PDA

View Full Version : 3-stage cascade







before
21-06-2007, 04:12 PM
Hi there, :)

First of all, let me introduce myself. I'm not a refrigeration professional and certainly not an expert among amateurs. :rolleyes: I'm a former phd in psychology; who has taken another occupational way. In fact, I'm working now as an executive for a small computer parts reselling company.

In my job, I'm used to perform CPU testing on a daily basis and of course overclocking has become an hobby. I've already owned a 2-stage cascade (R290 & R170) during the past year but the evaporating temp wasn't enough (about -100C unloaded, -91.5C at about 270w of load)... I wished for less. :p

So; my project was/is to add a third stage to that 2-stager unit.

The whole unit has already been assembled, but results aren't very good for the moment. Briefly, here's what it actually looks like:



Actual Cascade

1st stage

Condenser + Fan
Rotary cp 10,000Btu/h
Captube .8mm (1.9m)
HX 14 plates

R404a

2nd stage

Rotary cp 8,100Btu/h
Manual Oil separator
Desuperheater (huge condenser split into 2 parts, one for 2nd stage and one for 3rd stage)
Manometer for HP
Expansion tank
Pressostat
Deshydrater
Captube .8mm (2.05m)
Coaxial HX

R1150

3rd stage

Rotary cp 17,000Btu/h
Automatic Oil separator
Desuperheater (huge condenser split into 2 parts, one for 2nd stage and one for 3rd stage)
Expansion tank
Nanometer for HP
Pressostat
Deshydrater
Captube .8mm (2.30m)
Evaporator

Mixture: R50 + R1150 (71.5% + 28.5% at the moment; i've tried lots of amounts; but also only R50 and only R1150. R50 alone requires to much of HP; and R1150 alone isn't really good)


Observed problems

- Unsufficient hold of load of 2nd stage: SL can't stay frozen no matter the amount of R1150 while 3rd stage's running; and HX temperature gets hotter and hotter... HP diminushes accordingly. Supposed solution: substituing the compressor by a more powerful one.
- Redundant missfunctionning of 3rd stage: sometimes it works really well... but it may also fails to produce cold into the evaporator while making strange sounds into the compressor (sounds of liquid or oil; dunno). Also, it can start correctly and produce a satisfactory evaporating temperature (let's say -110C with R1150 only), but (1) the end of suction line doesn't freeze as it usually do (I'm thinking about a schrader gauge which is located at the end of the SL near the compressor), and (2) the load cannot be hold. Supposed solution: :confused: The matter is I don't understand what happens :D Couldn't it be a flash gaz issue? ...evaporation occurs inside the capillary tube and then can sometimes produce some cold into the evaporator but cannot hold the load correctly...

Things to work on / think about:

- Rearranging/replacing compressors relativily to their power: 24000btu/h on 1st stage, 17000btu/h on 2nd and 13000btu/u on 3rd (instead of 10000tbu/h, 8100btu/h and 17000btu/h)
- Thinking about making something like an SLHX into 3rd stage. In fact, it mostly depends on what the actual problem is: is it because of flash gas? If yes an SLHX can be the solution. Additionaly, also about flash gas, an exchange of heat into the suction line might be benefical for the R50/R1150 mixture.
- Thinking of replacing coaxial HX with a plate one.
- Thinking of increasing the lentgh of 3rd stage capillary tube.
- May be other things... :D

Thx for reading. I'm impatiently looking forward to read your comments and inputs (god knows that I need some... :rolleyes:)

before
21-06-2007, 08:20 PM
I've thought more on 3rd stage issue; and I'm thinking now that the "flash gas theory" was obviously wrong; it might be something related to deshydrater or/and captube. :)

chillin out
21-06-2007, 09:15 PM
Any chance of a picture or two?


Desuperheater (huge condenser split into 2 parts, one for 2nd stage and one for 3rd stage)
Don't quite know what you mean by this.
Why would you be using a condenser for the 2nd and third stage?

Or is it just me being over tired ....lol

Chillin:):)

before
21-06-2007, 10:22 PM
Hi buddy :)


Any chance of a picture or two?

Of course; I've tried to show one; but I need at least 15 posts prior to be able to post a link. :p


Don't quite know what you mean by this.
Why would you be using a condenser for the 2nd and third stage?


In fact, as you imagine the condenser isn't use to condensate here, but to desuperheat the HSP part just after the oil separator. ;)

chillin out
21-06-2007, 11:02 PM
PM me your image links and I will post them for you.

