A Few Lingering SS Phase Questions

I've been doing a fair bit of playing around with small compressors.

Now often I like to tune 350w into a system because it's nice to know that whatever you use, you're covered except in rare cases (like the 990x and such, at serious clocks people are saying those are pushing 400w at times) and as a builder, I like that.

If you're planning specifics that include lower wattage cpus and especially if you're the builder, that changes things. You can always rework a system to suit.

So, planning your compressor to the capacity and temp you want, you can go quite small and use little power.

I built a quad compressor 'all in one' system...

http://www.overclock.net/t/894146/quiet-quad-compressor-all-in-one-unit

Using the smallest compressors around. Things out of water coolers, tiny fridges, etc.

1/12hp the smallest test, 1/6hp the larger for cpu.

That cpu compressor at 1/6hp was the biggest draw, and 150w I think was what it pulled for power. Nothing much really.

Specification is your friend in these cases. If you find it's ashrae rating and go from there, you can estimate your result pretty well. Maybe you won't get the 'perfect world' result that ashrae spec says, but if you size your captube close, have your condensing temp at or better than spec, then you can be pretty accurate.

So, you want -25c for instance?

Get a compressor capable of 134a operation (they're quieter than the higher pressure compressors, on average) at an ashrae of 150w at -23c.

Using a colder refrigerant, you can increase that capacity to 250w at -23 to -25 since it's the pressure it's capable of maintaining, not the temp itself, and that pressure will translate into the -23c ashrae rating pressure.

So if -23 at 150w is 2psig on your compressor, yeah? Then you move to 404a and 2psi is suddenly -43c. You increase your capacity to 250 (and this is where it's 'controlled guesswork') you assume an increase in pressure to around 15psi, maybe 20 at most. At that pressure, with 404a, you're seeing your -30c to -25c range you wanted.

So that's a pretty small compressor too. I'd have to redo my testing to get real numbers, but for the sake of 1-2 light bulbs worth of power, you could run a moderate temp on your cpu with phase.

Small compressors are pretty good for efficiency but as always, the lower the back pressure (or colder the temp depending how you look at it) the lower the cop on average. At least 2-1 for the most part, get into AC pressures, can see much better, 3/4 to 1.

So 150w giving you 250 to 300w. I think honestly I was lower than 150w with that compressor, and was pushing 300w+ in my testing.

The smaller compressors weren't as good at higher load, but would do 200w. If you could live with the temp results on a sub 100w compressor (power) then it's not a lot of power cost, or issue with any plug.

Went back for a read, guess the higher load result on the gpu compressor was 260w at 2c, with est. cpu temp of around 23c max. full load.

That's around 1/10 to 1/8hp. Those tiny compressors don't use more than 100w on average. TEC won't keep up to that kind of performance with that kind of temp and that little power use.

So yeah, otherwise that whole test bed was propane. For a test gas it's ideal, more or less a 'throwaway' gas (I re-use it, waste not want not) and no cert. needed. So the gas was pretty much free. The compressors were all pretty much free (if you pay more than 10 bucks for a water cooler used, you're not trying) and just the condensor I paid any amount of money for. One for the whole thing split up though.

So yeah, efficiency is great, just need planning for what you need to do with them and you can use small compressors that use really low power.

I'm not in the US though, can't comment on your cert's over there.

But avoid R410 for this stuff. Unless you get a 410 compressor it won't work nice in non-410 compressors, pressure ratio's too high.

Gray

Even at your higher wattages TEC's can work, and be a good alternative. Low load is always where they shine, those tiny portable fridges and an excellent example of where compressor cooling isn't a better option.

Just that if you get a really nice build going with a small compressor it shows how much more you can do with very little power.

I couldn't squeeze any more than around the 250w of decent result on those tiny ones though. I think 1/8hp is about where the ability to get useful 300w+ temps is. Anything under that, may as well use water

But I think a 24/7 folder, SB or something in a 4 core, should max out at 200-250ish OC'd. So an evap temp of 0c or so, cpu around 20c full load, should give enough OC to help account for the increase in OC and something like folding production. You've always got to account for cost I guess, and if it's cheaper to have 2 folding PC's than to have one on a cheap-to-run fridge with a better OC, then that answers that.

