Building a 100 Year Refrigerator

[marx290]

Haven't been on the forum in some time; I have been busy. Began a project in June to construct a modular, long lasting refrigerator. It has gone through many iterations, and there are many more to come. The refrigerant is propane. All heat exchangers are intended to be passive. The evaporator is currently in a freezing brine tank to store cooling power; that is likely to be part of the final design. The compressor is about 1/10 hp, 120VAC, but it will eventually be changed to a DC model. As I work on improved performance and develop better manufacturing techniques, the machine will become more robust and easier to work on. I'm pursuing an aesthetic quality as well; it will be installed in the top of a hand built wooden refrigerator cabinet some day.

The system runs 1 to 1.5 hours at a time, and 3 to 5 times per day.

I have experimented with several different ideas including two-phase thermosiphons and paraffin phase change materials in the condenser side. All had marginal success, but were scrapped due to either limited benefit or unnecessary complication.

I thought this was a good place to share some pics as I just built a new condenser yesterday, and I'm going to take a week or two off to get some reading done. I may start a side project which is an intermittent absorption ammonia refrigerator.

Thanks for taking a look.

This one one of my first jaunts into capillary tubing, Charts to reference guidelines for lengths don't normally include r290, so I was kind of guessing at first. Unfortunately, I guessed far too large of a bore, and wasn't understanding why cutting the length down wasn't getting me anywhere. Not to mention the high compressor amp draws and ridiculous discharge temps. It was a troubled time. I'm in better shape now. I believe that r404a is a reasonable comparison for sizing cap tubing for r290.

This setup was originally built to cool a brine. The brine was used to test a two phase thermosiphon. I was successful in chilling a bucket of water remotely using this thermosiphon containing propane as a working fluid. Think of a thermosiphon and a heat pipe wrapped into one. Later, I cut a hole in a small minifridge to begin a modular cooling unit project.



I got a little fancier with this design. Back in June of this year, I decided I would begin to actually construct a machine I have been thinking about for a long time. It is intended to eventually be a modular cooling unit which can be installed in a a variety of refrigerator cabinets or in a custom made wooden cabinet. The design principle of dropping it into the top has obvious parallels to the GE Monitor Top of the 1920s and 30s, as well as a few other manufacturers at the time who were making their machines more easily serviceable; it was a different time. I also enjoy the obvious benefit of discharging the heat at the top, where it belongs. The evaporator underneath is immersed in a strong solution of propylene glycol and water- intended to act as thermal mass. Later versions used a much weaker solution what begins to freeze at about 26 F. This has the effect of holding the cabinet temperature down for several hours.




Limited surface area of the brine tank was resulting in poor temperature recovery when opening the door or adding heat load. Rather than modify the tank (Like I would do later on), I chose to experiment with a larger plate heat exchanger. I had two aluminum plates and no way to easily solder copper to them, so I just made several serpentines with the tubing and sandwiched it between the plates using some bolts. This evaporator was quite effective at pulling the cabinet temperature down rapidly, but lacked the holdover of a some thermal mass. I have considered adding some phase change material to a plate evaporator like this, but to date I have not done so.



I left the evaporator alone for a while to focus on the condenser. I have been a long admirer of any kind of static heat exchanger. I am trying to build as many of the components myself, and an aluminum finned air cooled coil just doesn't fit the bill. I was going to have some fun with this. I believe technology should be beautiful craftwork, I tried my hand at some fancy copper work. The hyperboloid shape was chosen because I simply like the shape. Figuring out how to make it was another matter. I drove from Portland, OR to Portland, ME to deliver a car, and along the way I worked this weird thing out. What I liked about it was that the hyperboloid curve was not cut, it just is a natural result of these straight lengths of copper arranged in this way. Pain in the *** to build though, and future versions will be assembled differently.





This is a new evaporator and brine tank. There is simply a coil of 1/4" copper tubing in the bottom of this 2" half steam table pan. To increase surface area I soldered stainless steel fins to the bottom and sides. the tank is filled with a weak solution of propylene glycol and water which begins to freeze at about 26 F. This will hold the temperature in the cabinet for many hours.


Because the machine now runs for perhaps an hour every 4 to 8 hours (depending on ambient temperature), then all those hours of heat load need to be discharged in about an hour. I like this type of cycling- more or less, as the evaporator is better fed when it runs for a longer period of time rather than short cycles. I decided to try a phase change material on the condenser side as well. I figured if I could store cooling power in the brine, I could store discharge heat which could be dissipated over several hours at a lower temperature. Didn't exactly work out that way.

