Johann's chiller [Archive] - XtremeSystems Forums

johann

11-21-2005, 06:17 AM

I picked up two beercoolers and decided to make a chiller with them. Basically I ripped out the big res/hx and built both compressors into one casing and made a heat exchanger for it.

Compressors are 2 x Tecumseh 1/10hp (I think)
HX is not tube in tube because I had no tubing big enough so I decided to make the best with what I had here. It may not be so effective but I think it should do the job.

Evaporator coils are 2m of 1/4 tubing and the liquid loop is around 1.5m of 3/8 with 1/2id barbs brazed onto it, The evap tubes was coiled closely around the 3/8 tubing and then brazed solid.

Gas is R404A
Cap tube is 10 feet of .028
Pump is yet to be bought
Dimensions is 14"(w) x 14" (d) x 12" (h)

Heres some pics:

Compressors mounted:

http://emerald.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/bare.jpg

I had to Dremel a hole for the second condenser:

http://emerald.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/dremel.jpg

Making and insultaing the HX:

http://emerald.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/chillerhx.jpg

http://emerald.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/hxfoam.jpg

http://emerald.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/hxinsulated.jpg

HX brazed on and insulation done with a quick testrun:

http://emerald.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/assembly1.jpg

http://emerald.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/chillercold.jpg

http://emerald.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/insulate.jpg

Unit in his housing:

http://emerald.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/housing.jpg

I have not done any testing yet, since I have no pump but Im working on that. Ill put my 220watt load tester on him to see what it can do.

mtb856

11-21-2005, 06:30 AM

Looks good..... the HX looked a little ghetto at first, but it turned out nice. Should be interesting to see the performance that you get from it :)

johann

11-21-2005, 06:48 AM

Looks good..... the HX looked a little ghetto at first, but it turned out nice. Should be interesting to see the performance that you get from it :)

thanks, Yeah the HX is actually a bit ghetto. If it doesnt work well Ill replace it with a coaxial one later

RussC

11-21-2005, 11:07 AM

Hi,

Nice system. I like the ingenuity of taking the beer cooler to a new system. Very nice.

I have a couple of questions, why is the cap tube so long? I guess I don't understand the hookup? Does each compressor drive an individual cap tupe to the evap? If so, I believe your evap is undersized for an equivalent 2/5hp system. You'll have to under fill the system w/refrigerant if thats the case. ~6.5ft of 1/4" copper seems small for that capacity of compressors.

Also, Im not sure if wrapping the captube around the evap is correct? Usally its wrapped on the evap outlet to control the amount of refrigerant into the evap much like a TEV? That part of the theory Im a little fuzzy on though.

I have small criticism, I not a fan of your fan layout. The compressors block a fair amount air flow to the condensers. But, size can be a factor.

Brazing the water lines to the evap lines should be pretty good, as long as the insulation is sufficient to properly shield the system together for proper heat transfer. Thats the key.

I look forward to your cooling results, great work overall.

RussC

johann

11-21-2005, 11:37 AM

Hi, thanks for the reply!

I made the captube as long as Gary suggests here: http://www.xtremesystems.org/forums/showthread.php?t=29694

I thought that would be a good starting point, Ill see how it does and adjust it if needed.

There are two evaps of 6.5ft each being fed by 9ft of captube. The systems run completely seperate from each other. I only vac and charge them together, that way they run exactly the same charge and pressure etc.

You say the evaps seem small, but the compressors are only 1/10hp? The load on them will be pretty high. Ill see how the compressors run and if theres liquid returning to them then Ill adjust the charge accordingly. This is my first chiller I build so I will change things if needed. Ill get the pump and load tester set up and see what she does, then see where changes are needed.

Wrapping the captube around the evaps is for extra subcooling on them, I dont think it makes a difference in flow. Better cooling there than in ambient.

The airflow comes from the front through the condensers and then over the compressors and out the back, I dont understand what you mean by the compressors are blocking the airflow to the condensers?

I charged it to 10psi static and removed the lines simply to see if it runs fine, Ill play with the charge later but it looks good sofar, I stuck the temp probe as deep inside the waterline as possible for measurement. Pulldown to -30c is really quick, I was amazed at how quickly it pulls down. Once turned off temp takes about 90 minutes to reach ambient again so I think the insulation is working well.

Im excited now to see how it does in the load testing, its doing better than expected sofar!

Thanks for the helpful comments RussC! :)

RussC

11-21-2005, 12:37 PM

I see,
OK, separte systems.

My point on the fans is just they need some area to flow air. If there blocked by the compressors, then it will be reduced. Im in the process of changing mine right now, more on that later. Again, if you need the space, then so be it. But overall not the best to me....

On the cap tube, OH MY, mines way short according to that thread on sizing :eek: Im running 58", by those calcs, I should be at double that with my 0.031" tube on Propane. Its in direct conflict with the Supco chart that comes with the cap tube??????

RussC

johann

11-21-2005, 12:58 PM

That guide by Gary is quite accurate, I think 58" is way too short.. Temps will be really low, might be why you are having problems with frosting on the compressor as it doesnt restrict the flow enough and all the propane doesnt boil off..

