CryoBUG is Back

[mytekcontrols]

Hey it seems like things were getting a bit dead around here so I thought I'd try an liven it up a bit. And yes I'm back on the CryoBUG project, only this time with a new twist.

So what's up?

I stumbled upon an old patent by Andrija Fuderer. Actually I rediscovered it with new eyes, because over the years I've glanced at it more than once, but never really saw it for what it was. And believe me when I say it is a jewel. The patent was issued in 1965, and describes what at first glance appears to be a basic single stage autocascade. However upon closer inspection, it can be seen that there is much more to it. In his primary design, there are a series of sub-cooling stages being used in combination with what I call the Auxiliary and Cascade Condensers. He also goes on to discuss the possibility of utilizing something similar to a fractionating column for phase separation. But the real beauty of what he proposed was basically the idea of keeping it simple, using only a few refrigerants (as little as two) and intentionally picking out refrigerants with as much as 100C difference in boiling points.

Anyway I looked at what he had done, and created my own streamlined version as can be seen in the image below. For much more detailed information on what I am doing please visit my blog (Yep I finally got my site back up and operational). We can also keep the discussion going here as well.



Best news is that today I came pretty close to matching Fuderer's results on my first test. Nearly -110C with only R-134a and R-14.

[n00b 0f l337]

As always a good looking setup. I wish I was home in my workshop to try it out.
It seems like the normal one stager we've always discussed with a subcooler tacked on that isn't fed by anything but suction gases. So effectively a single stage and a suction line heat exchanger.

[mytekcontrols]

As always a good looking setup. I wish I was home in my workshop to try it out.
It seems like the normal one stager we've always discussed with a subcooler tacked on that isn't fed by anything but suction gases. So effectively a single stage and a suction line heat exchanger.

I think the key to making this work for very cold temperatures, is to keep the evaporator load small, otherwise the SLHX aspect probably wouldn't be very effective.

It was quite exciting today when I fired up my prototype for the first run. I wasn't really all that sure it would even work. Funny thing is it worked better then I expected, especially considering the small amount of R134a in use (100 grams or barely over 3 ounces). In fact I was watching the phase separator temperature, and it didn't get much below zero C, a pretty good sign that I was running out of condensate to expand on the way back to the aux condenser. So my thought is to boost the R134a up a bit, and lower the amount of R14 since I was doing a good job of flooding the evaporator. I also need to make room for the eventual addition of Argon in order take the system down even colder.

[Gray Mole]

Well it's sort of an Autocascade, and sort of not...
I have an AC window unit i got cheap and thought I'd play a bit.
Really the inspiration is Drewmeister more than anyone else. He's had a system for blending gasses that's given some amazing temps, and working in the science and medical cooling field of cooling, has the gasses and experience to make it happen.
There was a time that I wasn't a 'believer' but I realised that any limitations I placed on his work, I place on my own.











Took around 4 min to get the initial cooling working well, then a steady drop over the next 5-10 min to around -90c. Around 5 min after that, to -100c.
So this is a 1/2hp rotary, and the blend is 134a/290/1150. Enough of the lower pressure gasses to get started, and enough Ethylene to maintain the low temps.

Not pretty or even really well organised, just brazed in what was needed without a lot of fuss. Not going to be a finished unit how it is.

No load tests yet, but I'll clean up the evap and all that, get a bit better insulation, and go from there. I don't expect much, but if I can get even a 20% Ethylene 'cooling blend' to work, then that's enough to pull the temps down on load dramatically.

Currently around 1/3 of each gas, but I don't expect the Ethylene to remain at that level when it's entering the evap. Hard to explain exactly, but I am hoping that it can stay over 15% fully condensed Ethylene for actual evap cooling.

