Single-stage vs. autocascade, explain it to me :)

[Stigma]

Hey all,

I'm trying to get a better understanding of the differences between an autocascade system and a single-stage.

What I know is that both use a single compressor unlike "regular" cascade systems which is (from my understanding) basically just single-stagers with different gasses serial-connected to eac hother to produce lower temperatures than what is possible with a single-stage. After that I am a bit fuzzy on details, so I'm hoping a friendly soul here at XS can explain them to me :)

More specifically here are a few questions I'd like answers to:

- How can you cascade with only a single compressor? I know it involves using different gasses that separate or something, but how exaclty is that done in broad terms?

- Aside from build complexity and cost issues, what are the drawbacks of an autocascade compared to a SS if the specs of the system (compressor power, condenser ect.) are the same for both? My own suspicion is that the tradeoff for the potentially higher temps comes at the cost of reducing the totla heatload it can handle, is that right?

- How much more involving is it to make an autocascade compared to a SS? I mean both in terms of the work-hours to build and in the material costs for a finished unit. I obviously don't expect exact numbers here, but a ballpark estimate sorta thing.

For the purposes of explaining these things to me in a way I can understand, I can say that I already have a good understanding of how a single-stage works (thanks to the XS forum ;) ), so explanations need not be completely layman style ;)

Thanks in advance for any answers recieved.

-Stigma

[[XC] 2long4u]

An auto cascade uses one compressor to compress a mix of 2 or more gasses. The compressed gasses go through the condenser condensing one gas. It passes through a phase seperator where the condensed gas evaporates condensing the other gas. Then that condensed liquid goes into the evaporator. I may have a few things incorrect, but I'm not completely wrong.

[tim-]

- How can you cascade with only a single compressor? I know it involves using different gasses that separate or something, but how exaclty is that done in broad terms?

an autocascade have compressor and atleast 2gases. we make it easy and say it have 2 gases. gas A and gas B.

both gasses gets compressed by compressor. go thrue the condenser. Gas A condens to liquid, Gas B is still gas. Thise goes to phase sep. Liquid (gas A) falls on bottom of phasesep and gas (Gas B) go to top. A pasesep have one inlet in middle and 2 exhaust one at bottom and one in top.

Gas A Goes from phasesep bottom to an heatexchanger by expansionvalve or captube. Here gas A evaporates and gets cold (~-30 -50*c) and then it goes back to compressor.

Gas B goes from phasesep top (as gas) into the heatexchanger as well. But other "side" to exchange and get cold. Here the second gas condense to liquid.

This liquid goes via a captube or expansionvalve to evap and evaporates. and then back to compressor as well.

if you have more stages you got more heatexchanger (one for each gas you want to condense).

this was a simple explanation.. search and you will find more and better explanations.

gas A and gas B need to have different boilingtemps and you need to be 100% sure the firststage gas (Gas A) will condens in condenser and secondstage gas (gas B) will condense in heatexchanger without condense in condenser. If firststage gas doesn't condens in condenser you will not have any capacity too cool gas B in heatexchanger. It's plenty of work to tune an autocascade proper. for example is r290/co2 common in cheaper autocascades and some seems to get them work okay (lukeXE is proberbly one of the best on co2 autocascades). but there is some issues with dry ice forming inside co2 suction as it often hits' critical tempture/pressure in autocascade. r23 is a better choise as second stage gas but cost lot's more then co2 but still not any crazy prices compare to whats used, r1150 may work but nothing I have been dealing with.



- Aside from build complexity and cost issues, what are the drawbacks of an autocascade compared to a SS if the specs of the system (compressor power, condenser ect.) are the same for both? My own suspicion is that the tradeoff for the potentially higher temps comes at the cost of reducing the totla heatload it can handle, is that right?

it's hard to say, but one thing is for sure a regular vapochillunit will not work proper as autocascade, to small. you need to go up on same size of compressor (or a bit bigger) and condenser and such to make a cascade with separated stages (with 2 or more compressors). everything depends on what you're aiming for. a cascade with too small compressor, condenser, hx and such will give you worse temps then a SS, but a proper sized autocascade will give you much better temps.

[/b]- How much more involving is it to make an autocascade compared to a SS? I mean both in terms of the work-hours to build and in the material costs for a finished unit. I obviously don't expect exact numbers here, but a ballpark estimate sorta thing.[/b]

well material cost depends, I wouldn't build with coil Hx, I always choose plateHX for heattransfer. a 12plate hx cost ~45euro and is a good size for our applications.. almost every cascade uses those today in simular size. then you need a phase sep (cost a ~40-50euro) for a homemade one as copper is very expensive today (thanks china ;)), two captubes/valves and more piping is need. also a bigger condenser would be a good suggestion..in time you have 2 captubes to adjust instead of one if you choose to use that, if you have valves they will raise price.. then you have to put atleast a couple of hours to play with gas ratio to get the best possible.. this needs to be done under load.

do you want better answears.. let me and everone else know.

regards
Tim

[slavic1705]

Ok I'm going to give this a go...

http://img212.imageshack.us/img212/8242/autocascadebnumbzx7.jpg

1. Both gases are mixed and pumped to the condenser.
2. The first stage gas condenses and due to gravity flows to the bottom of the condenser. The second stage gas will not condense and is still a hot gas that flows to the top of the condenser which leads to the separator.
3. If there is any stage 1 gas left once we get to the separator it is turned into a liquid and falls to the bottom while the hot stage 2 gases stays at the top.
4. The liquefied stage 1 gas is fed through a filter/dryer and into a metering device (cap tube).
5. The stage 1 gas is released into the first evap (what’s that called?) and the gas boils off causing the 2nd stage gas in the coil to finally condense.
6. The second stage gas is finally fed into a metering device (again a cap tube) which runs to the final evap and boils off at a very low temp. Causing our CPU to be very happy.
7. Everything is fed into the suction line and pumped back to the compressor to start all over.

