Strange TXV setup

[Chef]

Moving the bulb to the freezer would solve the charge migration problem.

Then we would see if the other plates are indeed SH cooled.

MF has a good idea here and it may still allow SH cooled plates later in the system and avoid the migration.

So we now have a new fridge term - SH cooled - thats going to be tricky to explain but it is a real phenomena!!

Chef

[anna.savvy]

Nice representation.

[mad fridgie]

If the compressors are in series they must run at he same RPM to get the same flow rate but they will have half the pressure ratio.

If the compressors are in parallel then they will need to be run at half speed and and they will still have the same high pressure ratio.

Its a thought but not sure its the best solution.

Chef

Car AC comp have avery short piston stroke (especially swash plate), so pressure difference becomes very important as far as volumetric efficiency.
There is a gap between the piston and the discharge valve, this is filled with compressed gas, on the down stroke (suction) no gas is drawn in until the higher pressure gas expands to meet the suction pressure.
So you actual mass flow through the compressor is likely to be quite low. Low volumetric efficiency.
Going 2 stage, reduces the the pressure drop, increasing volumetric efficiency, thus increasing actual mass flow. So slowing the speed down will bring actual mass flow back to original amount. Slower speed less mechanical stress, lower discharge temps.Increased life!!! and is actually more efficient as far as power goes.
(forgot to say interstage does need cooling)

[Gary]

So we now have a new fridge term - SH cooled - thats going to be tricky to explain but it is a real phenomena!!

A new term for a bad idea. High compressor inlet SH is a real phenomena.

[mad fridgie]

SH cooling is not new, the term we use is "useful superheat" verses non useful superheat (heat absorbed say down the suction line, and can also can be expressed as suction line pressure drop)

[Chef]

MF - I understand the prblems with the swash plate compressors and the volumetric efficiency thats why thet fail a lot but if you run slower then the volumtric flowrate will drop.
These units are are 140cc per rev so if we slow it down it will be less volume and less mass flow.

The pair of compressors must run at the same speed!

Well actually the first one must run at the same speed and the second one must be smaller and designed to run with the output of the 1st compressor - seems quite tricky but solvable.

Chef

[Chef]

Nice representation.

Good to meet you anna.savvy and thanks for joining in.

Please go ahead and ask or contribute as you feel, sometimes the discussions go quite quickly and we gloss over stuff so please interrupt and ask for clarification - might even help us as well.

Chef

[mad fridgie]

MF - I understand the prblems with the swash plate compressors and the volumetric efficiency thats why thet fail a lot but if you run slower then the volumtric flowrate will drop.
These units are are 140cc per rev so if we slow it down it will be less volume and less mass flow.

The pair of compressors must run at the same speed!

Chef

Reducing the pressure differential across the compressor will increase the actual mass flow (refrigeration effect), thus to keep the mass flow the same, speed must be reduced on each compressor.
Do the 2 compressor need to run at the same speed,
"no" , but for ease "yes"

[mad fridgie]

Taking some estimates, using 2 of the same type of compressor (as you are presently using)
Low stage 50% speed of present speed
High stage 30% of present speed.

[Chef]

Reducing the pressure differential across the compressor will increase the actual mass flow (refrigeration effect), thus to keep the mass flow the same, speed must be reduced on each compressor.
Do the 2 compressor need to run at the same speed,
"no" , but for ease "yes"

OK if the volumetric efficiency improves because of the lower pressure ratio the compressor can be run slower by the ratio of the efficiencies at the 2 pressure ratios.
It may be at best 20 %. (probably more like 5%)

So I agree it will go a bit slower. But the second compressor will now be getting gas a higher pressure and so the volume will be lower and it will need to be a smaller compressor. If the first compressor is 140cc per rev the second might be around 50cc per rev. The problem here is there are not many AC car units of this
size? In fact I do not know of any?

Double stage compression with inter cooling is a fascinating subject but I am not sure its right for this application - yet.

