Rotary 2 stager with metal case and some extras

Hi

I want to show You mine latest cascade, it's 2 stager based on 2 rotary compressors, plate HX, Danfoss TS2 valve, Danfoss OUB1 oilseparator, some extras like four REFCO panel gauges or dixell controller and .... metal, BLACK case :rehab: . We like that kind of cases, do we ? :fact:

Specs :

compressors: two Sanyo C-1RV137H11AA rotaries
evap: steeper by piotres
XV: Danfoss TS2 / captube 0.8 mm
HX: plate HX , 12 plates
condenser: 2 row, alu-cu, cooled by 4 x 120mm/12 V NBM fans
refrigerants : R507/R1150
oilseparator : Danfoss OUB1
safety : cut-off
case : metal case, 46 x 46 x 43 cm, black paint
others: 4*REFCO panel gauges, Dixell XR10C controller (turn on 2nd stage automatically when 1st stage is enought cold), homemade receiver before Danfoss TS2, big 230 mm/230V fan on the back wall of case

http://www.members.lycos.co.uk/piotres2/cascade2/1.jpg

:woot:

http://www.members.lycos.co.uk/piotres2/cascade2/2.jpg

http://www.members.lycos.co.uk/piotres2/cascade2/3.jpg

Suction line is braided one, 18 mm thick insulation and red sleeve :D .

http://www.members.lycos.co.uk/piotres2/cascade2/4.jpg

Sorry for some dirty old insulation on compressors' accumulators :slapass: .

http://www.members.lycos.co.uk/piotres2/cascade2/5.jpg

http://www.members.lycos.co.uk/piotres2/cascade2/6.jpg

http://www.members.lycos.co.uk/piotres2/cascade2/7.jpg

Finally I've add shock absorber to 1st stage suction line (called "anaconda" here in Poland ;) ) and add also second drier to 2nd stage :) .

http://www.members.lycos.co.uk/piotres2/cascade2/8.jpg

http://www.members.lycos.co.uk/piotres2/cascade2/9.jpg

Some first run digits ;) .

http://www.members.lycos.co.uk/piotres2/cascade2/10.jpg

http://www.members.lycos.co.uk/piotres2/cascade2/11.jpg

http://www.members.lycos.co.uk/piotres2/cascade2/12.jpg

http://www.members.lycos.co.uk/piotres2/cascade2/13.jpg

http://www.members.lycos.co.uk/piotres2/cascade2/14.jpg

Final charge, no load :)

http://www.members.lycos.co.uk/piotres2/cascade2/15.jpg

And here loaded with 225W :cool: .

http://www.members.lycos.co.uk/piotres2/cascade2/16.jpg

http://www.members.lycos.co.uk/piotres2/cascade2/17.jpg

http://www.members.lycos.co.uk/piotres2/cascade2/18.jpg

We have tested that cascade on E6700 till now, and we got around -98-99*C loaded, and -100-102*C idle on evaporator with 4.7 GHZ / 1.7 V (end of bios in stock asus P5B DLX :( ) .

I hope You like that project :) .

Thanks
Regards
Peter

VERY VERY NICE! Why is it your last cascade?

Quote:

---

Originally Posted by Noob-ftw

VERY VERY NICE! Why is it your last cascade?

---

sorry, I mean latest ;) hehe

+ case
+ smaller than my cascade :]

- poor temps :]

very nice work piotres. Thats some tight work there.
Almost at small as my cascade, but colder ;)

Very good looking setup there Piotres. Nice case and all. I would have expected a bit colder temps tho.

EDIT: Whats load temperature on HX, it might be a bit undersized.

After seeing a few units and building a few you learn to easily see how things should naturally connect together :)

Great build, was the shock absorber effective in cutting down the vibration from the compressor?

very nice work piotres, good job mate.

Quote:

---

Originally Posted by n00b 0f l337

Very good looking setup there Piotres. Nice case and all. I would have expected a bit colder temps tho.

EDIT: Whats load temperature on HX, it might be a bit undersized.

---

Hi

HX outlet with 225W load is -33-35*C, but 2nd stage hiside in max 15.5 bara, so I think that's not the problem :) . Cold_ice use same plate HXes in cascades .

