View Full Version : R410A captube sizing for 2.5 bar / 35 psi pressure drop

24-09-2003, 04:21 PM
Another weird problem to drive you guys nuts ;)

I am currently in the process of building a cascade to chill my CPU. A picture says more than 1000 words, so here is one:


There are some small errors in this drawing, but it gives you an idea of what I am trying to accomplish.

High stage will be running R507, evaporating at -35C .. -40C. Therefore, low stage condensing will be around -30C. Low stage refrigerant choice is not fixed yet due to obtainability problems and cost. However, it will either be R170 (ethane), R508b (DuPont SUVA-95, the most preferred choice) or R23 (not preferred due to high discharge temps)

To get some experience and to get the high stage working correctly, I intend to use R410A in the low stage before making the switch to one of the 'real' low-stage refrigerants. Why R410A? It features the highest volumetric efficiency at < -65C / -85F temperatures of all refrigerants I can find in an average ACR engineer's truck. I can also recover R410A back into a bottle, allowing me to play with the charge without dumping refrigerqant into the atmosphere. R170/R508b/R23 pressures are too high for that.

But since R410A is only an intermediate step used to gain experience, I would like to run the high stage as if I really needed the low condensing temperatures.

OK, that was the introduction. Now the problem: I need a capillary tube for expansion of the R410A, but the conditions are slightly unusual.

Assumed conditions:

- SCT: -30C/-22F (2.75 bara / 41 psia)
- Subcooling: 5K
- SST: -65C/-85F (0.51 bara / 7.6 psia)
- Superheat: 5K
- Qload: 150W

Which translates to:

- Pressure drop over the captube: 2.5 bar / ~35psi
- Refrigerant mass flow: 2.16 kg/hr (4.76 lbs/hr)

Does anyone has any idea about what size/length captube I would need?

24-09-2003, 06:12 PM
I'm thinking this can only be calculated with the SWAG method (Scientific Wild-Assed Guess), so here's my SWAG:

5 foot of .049

25-09-2003, 08:56 AM
I happen to just have some 1.2mm captube, which translates to .047. Close enough.

On what side of the balance do you think those 5 foot are? Too restrictive, or the opposite?

After all, I still have the needle valve on the evaporator (http://www.icecoldcomputing.com/text/show_page.php?id=66) to finetune the thing.

25-09-2003, 10:39 AM
I am guessing this would be sort of in the ballpark. What side of the balance? Who knows?

25-09-2003, 04:27 PM
I am guessing this would be sort of in the ballpark. What side of the balance? Who knows?

I don't. You don't. But I hope someone does. My personal estimate (based on Supco and Honeywell charts, correcting for pressure drop over the tube) was about 6 foot of .047". Your 5 foot ballpark is close, and I just wanted to check if you thought 5 foot would be on the short side or not.

25-11-2003, 11:01 AM
so just in case i am thinking way along the wrong track here... your building a cascade system to cool your cpu? as in the cpu in your computer?

25-11-2003, 11:16 AM
That's right.

BTW: to elaborate further on this subject: I have tried 5 foot of .047" captube, and it was WAY too unrestrictive.

Currently I am running with 7 foot of .028" , which is also not restrictive enough. As soon as I try to get the subcool from the low stage up by increasing the refrigerant charge, the suction line frosts all the way back to the compressor.

When my new apartment is ready, I will continue to improve the captube sizing.

25-11-2003, 11:22 AM
hehe.... i had a thought in my head for about 5 seconds of trying something along the lines your taking...

then i though of the noise of a compresser... brrrrrrrrrrrrrrr sitting next to me, maybe sit it outside.... nah whole in the wall.... too costly to fix

so thats as far as it went.

what the hell are you using for the evaporator?

25-11-2003, 11:33 AM
My evaporator (http://www.icecoldcomputing.com/text/show_page.php?id=58)

You might also be interested in the story I wrote (http://www.icecoldcomputing.com/text/show_page.php?id=68) about the cascade progress.

Noise? Ever heard a Danfoss?

26-11-2003, 08:18 PM
to calculate a cap tube, check www.supco.com look for captube computer model CTC3.


27-11-2003, 09:05 AM
I cannot find pricing information on that computer, but I think it is too expensive for single use.

Is there really nobody who ever poured Wolff's equations based on the pi-theorem into a free (or cheap) software program? I am missing information, otherwise I would have done it already a long time ago.