Chillin:):)

The MG Pony
22-06-2007, 02:33 AM
You got it back wards! The highest power compressor needs to be on the 1st stage as that is where the most heat has to be removed.

12,000 1st
10,000 2nd
8,100 3rd

Use a TXV on the first stage, and if you can get one in the corect range on the second as well.

before
22-06-2007, 08:33 AM
PM me your image links and I will post them for you.

Chillin:):)

Thank you ;)


You got it back wards! The highest power compressor needs to be on the 1st stage as that is where the most heat has to be removed.

12,000 1st
10,000 2nd
8,100 3rd

Use a TXV on the first stage, and if you can get one in the corect range on the second as well.

Yep; I'll rearrange compressors by power (and using bigger ones so that it'll be built with 24000btu/h, 17000btu/h and 13000btu/h.)

In fact, I've firstly thought it should be nice to connect actual huge one (17000btu/h) on 3rd stage to help improving the use of R1150 or R50. It was a kind of experiment. :)

The MG Pony
22-06-2007, 09:40 AM
more power will not make it better, just waste more energy. frankely the one you have has more then enough reserve capacity, might as well make a cascade chiller and cool the whole thing with that much power.

before
22-06-2007, 10:25 PM
more power will not make it better, just waste more energy. frankely the one you have has more then enough reserve capacity,

Ok; I find it quite difficult to estimate the power needs.

If possible, evaporator must stay below -110C even with about 400w to 450w of CPU load (and of much lower with less CPU load).

As for now, it seems me 8100btu compressor wasn't enough for 2nd stage because it wasn't able to hold appropriate HP no matter the amount of R1150.



might as well make a cascade chiller and cool the whole thing with that much power.


That's a good idea. :)

norseman
26-06-2007, 07:14 AM
Hi there, :)

First of all, let me introduce myself. I'm not a refrigeration professional and certainly not an expert among amateurs. :rolleyes: I'm a former phd in psychology; who has taken another occupational way. In fact, I'm working now as an executive for a small computer parts reselling company.

In my job, I'm used to perform CPU testing on a daily basis and of course overclocking has become an hobby. I've already owned a 2-stage cascade (R290 & R170) during the past year but the evaporating temp wasn't enough (about -100C unloaded, -91.5C at about 270w of load)... I wished for less. :p

So; my project was/is to add a third stage to that 2-stager unit.

The whole unit has already been assembled, but results aren't very good for the moment. Briefly, here's what it actually looks like:



Observed problems

- Unsufficient hold of load of 2nd stage: SL can't stay frozen no matter the amount of R1150 while 3rd stage's running; and HX temperature gets hotter and hotter... HP diminushes accordingly. Supposed solution: substituing the compressor by a more powerful one.
- Redundant missfunctionning of 3rd stage: sometimes it works really well... but it may also fails to produce cold into the evaporator while making strange sounds into the compressor (sounds of liquid or oil; dunno). Also, it can start correctly and produce a satisfactory evaporating temperature (let's say -110C with R1150 only), but (1) the end of suction line doesn't freeze as it usually do (I'm thinking about a schrader gauge which is located at the end of the SL near the compressor), and (2) the load cannot be hold. Supposed solution: :confused: The matter is I don't understand what happens :D Couldn't it be a flash gaz issue? ...evaporation occurs inside the capillary tube and then can sometimes produce some cold into the evaporator but cannot hold the load correctly...

Things to work on / think about:

- Rearranging/replacing compressors relativily to their power: 24000btu/h on 1st stage, 17000btu/h on 2nd and 13000btu/u on 3rd (instead of 10000tbu/h, 8100btu/h and 17000btu/h)
- Thinking about making something like an SLHX into 3rd stage. In fact, it mostly depends on what the actual problem is: is it because of flash gas? If yes an SLHX can be the solution. Additionaly, also about flash gas, an exchange of heat into the suction line might be benefical for the R50/R1150 mixture.
- Thinking of replacing coaxial HX with a plate one.
- Thinking of increasing the lentgh of 3rd stage capillary tube.
- May be other things... :D

Thx for reading. I'm impatiently looking forward to read your comments and inputs (god knows that I need some... :rolleyes:)

I will suggest that you measure and control the HX temperatures by sensors applied to them. The systems I have been into as cold traps and -152c ULT freezers use thermostates to hold the HX at the best working temperature by switching the compressors on and off. The oil quality and type of oil separators is also critical. I have good experience by coalecent types and may be two in series at the lowest stage. Some units use a cap. tube heater at the lowest too which cut in by a timer for a short periode every 4-6 hour to flush out the oil trapped.
Great to see that you try a real cold one.