Regards
Peter

good job!

Why you added filter before and after HX?

Quote:

---

Originally Posted by Blaster

good job!

Why you added filter before and after HX?

---

Because first I had thought I had problems with freezing of humidity/oil, but that had been nonsense, nothing like that never happened ;) .

Nice cascade :eek: - where is your insignia? :)

Quote:

---

Originally Posted by Blaster

good job!

Why you added filter before and after HX?

---

i also do the same, never have too much filtering on second stage ;)

Very nice i like alot.

The TS2 TEV, isnt that just for R404a refrigerant and its the TES2 that is for R404a/R507a ?

Just curious as i have a TES2 and TS2 here and the TS2 is only labeled for R404a?

David

Great job my friend!

TS2 and TES2 are both for R404a/R507. TS2 is internally equalized and TES2 external.

How hot do those rotaries run?

Quote:

---

Originally Posted by n00b 0f l337

How hot do those rotaries run?

---

100-110*C compressors' top, so to hotest place :) .

Ah nice, clever use of what looks to be brass barbs in the oil sep too.

What's the captube length?

Quote:

---

Originally Posted by Clemmaster

What's the captube length?

---

250 cm/0.8mm :)

On both?

Haha look again n00b, on 1st stage there is TXV ;) .

Duh, forgot. ;)

Quote:

---

Originally Posted by cold_ice

TS2 and TES2 are both for R404a/R507. TS2 is internally equalized and TES2 external.

---

Thanks Patrick ;)

Quote:

---

Originally Posted by piotres

100-110*C compressors' top, so to hotest place :) .

---

Isn't this too much? Reliability may be affected.

What about sound level? Is it acceptable for the 24/7 use? And overall, is it possible to use that unit for 24/7 operation?

I have a feeling she's loud, might just be the 120mm fans, or the two rotarys, but its a feeling. Might just be me feeling the vibration of him using the unit over there in Europe ;)
As for compressor temperature, I think they could use a bit more cooling, and lower compression ratio, but thats less performance. So its a balance.

Guys, its a cascade! Its not to be used 24/7, its not built for that.

Of course its loud, all rotarys builds are loud.

Blaster
I thought rotaries and scrolls supposed to be more quiet than pistons?
Less moving parts and so on..

If i were to build 24/7 cascade chiller, what compressors should i choose?
Provided i want acceptable sound level and reliable operation.

There's nothing more quiet than danfoss piston's compressors for a power-rating. The SCxx are quiet, the noisiest may be scrolls compressors (I had 2 scroll compressors at school that were very noisy :D).

What coolant would you use with a cascade chiller tell me?
Under -40° you must do your own coolant with ethylenglycol and I'm sure about its viscosity under theses temps ...

Be aware that you can't use ethylen for that purpose but R23/R508/R170 as you'd be working with very high pressures (or maybe use a little compressor on 2nd stage with very short captube but that's not advised)

Clemmaster
It is offtopic here, i am afraid, but liquid will be pure ethanol pumped by positive displacement pump (micropump series probably). The pump is the main problem, actually, it is rated to -46C. I wonder will it work at -90-100C?

Lets get back somewhat to the topic, i'd like to ask is it possible/advantages
to use electronic valves on the second stage?
I want to use electronic valves with stepper motor everywhere and largely oversized condenser on first stage and plate HX on second.

I don't think 110*C is too much for rotaries . I've heard that 130*C is safety limit for them :) .

Like Blaster said - almost all rotary build are loudly . But cascades can be used 24/7, why not :) .

About Danfoss SC series and loud ..owner of that cascade : http://www.xtremesystems.org/forums/...highlight=tiny said me that single-stage with SC21CL and 230v fan on condensor has been more loudly that this cascade, so 2 rotaries ;) .

Thanks
Peter

Heh, as long as the thermal overload doesnt trip.

Hey peter,

i do not advise to use that shock absorber in that position.

It's ment to be used straight afaik.