Chef


size

[mad fridgie]

maybe, you could give the volume of each eutectic plate (freezing point for each would be nice), then I can calculate duty.
Also what speed is the comp running at?
how long does it take to freeze (from a warm start)

[mad fridgie]

OK if the volumetric efficiency improves because of the lower pressure ratio the compressor can be run slower by the ratio of the efficiencies at the 2 pressure ratios.
It may be at best 20 %. (probably more like 5%)

So I agree it will go a bit slower. But the second compressor will now be getting gas a higher pressure and so the volume will be lower and it will need to be a smaller compressor. If the first compressor is 140cc per rev the second might be around 50cc per rev. The problem here is there are not many AC car units of this
size? In fact I do not know of any?

Double stage compression with inter cooling is a fascinating subject but I am not sure its right for this application - yet.

Chef


size

Is the point above based upon adiabatic efficiency, which is correct, but volumetric efficiency, I think you will be surpised, how poor your compressor is actually performing at present working condtions.

[Chef]

maybe, you could give the volume of each eutectic plate (freezing point for each would be nice), then I can calculate duty.
Also what speed is the comp running at?
how long does it take to freeze (from a warm start)

The freezer plates are 5 litres each and freeze at -15 to - 20C, the fridge plate is 5 litres and freezes at around 5 litres and freezes at -6C and the chiller is 4 litres and freezes at -4C

The compressor runs at 1800RPM - no variation whatsoever.

From warm it takes a while to freeze - data to hand shows about 2 to 2 1/2 hours from 32C down to -15C but that is without any box load as the units are cooled down before stuff is loaded - with a full load it could be much longer. Otherwise its a max of 1.5 hours a day to keep everything frosty.

Chef

[mad fridgie]

The freezer plates are 5 litres each and freeze at -15 to - 20C, the fridge plate is 5 litres and freezes at around 5 litres and freezes at -6C and the chiller is 4 litres and freezes at -4C

The compressor runs at 1800RPM - no variation whatsoever.

From warm it takes a while to freeze - data to hand shows about 2 to 2 1/2 hours from 32C down to -15C but that is without any box load as the units are cooled down before stuff is loaded - with a full load it could be much longer. Otherwise its a max of 1.5 hours a day to keep everything frosty.

Chef

Thanks, doing some very BASIC calcs, I think you are running at 30% volumetric efficeincy (very low)

[Gary]

I would go with the parallel circuits (second TXV, EPR). This is a common configuration where two or more temps are maintained off a single machine. The EPR prevents the higher temp circuit from dropping below its preset value.

And I suspect getting some cooling back to the compressor (lower inlet superheat) will greatly extend its life.

[monkey spanners]

If they are only running 1.5hrs a day but fail after a year, thats not many hours... about 18 days in fridge years!

[mad fridgie]

Peak load is around 1Kw,
Comp if 100% volumetric efficient should produce 3.7Kw (at stated conditions)
So actual Vol eff. is 27%, actually lower if you consider that peak load includes pulldown, so when just topping up vol eff. is likely to be 20%ish
SH cooling, is around 100watts, slightly lower than required, indicating that there is indeed a little bit of liquid carry over (not much) Which is likely to boiled off within a very short period in the fridge plate.
So the question is, when the bulb is moved into the freezer and freezer load is 100% satisfied, the TXV is closed to its minimum, will this minimum flow allow sufficent liquid carry over to chiller the next 2 plates.
If not, then a simple change of the orifice to a larger size may do the trick.
(based upon R134a)

[Chef]

I would go with the parallel circuits (second TXV, EPR). This is a common configuration where two or more temps are maintained off a single machine. The EPR prevents the higher temp circuit from dropping below its preset value.

And I suspect getting some cooling back to the compressor (lower inlet superheat) will greatly extend its life.

It seems a reasonable idea, if you have an example or link where I can see how its done it may worth the effort.

We have started to pull out the genset today and replace it but I am not sure if a car compressor will fit the new unit till it arrives. If not I might have to fork out for a hemetic compressor of some type, which one I have no idea yet as it has to be 110v 60Hz alla USA style.