Norseman

Peter_1
28-06-2007, 09:11 PM
What you're trying to make is very difficult.
What we need is pressures (LP and HP) and temperatures of each stage, especially SH and SC of each stage.
Try to measure these.
This is really an application on the edges of what is possible with refrigeration in general

before
12-07-2007, 04:46 PM
Thanks to both of you ;)

I've rebuilt/reinsulated some parts of each stage. Now I can properly read what evaporating temperatures really are.

Currently refrigerants are R404a into 1st stage, R1150 and R1150 as well into both 2nd and 3rd stages.

Without load on 3rd stage evaporator (instead of inapropriate insulation of the block); I got -46C for the first HX, -94C for the second one and about -120C for 3rd stage evaporator.

I've also looked for unloaded temperatures, 1st stage did -56C (with lower stages turned off) and 2nd stage did -99C after switching of the lowest one.

Ambiant temperature is about 26C.

Concerning the pressures, LSP are well fitted on what one could expect from these temperatures, i.e., manifolds gave me about 0.2xbar on 3rd stage and a bit more than a bar on 2st stage

HSP pressures are respectivily about 14bars on 2nd stage, and 4bars on 3rd. (statics are 16bars and 7bars). (I didn't look at 1st stage pressures).

I'm now waiting for a 24,000btu/h rotary compressor to replace 1st stage 10,000btu/h; and I'll replace 2nd stage 8,100btu/h with a 17,000btu/h which is impatiently waiting in my garage.

:)

before
19-07-2007, 12:10 PM
I've just ordered some R14 to fill in 3rd stage with what I look to be a more appropriate refrigerant for this stage.

What do you think about 2nd stage refrigerant, is R1150 a very good one knowning R14 will be used into 3rd stage? Or, may be, R23 or R170 should be better?

:)

norseman
19-07-2007, 01:37 PM
I've just ordered some R14 to fill in 3rd stage with what I look to be a more appropriate refrigerant for this stage.

What do you think about 2nd stage refrigerant, is R1150 a very good one knowning R14 will be used into 3rd stage? Or, may be, R23 or R170 should be better?

:)
I have good experience with the oil return by HC types so the R1150 is good. It will get you colder than the R170. The R23 is an oldish who gives
very high discharge temperautures. I have seen a lot
of burned cascades using that one. In the same area
the 508B would work much colder on the discharge.
Just a thought, but if you buy another small compressor and make the 4. stage, you could run that one by the HC R50 (Metan) who gives around
-161c. Not bad? The use of some "storage" tanks on the suction side would give some more capacity.
Seen them operated by timers, very thiny capillary
tubes or temperature control.
Good luck from Per

before
23-07-2007, 09:01 AM
Thanks for your reply Norseman :) I'm gonna stay with R1150 into 2nd stage ;) Concerning a 4th stage, -161C is little bit too cold because CPU are usually cold-bugged after let's say -135C.

Ok then; I've substituted 2nd stage 8100btu/h compressor by a 17000btu/h. Drop in temperature is very significant: evaporating point into HX 2/3 is now at -102C while 3rd stage running (about -105C to -106C unloaded). This is nearly 10C of gain with a bonus: succion line is now frozen until the ACL.

So; this makes me thinking about how colder will be both HX after replacement of 1st stage 10000btu/h compressor by a 24000btu/h one?

Evaporating temperature of 1st stage is currently -44C while the 2 lowest stages are running. But it'll be significantly lower with a bigger compressor and so will be the evaporating temperature of 2nd stage...

I'm afraid that R14 alone into 3rd stage will be impossible to use. It'll be too easy to condensate... and it might require me to try a mixture of R50/R14... or to use R290 into 1st stage instead of R404a. What do you think about that?

before
24-07-2007, 02:06 PM
I've got a weird matter guys...

I've charged 3rd stage with R14 and everything is fine down to about -125C to -130C. Then, it stops going down anymore and slowly gets hotter and hotter up to let's say -60C or something to finally recover and going down again. It does it everytimes.

That's make me thinking about a problem with oil... What's you opinion?

Do think adding a small amout of R290 can help?

:)

before
24-07-2007, 04:41 PM
R290 was a good solution. 3rd stage is now stable at -131C. :)

Javabones
24-07-2007, 05:48 PM
Have you done something to reduce the oil viscosity on the 2nd & 3rd stage? Norseman had it right questioning the oil. This may explain the inconsistency.

