PDA

View Full Version : Minimal required R134a speed in suction line







DaBit
02-05-2002, 04:34 PM
Hi guys,

I have built an R134a based water chiller which should cool a water/methanol mixture, which in turn should cool the microprocessor in my PC. I am currently able to reach -17 degrees C under load, but I have the idea that the evaporator is not performing optimal.

Some more information:
- Compressor: Danfoss NL11F
- Condenser: an old steel R12 unit well capable of handling the load
- Capillary: 2.4 meters of 0.8mm ID copper tube
- Evaporator: 6.5 meters of 1/4" copper pipe wound in a spiral, submerged in the water/methanol mixture.

I suspect that the 6.5 meters of 1/4" piping submerged in liquid just gives too much pressure drop. With 100kPa pressure at the compressor's service line the liquid temperature does not go below -16 degrees C (R134a boils at -26 degrees C at that specific pressure). I also suspect that liquid refrigerant is blown through the evaporator before it has the chance to evaporate.

If I want to build a better evaporator, I need to know a few things:

- What must be the minimal speed of the (gaseous) refrigerant in the suction line to ensure sufficient oil return? I assume these values are different for both vertical and horizontal sections.

- For a total load of 200W, how large an evaporation surface (copper tubing) do I need? I understand that this depends on the allowable dT between liquid and boiling R134a, but are there guidelines? The lower the dT, the better.

- Any other considerations for the new evaporator? I plan on building a coaxial heat exchanger type of evaporator (a thin refrigerant line within a larger pipe where coolant flows along the line)

Prof Sporlan
02-05-2002, 07:23 PM
I suspect that the 6.5 meters of 1/4" piping submerged in liquid just gives too much pressure drop...... I also suspect that liquid refrigerant is blown through the evaporator before it has the chance to evaporate.
The 1/4" OD tube may not be too bad for 200 watt duty. A superheat reading at the compressor inlet would be in order though.


What must be the minimal speed of the (gaseous) refrigerant in the suction line to ensure sufficient oil return?
Rule of thumb would be something like 700 ft/min on a horizontal run, and 1500 ft/min on a vertical riser. Some may suggest you can live with numbers a bit lower than this.


For a total load of 200W, how large an evaporation surface (copper tubing) do I need? I understand that this depends on the allowable dT between liquid and boiling R134a, but are there guidelines? The lower the dT, the better.
I would seem you're operating at a 10C TD (-16C - (-26C)), which doesn't seem too bad for a bare tube evaporator. First, check the superheat. If it is also at 10C at these conditions, you should try adding refrigerant

DaBit
02-05-2002, 08:41 PM
The 1/4" OD tube may not be too bad for 200 watt duty. A superheat reading at the compressor inlet would be in order though.

I have only inaccurate temp reading devices available right now (I am not a pro, just a computerfreak with two right hands wanting some extra speed), but superheat is about 4K. This is probably due to the capillary<->suction line heat exchange. The temperature of the gas leaving the evap is about the temp of the liquid.

Also, Danfoss recommends 3/8" suction line tubing to prevent performance loss. I would have used 3/8" if I knew before that the final compressor was the NL11F. I used a second hand NL7F before, but this one broke down due to acid in the compressor.


I would seem you're operating at a 10C TD (-16C - (-26C)), which doesn't seem too bad for a bare tube evaporator. First, check the superheat. If it is also at 10C at these conditions, you should try adding refrigerant

It's not bad? Hmm, during my entire watercooled CPU projects I got used to TD's in the order of single degrees. Also, the bare tubes are submerged in water, which is highly capable of removing heat (or cold).

Refrigerant charge seems OK. I tuned it over the past few days adding a little bit and removing a little bit until the temp dipped.

if 10C is normal for a bare-bone evap, how would I improve the evaporator so I can use those extra degrees? copper to water surface area of this evaporator is already almost 0.1 sq. meter.

zolar1
02-05-2002, 10:59 PM
Just a spoonful of liquid dish detergent might help?

Salt is corrosive, but maybe try a little sugar to make the density different?

Have you tried an air pump for an aquarium to make bubbles to circulate the water?

The water cooled part is rather neat. Could you give some details on it's construction and location, including the number of litres of water it holds?

Since you're using water for cooling, assuming it's some sort of vat or container, have you considered the water might be getting warm and your efficiency lower?

Do you put ice in the water?

Could you use 3/8" aluminum tubing for your condenser?

Have you tried using Propane as your refrigerant?

zolar1
02-05-2002, 11:01 PM
You could also use copper screening material and solder it onto the copper tubing - like adding a flap or sorts.

Gary
03-05-2002, 02:14 AM
I have only inaccurate temp reading devices available right now (I am not a pro, just a computerfreak with two right hands wanting some extra speed), but superheat is about 4K. This is probably due to the capillary<->suction line heat exchange. The temperature of the gas leaving the evap is about the temp of the liquid.


You have too much refrigerant in the system. Remove refrigerant slowly until the superheat near the compressor inlet is about 10-15K, then let us know the new temp readings. Don't be surprised if the line temp drops as you remove refrigerant, before it then rises to the higher superheat. This is typical of gross overcharge on a fixed orifice system.

zolar1
03-05-2002, 04:08 AM
Have you gone to www.overclockers.com and checked out their suggestions?