Its the obvious time to do a complete refurb but it will mean designing it without the chance of any trials and testing along the way, always tricky with such a kit of bits to work with.

As monkey spanners said it is about 18 (or 23) days of running for each year of operation which is pretty awful when it gets put like that.

Chef

[Chef]

Peak load is around 1Kw,
Comp if 100% volumetric efficient should produce 3.7Kw (at stated conditions)
So actual Vol eff. is 27%, actually lower if you consider that peak load includes pulldown, so when just topping up vol eff. is likely to be 20%ish
SH cooling, is around 100watts, slightly lower than required, indicating that there is indeed a little bit of liquid carry over (not much) Which is likely to boiled off within a very short period in the fridge plate.
So the question is, when the bulb is moved into the freezer and freezer load is 100% satisfied, the TXV is closed to its minimum, will this minimum flow allow sufficent liquid carry over to chiller the next 2 plates.
If not, then a simple change of the orifice to a larger size may do the trick.
(based upon R134a)

I get about 3.5Kw of cooling as well and compressor seems to use a bit more than 1Kw - may be 1.4 or 1.5 Kw. I looked at the specs for the compressor and at a compression ratio of about 10 it has a volumetric efficiency of just 60% which seems very low and could be why its failing early.

I have no way to test the bulb in the freezer idea as we have just pulled out the genset today- more work I fear.
I have no idea if the plates further down stream will cold if this mod is made.
But just looking at the SH cooling available it seems way to small to be effective and is probably not the best option to rely on for controllable cooling.

Chef

[Gary]

It seems a reasonable idea, if you have an example or link where I can see how its done it may worth the effort.

http://www.sporlanonline.com/90-20.pdf

[Gary]

This might be better:

http://www.sporlanonline.com/90-21.pdf

[mad fridgie]

With the new information, I would suggest you change the refrigerant to R404a (new hermetic LT compressor rated to -40C), change TEV (ext equ), place bulb in freezer, remove liquidline/ suction line heat exchanger (already enough superheat). Leave system as is.
size comp at 1kw -20C SST 20C SH
Reason, at the end of pull down, SST will drop, so will SH, the lower superheat, will freeze the fridge plates.
Ther freezer will get colder, this is never a problem, the chances of actually reaching -40SST are remote, so equipment all ways in spec. Reliable. Plus system at lower temps still in positive pressure

[Chef]

Gary - The pdf showing the 2 TXV's and the EPR's is very good and explains it. I get the idea now and will have to think on this for while.

BUT

Now I have 2 seperate systems but they are piped as a single system. So the freezer plates have 1/2" and 3/8" pipes in parrellel through both plates and they then join the fridge which has 1/2" pipes and the chiller has 5/8" pipes. So now one TXV will run the Freezer and one the fridge and chiller, lower flows in each one, oil return? Maybe a complete repiping of all the plates - not really a option as the piping is mostly not removeable without a chainsaw, its built in.

Chef

[Chef]

With the new information, I would suggest you change the refrigerant to R404a (new hermetic LT compressor rated to -40C), change TEV (ext equ), place bulb in freezer, remove liquidline/ suction line heat exchanger (already enough superheat). Leave system as is.
size comp at 1kw -20C SST 20C SH
Reason, at the end of pull down, SST will drop, so will SH, the lower superheat, will freeze the fridge plates.
The freezer will get colder, this is never a problem, the chances of actually reaching -40SST are remote, so equipment all ways in spec. Reliable. Plus system at lower temps still in positive pressure

Ah thinking out the box again MF. the R404 has some merit but as we already have to carry R22, R134a and R** another 13Kg gas bottle is a stretch and we can only get the 13Kg bottles here.

The idea of having the freezer ridiculously cold is a nice one but I would have to be very convinced the SH cooling was going to do the job. Move away from the orifice and critical charge and may never work again.

As I said in the post to Gary, most of this stuff is built in ( ie part of the structure) and cannot be removed or easily changed so the solution has to be both technically possible and hopefully elegant.