It's autocascading obviously, but not with phase separation as such. More of an enhanced blending system within the captube.
Wish I'd gotten pic of how it's been brazed, but the manifold in the pics is the start. The return gas uses it as the suction line heat exchanger, and then captube is a 2 part setup.
One end got crushed flat in a vise and brazed,and the other end I slid the coiled captube and 1/4" captube join/mixture chamber setup and crushed and brazed that with the inlet 1/8" and outlet 1/4" pipes sticking out.



That's a mockup of what it looks like under the insulation.
The pipes I used in the actual heat exchanger setup were cleaned and slightly 'scuffed' with emery cloth and cleaned again. Trying to get the heat transfer as good as possible in there.

The Captube is coiled enough to slide into the SLHX, and 2 kinds.
First length is 1/8", around 7'. Then it's brazed into a 6" piece of 1/4" copper pipe. Then it carries on as .031" captube, around 3'.
The 1/8" and 1/4" are coiled and brazed into the SLHX and it's a fairly tight fit.
The 31 is just wrapped, and I'm thinking of brazing that to the pipe.
No flexline for this though. No intention of this being a 'unit' just some experimentation. The gas blend is pretty standard to be honest. Started with a 40psi charge of 290. Then up to 80psi with 134a. After that, around 150psi total static with the 1150.

I added another small amount of R290, but my static is still around 130psi with a cold suction line, more likely 150psi still with it warm.
But I found a good blend, and the 6" piece of pipe seems to be working the way I'd hoped, making a chamber for the liquid low pressure stuff to help the 1150 condense, and then travel into the 31.
But the numbers speak for themself so far.
I'll be playing a bit here and there with this, try to find a blend of gas and a captube setup that really makes the high pressure blending work for me.
Don't know if I'll be able to get the Ethylene to stay condensing well enough to be part of the gas that's actually cooling with higher load, but it's a fun experiment.
I enjoy the research a lot more than the 'real work'
Gray

so far it's really only around 250w, and -45c.

Which is something, but really there's not a great deal of load capability, and not a huge drop in temps from a well made SS.

Though with it being a 1/2hp compressor, there is some interesting potential.

I've built Auto's before, and the one thing that always bothers me is the long wait.

However, Ethylene is proving to be far too aggressive to jump to.

So I'm still looking for a stable blend that'll go where I'd like, remaining within the realm of SS methodology.

I've been doing a lot with enhancing subcooling lately, and I think that's the only reason my result is as good as this.

Oh, little background on the load tests...

Started at 100w, just to get a feel for where it was.

Worked up to 250w, mainly adding 134a and 290 to keep the SLHX frosting through, minimum it was taking to make the whole thing 'load worthy' was around -15c measured centrally on the HX.

The HP is actually now quite resonable, being pretty diluted. 260psi on average, 275psi at a push.

when I go to over the limit of load, it's actually seeing a severe drop on the Lowside pressure, and a very fast rise in temps.

When it's dropped below the load it can handle, there's a very strong rise in lowside, up to 25psi Low, and the highside goes to around 325. This lasts for about 15 seconds, then it's back to 10psi/275psi until it's cooled.

The result of -45c is at 8psi.

What I like, is that the cooldown now with the higher mix of 134a and 290 is amazingly fast. 45 seconds to -60 on 2-3 hour restarts, and 15 seconds or so on immediate (2 min.) restarts.

So I'll keep working this method until I know if it'll go anywhere.

I'm considering a multiple chamber system on the next, to see if it makes any real difference or not.

But the main thing is the gasses themselves. I need a more progressive setup to make very good use of them.

Ethylene may just be too much.

But a way around the Autocascade's 'classic' design is what I'm looking for, and this is as close to a workable solution I've found so far.


Gray

Additional Comment:

Oh, and the last time I was trying this, was with R23.

It's too fat. I couldn't get anywhere with a consistent condensing design. I don't know if this would help but I think a lighter gas is more likely to actually condense in this system.

So Ethane is another logical choice, and very likely the best one for it's lower temp but very light mole weight.