But my thinking was 'How small?' and the idea was to get low powered overclocks. If the overclock is cheapish to run, it's worth doing a folder with it. That extends into normal use too, but someone had said that phase just wasn't worth folding on. Power costs outweigh any OC work benefit. So I tried this out to see where I could go with it

And I think it's a step. Would be really interesting to run something folding on it that you know your workload, know your power cost, and see where it would stand with the OC + phase power added to it, to work out the % gained or lost, vs. speed/cheapness of having just the 1 folder instead of starting a farm.

But if you do it right, propane's pretty safe. Got to vacuum out well (air + fuel = risk) and leaktest really well but it's a -42c gas, close to 404a and a lot cheaper. I use it to test most things before I gas them with anything else, just in case there's issues, moisture, hard to notice leaks, developed microleaks, etc. In Canada it's 500+ for a bottle of gas, as much as 800 for some of them. So I like to be totally sure. 134a is good too if you're happy with moderation. It's a -26c gas and easy to get without cert's.

In the US it's a lot more reasonable for gasses, so if you do get an unplanned loss it's not quite as hard to take.

I don't recommend anyone trying to circumvent getting qualified, there are reasons for it, but at the same time the test for gas certification isn't meant to be so you can build or use the gas, just to 'hold' it. Like a guy working at a place that sells gas. The test is easy and means nothing.

More than that though, AC/R work pays well, so if anyone's young, not sure what to do, and likes it, look into 'real' training'. Mostly HVAC is pretty boring unless you get a diagnostic job, or something R&D or custom. But it pays well either way, and I find it interesting at times. I got into transport and in that you do a lot of diagnostic work. Not a lot of research type jobs for ACR here

(the usual disclaimer when I talk about this stuff and people who aren't properly qualified. It can be dangerous, isn't legal, and you should do what you can to obey your local laws)

Gray

Well it seems like Cali should be a good place to be an AC guy

First step, get your cert's for at least the basic stuff.

Get an Automotive AC course if they do one that's separate. GREAT summer money maker! Buy an old AC machine or just the bottle/recovery pump/vac pump/guages to do it, advertise, and get practice. Gas is cheap for 134a so you can make a lot at it even if you charge less. Get a car or a cheap little van and it makes it a 'mobile service' and callouts you can charge more for. Basic tool set from Sears or whoever and you're set. Cars are easy.

I do the AC stuff at work but I bought a machine for cheap that wasn't working right. Serviced it and replaced one pressure switch. It's made me 20g's over the last couple years at least, probably more. Just 1 per day for 3 months at 100 bucks each is 10k or so and I'd say around 80% is profit. Here it's a summer only thing. In your part of the world, closer to year round if it's a good year.

So there are ways to get going, and when you get some experience it's easy to get in to an apprenticeship for HVAC. The guys on the vans are always more likely to take on a guy who's clued in.

There. My recruiting speech
Edited by Gray Mole - 3/1/12 at 4:23pm

The old vapochill models used Danfoss 12/24v DC compressors and used the PC power to run them. They were good, but a bit small. The BD35 and BD50 back then. I tried to run 404a in the larger but it was tough. It ran but it was the condensing temp that killed it. Now, maybe I'd have more success. Experience goes a long way.

But usually you'll find that type in portable fridges. Larger ones really. The tiny ones are TEC. They cost a lot to buy new. Those portable fridges go for a fair bit too really.

If you could find an old Vapo you could use one. They're pretty long ago but there was a time they were out there and obselete really, and went for nothing. Came in a case usually. A really ugly beige one usually, occasionally black.

I remember there was a BD60 or 80 too. Never saw one but that'd be a good size. Cold when you want, low speed/low power when you need. They were brutal for price and likely still are.

COP isn't as good on those DC ones either but still better than 1:1



Keep in mind that 134a DC compressors sometimes don't play as nice with 404a. The BD35 danfoss would but the 50 didn't like it. No idea on the Samsung ones if you can, basically you try and see I guess

But there should be a spec sheet with dimensions and more details somewhere. I'll look when I get a chance too. Should be similar to the Danfoss sized DC for initial reference but really I thought they'd have that on the main DC compressor page

Im sure you posted it somewhere but I hate to read long posts?

What are you trying to do build a small phase system.