The dual TXV approach proposed by Gary would be fine except for my reservations over the lower flows in each section.

But keep thinking MF thats what your good at.

Chef

[mad fridgie]

You can get an LT R22 compressor, similar principles (not much info available, as R22 not used here anymore) Worst case if does not work, you have to move probe, set then high superheat setting.
Only thing i would do is fit a high discharge thermostat.
It will work!

[Gary]

Gary - The pdf showing the 2 TXV's and the EPR's is very good and explains it. I get the idea now and will have to think on this for while.

BUT

Now I have 2 seperate systems but they are piped as a single system. So the freezer plates have 1/2" and 3/8" pipes in parrellel through both plates and they then join the fridge which has 1/2" pipes and the chiller has 5/8" pipes. So now one TXV will run the Freezer and one the fridge and chiller, lower flows in each one, oil return? Maybe a complete repiping of all the plates - not really a option as the piping is mostly not removeable without a chainsaw, its built in.

Chef

Liquid rich mixtures carry oil very well. Superheated vapor does not.

Wherever the line is sloped downhill, moving the oil is not a problem. It simply rolls down the hill.

So... oil return is only a concern in the suction line and only where the oil is being moved uphill.

I tend to think that if oil return has not been a problem in the past with the two plates running only superheated vapor, then it will probably not be a problem in the future, despite the decreased flow.

I can't see the system from here, so you will have to be the judge.

[mad fridgie]

Option 2 "Double expansion"
New TEV (ex equ), feed fridge plates first, then to freezer plates, install hand expansion valve (needle valve) where present TEV is. "Install pressure port either side of HEV"
TEV bulb on freezer plate outlet.
Set hand expansion valve so inlet pressure is -6C.
Effectively as far as control of the TEV the fridge plates and hand expansion valve) are acting as a distributor (pressure drop) You could use an EPR instead of the HEV, but not sure what will happen over the long term to the valve seat with a continious liquid vapour mix.
Only mod to cabinets is a new suction suction from the freezer.

[mad fridgie]

There is awide range of semihermetics, that will run down to below -40C on R22 (head cooling fan required), good for transport, no internal springs to worry about, only problem $$$$$$$$$$$$

[desA]

I get the point about the freezer plates being the controlling part and the suction would drop quickly if we did not turn it off, but the whole point seems to be diaphragm is stuck in the freezer and so the bulb will never get used in the control.

I know I will have to move the TXV to another location to make it work and as you say get a bigger receiver.

Or maybe just leave it as it is and continue to do manual control? Nah - too much of a challenge really.

Chef

Under these conditions, could the TXV be replaced with a manual valve?

[Gary]

Under these conditions, could the TXV be replaced with a manual valve?

Sure... but it wouldn't improve anything.

[desA]

Sure... but it wouldn't improve anything.

Sure, but setting a TXV in a place sure to create bulb charge migration would be a bit inefficient anyway - too many unknowns. Why go to the expense with a system which must surely be operating well outside manufacturer's design envelope.

Either place the TXV in a sensible location, or change to cap tube, or manual valve.

[Gary]

Sure, but setting a TXV in a place sure to create bulb charge migration would be a bit inefficient anyway - too many unknowns. Why go to the expense with a system which must surely be operating well outside manufacturer's design envelope.

Either place the TXV in a sensible location, or change to cap tube, or manual valve.

This is a very bad design. If we reposition the TXV, or switch to a cap tube or hand valve, this woud still be a very bad design.

I am surprised that it works as well as it does. I am not surprised that it eats compressors.

[desA]

There have been several recent threads that have made me think about this setup in a new light.

The system has 3 boxes each with holdover plates, the refrigerant passes through each one in turn and running temps in each box are:-
Freezer -10 to -12C (2 plates at -20C)
Refrigerator 2 to 4C (plate -6C)
Salad Chiller 4 to 8C (plate -4C)

But the TXV is inside the freezer and the bulb is inside the Chiller so once the system's down to temps one must assume that the TXV will be sensing the diaphragm temperature and not the bulb's at some instant. Maybe as the system cools down further on cycle the bulb see temps lower than than the freezer?