For the rest you made a really nice cascade!!! :o

I've got a SC18B and it's quiet :)

Perhaps I'm being picky but it isn't a shock absorber at all, rather vibration dampener, and jort you are quite correct, they are meant to be straight and level. In the position OP has it in it has a greatly increased risk of forming leaks.

Quote:

---

Originally Posted by Xeon th MG Pony

Perhaps I'm being picky but it isn't a shock absorber at all, rather vibration dampener, and jort you are quite correct, they are meant to be straight and level. In the position OP has it in it has a greatly increased risk of forming leaks.

---

Tnx for the verification :toast:

Quote:

---

Originally Posted by Cronos

Blaster
I thought rotaries and scrolls supposed to be more quiet than pistons?
Less moving parts and so on..

---

In the size engine we use for cpu cooling piston are quieter, for high power, scrolls become more efficient and quiet, an equivalent scroll to this one would be quieter.

Quote:

---

Originally Posted by Cronos

If i were to build 24/7 cascade chiller, what compressors should i choose?
Provided i want acceptable sound level and reliable operation

---

depends what temp and capacity you wanted, and what you think its acceptable sound level
maybe best option would be to put chiller at distance

Congrats on the -100C, I must say your box is one fine looking machine!

Quote:

---

Originally Posted by piotres

I don't think 110*C is too much for rotaries . I've heard that 130*C is safety limit for them :) .
Like Blaster said - almost all rotary build are loudly . But cascades can be used 24/7, why not :) .
About Danfoss SC series and loud ..owner of that cascade : http://www.xtremesystems.org/forums/...highlight=tiny said me that single-stage with SC21CL and 230v fan on condensor has been more loudly that this cascade, so 2 rotaries ;) .
Thanks
Peter

---

110c is ment to be close to max running temp, but as the guys say cascades are for benching. If the compressor does get too hot they shut down anyways. Ya can normally get atleast 4hours benching though no trouble :toast:

Nice build Piotres!

Quote:

---

Originally Posted by Blaster

Guys, its a cascade! Its not to be used 24/7, its not built for that.

---

lol I remember FUGGER saying at an exhibition he went to he left his cascade running for several days (IIRC) and it ran perfect.

Those weren't rotarys either, and they were pretty well cooled if I recall. Triple SC12MLX I beleive.

Yeah I think it was at CES, so probably air conditioned in the room.

Eh not as much as the reciprocatings run at lower temperatures, and without accumulators they get plenty of cooling, and sometimes floodback. But even a 120mm fan can cool em down. I've found that most rotarys if you put a fan above them blowing down, they cool down quite well.

Ahh that's a point. I remember him saying when the unit was built the compressors didn't even get warm at all. The returning gas probably still cool.

Yep, though something like r11 would probably be used in the industry, or often in a autocascade that needs compressor cooling badly.

Quote:

---

Originally Posted by n00b 0f l337

Yep, though something like r11 would probably be used in the industry, or often in a autocascade that needs compressor cooling badly.

---

Nowadays, R123 is the substitute for R11 :fact: .
:D

Yes it is, but you use whatcha got ;)
Not to mention its pretty hard to find non-barreled r11 or r123, small quantities are more expensive and annoying.

NoL,

Or could it be that a 12cc LBP/MPB recip doesn't have a 1KW+ motor in a rotary designed for HBP A/C...

Ethylene is a hot discharge gas, you'll have a high superheat on ST2 and you've got a lot of motor power going into the system, it'll be hot.

Nice cascade, very compact :)

Tom

+1 Tom, with R22 ethylen is a gas that involve a very high discharge pressure, you must take care about that effect, that's probably the cause of such hot compressor on 2nd stage

Quote:

---

Originally Posted by Clemmaster

.. R22 ethylen is a gas that involve a very high discharge pressure, you must take care about that effect, that's probably the cause of such hot compressor on 2nd stage

---

Lower condensing temperature for 2snd stage HX? How to do this?

I've made mistake, that's "high discharge temperature" I wanted to say :p

And to decreas this temp you must reduce the captube lentgh but on such systems with same compressors on both 2 stages It would be even harder for the first stage to work properly :/

It's not just with r22 condensing ethylene - although larger compression ratio of a higher discharge pressure at the same suction pressure will increase motor power making it worse. Discharge temperature is from the gas behaviour, ethylene is just a hot discharge gas.