I chose petroleum jelly as a phase change material. The hydrocarbon chains of white petroleum are generally shorter than candle wax or canning wax, but longer than mineral oil. This makes for a melting point around 90F. 6 pounds of the stuff in a glass jar with 50 feet of 3/16" copper tubing made for my condenser. During the run, the hot gas condensing would melt the jelly into a clear liquid. But how to cool the wax before the next cycle? I chose to use a two phase thermosiphon- again with propane. Just for an added measure, I used an air cooled coil at the top. Several wraps of 1/4" copper tubing in the jelly was connected to the coil up top. Liquid propane vaporizes in the hot jelly and condenses in the air cooled coil up top. The system ran at pretty high head pressures and subcooling was a serious problem. On a positive note, the thermosiphon worked beautifully!


White petroleum is a poor thermal conductor- liquid or solid. I tried a more direct heat exchange between two cone shaped coils sandwiched together with some sheet metal cones. The lower one is the VC condenser. The upper one is the evaporator of another two phase thermosiphon. The air cooled coil of the last thermposiphon seemed to work so well, I decided to use the hyperboloid condenser this time.

Subcooling was still a major issue, but again the thermosiphon worked well.


Continued on a later post.

[jeff5may]

I love what you have done and how you are improving on everything. Not only from an efficiency standpoint, but also an aesthetic and rugged nature to everything. This is how great products come into existence.

Keep in mind how modern refrigeration condensers are built. For the actively air-cooled type, being compact isn't much of a problem. For the passively cooled designs, however, there must be lots and lots of surface area. The more surface area, the cooler the high side will reach equilibrium. Most of the manufacturers have designed their passive condensers to be able to digest the entire heat load at a certain topout temperature, just in case junior opens the door and the thing tries to cool the world all day long. As a result, duty cycle is maxed out, and the balance point is optimized for a hot garage in summertime. At "normal" operating parameters, the sizing is well into diminishing returns territory.

Man, that hyperboloid form looks marvelous. I would be tempted to run that frame as a condenser, without the coils around it. All the parallel flowing spokes should act as chimneys. It begs for something to be woven between the spokes for performance and visual appeal. Unique, for sure.

[IdleMind]

I'm glad you posted your project, I am especially interested in why you choose that compressor. Was it for longevity, efficiency, availability? I have some opinions about compressor efficiency and thought I would post them here for anyone to challenge.

I believe for low cost small displacements reciprocating piston compressors like the one you are using are more efficient than low cost rotary compressors. The oscillating piston in a rotary compressor has possible leak paths at both ends, on the OD where it contacts the chamber wall as well as where it contacts the vane. It could also leak around the vane. The reciprocating piston compressor can only leak at the piston OD. In larger rotary compressors I believe the total length of potential leak paths Vs the displacement volume becomes more favorable. I speculate that is why reciprocating compressors are used in refrigerators and rotary compressors are used in larger AC's. In fact I think the only advantage of rotary compressors is cost and possible less vibration.

Another difference is in a reciprocating compressor the cool low pressure inlet gas enters and fills the housing, in a rotary compressor the inlet gas goes directly to the pump and the housing is filled with the high pressure hot outlet gas.

I am working on small hot water heat pumps and am trying to determine if one type is more efficient, so far I have not been able to convince myself either way.

[stevehull]

Idlemind,

Interesting perspectives on rotary vs reciprocating compressors.

My intuition is the opposite as a reciprocating compressor has to overcome the piston inertia at each cycle. The rotary has no such issue as it always spins in one direction.

Secondly, I have been told that rotary compressors "seal in" early after initial use and that the leakage around the vanes gets less and less over time until it is virtually unmeasurable. Perhaps this is wishful thinking on someone's part . . . .

But the inertia issue is real and somewhere in a class many years (decades?) ago, I calculated the amount of advantage a rotary compressor had over a reciprocating one just due to inertia. It was surprisingly large.

Thoughts?


Steve

[jeff5may]

Rotary "rolling piston" compressors are the most efficient of the compressor types, but are more complex and difficult to overhaul than recips. Scroll compressors are less efficient, but are much quieter than the other two at equivalent capacity. This efficiency difference isn't large, and depending on capacity and speed range, each type trades places with each other for bragging rights. Multi-stage, intercooled recip compressors are more efficient than the other two, but are EXPENSIVE. Then again, they are easier to overhaul.

Way back in the day, all refrigeration equipment was expensive, most manufacturers put recip compressors in everything they made. During the jet age, much research was applied to the refrigeration industry. The manufacturers figured out the rolling piston compressors are less expensive to build, and are more tolerant of liquid ingestion than recips.

[stevehull]

Jeff - thank you. Your excellent comments prompted me to go back and relearn some compressor fundamentals.

Attached is short paper that I "thought" I had saved and I did. It goes over the scroll compressor history and briefly the construction. The comparisons of different compressor are listed and contrasted - but not rolling piston types. It is a quick read and I found it helpful.