May I also suggest you control via discharge gas through the evap, which will also keep velocity up and keep the evap clear of any oil that may settle.

Please keep us posted!

Javabones
24-07-2007, 05:52 PM
Ahhh... duh... read the 2nd page first before replying :-) Just saw the post(s) about R290. Good move!

before
24-07-2007, 11:11 PM
Thank you! :)

Do you suggest me to add some R290 into 2nd stage as well? I didn't get any strange misfunctionning like today's one of 3rd stage so far. By the way, being blind doesn't mean that nothing is happening... :rolleyes:

:)

cascade tech
06-08-2007, 10:34 PM
Hi i only work with 2 stage cascades and liquid nitrogen and stratosphere chambers.
You might want to try using a LN2 system rather cause you dont have to be a design engineer to make something cold with that stuff , but be careful its deadly.

If you are gonna stick with the 3 stage design , i think you will need.

- Back pressure regulators on all the compressors .

- Hot gas bypass and liquid injection on all the stages .This ensures that you dont start slugging liquid when the load drops off from the higher stages and liquid is always available for when they do get loaded .Obviously you will need three contollers to control the temprature of each stage.

- a timer or stat that lets the cascade get cold before the second stage comes in.I imagine this would apply to the third stage.

-Blow off tanks on the second and third stage are probably also a good safety measure and should be fitted with blow off valves.

- A very good cabinet , maybe even at a negative pressure so that theres no air to condense.This may sound overboard but i work on stratosphere chambers and its basically just a chamber with a thick sealed metal liner attached to a big vacum pump.Your PCB may become a block of ice otherwise.

- Expansion valves are the way forward or try using M.O.P valves if you wanna eliminate the crankcase regulators. Capillary selection is for people who have loads of time to make mistkes :)

All the best and be safe the pressures can get a bit high in cascades.

Rob S
16-08-2007, 05:45 AM
I believe 290 will liquify at that temperature... Can't find the data to be sure tho.

I would use 404A gas... which in a gas form is 125 and 143a.

before
16-09-2007, 05:15 PM
Thanks for your suggestions.

During this summer, I've recieved a first 23,400btu/h compressor for 1st stage that was DOA and then received a second one in perfect condition. By the way, the high stage condenser has became useless and the unit needs a larger one.

Anyway, I think of taking more time to finish this project and add additional parts.

Best 3rd stage temperature so far was -137C using a R14/R50/R290 mixture.


Hi i only work with 2 stage cascades and liquid nitrogen and stratosphere chambers.
You might want to try using a LN2 system rather cause you dont have to be a design engineer to make something cold with that stuff , but be careful its deadly.

Hi :) I'm used to perform CPU benchmarking with liquid nitrogen but I prefer a phase change unit to do the job.

LN≤ requires much more attention and is much more expensive. Filling my 25l dewar costs my compagny about 280Ä (i.e. USD390). Building a cascade (even a 2 stager) is a most important one-time-payment, but once finished, it doesn't need more money (okay, despite of the electricity bill).



If you are gonna stick with the 3 stage design , i think you will need.

- Back pressure regulators on all the compressors .


Interesting idea. I've never used one; but it looks clever.


- Hot gas bypass and liquid injection on all the stages .This ensures that you dont start slugging liquid when the load drops off from the higher stages and liquid is always available for when they do get loaded .Obviously you will need three contollers to control the temprature of each stage.

Ok :)


- a timer or stat that lets the cascade get cold before the second stage comes in.I imagine this would apply to the third stage.

Thermostat would be nice but, I can do it manually by looking at both temperatures and pressures.


-Blow off tanks on the second and third stage are probably also a good safety measure and should be fitted with blow off valves.

Aren't the pressostats sufficient for safty purposes?



- A very good cabinet , maybe even at a negative pressure so that theres no air to condense.This may sound overboard but i work on stratosphere chambers and its basically just a chamber with a thick sealed metal liner attached to a big vacum pump.Your PCB may become a block of ice otherwise.


Oh yea, it would be ideal; but a good insulation of the PCB and other cold parts is highly sufficient. ;)


- Expansion valves are the way forward or try using M.O.P valves if you wanna eliminate the crankcase regulators. Capillary selection is for people who have loads of time to make mistkes :)

You're right. I'm hardly thinking about removing those captubes from all stages and replacing them by TXV for 1st and 2nd and CPEV for 3rd.

before
23-09-2007, 01:50 PM
Thx to my 15 posts, I can now add some pictures to illustrate a bit the thread. :D

Here are some "old" results acheived before summer.