What Clemmaster is saying is that by reducing the capillary restriction (-length or +diameter) and also increasing the ethylene charge you can reduce the superheat on ST2 suction, which will lower the discharge temperature proportionately to the reduction in suction temperature of ST2 in theory.

The problem is that with such a large motor it might be a loosing battle, rotaries for A/C just aren't meant to run like this.

But they work great for a few hours at a time, making ideal benchmark cascades :)

Tom

The first tests results (E6700 L629B875 @ ~1,63V real, as mentioned in first post):
10.750s 1M @ 4,73GHz
http://img411.imageshack.us/img411/2...0750qi7.th.jpg
11m25s 32M @ 4,65GHz - "the -99.4°C run" (the same run as in photo with thermometer)
http://img411.imageshack.us/img411/5918/32mqq8.th.jpg

Nothing to post in the xtreme overclocking section yet (I hope we will continue in near future with Commando or QuadGT 1,85V+, aiming for sub-10s 1M ;)). We were very short of time (hmm... real benches for 90 minutes or less) and it was far from optimized ;), 1066 strap but {4-4-4-4}-5-35-10-10-10-10 as I didn't know anything about the mobo we were using etc.
"We" = JMKS, piotres and promian (owner of this E6700 + P5B-Dlx + Ballistix).
It was the first ever subzero in my life and, to add more, my first ever benches with cooling better than air :). Wonderful experience and IMHO also wonderful cascade :D.

Quote:

---

Originally Posted by SoddemFX

It's not just with r22 condensing ethylene - although larger compression ratio of a higher discharge pressure at the same suction pressure will increase motor power making it worse. Discharge temperature is from the gas behaviour, ethylene is just a hot discharge gas.

What Clemmaster is saying is that by reducing the capillary restriction (-length or +diameter) and also increasing the ethylene charge you can reduce the superheat on ST2 suction, which will lower the discharge temperature proportionately to the reduction in suction temperature of ST2 in theory.

Tom

---

Isn't the discharge temperature also depends on condensing temperature?
Lowering condensing temperature we can lower discharge temp.
Another way would be to apply subcooling to liquid ethylene.

It plays a little role but the main depending cause is, as Tom said, the superheat on ST2 succion and the ethylen quantity (that's why by using long captube with ethylene involve a very high discharge temperature as the quantity must be higher, and it involve a high static pressure too). That's also the problem you get when using ethylen on both 2nd and 3rd stage, you have to shorten the captube on third stage as the condensing pressure on 3rd is very low (>5bar), but if you put the most powerfull compressor on this stage the 2nd and 1st stages won't work the best.

Quote:

---

Originally Posted by SoddemFX

It's not just with r22 condensing ethylene - although larger compression ratio of a higher discharge pressure at the same suction pressure will increase motor power making it worse. Discharge temperature is from the gas behaviour, ethylene is just a hot discharge gas.

What Clemmaster is saying is that by reducing the capillary restriction (-length or +diameter) and also increasing the ethylene charge you can reduce the superheat on ST2 suction, which will lower the discharge temperature proportionately to the reduction in suction temperature of ST2 in theory.

The problem is that with such a large motor it might be a loosing battle, rotaries for A/C just aren't meant to run like this.

But they work great for a few hours at a time, making ideal benchmark cascades :)

Tom

---

That's not what I meant, I don't speak very well :D

What I wanted to say is that R22 and Ethylen are both refrigerant (not using them together on cascaded systems, but R22 on a single stage for example) that involve a high discharge temperature.

Quote:

---

Originally Posted by Clemmaster

It plays a little role but the main depending cause is, as Tom said, the superheat on ST2 succion and the ethylen quantity...

---

High superheat on ST2 is caused by high condensing temperature in ST2 HX.
The liquid ethylen is intering evaporator at a temperature much higher than
saturated evaporator temperature, resulting in low efficiency and high superheat.
Reduicing condensing temperature and/or using subcooling we can reduce superheat and increase overall efficiency.