However, it works, but is very susceptible to charge.

I've already got the message that it may be better to get the TXV say into the chiller but why does it work as it is? Any ideas

Chef

Oh - its R134 and running Dx = 125PSI and Sx = 2PSI

Chef, have you perhaps looked at the sequential cool-down curves for each of these chambers?

I wonder if the original designer may not have used the chamber differential temps & locations to force the bulb to provide a level of progressive control. After a certain point, charge migration should ultimately allow the system to run as an orifice-type.

These temp-time plots could provide some very interesting insights, I'd think.

[Gary]

Chef, have you perhaps looked at the sequential cool-down curves for each of these chambers?

I wonder if the original designer may not have used the chamber differential temps & locations to force the bulb to provide a level of progressive control. After a certain point, charge migration should ultimately allow the system to run as an orifice-type.

These temp-time plots could provide some very interesting insights, I'd think.

Are you assuming this was professionally designed? Seems doubtful.

As I see it, the original intent was to equally feed/freeze all of the plates. It worked out with two different temp ranges because the bulb charge migrated.

Dumb luck.

[desA]

Hahaha... could be.

[Chef]

To save lots of quotes I will just answer all the comments in one go.

The system was originally designed and marketed by a USA company - still in business I believe - as a 2 plate system. The 2 plates in the freezer, this was probably installed by someone else and they put the TXV in the freezer and had some little temperature controlled plastic window things that were supposed to open and close and use cold from the freezer to keep the fridge and chiller cold. (they called it spill over)

Well that was obviously never going to work as the size of the plates was too small for all of the 3 box's.

So we added the fridge plate and kept the chiller with the spill over system. Later we added the chiller plate, sized to keep the temp in the box correct for the salad stuff. Then later to enhance performance the liquid/suction exchanger was added. This helped a lot as the charge could be increased and get better cooling but of course I only know that with the hindsight from this thread.

Then a desuper was added to stop the discharge line melting stuff. The TXV was changed to a Danfoss and with trial and error the #4 was chosen as the best performance. Again hindsight shows it to be an orifice system and thats why it was sensitive to the assembly number.

So in answer to your question - it was not professionally designed and underwent many many changes and tweaks to get it to work and now I understand what is going on (mostly anyway) I agree it is amazing it ever produced cold at all but with all the tweaks we did to get it running over the years it is working.

However it is wrong and needs a proper sort out.

As regards the pull down data - I have detailed pressure, temp and time plots for the freezer only and will post it later, the fridge and chiller are simply verified as being cold enough when the plate is frozen all the way across - simply feel it by hand but this always occurs after the freezer is down to temp ie critical charge.

Was it dumb luck it worked - probably but the tweaking made sure it worked as good as it could.

Compressor failures? not sure yet but LBP application and low suction with high dp across the compressor probably caused it.

So I have a schizophrenic system that has psychopathic tendencies. Brilliant.

Chef

[desA]

So I have a schizophrenic system that has psychopathic tendencies. Brilliant.

Goes with the part of the world you live in...

[Gary]

Then a desuper was added to stop the discharge line melting stuff.

Ahhh... another piece of the puzzle.

Discharge temp is directly related to compressor inlet superheat. High discharge temp indicates high inlet superheat which causes very high temp at the discharge valves inside the compressor. This cooks the oil, destroying its lubricating properties... and that explains the compressor failures.

The desuperheater treated the symptom, not the cause. The cause was/is high compressor inlet superheat.

The second TXV with EPR will allow you to regulate the fridge/chiller temp independent of the freezer temp. The fridge/chiller temp is dictated by the EPR. It pulls down to the desired saturation temp and then stays right there while the freezer pulls down to temp. The cold vapor coming back, no longer being over-superheated by the fridge/chiller plates, keeps the compressor and its oil cool.