I'm actually find a small success with Ethylene, more so than I ever had with R23. Hydrocarbons are really light, and really like mixing. R23 was a great Autocascade gas though Seems to like to stay separate, which makes phase separation so much easier.

Ideally, I need a good 2-3 way mix. More of a 'ultra long glide' than a 'step from one to the next' and I think the hydrocarbons are a way to get there.


Gray

Double posting again...

Additional Comment:

Well a last little note to wrap up.

Did the final tests on it, some variable load and after a day of sitting, I wanted to get an idea of how it would behave as an actual 'unit' in operation.

Cooldown with the blend being around 20% Ethylene is now phenominally fast. 10 seconds or less to -20, so the 134a is condensing almost instantly.

Bit of a hang there, then it goes to -50 in another 20 or 30 at most.

Then it's a slower cooldown to -60, another minute. The compressor, when cold, isn't dumping heat and it's not getting the pressure from that heatdump. Only reason I can see for it working a bit differently.

Ultimate cold temp is around -70, take a few minutes to get there.

Some variable testing gave odd results...

-55c at 100w, and the pressure from no load to 100w is marginal if anything.

-40c at 200w

-45c at 250w

Odd, but I think the extra pressure from the heatload give the Ethylene a bit of a boost. Only reason I can think of.

So I can see where I'm at, and where I'd like to go with it.

Ethane (I hope) will blend a bit easier, and I suspect will condense at a marginally more consistent level, and lower high side pressure. Not a lot, but enough to smooth transition I'm hoping.

Will stick to the 20% figure for now, with equal parts of 134a and 290 again.

So that's the end of this iteration of the project. Some very cool data Not a 'complete' success but I didn't expect that as such.

Knowing that I can get the HP gas to condense in a single captube without phase separation I just need to keep working it, and find a more suitable mix.

So the next one, will most likely keep the size similar on the chambering, but split it into 2 or 3 parts.

I'm likley looking at 2 x 3' 1/8" captube sections and 1 x 3' .031" section with 2 chambers of no more than 3-4" long.

If it changes anything, I'll know how it's affecting the system. If it changes nothing, I may look at a longer section of larger captube, possibly in the .050 to .060 range. Longer and more subcooling.

Next one will have a fully internal captube run. Should have done that this time really, instead of running the .031" externally.

really low HX temp should get the R23 to condense nicely though. I really struggled with R23 but subcooling wasn't something I was doing as much with before. I was a lot less experienced then. Years help, though my hair has suffered with age :

I have thought of going with R600/R290/Ethane to see if some kind of near Azeotrope could be formed but I doubt it.

I can't find any kind of -60ish hydrocarbon in the encyclopedia but I'll keep looking.

-40ish to -80ish is a big step, but I'm hoping that a hybrid of sorts will form in the mix, even a small one could be a gas to mate the 290 up to ethane.

We'll see. If I have any success with it, I'll be sure to share that. Though I always feel nervous about anyone else trying what I do, if they end up doing damage I feel somewhat responsible.

So yeah, next step is Ethane, a more integrated captube for complete subcooling of the length, and more of a 3/4hp or slightly higher powered compressor.

I hope it works. -60 at 300w is my 'target' for now, and anything more would just be a bonus.

[Gray Mole]

Sorry about vomiting that whole thing in one shot, but I'd done this project a year ago give or take.

Same idea, but I went with a small chamber and a 2 stage captube system.

It was a great very low load -100c setup. Would be ideal for a very compact sealed fridge for tissue sample/stem cell storage or the like.

Anything over 50w and it didn't have the jam to keep the Ethylene going though. I was thinking about more chambers, and perhaps larger interstage captube boiling chambers.

Maybe 2-3 suction HX's, each with a 2 part captube system. Perhaps a 2-3" join in each captube of maybe 3/8" pipe for a small boiling chamber to amplify the condensing of the next gas in line?

I never got back to it, though the project is still sitting there.

