Killer Single Stage - Turns Autocascade 19300 Btu rotary build

[bartx]

EDIT: I decided to make autocascade instead SS. See next posts for details.

New build, this time I'm building really strong SS for myself. I'm aiming in -60/70C area, so it should be nice alternative for DI.

Used parts:

Evaporator: 42mm stepper rev.2 http://www.xtremesystems.org/forums/...Stepper-rev.-2
Compressor: Daikin YZG-56RBY1 19300 btu
Condenser: Roen Est 1900W with 25W Ebm Papst Fan
Expansion valve: 0.9mm captube
Refrigerant: R507
Baseplate: 1.5mm stainless 500x400mm, on wheels

So far, all things are mounted to the plate, most piping is also ready, I've encountered small troubles, because flex hose turned out to be damaged and I have to order a new one and replace. I'm also planning to make GPU mount, together with my water chiller it should be nice combo to cool whole platform. I'll keep you updated

[runmc]

Looking good.

[Buckeye]

Oh nice, a new Bratz machine !

[Splave]

love these threads good luck

[Gunslinger]

woot for the Daikin compressor!

[n00b 0f l337]

Very nice, although your expectation may be unreasonable. You may get solidly into the -50's.

[Buckeye]

I was going to say... I thought -60 to -70 range was auto cascade area.

[Xeon th MG Pony]

or conventional cascade, or multistage single stage systems need to keep below a certain compression ratio befor you need an intercooler and a seconday compressor in series.

[bartx]

Very nice, although your expectation may be unreasonable. You may get solidly into the -50's.

My last unit (13300 btu rotary) was doing -52C @ 260W load and -58C idle, so I don't think that -60C is unreasonable -70C maybe sounds a bit optimistics, but maybe idle or slight load like single core Sempron... we'll see

[Praz]

Not sure how long the compressor would last operating in the vacuum range necessary for those temperatures.

[Kevin Hotton]

bartx,

I would highly recommend you build an autocascade. The -70C design I presented would be perfect for what you want. I have been using plate heat-exchangers from duda diesel. They can special order ones with copper sweat ports. I especially like their B3-36a models (long and thin). I have built an autocascade using 16-plate models with a 24cc/rev 25000 Btu/hr rotary -- It has 750W duty at -58C and is used to liquefy CO2. With your smaller 19300 Btu/hr compressor I would order 14-Plate and 5/8 OD copper sweat. You would need two HX's (one for the auxilary condenser and one for the cascade condensor) and then buy/make the special evaporator needed for CPU cooling. This will save you all the trouble of bending a long tube-in-tube copper heat-exchanger. You should (with n-butane and ethylene) get over 500W cooling at -70C and no compressor overheating.

Kevin

[n00b 0f l337]

If you got -52C with a slightly smaller compressor, I think you'll get around -54, -55C with this.
I think you are WELL past the point of "diminishing returns" on this one.

[bartx]

bartx,

I would highly recommend you build an autocascade. The -70C design I presented would be perfect for what you want. I have been using plate heat-exchangers from duda diesel. They can special order ones with copper sweat ports. I especially like their B3-36a models (long and thin). I have built an autocascade using 16-plate models with a 24cc/rev 25000 Btu/hr rotary -- It has 750W duty at -58C and is used to liquefy CO2. With your smaller 19300 Btu/hr compressor I would order 14-Plate and 5/8 OD copper sweat. You would need two HX's (one for the auxilary condenser and one for the cascade condensor) and then buy/make the special evaporator needed for CPU cooling. This will save you all the trouble of bending a long tube-in-tube copper heat-exchanger. You should (with n-butane and ethylene) get over 500W cooling at -70C and no compressor overheating.

Kevin

It's really interesting idea - now I'm really thinking about morphing it to autocascade I'll need only heat exhanger (I was thinking about 10 plate exchanger, exact the same which is working in my chiller) and phase separator, which can be made from piece of 42mm copper pipe with two caps on it.

I've never made autocascade, but I'd like to try What gases do you suggest to charge? I've got R507, R290 (not 100% pure, but should do the job) and CO2.

[tiborrr]

This is a waste of a good compressor for a single stage unit, Bart.

I did a 250W capable LPG (cooking gas, filtered and dried) / CO-2 autocascade that held -60?C loaded. You may find links in my sig. If someone can give you a squirt or two of R-1150 that would be even better.

I don't think plate HX will work though, it has too low pressure drop. It needs to be more restrictive, more narrow. I would recommend tube in tube although it's a lot of work to do it.

[Crabby]

That is one heck of a compressor! I've never heard of Daikin so I looked it up, whoa! What did you pull it from?

[felix_w]

That is one heck of a compressor! I've never heard of Daikin so I looked it up, whoa! What did you pull it from?

Carrier & Daikin are the two top HVAC brands, afaik...

[Kevin Hotton]

bartx,

Plate heat-exchanger (HX) will definitely work, but not just any plate HX. That is why I specified the long and thin model B3-36a models (18.2 inch long and just 2.9 inch wide. The reason you need long and thin is because of the large temperature difference from end to end of the HX. With a short length HX the thermal conduction in the metal will degrade performance excessively.