[mad fridgie]

Option 2 "Double expansion"
New TEV (ex equ), feed fridge plates first, then to freezer plates, install hand expansion valve (needle valve) where present TEV is. "Install pressure port either side of HEV"
TEV bulb on freezer plate outlet.
Set hand expansion valve so inlet pressure is -6C.
Effectively as far as control of the TEV the fridge plates and hand expansion valve) are acting as a distributor (pressure drop) You could use an EPR instead of the HEV, but not sure what will happen over the long term to the valve seat with a continious liquid vapour mix.
Only mod to cabinets is a new suction suction from the freezer.

Forgot to say any refrigerant

[Chef]

The desuperheater treated the symptom, not the cause. The cause was/is high compressor inlet superheat.

Yeah I was kind of thinking along this line. The outlet temp is very high for 125PSI condensing pressure, I think a hermetic close the evap output (to stop water dripping) is an idea, also to consider with the EPR. Put both TXV's in the chiller (somewhere?) and it may be possible to change the chiller from 5/8 to 1/2" so helping the flowrates.

Not sure about the suction/liquid exchanger?

Chef

[Gary]

Yeah I was kind of thinking along this line. The outlet temp is very high for 125PSI condensing pressure, I think a hermetic close the evap output (to stop water dripping) is an idea, also to consider with the EPR. Put both TXV's in the chiller (somewhere?) and it may be possible to change the chiller from 5/8 to 1/2" so helping the flowrates.

Not sure about the suction/liquid exchanger?

Chef

The freezer TXV should stay in the freezer. Just move the bulb to the freezer.

[Gary]

I think a hermetic close the evap output (to stop water dripping) is an idea...

I have no idea what this means. Can you rephrase this?

[Chef]

At the moment there is a long run of suction line to the compressor on the engine but if I use a hermetic unit then it can go closer to the evap outlet and so with its new low temp suction will not cover everything in dripping water, even to insulate will help maybe but they always want to sweat somewhere along the awkward suction run.

[Gary]

At the moment there is a long run of suction line to the compressor on the engine but if I use a hermetic unit then it can go closer to the evap outlet and so with its new low temp suction will not cover everything in dripping water, even to insulate will help maybe but they always want to sweat somewhere along the awkward suction run.

That makes sense. Would be much easier on the compressor, too. Will this be a problem for the condenser piping?

[RLK]

I am suggesting two parallel circuits.

The first circuit: liquid line>TXV1>Plate1>Plate2>suction line.

The second circuit: liquid line>TXV2>Plate3>Plate4>EPR>suction line.

The EPR maintains higher temp in Plates 3 and 4.

Gary, is correct standard practice is a split system with engine driven eutectic holding plate.

In addition to dual TXV’s a solenoid is a good feature in liquid line to refrigerator TXV controlled by thermostat located in refrigerator.

On holding plate system, suction pressure regulated by an adjustable low pressure switch set to stop compressor when freezer plates are frozen solid is used to control system.

These systems are generally water cooled so a high pressure switch set slightly above max operating temperature.

I assume the compressor is rated from 6000 to 12000 Btu at 2000 rpm when used in a High Back Pressure application. I also assume you are using a swash plate compressor like one of the Sanden models. In a properly designed refrigeration system these compressors have operated without failure for more than twenty years. Repeated compressor failures on eutectic plate systems are caused by:

Incorrect oil level, The oil cavities of these compressor are designed to hold less than 7 ounces oil, in these mutable plate systems additional oil is needed as oil and refrigerant velocities at low plate freezing temperatures is not adequate. Adding to oil return problem there is often many oil trap areas in eutectic plates.

Normal cycling of a eutectic plate system compressor runs one or two hours per day and off twenty two hours. During off cycle refrigerant and oil migrates into very cold plates remaining there until the next startup. Since there are no restrictions in suction side after TXV liquid in plates accelerates rapidly back to compressor. To reduce the possibility of compressor liquid flooding these systems need a suction line accumulator and TXV orifices that are sized to max plate Btu acceptance rather than compressor max capacities.