Don't know if you can see from the pics, but I think Mytek will see the basic idea. Different method but same basic idea.

I was going in the direction of a SS with Autocascading principles, but with a shorter cooldown than a conventional Auto.

I think that more gasses with varying and sequential BP's, and more geared towards low load and low temp, and you could make a refrigeration system that's very lightweight for storage and transportation of biomedical samples.

Cool to see someone else doing a similar project though

Gray

[Gray Mole]

Mine were really good.

Honestly, I didn't need a lot of Ethylene though.

I had to read back, but my max discharge pressure was 275psi, and that dropped with temps.

Compressor didn't get hot by Rotary standards.

I did see a rise to 325psi when I tried high load though :p which was why I was thinking Ethane instead for high load results. But it was for a few seconds only. Like a weird pulse.

But I did hold load with Ethylene. I was well below R290 temps throughout load. So as a 'Cryo fridge' is was an ideal setup with almost 'flash freezer' results.

In terms of smaller HP gasses for the application that it's really more ideal for, the charge of the HP gasses being lower would mean far lower pressures overall, and lower spiking if at all since a cryo fridge really doesn't apply high load whatsoever.

I figure anything 325psi or less on a 'real' compressor for R22 or better, it's no risk of lowering lifespan.

I think I saw 80c on the compressor at most, and only then when I was pushing the load. I was impressed. I think I got lucky

Gray

[Gray Mole]

Yeah, normally it really is. As a substitute for phase separators though, it's ideal. Oil control is a must though, since any oil issues are amplified.

Keep in mind though, that the flashing is very minimal, but you really want it for the next section since it will help the next section's condensing efficiency.

R14 is wicked for carrying oil and if you had to use it, I'd say a series oilsep setup would be a must.

I think it would work best with 2 captube sections per SLHX, and 3 SLHX's with the highest pressure gasses closest to the evap. Graduating the HX temp along with the captubes will minimize flashing.

You don't really want a full flash of the higher temp gas per captube interstage chamber, just an enhancement of the next captube stage's condensing. The flash effect, though marginal, will increase pressure marginally, forcing the next stage to condense more effectively, and lower it's temp just a little with it.

It worked better for my project than I'd hoped, I just didn't have the variety of gasses I needed for the gradual cooling effect that an autocascading principle to work well, and the step from propane to ethylene was a huge one.

Gray

[mytekcontrols]

Hey Mr. Mole thanks for sharing your most interesting system. Yes it does look like we are both experimenting with simplifying the AutoC approach. Your design reminds of a unit that Polycold produced called the Cryo Tiger. It used an SLX (or SLHX) placed within the customer's vacuum chamber, and some other tricks which I wasn't privy to, but in essence did away with the traditional AutoC's phase seps. The only problem was, and still is, that they keep eating up compressor's due to over heating problems. What kind of discharge temps were you seeing?

And yes the Cryo Tiger had a very rapid cool-down similar to what you mentioned.

[Gray Mole]

Oh, and just to be clear about what's different about my approach...

I'm using the SLHX method, and really one would likely do it, as long as there's ample room for the captube system to be internal...

But it's the captube interstage boiling method that I used, and I think it's why it worked so well. 5-10minute -100 cooldown was surprisingly good.

I recommend focusing on the caputbe layout.

Graduate the captube. The full length of the caputbe should still be close to what your target, but in stepped up sizing, and a chamber in between each section, so the the very slight pressure drop to each of the next sections of captube allows the 'warmer' gas to boil a bit, enhancing the condensing within the next.

So, in my next project (if I were to use multiple gasses) I would start at a short section of .026 - 1/4" x 2" pipe - .028 - 1/4" x 2" pipe - .031 - 3/8 x 2" pipe - .036 - evap

Or something like that. Each chamber with a slight pressure drop via the next larger captube's increase in size.

This should create a graduated drop in captube temp and therefore an increased condensing of the next gas in line.