I've substituted R1150 into 2nd stage with a mixture of R290/R50 and it works pretty good... :eek:
About -80C unloaded and -75C after turning on 3rd stage.

3rd stage :p

http://img510.imageshack.us/img510/5446/137og4.jpg

Mixture into 3rd stage
- 8 to 9bars of R14
- 2 to 4bars of R290
- 2bars of R50

That was too cold for the X6800 which cold-bugged at -121C speedfan with 400FSB. :D CPU wattage was about 276w.

http://img528.imageshack.us/img528/7632/120wu0.jpg

Some pictures.

http://img405.imageshack.us/img405/7122/p1020758copieyy5.jpg

http://img61.imageshack.us/img61/4575/p1020760copieyh6.jpg

http://img139.imageshack.us/img139/1624/p1020769copieqj7.jpg

http://img139.imageshack.us/img139/4141/p1020771copienq6.jpg

before
28-09-2007, 11:25 PM
I've brazed a huge desuperheater between first stage compressor and condenser. It looks uncommon but quite efficient.

Also, I've substituted first stage suction pipe by a larger one that fits with compressor suction side. BTW I've got some problems brazing the copper pipe with the HX... I don't know what the brazed plate HX is made of (steel, aluminium? It's an APV Invensys), but I can't get something good here; it's leaking everytimes. :(

Peter_1
29-09-2007, 08:05 AM
The PHE from APV is normally made of Stainless Steel, Solder these with at least 40% silver rods with flux.
I wonder if all the gasses find the proper way to flow through the proper tubes.
It looks a little bit untidy but I'm impressed about the result you've achieved.
I think many don't even have a clue how you can do this all.
Make sure that all your pressures and temperatures are within the manufacturers specs.
You also can check via pressure and more important SC and SH measurements if your 3 systems are performing efficient

before
29-09-2007, 09:20 AM
Thank you :)

I know, the unit doesn't look clean, I mean aesthetics wasn't the key and things were done in the easiest way.

Also, the polyurethane foam looks very dirty; it would have been better to prepare boxes for insulating the HX and put the foam into them.

before
12-02-2008, 06:10 PM
I'm considering making some changes to this unit.


Replacing actual condenser with a Luve 2.73kW
Replacing actual 2nd stage captube with a FACH CPEV
Replacing actual 3rd stage captube with a FACH CPEV (target temp range is -115C to -135C loaded; temp should be adjusted upon the need knowing most actual Intel CPUs start to cold-bug passing down about -115C)
Filling 3rd stage with R14 only
Making individual boxes for Heat Exchangers and fill them with polyurethane foamNot sure if it's a good idea to replace 3rd stage actual ShineYear oil separator with a Temprite 900, or a Danfoss OUB. I mean, something suitable for R14 is needed here.

Any suggestions guys? :)

before
13-02-2008, 02:42 PM
Maybe I'll change oil as well. How to remove it from a rotary? Putting it upsidedown? Or with a particular tool?

nike123
13-02-2008, 04:08 PM
Hi there, :)



Thx for reading. I'm impatiently looking forward to read your comments and inputs (god knows that I need some... :rolleyes:)

Gees, that system looks expensive, complicated and energy consuming.
Maybe it is cheaper and safer to buy some supercomputer on his nominal tact.:D:D

before
14-02-2008, 06:44 PM
Humm; you are quoting a more than half-year old post. :D This unit is working quite good actually, but time has come to make some improvements.

before
20-02-2008, 01:49 PM
So; I've ordered the following parts to cold_ice's shop (http://kaeltetechnik-shop.at). I'd like to thanks him for his good communication.

Luve 2.73kW Condenser + Incorporated Fan
3 x 30Gr Copper Filterdiers
2 x FACH PZ1 CPEV + 4 x Brazing adapters
4m of 2mm Captube
R14 Refrigerant
I've also asked Air Liquide for a bottle of R1150. I need some for the second stages of my 2 cascades... Shame on me! :p

Basically, specs will be as follows:

1st stage

Compressor: LG NP407P (23,400Btu/h; 40.7cc/rev)
Condenser: 2.73kW + Fan
Filterdrier: Copper 30gr
Expansion Device: 1.9m of 0.8mm Captube
Heat Exchanger: Invensys APV ParaBrazed HX (14pl)
LSP Service Valve: Schrader
Refrigerant: R290
Target Temperature Range: -50C to -60C
2nd stage