Quote:

---

Originally Posted by Cronos

High superheat on ST2 is caused by high condensing temperature in ST2 HX.
The liquid ethylen is intering evaporator at a temperature much higher than
saturated evaporator temperature, resulting in low efficiency and high superheat.
Reduicing condensing temperature and/or using subcooling we can reduce superheat and increase overall efficiency.

---

No, this is not true :)

High superheat is caused by either undercharge or by insufficient refrigerant entering the evaporator like you would get with a too restrictive capillary tube.

The attachment shows two ideal ethylene cycles, both with Tc = -40C, Te = -100C, Tsc = 0K, the only difference is the suerheat. The first cycle has a superheat of 30K, the second has a superheat of 70K, this is the only difference.

Because of this superheat the first cycle has a discharge temperature of ~70C, the second cycle has a discharge temperature of ~115C

These are both with ideal isentropic efficiency (Isen = 1) where the compression is at constant entropy i.e. follows the blue lines on the Log P-h diagram. In a small compressor the power dissipation through the case is enough to mimic this (Isen isn't 1 at all but the discharge temp behaves as if it was). Actually ssometimes the case cooling is even enough to actually drop the discharge temperature below that of ideal compression.

In a large compressor the discharge temperature behaves much more real, case power dissipation isn't enough and the compression process doesn't follow the blue lines, it slopes a lot, the discharge temperature is much much higher.

r22 isn't a hot discharge gas it's hotter than propane but it's not bad. The main thing is that you have the benifit of a first stage refrigerant that it's very easy to achieve lower superheat.

Tom

Quote:

---

Originally Posted by SoddemFX

No, this is not true :)

High superheat is caused by either undercharge or by insufficient refrigerant entering the evaporator like you would get with a too restrictive capillary tube.

---

Yes, formally.
If we increase refrigerant flow, we lower superheat. But how exactly this happen? Mainly through the increase in the saturated temperature, because the suction pressure is increased. But the change in suction return temperature is rather small, or at least is a small part of the superheat change.
And in this case the evaporator temperature will increase, which is not what we want.

Another way to lower superheat is to lower actual suction return. This can be achieved by reducing condensing temperature or/and liquid subcooling. The advantage is that evaporator temperature is not increased.

You increase refrigerant flow (mass flow) by decreasing the capillary restriction and usually increasing the systems charge.

If you lower condensing pressure with the same capillary tube you will decrease mass flow and increase superheat.

Quote:

---

Another way to lower superheat is to lower actual suction return.

---

:confused:

Superheat at the suction port for a single molecule refrigerant or azeotrope is saturation temperature minus suction gas temperature. In a zeotropic blend superheat is the dew point (end of evaporation) minus suction gas temperature.

Either way as long as the return refrigerant isn't in saturation, lowering its temperature will obviously lower superheat. This is what superheat is, so i'm not sure what you think is happening....?

When you have very high superheat the mass flow is not enough to effectively cool the load you have, increasing mass flow will obviously increase suction pressure - but if you're only currently supplying enough refrigerant to evaporate the in first little bit of the evaporator, with the gas being heavilly superheated throughout the rest of the evaporator then you didn't have enough flow rate anyway.

In the end you've gained evaporator capacity at the lower temperature.

Sorry for taking this so off topic piotres :)

Tom

Quote:

---

Originally Posted by SoddemFX

Sorry for taking this so off topic piotres :)
Tom

---

Yes, please, if this bothers the thread author or anyone else please say and I'll stop this conversation right away.

Quote:

---

Superheat at the suction port for a single molecule refrigerant or azeotrope is saturation temperature minus suction gas temperature....
Either way as long as the return refrigerant isn't in saturation, lowering its temperature will obviously lower superheat. This is what superheat is, so i'm not sure what you think is happening....?

---

Yes, i know this. So, you can lower superheat either by decreasing suction temperature or increasing suction pressure, right? :)

Quote:

---

If you lower condensing pressure with the same capillary tube you will decrease mass flow and increase superheat.

---

Sorry, i should have cleared this from the beginning -lets forget about cap tubes and assume we are using some more effective devices like TXV or even better electronic XV. We lower condensing pressure, we lower liquid temperature, but we don't change mass flow. Thats the idea.