Just my thoughts on it. I should patent it before it's too late

Gray

[n00b 0f l337]

I just cut apart a cryotiger before I left the shop.

Tricky indeed

[mytekcontrols]

Gray --- A much better description of the multi-chamber graduated cap tubes. But I must confess I'll need to stew on it awhile to truly understand how it really works. I'm use to the old school thought that flashing within a cap tube is bad. But I am also open to new ideas, although this old brain needs time to process.

Hey Adam --- I wish I had a Cryo Tiger cold head to dissect. Never got to see one really close, so some of it is a bit of a mystery to me. Can you enlighten us by any chance?

[Gray Mole]

Oh yeah, and before I pass out

Stick to low-mass hydrocarbons.

A graduated chambered pressure drop captube approach is faster, but far less effective than an phase separation approach. Gasses like butane, propane, ethane, ethylene, methane etc will give you WAY less grief to condense along the line, and put a great deal less strain on the compressor.

My thinking was something like 30% butane, 25%propane, 17%ethane, 15%ethylene, 5% R14, 8%methane

Or something in that region. Anything that you have to use as a 'bridge' like R14 with high mass, use a little less since it's 'fat' enough to need less of it.

I think that mix should give close to 50w of usable load, which is ample for a well insulated fridge.

The primary stuff being really low pressure should also shorten the compressor strain time. My cooldown on the initial -30ish was within seconds, and the pressure dropped phenominally fast along with it.
Gray

[mytekcontrols]

I already passed out, just talking in my sleep.

Yes I'll definitely have to give your ideas some more thought when I'm awake. But it'll also be taking a back burner until I've better explored my own project's possibilities first.

Thanks again for sharing your strategy, and have a good nights sleep.

[mytekcontrols]

Yeah I definitely get the idea now behind the interstage chambers, and yes very good oil control would be a must. In my design this shouldn't be a problem because I have a Temprite 340 as my phase separator for the R134a, which of course will virtually strip all of the oil out of the stream before it gets into the colder stage.

BTW I have an even larger step, R134a to R14 without any intermediary refrigerants. That's a good 100 degrees difference in boiling points.

[Gray Mole]

Looking forward to continuing this one

I'm curious to know what your cooldown times, pressure spiking and ultimate load capability was. I was focusing on higher capacity so 134 to R14 simply wasn't feasible.

But for a 134a - Ethane - R14 - Argon setup with under 50w capacity target in mind, I think that differential in gas properties wouldn't be nearly the issue it was for me with a 200w+ target.

Maybe I'll have to dig out the old project and do some revamping. I have Argon and ethylene here, maybe I'll see if I can make it at least give a 10w or less result

'Night

[mytekcontrols]

Initial cool down was rapid to about -20C, but then it slowed considerably taking nearly an hour to reach -100 C. The start up pressure head was barely over 300 psig, and the highest it got was 340 about 10 minutes into the run (with a 12 psig suction). After it hit -100C, the pressures mellowed, and the suction had come up to 25 psig. And after about 2 1/2 hours my head pressure was at 195 psig, with the suction still around 24-25 psig.

I really think the cool down can be accelerated by adding a bit more R134a, but I would also want to decrease the amount of R14 to keep the suction low after the system has completely stabilized. The evaporator flooded soon after hitting -100C, so this makes me believe that I can get by with less R14.

BTW all my initial tests will be done with no added heat load. Only seeing whatever static load is imposed by infiltration of heat from the outside world through the insulation. This is actually more load than the ultimate target application, that being radiant heat load within a vacuum vessel. So if I had to guess, I'd say I'm presently seeing maybe 5-10 watts.

[Gray Mole]

Yeah, passive heatload usually sits about 10w or less with just a smaller exposed area. Lessens once the 'snow' starts to insulate

That's really not that bad with R14.

