Two stage system troubleshooting

[nike123]

I have been involved recently in one project (as observer and if I could help with practical advice) which is pretty much undiscovered field for me.
Project is to make CASCADE cooling for freezer of medical samples to temperature of -80°C (if possible) with two Danfoss compressors. Details and diagram is in this PDF (http://www.mediafire.com/?cobjb64a0pvb195) document.
What we achieved so far is lowest temperature of -65°C in box measured with (i think) Eliwell thermostat and probe.
Now we have some problem and it manifest itself after some time temperature is around -45 to-50°C and than without obvious cause goes up to -25°C and stay there with heavily frosting on low stage compressor.

We contacted constructor of equipment and he said that problem may be in too short capillary.
Also we switched from original R507 refrigerant in high stage to R404 (it was unavailable to us and it will be for few next days).
Constructor also said that R404A has some problem with component separation at that low evaporation temperatures and that also could contribute to our problem.

We also wait for digital gauges for measurements so do not ask me now for detail numbers since they are not availiable.
Refco Digimon:eek: is useless in negative relative pressures and I waiting for Monday evening to get hold on Digicool to make accurate measurements.

What is your taught on design and experienced problems on this mater.
For downloading schematic and details, ask me for password on Private Message (http://www.refrigeration-engineer.com/forums/private.php?do=newpm&u=13885).

I am interested what is happening with oil on that temperatures and capillary tube diameter in relation to oil viscosity.
Also, does R410 have enough oil carrying capacity at that temperatures and if it need some addition (like propane or something else)to transport oil properly.

Please keep details and drawing private as this is not published or patented yet. Access to details and drawing will be available only to good known members of this great community.

Cheers!

[nando]

to go to -80°C, is it not better to work with ethyleen and propyleen?
To have this temp you will go to -90°C evaporatortemp.

[nike123]

to go to -80°C, is it not better to work with ethyleen and propyleen?
To have this temp you will go to -90°C evaporatortemp.
This is direct expansion system like this one (http://www.mediafire.com/?2wfkeaccq9d3km8).

[mad fridgie]

Not to much detail.

Have check the cap sizes, have they been sized for split load, not each for full load? (each either to short or to big a diameter)

[nike123]

Not to much detail.

Have check the cap sizes, have they been sized for split load, not each for full load? (each either to short or to big a diameter)

That is all what i have! 0 in TXV designation mean orifice size is 0.
If I understand corecty your question, box is made with two identical evaporators of pasive type like in regular fridge (behind inner aluminum wall embeded in PUR) working together and capilary is sized for split load.
What sould be designed superheat of low stage?
Also what superheat I should expect at -50°C?
Also, on high stage, how mutch should liquid be subcooled in liquid subcooler in order to low stage have normal operation?

[Aik]

Also we switched from original R507 refrigerant in high stage to R404 (it was unavailable to us and it will be for few next days).
Constructor also said that R404A has some problem with component separation at that low evaporation temperatures and that also could contribute to our problem.

Acording to ASHRAE 2006 (Ultralow-Temperature) in cascade systems:



Typical refrigerants for the high-temperature circuit include R-22, ammonia, R-507, and R-404a.

What's refrigerant on low stage system?

[mad fridgie]

That is all what i have! 0 in TXV designation mean orifice size is 0.
If I understand corecty your question, box is made with two identical evaporators of pasive type like in regular fridge (behind inner aluminum wall embeded in PUR) working together and capilary is sized for split load.
What sould be designed superheat of low stage?
Also what superheat I should expect at -50°C?
Also, on high stage, how mutch should liquid be subcooled in liquid subcooler in order to low stage have normal operation?
It looks like the sub cooling is only there to reduce the pressure drop across the high stage evap (no direct increase in performance).
Re superheat, on low side will be high on pulldown, and reduce to close to 0 a design load (evap will be flooded)

[nike123]

what's refrigerant on low stage system?


r410a
.........

