bottom end of the battery. The porous piece would then fit inside it and glue against the bottom end of the battery. I designed and printed this. It wasn't quite right, and the printer didn't bother to print my tiny support pieces and so messed up the area above the electrode slot (printing over air). I adjusted the dimensions and the supports changed the support for the graphite "+" tab in OpenSCad for next time.

The "new ideas" flat cell, bottom and inner porous 'shelf'.
* The "+" current collector (with punched holes) would be placed
in the bottom with a connection tab sticking out one corner.
* The "+" electrode would be fitted into the 'dish' in the porous
inner piece, then the inner piece would be pressed into the cell,
with the electrode against its current collector.
* Then the plastic pieces would be glued together, (including the
inner posts) with a press pressing them together.

Once the "+" electrode is done and the pieces are glued together,
the zinc "-" electrode would be placed on top of the porous divider.
Then a cover (3D printed or a simple sheet) would be glued on top.
(Hmm... a filler hole... in the cover?)

   Another thing that occurred to me was the ABS isn't the world's stiffest plastic. The thin porous face was definitely going to bulge between the supports. PLA and nylon didn't work, but perhaps something else might... What other printer filaments were there?

(25th) Internet service having returned, I finally checked out filaments. There was a new(?) one called "PVB" - polyvinyl butyral. Unlike "PVC" it only had carbon, hydrogen and oxygen so it should be environmentally benign to burn for disposal, and it was said to be stronger, more rigid, tougher (break resistant) and more heat resistant than PLA - which is itself stiffer than ABS. And it looked from the printing parameters, similar to PLA, that it might not need an oven enclosure or a heat gun aimed on it during 3D printing. (It's the plastic with which car windshields are laminated for safety, and apparently it sticks to glass - presumably if it's put on hot. The Wikipedia article said it had few other uses owing to processing difficulties and didn't mention it comes as 3D printer filament. So I added that in.)
   I ordered a roll. It definitely costs more - around 5 to 7 times the price of PLA. I trust that still won't be a big part of the overall cost of the battery. Will I have to put something on the printer's glass bed to keep the object from sticking to it even after it cools? Also, nothing says it's resistant to an alkaline environment, so I'm taking a gamble on that. (I was surprised to find that PLA disintegrates in the battery.)

(26th) I found a write-up on PVB by a company making the filament. They seem to think PVB brings 3D printing to "a new level". It seems very promising. Someone at one manufacturer says "PVB really is a dream 3D printing material ..."

Here is a Write-up on PVB (long, wordy excerpt chopped mercilessly to essentials):


"PVB printing filaments ... considerably improve processes and quality in 3D printing ... compared to PLA and ABS filaments ... while simultaneously preserving the environment."

Cupro-nickel Current Collector?

(20th) When I was using just KCl salt for electrolyte, both copper and nickel oxidized. It seemed I couldn't use any metal as a current collector. In using weak KOH alkali, making it about pH 13, I noted that cupro-nickel seemed to work and not noticeably degrade. Now I was using a mix of KCl and KOH, but it was still about pH 13. How would cupro-nickel fare in that electrolyte? Had I not tried that? Probably not. It would be easy to try out in the present test cell (which I had just abandoned because the graphite current collectors kept breaking with the sharp bend - but had not yet thrown out), so I cut a piece to size and put it in.

   I noted that the nickel-manganates 'brick' with plaster in it was still in one piece, even if looking rather fuzzy around the edges. It didn't have that tendency of all the previous ones to break into bits.
   I had hoped (and rather expected) that currents would go up with a metal electron collector, but they seemed about the same as with the graphite.

   After a few hours charging I ran a short load test, just 20 minutes. It wasn't dissimilar to previous ones with a graphite collector. Again checking the pressure, I had a fair bit of weight sitting on the cell, but once every minute I leaned on it. With the extra pressure the load voltage would go up, 70, 50 ,40 millivolts. And drop down again when I ceased. Likewise, in later short circuit tests, without leaning on it it might yield 370 mA dropping off to 260 over 10 seconds. If I pressed my weight on it, it would hit 500 mA and only drop to 470. Obviously the pressure, probably mainly of the brick being pressed against its current collector, was a key factor in a successful cell. It becomes apparent why battery makers cram every bit of active material that can be fitted in into round cans that can take a lot of pressure and strongly press everything together.

 Cupro-nickel current collector after several
              days of use: still shiny!

   I ran and cycled the cell for a few days. Performance seemed to be dropping. Was it the CuNi corroding? There were other possibilities. How was the zinc? It looked okay. But the little cover was off and the cell was drying out. Then in the test leads, an alligator clip fell off, and another seemed to have a bad connection. So I didn't finish disassembling the cell to look at the cupro-nickel sheet at the bottom that time, but performance was still down and I did later. It looked perfectly clean and metallic, unchanged except for a slightly more coppery color in one area, but not as much as the one in simple pH 13 alkali. And it had run many cycles. It worked!
   That also says that the monel (or other cupro-nickel) powder should work as a conductivity additive after all. Might it be better than graphite/carbon black powder? That'll be something to try out. And maybe not so much plaster?
   Another thing to try out is gold plating on the cupro-nickel sheet. This would be only to improve surface conductivity (and hence hopefully current capacity), the gold could be very thin, and it wouldn't matter if the coverage had some gaps. In fact it should be easy to try that out in the present test cell.

Haida Gwaii, BC Canada