Still think it's just a bit too aggressive. Though knowing your result I think that maybe a 134a -> Ethane -> R14 or something as a 'bridge' would make your cooldown a lot faster and pressure spiking less high.

An hour is a long time though, tough to say if it's too much or too little R14 but it's really 'fat' compared to the hydrocarbons I used. That's why I recommended them for this type of project. Condensing them has always been so much easier.

I had to read back, since it was around a year ago, and I was using a pretty high % of Ethylene but because it's such a light gas I could get away with it.

Beyond that, I had 134a and 290 to work together to get the Ethylene going.

I'd be interested now to try what I've done, but just rework the system into 2 x SLHX primary chambers, and have the 2 x Captube sizes within each chamber, running 134a/290 in the first one, and Ethylene/Argon in the second (though it really doesn't work that way, more of a mix throughout)

Maybe some day, a combination as well. Use that captube setup, and have phase separation in the first SLHX chamber so that the Argon mix is more pure?

Something I didn't have in my initial prototype that would greatly help with the 'spiking' on pressure was a heated 'boiling tank' for compressor protection.

The idea was to create an environment of gasses that blend slowly over the length of the captube system, that eventually become a 'pure liquid' phase blend of all of the gasses used, then when it's down to low temp, ONLY the HP gasses boil within the evaporator, and the lower pressure gasses are returning in primarily liquid phase, boiling slowly through the SLHX's.

This creates the danger of liquid return to the compressor and the source of the occasional pressure spiking I saw. I think that a tank of sufficient volume to boil any remaining liquid, heated by HX from the discharge, would nullify the spiking effect.

The only way for this to work properly is to have full liquid at the captube exit, and the R14 (or whatever your final gas is) to be the only gas that's actually boiling in the evap, so maybe an evap that's got a fairly open chamber design would be ideal so there's room for the R14 to boil AND the rest to be mostly liquid in the time it takes to get out of the evap.

More brain vomit to digest

Gray

[n00b 0f l337]

Next time I'm home Michael I'll take pictures of what's left of it. I got it for $25 on ebay because it was bent to hell at the connector ports.

Going from memory.
Wonderful tube in tube work.
I noticed only one capillary though, at the end to the face evaporator, the inner tube of the HX would simply be "pinched" and a tiny hole present in the side of the pinch. Looks as if they designed it to simply bleed off a little refrigerant at this restriction, but I could not find another capillary line.
I believe the vacuum sealing really allows them that nearly perfect adiabatic heat transfer, so exactly like Joule-Thomson effect.

[Gray Mole]

Here's a crappy paint drawing of what I mean...



Likely would be better to have the heating of the expansion tank after the oilseps instead of before, just my lack of paint skill planning, ended up being easier to draw like that.

Discharge muffler would help a lot with the initial pressures I think.

Gray

[mytekcontrols]

Adam quite a steal of a deal on that Cold Head. Since that is the "working" part of the design, it rarely goes for so little, because most people can glom up something for the condensing unit side. Yep I knew it had to be a JT system, but suspected that there was a bit more to it. Speaking of the condensing side; it is nothing more then an air-cooled condenser, fan, rotary compressor, coalescing oil sep, and what's called an absorber. The absorber soaks up any remaining oil after the oil sep and holds on to it. since the amount of residual oil is so small, the absorber lasts for a long time before needing to be changed. Same technology as used to drive Helium Cryopumps.

Gray, yep that diagram looks like what I had in my mind after our discussion last night. I think another option would be to substitute a single Temprite 340 for your two series connected oil seps (and move it to the other side of the condenser), then feed a cap tube from it to the top of your first SLHX (make it into a cascade). Even without an auxiliary condenser, it should help out without sacrificing too much cool down speed.

[n00b 0f l337]

Yeah that's why it interested me. I have a habit of finding things on ebay and asking for internal pictures of the polycold units. Sellers don't mind showing, and you can get an idea for products that way. There was 3 or 4 of these pinches and divot holes in the main double spiral, then finally the captube. And the holes were quite tiny, only noticed them because I was pumping N2 through the coil and noticed "colder" areas to the touch.
Also fairly restrictive I'd say.