[nike123]

It looks like the sub cooling is only there to reduce the pressure drop across the high stage evap (no direct increase in performance).
Re superheat, on low side will be high on pulldown, and reduce to close to 0 a design load (evap will be flooded)

If we have 0K superheat at evaporator and short pipes to compressor, dont we have danger of liquid in compressor?

[mad fridgie]

If we have 0K superheat at evaporator and short pipes to compressor, dont we have danger of liquid in compressor?
You normally have your caps in contact with suction, to increase superheat at the compressor.
It is critical charge, so liquid should not return to the compressor.
What temp do you want in the chamber?

[nike123]

You normally have your caps in contact with suction, to increase superheat at the compressor.
It is critical charge, so liquid should not return to the compressor.
What temp do you want in the chamber?

There is liquid receiver in both circuits therefore I don't think it is critically charged.
We aim at -80°C but -70°C would satisfy.

[mad fridgie]

There is liquid receiver in both circuits therefore I don't think it is critically charged.
We aim at -80°C but -70°C would satisfy.
You do not have a liquid reciever with caps systems, remove!, you can have a suction acc.

[nike123]

You do not have a liquid reciever with caps systems, remove!, you can have a suction acc.

Could you explain reasoning?

Becaose of this?


Therefore, when the compressor stops, the condenser empties into the evaporator, which fills with liquid.


If a liquid receiver is introduced, its contents could migrate into the evaporator when the compressor stops and fill it completely, especially if the evaporator is 'cold'. There would then be extensive liquid slugging when the compressor starts up.


A liquid receiver is therefore never installed in a system fitted with a capillary expansion device.

http://www.kotza-international.com/site/gb/ccn.html

[mad fridgie]

Could you explain reasoning?
A cap system relies on genuine sub cooling during the process, a reciever which is fed from a free draining condensor does NOT have liquid sub cooling (excluding the liquid head in the reciever).
Also the cap pressure drop is reliant on the flow, pressure drop and sub cooling, the reciever muffles the balance, as the working conditions change.
(stated in practical sense)

[mad fridgie]

And that as well.

[nike123]

Thanks mad fridgie!;)

[mad fridgie]

Pleased i could help, and thanks for all the assistance that you give to others, on the this furom

[Aik]

r410a
.........
:D:D:D
It's very funny, evaporating pressure below than atmosphere pressure...it's wrong.


A cap system relies on genuine sub cooling during the process, a reciever which is fed from a free draining condensor does NOT have liquid sub cooling (excluding the liquid head in the reciever).

Very interesting... what do you think about this article?
http://www.sporlanonline.com/February%2010%20Cold%20WAR.pdf


Could you explain reasoning?

Becaose of this?
IMHO, you may add solenoid which stops flow when compressor stoped and leave receiver.

[Aik]

Acording to ASHRAE 2006 (Ultralow-Temperature) in cascade systems:



For the low-temperature circuit, a high-pressure refrigerant
with a high vapor density (even at low temperatures) is chosen. For many years, R-503, an azeotropic mixture of R-13 and R-23, was a popular choice, but R-503 is no longer available because R-13 is an ozone-depleting chlorofluorocarbon (CFC). R-23 could be and has been used alone, but R-508b, an azeotrope of R-23 and R-116, as superior properties, as discussed in the section on Refrigerants for Low-Temperature Circuit.

[DTLarca]

Very interesting... what do you think about this article?
http://www.sporlanonline.com/February%2010%20Cold%20WAR.pdf

That article puts in a condensed fashion the arguments for subcooling in receivers, including free-draining receivers, that I have been pushing for years on these discussion forums - especially alt.hvac and sci.eng.heat-vent-ac - Andy participated in the some of them. I wonder if it is a coincidence that he calls the engineer in that article "Mark"?

[nike123]

And that as well.
What is not showed in diagram is that non-return valve is in line which connect receiver and condenser.
Therefore, I suppose, that migration is prevented by that!

Edit: I am not sure where is placed non-return valve.

[nike123]

Pleased i could help, and thanks for all the assistance that you give to others, on the this furom

That is why we are all here, to share what we know!;)

[nike123]

IMHO, you may add solenoid which stops flow when compressor stoped and leave receiver.