The engine drive refrigerator compressors that I know of are operated at an rpm that matches maximum Btu that plates can absorb. On a system where the aggregate of all plates eutectic solution is 15 gallons, normal compressor operating rpm is 900 to 1200 with a max short duty cycle or 1600 rpm.

[Chef]

RLK - many thanks for your info.

The system has a low pressure switch but is normally always manually controlled, even so it does drop out if the suction goes past 0PSI. The high pressure switch is set to allow initial pulldown from warm and is about 225PSI.

It is water cooled and the condenser seems fine as the liquid out is just a few degrees above water temp.

The unit is a TM13 which is similar to a Sanden SD5. It is rated at 14,000BTU at 1800RPM for HBP applications and we calculate we are getting about 11,000 to 12,000 BTU at our conditions. The eutectic capacity is 8.6 USgal in all 4 plates

The max BTU acceptance of the plates is unknown at the moment as the coil arrangement inside the plate is unknown and some are sealed so we can never open it to find out. The only way would be to measure the pipes internal volume once the system is apart but it is an approximate method only.

The compressor failure by oil hold up is possible but whenever we have checked it the oil quantity seems OK for the unit - normally we test about 2 or 3 oz in the compressor which is pretty close to the specs for an SD5. However we always test the oil immediately after a cycle so the unit should have the correct oil returned to it and maybe after many hours of off cycle things change. Dont know the answer to that one.

Also it appears we are running with a SH in the range of 30 to 40C (nominal is 35C) whereas the quoted SH for the compressor is 20C at the rated load.

Gary points out this will cause high discharge temperatures (which we do indeed get) and this in turn is a possible cause for failure.

Poor oil return (unproven but possible) and egg frying hot - looks like a recipe for short lived compressor.

Thanks again for your information.

Chef

[desA]

^^@RLK: Excellent post. Very informative.

[mad fridgie]

RLK - many thanks for your info.

The system has a low pressure switch but is normally always manually controlled, even so it does drop out if the suction goes past 0PSI. The high pressure switch is set to allow initial pulldown from warm and is about 225PSI.

It is water cooled and the condenser seems fine as the liquid out is just a few degrees above water temp.

The unit is a TM13 which is similar to a Sanden SD5. It is rated at 14,000BTU at 1800RPM for HBP applications and we calculate we are getting about 11,000 to 12,000 BTU at our conditions. The eutectic capacity is 8.6 USgal in all 4 plates

The max BTU acceptance of the plates is unknown at the moment as the coil arrangement inside the plate is unknown and some are sealed so we can never open it to find out. The only way would be to measure the pipes internal volume once the system is apart but it is an approximate method only.

The compressor failure by oil hold up is possible but whenever we have checked it the oil quantity seems OK for the unit - normally we test about 2 or 3 oz in the compressor which is pretty close to the specs for an SD5. However we always test the oil immediately after a cycle so the unit should have the correct oil returned to it and maybe after many hours of off cycle things change. Dont know the answer to that one.

Also it appears we are running with a SH in the range of 30 to 40C (nominal is 35C) whereas the quoted SH for the compressor is 20C at the rated load.

Gary points out this will cause high discharge temperatures (which we do indeed get) and this in turn is a possible cause for failure.

Poor oil return (unproven but possible) and egg frying hot - looks like a recipe for short lived compressor.

Thanks again for your information.

Chef

Not sure your calcs are correct, at rated 7.2C SST, you get 14000BTUs, but you are running at -20C SST, you would be lucky if you are getting 5000BTUs,
Your total load presuming 100% water (of course is not) would be 16100 BTU (allowance in SHC, to cover cabinet)
You said pulldown time 2.5hrs, so averaged duty is 6440 BTU, you would expect this as SST would start higher then drop towards the end, to below 5000BTUs
Very high discharge temp has also been caused by the introduction of the suction to liquid line heat exchanger. ( addition SH on and above what is leaving the cabinets)
It was indicated that major changes were not applicable to this system.
If that is not the case, then 2 seperate TEVs and 2 liquid sol valves, fridge off on plate stat. The temperature of the refrigerant is not important?
The freezer sol valve, off when comp goes off (stops liquid migration) Do not worry about pump down, as trying to set a very low differential on this application would be difficult. Because you are using aphase change, OFF cycle pressure will remain constant. Only when completly melted will pressure rise, to late!