Please correct anything that is wrong. :shrug:

[Big SturL]

You got it right! :D

It could also be an idea to implement an SLHX before the phasesep (or is another location more optimal?) to gain better subcooling.

[Stigma]

Man, I just have to say, XS rocks ;) That was all very educational.

I get the impression that the difference in material costs between a "overkill" SS and an autocascade isn't necessarily so great, but buildtime/effort is much greater.

The only question I feel I wasn't addressed was "Is a probably tuned autocascade capable of carrying as bit a heatload as the SS with the same basic specs and still reach lower temps?" If anyone wants to tackle this one with a gut estimate that would be great, but thanks a ton for the already great answers guys :)

-Stigma

[Freddie123]

There have been auto cascades that have done -100 with a single compressor. Though it was rotary. I'm fairly sure with a decent sized hermetic (SC18/21CLX) you could get much lower temps then that of a SS with the same compressor

[Clemmaster]

SC18/21CLX aren't decent sized for that. GS34CLX could do wonders ^^

[Stapler]

Man, I just have to say, XS rocks ;) That was all very educational.

I get the impression that the difference in material costs between a "overkill" SS and an autocascade isn't necessarily so great, but buildtime/effort is much greater.

The only question I feel I wasn't addressed was "Is a probably tuned autocascade capable of carrying as bit a heatload as the SS with the same basic specs and still reach lower temps?" If anyone wants to tackle this one with a gut estimate that would be great, but thanks a ton for the already great answers guys :)

-Stigma

Don't forget gasses for low temps are expensive, so there is that added cost. I think the best auto temps are from Blaster's which did 155w at -87C, but I think he might have improved that.

[Stigma]

Thanks that pretty much answers it :)

Is it as safe to run an autocascade in daily use (not 24/7 tho) as a SS? If it is, then sometime in the future if I upgrade my cooling I might look into having en expert build me one, if I'm feeling rich that is hehe ;) Until then my sweet "little" overkill SS will hopefully serve me long and well ;)

-Stigma

[n00b 0f l337]

Yes it is, better oil return then a Cascade for sure.
Of course it has to be built properly ;)

[runmc]

Is it as safe to run an autocascade in daily use (not 24/7 tho) as a SS?

No Way near as safe!!! Don't believe everything you read:shakes:

[n00b 0f l337]

Well thats not true Ron. Safety is based on how its built. 100% on how its built, and who builds it, and integrated safety features. To say otherwise makes me wonder why we all have cars and fly in airplanes.

[runmc]

Well thats not true Ron. Safety is based on how its built. 100% on how its built, and who builds it, and integrated safety features. To say otherwise makes me wonder why we all have cars and fly in airplanes.

Well maybe you could show me an autocascade that was built for processor cooling on this forum that meets the qualification your speaking of. :up:

[n00b 0f l337]

Haven't seen one yet Ron but might happen soon enough. ;)

[Clemmaster]

Auto-C Is one of the most dangerous phase unit, far more dangerous than cascaded systems ... (especialy when using CO²)

[n00b 0f l337]

I think it's very hard and expensive to build a very safe auto cascade.
Safe might not be the right word either, safe could be a pressure cut off on high and low, sensors on the stages with calibrated temp ranges, and relief valves with blow offs to expansion tanks. Because its so safe it might not be able to run in any condition over 50F or something insane. So I think properly built can be taken as safe, or over safety.
Hurtling thru the skies in a projectile of metal and then using friction to stop you doesn't exactly sound safe, it just now sounds reasonable.

[runmc]

Is it as safe to run an autocascade in daily use (not 24/7 tho) as a SS?

Back to the mans original statement.


No Way near as safe!!! Don't believe everything you read

Remember he is comparing and autocascade using a high pressure gas with a single stage using a low pressure gas. :idea: :idea: :idea: :hehe:

[n00b 0f l337]

Okay as far as units built here so far, single stage is safer. But as far as possibilties go, autocascades can have even lower pressures then single stages.

[n00b 0f l337]

Yep a control system with solenoid to expansion tank can really lower the startup pressure as well!
Proper design is the key to sucess.

[Clemmaster]

Don't belive the discharge pressure is the same on both stages, you have to put at least 4 pressure gauges on auto-C and to control them regularly when running. The most dangerous part of an auto-C is the 2nd stage liquid line, any temperature increase on the line (for any reason) can become really dangerous, and you won't see it on a simple pressure gauge placed before the phase-sep ;)

[n00b 0f l337]

Thats an odd statement. By that your saying ever Autoc thats used commercially must have a technician who knows exactly how it works standing next to it all the time.
I dont think so.

[Clemmaster]

You can't compare it like you do, we don't have the same qualifications, the components and safety devices neither...