I missed to draw non-return valve in that diagram.

[Aik]

I missed to draw non-return valve in that diagram.
I mean solenoid before cap system...which stops refrigerant migration from receiver to evaporators (low stage system) through caps when compressor stoped.

[nike123]

I mean solenoid before cap system...
I am not there at moment but I will check later where is non-return valve positioned.

[mad fridgie]

That article puts in a condensed fashion the arguments for subcooling in receivers, including free-draining receivers, that I have been pushing for years on these discussion forums - especially alt.hvac and sci.eng.heat-vent-ac - Andy participated in the some of them. I wonder if it is a coincidence that he calls the engineer in that article "Mark"?
The general argument abot sub-cooling is more related to where you measure the pressure. The common method is to measure compressure discharge pressure, then measure liquid temp, this would give the impression of the amount of liquid sub cooling, what is shown in the article is that the discharge pressure and liquid pressure are the same, on free draining air cooled condensors and recievers, that there is likely to be a significant pressure drop, so the actual sub cooling is significantly less than thought. This can be seen with many of the "R4" based refrigerants, how many of you have struggled to get a clear sight glass, with glide small pressure changes allow for slight evaporation, with being able to recondense.
Also the type and style of the reciever is important, how does it flow, is the boundary affected.
This can be a very complex area.

[mad fridgie]

:D:D:D
It's very funny, evaporating pressure below than atmosphere pressure...it's wrong.


Very interesting... what do you think about this article?
http://www.sporlanonline.com/February%2010%20Cold%20WAR.pdf


IMHO, you may add solenoid which stops flow when compressor stoped and leave receiver.
If we look at the drawing in the above attachment,
we can see a number flaws with the articles argument.
1; The sub cooled liquid from the condensor pass direct through the saturated vapour, "what an excellent energy transfer method" the liwuid would warm and the vapour would condense.
2; If the thermal length is too short for the above to occur fully, the liquid stream will cause a wave/ripple on the top of the liquid held within the reciever, breaking the surface boundary layer, enhanceing the heat transfer between the liquid and the vapour, whilst also reducing the natural stratification.
3; Even at this point, assume the liquid is highly sub cooled (which i believe it would not), the sub cooled liquid has to pass through the body of the liquid in the reciever, which at the top is at saturation, mixing and temperature equalibrium will occur.
Hence little actual sub cooling at the reciever outlet. (as stated in my earlier post about pressure drop)
In an ideal world, your reciever would be vertical with a reasonable head of liquid. The sub cooled liquid would enter the bottom of the reciever and would be sized that no mixing occurs. The naturally stratisfied liquid/vaour would insulate the sub cooled liquid from picking up large amounts of energy from the vapour. And you then would indeed have genuine sub cooled liquid leaving the reciever

[DTLarca]

If we look at the drawing in the above attachment,
we can see a number flaws with the articles argument.
1; The sub cooled liquid from the condensor pass direct through the saturated vapour, "what an excellent energy transfer method" the liwuid would warm and the vapour would condense.
2; If the thermal length is too short for the above to occur fully, the liquid stream will cause a wave/ripple on the top of the liquid held within the reciever, breaking the surface boundary layer, enhanceing the heat transfer between the liquid and the vapour, whilst also reducing the natural stratification.
3; Even at this point, assume the liquid is highly sub cooled (which i believe it would not), the sub cooled liquid has to pass through the body of the liquid in the reciever, which at the top is at saturation, mixing and temperature equalibrium will occur.
Hence little actual sub cooling at the reciever outlet. (as stated in my earlier post about pressure drop)
In an ideal world, your reciever would be vertical with a reasonable head of liquid. The sub cooled liquid would enter the bottom of the reciever and would be sized that no mixing occurs. The naturally stratisfied liquid/vaour would insulate the sub cooled liquid from picking up large amounts of energy from the vapour. And you then would indeed have genuine sub cooled liquid leaving the reciever

I think your arguments are good arguments for your position. But I believe I have better arguments for my position. I have added the issue to the queue in my list of future articles :)

[mad fridgie]

I think your arguments are good arguments for your position. But I believe I have better arguments for my position. I have added the issue to the queue in my list of future articles :)
Look forward to it?