RussC

11-21-2005, 03:18 PM

No,
Im not a smart person. I opened my mouth before I knew all the info. 10' of cap is used with the smaller compressors, and by the Supco chart also. I have a larger compressor, ~9/40hp unit. Since my unit is much larger, the cap can be smaller to flow more refrigerant and still hold the proper pressure.

My frosting issues are totally charge related. I always over charge, then back off, its been easier that way. Ill have an update on my system later tonight :)

RussC

That guide by Gary is quite accurate, I think 58" is way too short.. Temps will be really low, might be why you are having problems with frosting on the compressor as it doesnt restrict the flow enough and all the propane doesnt boil off..

johann

11-21-2005, 11:33 PM

I dont know, that cap tube still seems very short, I dont think compressor size plays that much of a role.

Perhaps its worth asking someone like chilly1, he will be able to confirm that

SexyMF

11-22-2005, 12:32 AM

Nice work johann. Good and compact.

What have you used to insulate the foam?

RussC:

I'm using 1.2m of .7mm cap on a 1/3HP compressor.

johann

11-22-2005, 12:36 AM

Nice work johann. Good and compact.

What have you used to insulate the foam?

RussC:

I'm using 1.2m of .7mm cap on a 1/3HP compressor.

Thanks! I used Armaflex tape to insulate the foam on the hx. IMHO its the most versatile insulating material, you end up using it on everything

RussC

11-22-2005, 12:53 AM

Right,
The chart that comes with the cap tube says for my size compressor and low temp, 60" of 0.031". So thats what i did. The chart also has much longer lengths for the smaller compressors. Like Johann has. So theres obvious differences by temp and compressor size, according to Supco.

RussC

Nice work johann. Good and compact.

RussC:

I'm using 1.2m of .7mm cap on a 1/3HP compressor.

johann

11-23-2005, 02:59 AM

OK I have hooked up everything and tuned it with my 185watt load tester. It seems to work OK sofar.

The system takes around 1.5 litres in total, I used 50/50 antifreeze/de-ionised water, seems to work well and pump does not seem to slow down at that temp.

Im glad I can achieve these temps with little liquid, it pulls down really quick and the system is nice and compact. Personally I see no point in having a chiller that needs an hour to pull down temp. Theres still more than enough liquid in here to hold fluctuations in CPU temp really well.

Ill leave captube alone, seems to be fine a good balance between temp, load capacity. Also in real life with around 150watts heatload CPU should run loaded at around -5c to -10c which is fine.

Only thing I might still do is recharging it as this is still my test charge, I only pulled a vac for around 1 hour so pulling a proper vac and recharging might produce a decent enough temp change to make it worth the time spent.

Loaded lowside pressure is 12-13PSI

Pulldown is nice and quick

Idle liquid is -25c

Under load liquid is around -11c
CPU temps is around 0c

Unloaded:
http://brown.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/chillerunloaded.jpg

Loaded:
http://brown.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/chillerloaded.jpg

Loaded lowside:
http://brown.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/loadedpressure.jpg

johann

11-23-2005, 07:36 AM

I did some finishing on it

Added insulated tubing for the water loop with thru ports to connect tubing on the back
Put some grills in front of the condensers
Finished electrical wiring

pics:

http://brown.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/chillergrills.jpg

http://brown.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/chillertubing1.jpg

http://brown.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/chillertubing2.jpg

http://brown.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/chillerinside.jpg

dward3

11-23-2005, 09:47 AM

Nice work Johann. :toast:

RussC

11-23-2005, 11:28 AM

Johann,
Very nice system. You've done a great job. Looks very professional. Mine won't look nearly that nice in the end.

Um, back to the cap tube, I believe you can't ignore the compressor size in the cap tube calculation. The link you posted to the tube thread was in regards to a specific compressor, essencially a Mach I or II, and even to the Vapor series. These all run a similar size compressor, so you can use that formula for that size of compressor. The issue is that the larger the compressor gets, the more refrigerant it can flow per time period ie work it can do, or hp, watts, BTU kcal ect. If the cap tube is too long, the front pressure is too high and resricts flow, essencialy throttling back power. If the cap is short, low front pressure but high flow, giving a higher work output(there other issuse here though of evap flooding and condensing issues at low front pressures). Do you see what I'm saying?

In your case, the smaller compressors flow less, so the cap has to be longer to give a higher restriction with less flow(than a larger compressor), that way it can condense the refrigerant correctly.

If you look at the charts for cap, the higher hp compressors all have shorter and bigger cap tube, and thats correct. Becuse of higher flow, they can hold the same pressure with shorter/bigger cap tube. Again, I don't think that formula works for the larger compressors, it has to be adjusted accordingly.

In my case, the cap is still to short for propane, I set it up for R134a, but I had propane in it, so it wanted to be ~12" longer.

Again, nice system, look forward to some benching.

RussC
(holly crap, I think Im learning this stuff finally)

Unseen

11-23-2005, 01:55 PM

very nice bro :)

johann

11-24-2005, 03:52 AM

Thanks everyone :D

RussC, be sure to keep us updated on your captube lengt experience, it will be interresting to know what you find works best for you.

