Lower condensing temperature for 2snd stage HX? How to do this?Quote:
Originally Posted by Clemmaster
Printable View
Lower condensing temperature for 2snd stage HX? How to do this?Quote:
Originally Posted by Clemmaster
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.
Isn't the discharge temperature also depends on condensing temperature?Quote:
Originally Posted by SoddemFX
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.
That's not what I meant, I don't speak very well :DQuote:
Originally Posted by SoddemFX
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.
High superheat on ST2 is caused by high condensing temperature in ST2 HX.Quote:
Originally Posted by Clemmaster
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 :)Quote:
Originally Posted by Cronos
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
Yes, formally.Quote:
Originally Posted by SoddemFX
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.
:confused:Quote:
Another way to lower superheat is to lower actual suction return.
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
Yes, please, if this bothers the thread author or anyone else please say and I'll stop this conversation right away.Quote:
Originally Posted by SoddemFX
Yes, i know this. So, you can lower superheat either by decreasing suction temperature or increasing suction pressure, right? :)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....?
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:
If you lower condensing pressure with the same capillary tube you will decrease mass flow and increase superheat.
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.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.
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.
Yes, thats right. R410A LBP Compressors have a connector for evaporating directly on the cylinder head. (Because of the very hot compressing end temperatur)Quote:
Originally Posted by Clemmaster
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