[nike123]

Mark and Mad, I love to see your argumentations!:D
Especially if they are in connection with my system problems.

[nike123]

:D:D:D
It's very funny, evaporating pressure below than atmosphere pressure...it's wrong.


I don't see problem there if system is hermetic.

[Gary]

Does the low stage have an oil separator? What kind of oil? Does it have some oil carrying fluid (propane, pentane, R12, etc.) added?

Generally, as a substitute for low stage subcooling, I measure the interstage temp difference (low stage SCT minus high stage SST).

[mad fridgie]

I think your check valve, might be on your second stage compressor discharge, to stop migration into your comp head.
"Have you not read what Nike wrote?" "Well I thought I did" "Read again" 'I have and I see your point, The check valve is between the reciever and condenser"
How ever this would not stop migration to the evap, and would reduce ever so slightly migration to the head.

Mad and Mad (having an argument)

[install monkey]

normally r290 or r600 for cascade units to reach -80

[Aik]

If we look at the drawing in the above attachment,
we can see a number flaws with the articles argument.
1; The sub cooled liquid from the condensor pass direct through the saturated vapour, "what an excellent energy transfer method" the liwuid would warm and the vapour would condense.
2; If the thermal length is too short for the above to occur fully, the liquid stream will cause a wave/ripple on the top of the liquid held within the reciever, breaking the surface boundary layer, enhanceing the heat transfer between the liquid and the vapour, whilst also reducing the natural stratification.
3; Even at this point, assume the liquid is highly sub cooled (which i believe it would not), the sub cooled liquid has to pass through the body of the liquid in the reciever, which at the top is at saturation, mixing and temperature equalibrium will occur.
Hence little actual sub cooling at the reciever outlet. (as stated in my earlier post about pressure drop)
In an ideal world, your reciever would be vertical with a reasonable head of liquid. The sub cooled liquid would enter the bottom of the reciever and would be sized that no mixing occurs. The naturally stratisfied liquid/vaour would insulate the sub cooled liquid from picking up large amounts of energy from the vapour. And you then would indeed have genuine sub cooled liquid leaving the reciever
From the physical point of view I totally agree with you, but this article has other opinion and I don't undestand it's explanation...I don't know where is the truth...:confused:

I think your arguments are good arguments for your position. But I believe I have better arguments for my position. I have added the issue to the queue in my list of future articles :)
It would be better to look :)

[Aik]

I don't see problem there if system is hermetic.
Of course there is no problem if system is hermetic, but if system has "leakage" then moist air migrate to system because atmosphere pressure higher...

[nike123]

I think your check valve, might be on your second stage compressor discharge, to stop migration into your comp head.
"Have you not read what Nike wrote?" "Well I thought I did" "Read again" 'I have and I see your point, The check valve is between the reciever and condenser"
How ever this would not stop migration to the evap, and would reduce ever so slightly migration to the head.

Mad and Mad (having an argument)

To clarify this, non-return vale is positioned between compressor and condenser.

[Gary]

What we achieved so far is lowest temperature of -65°C in box measured with (i think) Eliwell thermostat and probe.
Now we have some problem and it manifest itself after some time temperature is around -45 to-50°C and than without obvious cause goes up to -25°C and stay there with heavily frosting on low stage compressor.


This would seem to indicate oil logging problems, oil coating and insulating the inner surfaces of the evaporator. If the evaporator is insulated, then we might expect floodback to the compressor. Unless/until the oil return problem is resolved, other problems cannot be properly identified.

[Gary]

Even the best of oil separators will allow some tiny percentage to pass through. If that oil remains in the evaporator, then over a period of time sufficient oil will accumulate to cause logging problems. For this reason some small amount of oil carrying fluid must be added to ensure oil return, even with a good separator.