I have decided to perhaps take around 12 inches of mine to see if it increases load temp.

expansionvalve

11-24-2005, 08:11 AM

looking gooood that man, a nice little toy to play with!

RussC

11-24-2005, 08:46 PM

I don't want to hijack your thread, but Ill be using 72" in the new system. I wont be going back to the R12/R134a class refrigerants. It will all be propane/R22/R507 from here on out, so higher pressure and longer cap tube, for the size compressor I have. I should have the revaped system running I hope on Sunday.

I'll have updates on my "Well I have Frost" thread.

RussC

Thanks everyone :D

RussC, be sure to keep us updated on your captube lengt experience, it will be interresting to know what you find works best for you.

I have decided to perhaps take around 12 inches of mine to see if it increases load temp.

johann

11-25-2005, 12:11 AM

I look forward to the results.

I use R404a in everything, I have R22 aswell but have never ended up using it in anything.

johann

11-25-2005, 05:50 AM

Small update, chiller has been running loaded for about 24 hours now, nice and solid, temp range between -12c and -7c on the liquid and around -2c to 3c on the CPU, this is a 185watt load.

I suspect with a more realistic load of around 130-150 watts it should do between -10c and -5c on the CPU core.

Also flow is very low at this temp, Im running a 66/33 mix of Antifreeze/de-ionised water. With a better pump and more suitable liquid I think temps will be a few c lower.

Pic of the bench setup, I used a 1 gallon res because its the only way the pump works propperly, when inside his res. My findings when using a 1 gallon res is it affects temps negatively. Results with much smaller res was alot better. IMO the less liquid in the system the better temps will be. Obviously temps are more stable with bigger res but definately higher overall.

Pic of test setup:

http://brown.srv2.com/~sokkiesc/Fotos/Phase_Change/chiller1/chiller1running.jpg

Now next step is finding a new home for him :) I have another more much powerfull setup here that I want to use for personal use. I will build him next. Not sure yet if I want to go DD or chiller, still to decide :banana:

SexyMF

11-25-2005, 11:21 PM

bazx

11-26-2005, 12:17 AM

very nice job very neat and tidy

i took a differant beer cooler

beercooler (http://www.ocforums.com/showthread.php?t=420654)

johann

11-26-2005, 08:15 AM

With -20 coolant I can only get +7 on my CPU. I'm not sure about my waterblock mating with the CPU. I suspect its contaminated with the vaseline. I know the block is rather wrecked inside as it has been running with unfiltered tank water for ages (I'm rural) and has corroded badly.

Should I expect a negative temperature on the CPU?

I found that the difference in temp between the load tester and the liquid stays around 10c, so whatever temp the liquid is, the load tester is around 10c warmer. The probe is located in a small hole drilled in the aluminium from the side, pretty much 5mm below the evap and 5mm above the Resistors

Please bear in mind that the temps are taken with a digital thermometer from inside my load tester, the temp readings coming from the CPU may be completely different. Im not sure but I think the CPU temp reading comes from inside the core.

Maybe the old waterblock does not do such a good job or contact to the CPU isnt good.

very nice job very neat and tidy

i took a differant beer cooler

beercooler (http://www.ocforums.com/showthread.php?t=420654)

I have read that thread on your beercooler, its a much bigger model than what I had to start with. Your project is very nice with the aircon and everything, I would chuck that big res and heat excanger and replace it with a new one, the problem is with so much liquid, it takes too long for my liking to pull down to good temps. Also I would regass it with R404a which produce much lower temps and have increased load capacity.

bazx

11-26-2005, 12:19 PM

I have read that thread on your beercooler, its a much bigger model than what I had to start with. Your project is very nice with the aircon and everything, I would chuck that big res and heat excanger and replace it with a new one, the problem is with so much liquid, it takes too long for my liking to pull down to good temps. Also I would regass it with R404a which produce much lower temps and have increased load capacity.

i have a bottle of 402a that i have ordered to regas a moded vapo

how would this gas do against the 404a?

RussC

11-27-2005, 03:05 PM

R402 and R404 are very similar, from what I can see from the PT charts. Should work fine if R404 is OK.

That system you got going is AWSOME. Nice work, that is very impressive. Please keep us posted of your results, or better yet, make a topic here so we can stay abreast of your progress.

RussC

i have a bottle of 402a that i have ordered to regas a moded vapo

how would this gas do against the 404a?

johann

11-27-2005, 03:26 PM

I believe R402 is marginally better than R404A, although I think pressures and heat produced is higher.

bazx

11-28-2005, 12:02 AM

thanks for your input

this stuff is very expensive here in the uk so it is important i get the most use out of it

@johann what blocks are you going to use?

baz

johann

11-28-2005, 02:18 AM

thanks for your input

this stuff is very expensive here in the uk so it is important i get the most use out of it

@johann what blocks are you going to use?

baz

Swiftech MCW6000/6002 for various reasons:

1) Square simple all copper design, no places to leak and easy to insulate
2) In my experience, the best performer under low flow conditions, I find flow is much lower at the low temps, but the block still work well

baz, how much was the R402 here in the UK?