Compressor: Matsu****a 2K32 (17,000Btu/h; 32cc/rev?)
Oil Separator: Homemade by JP
HSP Service Valve: Schrader
HSP Manometer: REFCO
Filterdrier: Copper 30gr
Expansion Device: FACH PZ1 CPEV + 1m to 2m? of 2mm Captube
Heat Exchanger: Homemade Coaxial by Pgeo
LSP Service Valve: Schrader
Refrigerant: R1150
Target Temperature Range: -80C to -90C
Others: Desuperheater, Johnson Controls Pressostat
3rd stage

Compressor: Sanyo C-RV167H11AB (13,000Btu/h; 16.7cc/rev?)
Oil Separator: ShineYear Automatic
HSP Manometer: REFCO
Filterdrier: Copper 30gr
Expansion Device: FACH PZ1 CPEV + 1m to 2m? of 2mm Captube
LSP Service Valve: Schrader
Refrigerant: R14
Target Temperature Range: -115C to -135C (adjustable upon the need / limitations of hardware)
Others: Desuperheater, Johnson Controls Pressostat, Pgeo's "Marguerite" Evap + succion line, and 2 x Homemade Piotres' Expansion Tanks

fridgey
09-03-2008, 02:17 PM
Gíday :rolleyes:, Firstly Iíd like to say Iím impressed with what you have achieved so far. I stumble across this just randomly searching and find what you are doing from a refrigeration point of view to be very interesting. As a service technician who regularly works on cascade systems usually -80C two stage Revcoís, Herasus and the like and a few -150C Revco multi-cascades. Have you heard of them? Very interesting, the crux is that they only have 1 compressor, 3 heat exchangers and a mixture of about 6 different refrigerants inside and achieve -150C. How itís done is a little more complicated.

Anyway back to your 3 stage cascade and some advice on compressor from what I can tell you are using rotary compressors, these types of compressors are design for use on mostly air-cons, were they operate at medium to high back pressures with a larger volume of refrigerant being pumped through the compressor. In your application the compressors are operating at a very low back pressure sometimes even in a vacuum with a small volume of refrigerant being pump through the compressor, they were never designed with this type of operation in mind. Not to say that they wonít do it, they will. Itís just that they wonít last a long time doing it. My advice to you is to use low back pressure (LBP) reciprocating compressor. They are designed to run on LBP and have smaller volume of refrigerant pump through. They are used in most 2 Stage Cascades, and are much happier running at your operating conditions. With the capacity of the compressor for each stage they should be roughly about the same size with maybe the first stage a little larger. I know that most of the mass produces -80C freezers have the same compressors on both stages. When you think about it make sense as your trying to transfer the same amount of heat through each stage, providing things are balanced. This way you get better COP from your compressors and use less power to run it.

Refrigerant wise havenít had any experience with R1150 so I canít comment. But in the inertest of keeping the stages balanced I donít know if it is the best refrigerant for the second stage. I would suggest either R170 or R508B this will keep your compression ratios similar for each compressor.
1st Stage -40C R507 vs -40C R290
2nd Stage -80C R508B vs -95C R1150
3rd Stage -120C R14 vs -120C R14

There are so many other things to consider with a cascade system especially for reliable operation hope my few are useful. Love your work though; what youíve achieved by throwing your hat at it is pretty impressive considering that you donít come from a refrigeration background.
Cheers :D

Peter_1
09-03-2008, 02:31 PM
Fridgey, what you're talking about - mix of gasses - is an autocascade and was covered already several times in our own RE section CPU overclockers.

before
09-04-2008, 08:01 PM
First of all; I'd like to thank you so much for your long and friendly post.


As a service technician who regularly works on cascade systems usually -80C two stage Revcoís, Herasus and the like and a few -150C Revco multi-cascades. Have you heard of them? Very interesting, the crux is that they only have 1 compressor, 3 heat exchangers and a mixture of about 6 different refrigerants inside and achieve -150C. How itís done is a little more complicated.
Auto-cascade is very interesting indeed.