[mad fridgie]

From the physical point of view I totally agree with you, but this article has other opinion and I don't undestand it's explanation...I don't know where is the truth...:confused:

It would be better to look :)
If we look at the article, and having an open mind, I will attempt to explain what they mean.
Where liquid and vapour meet "boundry layer" we have some surface tension "imagine a thin layer of another material" the surface is a bit stronger than the liquid below and the vapour above. So we need a liitle move force (breaking the layer) to get transfer of energy.
Also this layer attracts non moving particles of both liquid and vapour refrigerant. The combination acts as an insulator (same as insulation in cold room) and aids stratification
Hence we could have quite a big difference in temperature between the bottom of the reciever and the vapour at the top. In the example I do not believe this happens. It is upto you to decide which will actually happen.
Being practical, I would suspect that both are possible and/or a bit of both, All I say is that if you want to know genuine sub cooling, you measure your liquid pressure at the point you measure your temperature, and do not rely on your compressor discharge pressure as being the liquid pressure.

[nike123]

Does the low stage have an oil separator? What kind of oil? Does it have some oil carrying fluid (propane, pentane, R12, etc.) added?
NO!
POE!
NO!


Generally, as a substitute for low stage subcooling, I measure the interstage temp difference (low stage SCT minus high stage SST).

I was not been there today! Project is on hold for few days.:mad:

[nike123]

This would seem to indicate oil logging problems, oil coating and insulating the inner surfaces of the evaporator. If the evaporator is insulated, then we might expect floodback to the compressor. Unless/until the oil return problem is resolved, other problems cannot be properly identified.

Yep, that was what cross my mind.
Gary you have PM!

[nike123]

Of course there is no problem if system is hermetic, but if system has "leakage" then moist air migrate to system because atmosphere pressure higher...

I don't want to change anything if constructor of unit is not allowing. Let this to be his child!;)

[DTLarca]

All I say is that if you want to know genuine sub cooling, you measure your liquid pressure at the point you measure your temperature, and do not rely on your compressor discharge pressure as being the liquid pressure.

I have come across very low footprint air cooled water chillers - also very noisy - that have a condenser refrigerant side pressure drop - at full load - of 1.2bar. That's like say 17 bar at the compressor discharge and 15.8 at the condenser outlet. Talking R22 here you can see easily the degree to which an apparent subcool could mistaken a young tech measuring pressure at the compressor discharge.

On such chillers the best place to connect the fan speed control pressure reference is on the liquid line. Not just because ultimately that is the pressure we are trying to maintain by controlling the fan speeds - the TEV capacity maintaining liquid line pressure - but because if controlled via the discharge then when the compressor unloads to say 25% the condenser pressure drop will have dropped to 1/16th of 1.2bar meaning that to maintain the discharge pressure at the compressor the fans would have to slow down enough to raise the liquid line by 1.125bar as a consequence. Thus the available TEV pressure drop would have increased at the time when we do not need as much capacity out of it and would prefer it to have reduced in capacity so as to have better TEV low superheat stability.

[nike123]

I Thus the available TEV pressure drop would have increased at the time when we do not need as much capacity out of it and would prefer it to have reduced in capacity so as to have better TEV low superheat stability.

Mark, could you explain is this related anyhow with original problem of this thread. I don't see relation. Please, show me if there is!

[DTLarca]

Mark, could you explain is this related anyhow with original problem of this thread. I don't see relation. Please, show me if there is!

It's not related - the moderator should package all these off topic posts and move them into a newly created thread with a matching title. This is rarely done - at least since 1995 this is how it happens - people come along and start discussions within discussions - I tried to prevent such occurrences myself back in 1995 and 1996 but gave up in the end.

If your compressors and cap-tubes are right then your problem is as Gary says - oil type and return provisions. This very topic has been discussed before either here on this discussion board or on one of the old Usenet groups - I forget where - but changing the oil type solved the problem.