Is there a similar paper on rolling piston types out there?

http://dms.hvacpartners.com/docs/100.../811-20065.pdf


Steve

[stevehull]

Here is another excellent pdf that shows failures in compressors and does show the rolling piston, scroll and true piston in diagrams.

Again, easy to read and lots of helpful visuals.

http://hvacrknowlagecenter.homestead...ompressors.pdf


Steve

[jeff5may]

Steve,

Both papers are super useful! I learned new things from both of them. I knew scroll compressors hit the market after recip and rolling piston compressors, but I didn't know it wasn't until the 1980's until they started showing up in manufactured units. The second document is all killer, no filler. The pictures show stuff you never want to see in real life. I just wonder what kind of noises those units made when they failed!

The OEM's are still inventing new applications of the scroll design. The latest innovation is the dual-stage scroll. Basically, the manufacturers have begun to use the intermediate-pressure ports in the scroll in a new way. Instead of simply opening the ports to the suction-side internally to reduce capacity, they have turned the ports into intermediate pressure taps. These ports are piped outside the compressor shell, enabling system designers to use the ports however they see fit.

The new Gree Terra series use the taps to inject cool gas into the compressor, increasing compression ratio while lowering discharge temperature. This enables their air-source outdoor units to dig deeper into cold-climate territory, providing full capacity at temperatures where other units cannot. Xringer installed a shiny new one of these this summer to replace one of his Sanyo units that lost a control board. It will be interesting to find out how the new design performs this winter. It gets pretty cold in Boston, and his 2 Sanyos had trouble keeping his house warm during cold spells without tripping circuit breakers and such.

Marx290,

Please post some words to go with the pictures. For example, what's up with the hourglass-shaped hx? And why did you put so many spacers on the cylindrical one? Are they heat spreaders, or the fasteners for some sort of cover? That's a huge heat exchanger for such a small compressor. Not so much the bulk of it, but the tubing diameter.

[marx290]

The next two are pretty recent. The hyperboloid condenser was neat looking, but it didn't have enough tubing or surface area to effectively dissipate the heat. Saturation temperatures on the high side were sometimes as high as 120 F or more. Subcooling was usually about 10 degrees. Circular construction is going to be a theme for this machine; I may eventually metal spin a brine tank for the evaporator which is shaped kind of like a bundt cake pan. The base construction and hole in the fridge will likely be round too, along with the bent laminated wood collar which will hold the controls and gauges, and will act as a cover to access the plumbing and wiring- if necessary.

This was originally supposed to have a sheet metal skin on the inside and out with the intentions of creating a draft for more effective cooling. That is why the posts are shaped and drilled that way. First attempt at assembly used 3/16" copper rather than 1/4". Lacing all those stainless steel posts (hand cut with shears and drilled on a press then bent with a cheap break) was exceedingly difficult. Imagine putting 16 keys on a key ring, but spinning them 16 times and feeding an additional hole with each one on each wrap. Think it would bind a bit? You would be right. I was able to complete the lacing, but it had a horrible taper and when I tried to fix it i just destroyed any symmetry it had.

I couldn't find anymore 3/16" tubing around town the next day, so I settled for 1/4" Drilled the holes out larger and even a little sloppy to make assembly easier. It has a slight inward taper bottom to top, so skinning it would have been difficult anyway. I decided to solder the copper to stainless posts and install it as is.

I'm happy to say that i have reduced the head pressure significantly, and improved subcooling to about 20 to 23 degrees. The ambient temperature in the shop varies dramatically, and It is getting a lot colder lately, so consistent data is hard to gather. I have tried a makeshift skin on the inside and out. It produced the draft I intended, but it wasn't enough I guess, and turned out to be a detriment to the performance. I like the exposed look anyway.

The next version will probably use 3/16" tubing. It will have a hyperboloid shape. I may split the 50 ft into two parallel runs, inside and outside coil, or I may run them in series. I will drill the holes, but then split them with throatless shears to give me fingers to apply the coils to and solder. More stainless steel surface area bonded to the copper would surely help.

[marx290]

Wow! A few days away from the computer, and a nice little conversation ensued. I don't have a strong opinion about reciprocating versus rotary. I would have to agree with SteveHull about the inertia point, and add that somewhat like scrolls they are in a compression stroke while in a suction stroke. I can see how they would be cheaper to mass produce. I take the side of whichever one is more reliable- regardless of efficiency. I say use a better built compressor and insulate the cabinet more.

I added some captions to the pictures to better explain what's going on there.

Jeff5May, I did use a large bore of tubing on the condenser. I explained why in the updated caption. It was supposed to be 3/16". I have my reasons for wanting to go smaller. What are your thoughts on it? It's about a 1/8 HP compressor.