I've found this pic about . Which refrigerants are used?

http://www.artisan-scientific.com/itemimages/Revco_ULT1340_View1.JPG



Anyway back to your 3 stage cascade and some advice on compressor from what I can tell you are using rotary compressors, these types of compressors are design for use on mostly air-cons, were they operate at medium to high back pressures with a larger volume of refrigerant being pumped through the compressor. In your application the compressors are operating at a very low back pressure sometimes even in a vacuum with a small volume of refrigerant being pump through the compressor, they were never designed with this type of operation in mind. Not to say that they wonít do it, they will. Itís just that they wonít last a long time doing it. My advice to you is to use low back pressure (LBP) reciprocating compressor. They are designed to run on LBP and have smaller volume of refrigerant pump through. They are used in most 2 Stage Cascades, and are much happier running at your operating conditions.

I've already heard about LBP compressors but those are pretty difficult to find for me. Actually, I can easily get rotaries; but LBP are rare (at least for me).

Of course, using rotaries can be a real matter while trying to get temperatures that correspond to very low succion pressures. For instance, with my other cascade, a 2-stager unit, the second stage filled with R1150 can do about -115C (~0.5bar) loaded. The result is interesting during a certain amount of time, but the compressor temperature is more than reasonable...

before
09-04-2008, 08:03 PM
Here are some pics of the revised unit.


Click the thumbnail to view full-size image
http://www.ixtremtek.com/images/Grosse%20Bertha/1a.jpg (http://www.ixtremtek.com/images/Grosse%20Bertha/1.jpg) http://www.ixtremtek.com/images/Grosse%20Bertha/2a.jpg (http://www.ixtremtek.com/images/Grosse%20Bertha/2.jpg) http://www.ixtremtek.com/images/Grosse%20Bertha/4a.jpg (http://www.ixtremtek.com/images/Grosse%20Bertha/4.jpg) http://www.ixtremtek.com/images/Grosse%20Bertha/5a.jpg (http://www.ixtremtek.com/images/Grosse%20Bertha/5.jpg) http://www.ixtremtek.com/images/Grosse%20Bertha/6a.jpg (http://www.ixtremtek.com/images/Grosse%20Bertha/6.jpg) http://www.ixtremtek.com/images/Grosse%20Bertha/7a.jpg (http://www.ixtremtek.com/images/Grosse%20Bertha/7.jpg) http://www.ixtremtek.com/images/Grosse%20Bertha/8a.jpg (http://www.ixtremtek.com/images/Grosse%20Bertha/8.jpg) http://www.ixtremtek.com/images/Grosse%20Bertha/9a.jpg (http://www.ixtremtek.com/images/Grosse%20Bertha/9.jpg) http://www.ixtremtek.com/images/Grosse%20Bertha/10a.jpg (http://www.ixtremtek.com/images/Grosse%20Bertha/10.jpg) http://www.ixtremtek.com/images/Grosse%20Bertha/12a.jpg (http://www.ixtremtek.com/images/Grosse%20Bertha/12.jpg)

before
09-04-2008, 08:20 PM
BTW I've got a weird problem with 3rd stage. Here's how it's running.

- Goes down quickly to lets say -45C (takes very few seconds)
- Stops going down
- Goes up to lets say -35C
- Goes down again not so fast to about -90C
- Stops (or goes down very very slowly) during compressor suc side is freezing (takes a while)
- Goes down to -130C (slowly but surely)
- Goes up to about -120C
- Goes down to -131C
- Goes up to -122C
etc.

What's wrong in your opinion? Could it be the HX between 2nd and 3rd stages which is not good enough? Evaporating temperature and flow of 2nd stage are great... something like -98C in / -98C out unloaded achieved very quickly. Also, adding more R14 into low stage doesn't make any improvement; it gets worst...

before
28-04-2008, 09:00 PM
I've replaced the coax HX by a brazed plate one (12pl).

The cascade works way better now. I've to find the right adjustment.

I couldn't resist to take a very quick pic.


Click the thumbnail to view full-size image
http://www.ixtremtek.com/images/AsusCommando/-138,1Ca.jpg (http://www.ixtremtek.com/images/AsusCommando/-138,1C.jpg)

ultralo1
02-05-2008, 11:01 PM
The ultrlo temp revco freezer is a mid 80s model that used R500 and R503.
Today they use R404a, R134a 1st stage and R508 and R404a second stage. There have ben rumors of switch blends back to AB oil and R404a, R290 1st and R508, R290 2nd stage.

fridgey
18-05-2008, 11:05 AM
How are you charging your second and third stages? I assume by static pressure. To get your charges right there will obviously be a lot of trail and error involved.

before
28-05-2008, 08:18 PM
Firstly, I've charged a bit of refrigerant into each stage and then adjusted under load. :)

I'm using this cascade frequently and get -110C and -120C under huge CPU loads.