I do want to discuss the principle of subcool occurrence with Mad Fridge - but I will start another thread for that - good luck :)

[mad fridgie]

I don't want to change anything if constructor of unit is not allowing. Let this to be his child!;)
It would seem that you are between a rock and a hard place, on this project.
When doing this type of R&D, it is the quantity and quality of data that you require (more the better) measure pressure and temp at each point in the system.
To start with do not worry about the end result.
Get the system to balance, and know your upper and lower limits.
Only make "one" change at a time.
As gary has clearly stated Oil will become a problem, but for R&D purposes can be flushed from the system after each system modification.
Use base principles,

[Peter_1]

Nike, add a little bit R134a into the system to rinse the cap tubes. The R134a remains liquid and flushes it.
Why R410a for the LP?

[Peter_1]

To clarify this, non-return vale is positioned between compressor and condenser.

Why an NRV? Every DP provokes a lower efficiency

[mad fridgie]

Why an NRV? Every DP provokes a lower efficiency
At these conditions (compressor being so cold), vapour can condense in the compressor head, and seep into the cylinders.

[nike123]

It's not related -
If your compressors and cap-tubes are right then your problem is as Gary says - oil type and return provisions. This very topic has been discussed before either here on this discussion board or on one of the old Usenet groups - I forget where - but changing the oil type solved the problem.

I do want to discuss the principle of subcool occurrence with Mad Fridge - but I will start another thread for that - good luck :)

OK, I appreciate your reasoning.;)

[nike123]

Nike, add a little bit R134a into the system to rinse the cap tubes. The R134a remains liquid and flushes it.
Why R410a for the LP?

I am there simply as possible source of valuable information. To me, it is much more important to gain some general knowledge on two stage cascade.
I don't have any influence on design and commission.
My primary goal is to gain some knowledge on cascades from concrete example from life.

[nike123]

Nike, add a little bit R134a into the system to rinse the cap tubes. The R134a remains liquid and flushes it.
Why R410a for the LP?
I don't know reasoning! I presume that is because of availability.

[nike123]

Even the best of oil separators will allow some tiny percentage to pass through. If that oil remains in the evaporator, then over a period of time sufficient oil will accumulate to cause logging problems. For this reason some small amount of oil carrying fluid must be added to ensure oil return, even with a good separator.
I have advised as you suggested!
We will see in next few days if advice is accepted.

[Gary]

With a good oil separator, only a small amount of oil will be flowing through the evaporator, thus only a small amount of carrying fluid is needed.

With no separator, large amounts of oil will be flowing through the evaporator, thus large amounts of carrying fluid will be needed... enough to substantially alter the pressure/temperature relationship of the refrigerant, resulting in higher evaporating temperature and higher box temperature.

Low stage oil return is a serious problem in cascade systems, which must be addressed.

[Gary]

It looks like the sub cooling is only there to reduce the pressure drop across the high stage evap (no direct increase in performance).


I would expect some reduction in flashing at the interstage, increasing the effective heat transfer surface area... but the same effect could be accomplished by slightly increasing interstage size. I'm thinking the reduced flashing is not worth the added complexity.

[mad fridgie]

I would expect some reduction in flashing at the interstage, increasing the effective heat transfer surface area... but the same effect could be accomplished by slightly increasing interstage size. I'm thinking the reduced flashing is not worth the added complexity.
Correct it comes down to how they have selected the heat exchanger

[Peter_1]

In the past, we made several cascades for Meissner coils (R501/R13, R404a/R23 and R1150/R1270) and we still service +/- 20 cascade and autocascades these days. Our experience is that you must have a very, very dry system. The smallest quantity of water in the system will, together with the oil clog your system.
As Gary said, an oil separator isn't always working 100%
What's in the Visio file? The same information as in the pdf?
Is this used as a CPU cooler.

[nike123]

What's in the Visio file? The same information as in the pdf?
Is this used as a CPU cooler.
No! It is used as freezer for medical samples.
File for download is PDF of diagram I made in Visio. That is actual diagram of system with some informations on it.
I will send you password for download at private message.

Other PDF is model from Coolpack.