Again, I have built and fully tested a morph'd 25,000 Btu/hr AC window unit using plate HX's and it works. The charge in this case was all non-flammable (R-123, R-22, R-23, and small amount of R-14). However, previous experience with autocascade designs have convinced me that flammable blends using n-butane, and ethylene (with perhaps some propane or propene) work better (higher cooling duty).

The main misconception that many people have about autocascades is that you can start with a common refrigerant (like R-22 or now maybe R-404a) and then just mix a much colder refrigerant (R-23, R-14) and it will work. I can tell you it will not work. You need to use something with a boiling temperature warmer than even R-134a (which is NBP -18C). That is why n-butane works so good (NBP is -0.5C). Blends with n-butane, propene, and ethylene work very well in autocascade in the -60C to -90C cooling range.

Kevin

[Kevin Hotton]

bartx,

If you do build an autocascade, you will need a fade-out vessel (to give space for the high vapor-pressure refrigerant to vaporize after shut-down to avoid building too much system pressure). I have used small receiver tanks from Refrigeration Research for this (specifically catalog model 5723 5"dia. x 12" length). This unit is about $75 and will take about one month to get (it seem they do not carry much inventory in this size). I have ordered four so far, so maybe they will begin to carry better inventory of this size.

Connect the vessel outlet with a captube (I use 0.031" x 12ft) to slowly feed-in the refrigerant charge during start-up. If you use this vessel there is a bit of retention of the lowest vapor pressure refrigerant (n-butane) in the very bottom below the dip-tube, so you need to slight bias your refrigerant charge higher in n-butane to acommendate this.

Kevin

[bartx]

bartx,

If you do build an autocascade, you will need a fade-out vessel (to give space for the high vapor-pressure refrigerant to vaporize after shut-down to avoid building too much system pressure). I have used small receiver tanks from Refrigeration Research for this (specifically catalog model 5723 5"dia. x 12" length). This unit is about $75 and will take about one month to get (it seem they do not carry much inventory in this size). I have ordered four so far, so maybe they will begin to carry better inventory of this size.

Connect the vessel outlet with a captube (I use 0.031" x 12ft) to slowly feed-in the refrigerant charge during start-up. If you use this vessel there is a bit of retention of the lowest vapor pressure refrigerant (n-butane) in the very bottom below the dip-tube, so you need to slight bias your refrigerant charge higher in n-butane to acommendate this.

Kevin

Thanks for your answers. I've found plate exchanger in good price, 20 plates, 315x73x52mm (costs about 70$ in my country). It's not that long as you mentioned, but still 50% longer than a standard one.

I've got question about phase separator. I'm planning to do make it from 200mm long copper pipe (42mm diam) with two caps. What has to be inside phase separator? It must be empty, or there must be some kind of steel mesh?

[Kevin Hotton]

Thanks for your answers. I've found plate exchanger in good price, 20 plates, 315x73x52mm (costs about 70$ in my country). It's not that long as you mentioned, but still 50% longer than a standard one.

I've got question about phase separator. I'm planning to do make it from 200mm long copper pipe (42mm diam) with two caps. What has to be inside phase separator? It must be empty, or there must be some kind of steel mesh?

A phase separator should have the inlet tube enter tangentially (to use centrifugal force to aid in the phase separation). I use stainless steel kitchen cleaning pads (without soap) brand was "Scotch-Brite" in the lower half of your phase separator (to stop the swiriling action of the liquid). Also, the vapor outlet tube (exiting the phase separator) should extend down into the phase separator to below the level of the tangential inlet tube.

Kevin

[bartx]

Phase separator is ready, I only need to braze it. Most of the unit is ready, I have to order plate exchanger and new filter, because I damaged the old one.

[Kevin Hotton]

Your phase-separator looks good. My only concern is the end caps seem a little short on length. Did you machine them? If you braze them on good, they should hold pressure.

Kevin

[Moc]

Also, the vapor outlet tube (exiting the phase separator) should extend down into the phase separator to below the level of the tangential inlet tube.

May I ask you to explain this further?

[Jack]

A phase separator should have the inlet tube enter tangentially (to use centrifugal force to aid in the phase separation). I use stainless steel kitchen cleaning pads (without soap) brand was "Scotch-Brite" in the lower half of your phase separator (to stop the swiriling action of the liquid). Also, the vapor outlet tube (exiting the phase separator) should extend down into the phase separator to below the level of the tangential inlet tube.

Kevin

May I ask you to explain this further?

Yes, I was wondering about that statement as well. Never heard of that before.
I always thought the vapor outlet was simply a hole in the top.

[Kevin Hotton]

It is not a critical design issue, but a refinement. If you look at how a conventional cyclonic separator is constructed, you will see that the vapor outlet is ducted to below the 2-phase flow tangential inlet. This is done to prevent any liquid from taking a "short-cut" out the vapor outlet before being centrifugally force to the wall, deswirled and drained out the bottom.

Kevin