[Chef]

Not sure your calcs are correct, at rated 7.2C SST, you get 14000BTUs, but you are running at -20C SST, you would be lucky if you are getting 5000BTUs,
Your total load presuming 100% water (of course is not) would be 16100 BTU (allowance in SHC, to cover cabinet)
You said pulldown time 2.5hrs, so averaged duty is 6440 BTU, you would expect this as SST would start higher then drop towards the end, to below 5000BTUs

I calculated the cooling using the volumetric flow rate of 0.0042m3/hr and a density of 5.1Kg/m3 to get mass flow of 0.021Kg/s.
Then from the PH diagram get H3 and so H4 and also H1 giving a deltaH in the evap of 150KJ/kg.
Kw=mass flow * deltaH = 3.15KW or 10,754BTU

Dont forget that the rated duty is air cooled condenser and discharge of 16bar so when water cooled the discharge pressure drops and the H3 moves to the left giving more cooling than you can deduce from the curves alone.

Pulldown time as I call it is from warm ie room temp and cycle time the period to cool from cold to very cold. The first is 2.5 hours and the second is about an hour.

Not knowing how much ice is still left in the plate when the cycle time is measured does not really give an accurate measure of the duty.
Anyway if all the ice was to be refrozen then 32litres of water at 333j/g yeilds 10.6Mj and at 3.1Kw that would yake 59 minutes so I think the numbers are quite correct but please check them and let me know if got someting wrong.

Chef

[mad fridgie]

I calculated the cooling using the volumetric flow rate of 0.0042m3/hr and a density of 5.1Kg/m3 to get mass flow of 0.021Kg/s.
Then from the PH diagram get H3 and so H4 and also H1 giving a deltaH in the evap of 150KJ/kg.
Kw=mass flow * deltaH = 3.15KW or 10,754BTU

Dont forget that the rated duty is air cooled condenser and discharge of 16bar so when water cooled the discharge pressure drops and the H3 moves to the left giving more cooling than you can deduce from the curves alone.

Pulldown time as I call it is from warm ie room temp and cycle time the period to cool from cold to very cold. The first is 2.5 hours and the second is about an hour.

Not knowing how much ice is still left in the plate when the cycle time is measured does not really give an accurate measure of the duty.
Anyway if all the ice was to be refrozen then 32litres of water at 333j/g yeilds 10.6Mj and at 3.1Kw that would yake 59 minutes so I think the numbers are quite correct but please check them and let me know if got someting wrong.

Chef

Moving imperial to metric, (working imperial as you uses BTU and gals)
You do not know how frozen the plate is when you restart the machine. But we do know the state when you start the system.
max weight of water 68LB (not quite true as there will be a minimum 7% air gap to allow for expansion, and reduction for volume of refrig pipes)
Start 100F 1BTU/Lb/F
Freezing Point 32F (not quite true but close enough) 144BTU/F (in real lfe is over varing temps) and would be less as contains an anti freeze
Final 0F 0.5BTU/Lb/F
so 68* 68= 4624
68*144=9792
68*32*.5=1088
Total=15504BTU (rounded up to 16100BTU, in my head lol)
Run time 2.5Hrs
Therefore averaged hourly duty 6440BTU (1.88KW)
At the start SST will be high, so actually duty is likely to be higher than the rated 14000BTUs, dowards the end duty i suspect is below 3000BTUs.
Yes you are correct your SCT is lower as well as the sub-cooling, but your SH is whey higher than design, (there is a bit of to and fro)
Your compressor calculations do not account for volumetric efficiency.(which is an unknown)