Quote:

---

When you have very high superheat the mass flow is not enough to effectively cool the load you have, increasing mass flow will obviously increase suction pressure - but if you're only currently supplying enough refrigerant to evaporate the in first little bit of the evaporator, with the gas being heavilly superheated throughout the rest of the evaporator then you didn't have enough flow rate anyway.

---

Gas superheated in the direct die by direct contact with evaporator? Sorry, i have my doubts. The way it is constructed and its very small size imply there should be very small thermal transfer from evaporator to the gaseous refrigerant. The high superheat comes from the fact that initial liquid temperature is very high, and the fact that, for low mass flow, there is low back pressure.

Ok i understand what you mean now :)

Assuming you dont have any pressure drop in the suction line then the only way you'll increase suction pressure is by increasing mass flow, so i guess it's the same point.

The main problem i can think of with the TEV is that a TEV will maintain constant superheat within it's range. If you decrease the liquid temperature then you gain capacity due to change in enthalpy and assuming the TEV isn't in saturation the evaporating pressure will drop (at a given load) therefore mass flow in the system must have decreased.

Despite the lower suction gas temperature the compressor discharge temperature will be virtually unchanged due to the larger compression ratio but less dense suction gas.

Lower absoloute liquid temperatures increase the nominal sized capacity of a TEV but as long as youre not in saturation, the TEV will adjust to match the capacity it needs to maintain superheat. It adjusts capacity by changing mass flow.

Evaporator exit gas is superheated in almost all capillary CPU single stages at full load, on a cascade it's usually worse. Try this - Under full load (200W or whatever) put a thermocouple about 6 - 8 inches from the evaporator on the suction line to avoid conduction from evap messing with the temperature too much. Make sure the suction line is insulated and see what the temperature is - it'll already be quite a bit above saturation and that heat has had to have come from the evaporator. Try it :)

Tom

Don't forget, we don't need the same compression ratio as we have also lowered condensing pressure!

Either way, i think there should be some path to improve efficiency and lower superheat and discharge temperature by using more intelligent control devices, mainly electronic expansion valves, and other means, like subcooling. This may be not as important for the small load 200-300W, but i need 500-600W, so its critical for me!

But realistically how much can you lower condensing pressure of the ethylene in a two stage?

Subcooling's over emphasised, you can subcool to hell and back but in the end you're looking at +20% capacity and you'd need a lot of work to get it.

If you want to be evaporating at -100C with 5-600W you'd want to be using three stages. It's possible with two but you won't find a second stage compressor which can do it. Maybe one of the SC21 twins could do it on paper but it'd be nasty...

Tom

Maybe Danfoss new GS26MLX 1HP compressor with autocascade on the first stage?

The discharge pressure has to do with pressure drop through both succion and discharge valves too, and it depends mainly of the viscosity of the entering refrigerent, that's maybe the difference between gases discharge temperature.

Tom, R22 is the first stage gas that involve the highest discharge temperature, even hoter than R410a do in the same conditions, that's why, as I've already said, in industry some systems have to put forced liquid return into the compressor, by evaporation of the R22 directly in the succion line, near the succion connector because of the too high discharge temperature and the compressor that reach its maximal temperature.

Quote:

---

Originally Posted by Clemmaster

Tom, R22 is the first stage gas that involve the higher discharge temperature, even hoter than R410a do in the same conditions, that's why, as I've already saud, in industry some systems have to put forced liquid return into the compressor, by evaporation of the R22 directly in the succion line, near the succion connector because of the too high discharge temperature and the compressor that reach it's maximal temperature.

---

Yes, thats right. R410A LBP Compressors have a connector for evaporating directly on the cylinder head. (Because of the very hot compressing end temperatur)

Hello

Sorry for diging, I just want to say that those cascade has been tested on QX6700 and QX6850 and on both it's making sth around that in idle :

http://strony.aster.pl/pmpcfg/PC/kie...1/CPU_load.jpg

And in max stress during SP2004 around -94*C :) .

All these temperatures during OC - around 4.5 ghz , over 1.7V .

Regards
Peter