The only matter so far is 3rd stage because I can't really play with the CPEV. I think that need a liquid reciever here. :)

Reflexive
20-12-2008, 10:01 PM
Very impressive system you've built, albeit as I'm sure you know, strangely piped. I try and autocad or 3d model in other forms my entire system before I even pick up the pipe bender. Try it sometime! The effort before makes beauty in the end.

Your results are great. If you ever have any problems with r14 on its own, a good method is an ounce or two of r600 or r290, plus a double volume oil separator. Something like the Temprite 901 or 902 will be helpful. Assuming your rotarys are not meant for this application and are simply charged with AB, try charging the oil separator with Zerol 150.
Also with r14, your going to want a nicely sized expansion tank for sure. Not sure if you've used one on this system. Oh and really your stuck to captube, so keep your liquid line very short, and attempt to get very good contact between your post-evaporator and capillary tube.

Hope some of that helps!
(Other option is to autocascade the final stage, you certainly have the compressor size, and you can reduce some load from your 2nd stage, something like R123, R22, R14, argon will allow you much lower temperatures with a subcooler. The r14 could reach loads towards -140C that way. R123 and R22 will chill the flow before it reaches your 2nd stages evaporator and help alot!)

before
01-02-2009, 12:09 AM
Thank you! :)

Well, I was very rookie (and didn't had so much tools) at piping when I started working on this cascade. I would probably do way better now, and to be honest I'm thinking about disasembling it to rebuild it sometimes. :)

I'm not using captubes anymore here but CPEVs at each level. Well, I'm using captubes only to connect 3rd stage expansion tanks to succion side (black cylinders on the left (http://www.ixtremtek.com/images/Grosse%20Bertha/9.jpg)).

I've already thought about using another oil sep for R14, but this ShineYear isn't too bad I think. The cascade can handle about -125C during hours without issues.

The only matter is, well at least I guess, the 2nd stage. I'm not sure R1150 is really appropriate here. I'm waiting for a bottle of R23; it might be better.

Refrigerants are R290, R1150 and R14. Well, R1150 is a bit hard to use here. It needs a quite low temperature to condensate, not easily achieved with R290 under load, and when condensated it goes down to a very low temperature, may be too low for R14 to work appropriately.

R290, R23 and R14 should be better.

:)

before
12-02-2009, 11:44 AM
I've done some work yesterday. I have now to finish re-insulating.
- R23 replacing R1150 at 2nd stage
- Expansion device entirely re-done at 3rd stage
- About -123C (probe) to -126C (mano) unloaded with 3/4 opened CPEV
- About -122C (probe) after 10mins with a 1200w hair drier next to the evap with 3/4 opened CPEV and massive flow into compressor suc

http://img8.imageshack.us/img8/8100/img8245zg4.jpg

http://img8.imageshack.us/img8/4209/img8248kl1.jpg

http://img156.imageshack.us/img156/7671/img8250xh2.jpg

http://img156.imageshack.us/img156/3905/img8251ou0.jpg

http://img187.imageshack.us/img187/7183/img8252lo7.jpg

http://img156.imageshack.us/img156/9467/img8253mb6.jpg

http://img8.imageshack.us/img8/9245/img8254yg4.jpg

Can't wait for testing under CPU loads.

before
16-02-2009, 07:27 PM
I've lost some R14 a small leak. Well, I've repaired it and moved back to R1150 at 2nd stage.

- Fully opened CPEV
- Evap at -131C probe (manometer indicates a bit less than that).
- Evap at -129C probe after 15mins heating with hair drier. (not a fine load tester BTW, but loosing only 2 degrees here is not so bad actually)

http://img293.imageshack.us/img293/466/img8316eg1.jpg

http://img293.imageshack.us/img293/2023/img8306pi2.jpg

The HSP of R14 is very low. 5bars (not 4, gauge is wrong).

http://img511.imageshack.us/img511/4200/img8307vn3.jpg

http://img293.imageshack.us/img293/5211/img8300ne3.jpg

sammyg69er69
17-02-2009, 10:17 AM
what is the pcb for ? have u ever thought about just using a helium system

before
17-02-2009, 12:20 PM
The PCB is to attached the evaporator over a CPU.

Well, a Liquid Helium sys would be way too low. Not all CPUs can do -130C or even -100C.

before
22-02-2009, 11:18 AM
I've started working on a HX to help condensing at first stage.

http://img244.imageshack.us/img244/3958/img8344.jpg