But an interesting concept. Makes you wonder about all the magic

[kang-China]

I think that ,if you use the R22(R404a) +R14 you can do the temp -110.
this is better than use R134 ,R134 the low temp -26

[Gray Mole]

I think that ,if you use the R22(R404a) +R14 you can do the temp -110.
this is better than use R134 ,R134 the low temp -26

Getting the temps closer can really help, but I think what made mine work as well as it did was the use of Hydrocarbons with low mass.

Hydrocarbon 'want' to blend, and the low mass keeps the pressure down a bit, as well as condensing just a bit more readily.
In this setup as opposed to a true Auto, you don't want the gasses to separate. Quite the opposite, you want an amalgamation of them so they'll be 'all liquid' blended, then at the end you'll see primarily the HP gasses boil off when the temp is low, and the LP gasses will go their way as liquid through the evap, boiling as the temp in the SLHX's increases. That's also why I like the multi SLHX format, since the temp will be higher, further from the evap. Increasing the temp in the sections that need it most.

I think though, that the result won't be as cold as the final gas in a hydrocarbon blend. More likely a compromise between the 2 final gasses if they 'stick' to each other well.

Mytek

Yeah, I had a feeling you'd pick up on my approach pretty quick. It's not hugely different to Auto formats anyhow, but with your experience I think you can take and use it with your variation and get a lot more from it than I did.

If you went with 2 distinct 'sets' of gasses and had a single phase sep with the multi cap approach I think that a high speed cooldown with Classic Auto load/temps might be achievable. That would be sweet.

My biggest peeve with Auto's is the 1/2 to 1 hour of waiting for a result. Cutting that into a 15-30min max for Argon-like results would be nice. Closer to 15 min would be seriously marketable too, but I wouldn't count on it until some testing was done.

Oh, and i don't know if I mentioned, but the twin phase sep was all about R14. Oil separation seems to be a nightmare with that gas, so a series oilsep configuration may be the best way to keep that in check, even though I've seen some serious oilseps slowly fail, the oil creeping through.



Gray

[mytekcontrols]

Oh, and i don't know if I mentioned, but the twin phase sep was all about R14. Oil separation seems to be a nightmare with that gas, so a series oilsep configuration may be the best way to keep that in check, even though I've seen some serious oilseps slowly fail, the oil creeping through.

Yep R14 isn't really an HFC (hydrofluorocarbons) like some people think, but more accurately a PFC (perfluorocarbons) or sometimes also simply called an FC (fluorocarbons). So although the HFC's are readily miscible with POE, I'm not sure if the same can be said for R14. Although with POE you'll stand a better chance, then with AB or Mineral oil.

At one of the places I work, it is standard practice to retrofit dual or even triple series fed oil separators in units that have been experiencing oil freeze-up problems.

[mytekcontrols]

I think that ,if you use the R22(R404a) +R14 you can do the temp -110.
this is better than use R134 ,R134 the low temp -26

Hello Kang -- although your statement is true, the purpose of using the R-134a was for 3 reasons, as follows:
1) I wanted to test out Fuderer's design and match as closely as possible his first suggested mixture, that being R12 and R14.
2) I want to keep all the refrigerants ozone friendly and non-flammable (HFC).
3) I want to avoid using a mixture in an AutoC that is covered by Brooks/Polcold's patent #6502410, which pretty much covers what would be an ideal mixture using HFC's.

BTW I made it down into the -130's yesterday by using a combination of R134a, R23, R14, and Argon. But I also started seeing a problem with my flow for CT #1 not being enough (check out my blog for more info). So I'll be doing a little modification to my hardware over the next several days.

Gray -- I'll have to play around with your idea at some point, but for now I need to continue pursuing my present course and see where it leads me.