DaBit
03-05-2002, 10:04 AM
Just a spoonful of liquid dish detergent might help? Salt is corrosive, but maybe try a little sugar to make the density different?


The mitrure I am usning right now is about as good as it gets for the application. It is a mixture of 30% methanol and 70% water. When everything is finished I will also add a bit of Red Line Racing WaterWetter. This will give me a bit more efficiency.

With this combo I retain as much of the good properties of water as a heat transfer fluid as possible, while the methanol lowers the freezing point of the mixture to approximately -30 degrees.

Adding salt of suger does not give me an advantage (I did check it). Also adding soad to reduce the surface tension of the water did not help a lot, except excessive foaming.


Have you tried an air pump for an aquarium to make bubbles to circulate the water?

No, I just pump it around using a Sicce Idra 1200L/h aquarium pump. The pump is not really up to the low temperatures and water/methanol, but I haven't been able to find anything better yet.



The water cooled part is rather neat. Could you give some details on it's construction and location, including the number of litres of water it holds?


Yes, of course. Everything from the beginning up to about December last year can be found on my website http://dabit.trybit.com. Check out the 'Watercooling' part and the 'Phase Change' (that's how computer guys call evaporation based refrigeration :) ) part.

The heat exchanger described there and the compressor used over there has died due to a litte leak in the evaporator tube. This allowed water to enter the compressor, which killed it. The fact that the compressor died was not that bad, since it's condition was already bad after all (it had been on a shelf with open ports for about half a year). From there on I started with a new heat exchanger and compressor.

Some pics:

This is the complete system which I use now, except for the condenser fan; this one has been replaced with 2 24VDC 120mm Papst fans with a temperature control, so they don't run faster than required, which improves noise level and maintains a relative constant pressure before the capillary tube.

http://www.arcobel.nl/~dabit/zooi/phasechange2_completesystem.jpg

This image shows the internal details of the heat exchanger:

http://www.arcobel.nl/~dabit/zooi/heatex2_spiral_topview.jpg

And here the same thing viewed from the bottom:

http://www.arcobel.nl/~dabit/zooi/heatex2_spiral_bottomview.jpg

Just for completeness: I use this airtight aluminium box to prevent condensation and freezing of the delicate microcomputer electronics. The thing is airtight, and moisture is removed from the air iside the box using silicagel:

http://www.arcobel.nl/~dabit/zooi/mobobak_17apr2001_voor.jpg

More images can be found in the http://www.arcobel.nl/~dabit/zooi/ directory. You can also visit the Dutch overclocking forum I visit on a regular base, and use a seach on topicstarter DaBit. Almost everything is there. The URL is http://gathering.tweakers.net/listtopics.php/22



Could you use 3/8" aluminum tubing for your condenser?
Have you tried using Propane as your refrigerant?


No, I wil keep my current condenser. It is very well up to the job, and with the fans at full speed the liquid leaving the condenser is just a single degree above ambient, while it entered at approximately 50-55 &deg;C.

We tried using (dried BBQ grade) propane in the previous compressor (NL7F), but it does not perform as well as R134a. This could be due to a different required dimensioning of the capillary tube, but I have no idea how to size the captube for propane.



You have too much refrigerant in the system. Remove refrigerant slowly until the superheat near the compressor inlet is about 10-15K, then let us know the new temp readings. Don't be surprised if the line temp drops as you remove refrigerant, before it then rises to the higher superheat. This is typical of gross overcharge on a fixed orifice system.

Hmm, I can do so, but then the temperature will rise. My current charge is the charge which gives me the lowest temperature at load. When I add or remove refrigerant, I see coolant temperatures (at load) rising.

Frosting of the suction line stops approximately 10cm before the compressor inlet. Maybe I should do a more accurate superheat reading first and come back with more accurate numbers.

Just a bit offtopic, but can someone explain me how the superheat is generated? Say that I am evaporating at -20 &deg;C in liquid which is -16 &deg;C, then the gaseous refrigerant entering the suction line will have a temperature between -20 &deg;C and -16 &deg;C. So, how can superheat be larger than 4K in this fictive situation?


Have you gone to www.overclockers.com and checked out their suggestions?

Yes, but there is a lot of rubbish out there on the Internet. overclockers.com does not really describe much of the technical details, and that's why I came here.

Anyway, I have to build another heat exchanger. The PVC shell of the current heat exchanger is not suitable for this application. I cannot stop leaks around the seals for a longer period due to the different expansion rates of PVC and copper. So, I will build a new heat exchanger, which I would like to be as efficient as possible.

Can you guys help me a bit with the design? This includes evaporator pipe sizing (I was thinking about using 3/8" piping), evaporator pipe length, etc.

zolar1
03-05-2002, 02:46 PM
You could try a steel coffee can as your shell.

I too like experimenting with different configurations. I made a simple heat exchanger for a domestic refrigerator. I varied the size, length, some of the shape, and connection locations before I got it to work satisfactorily.

Also, you might want to move your heat exchanger away from the condensor and fan. Blowing heat on the heat exchanger offsets the gains you were trying to get.
Why did you wrap the cap tubing around the dryer? Was it to keep the tuping from getting in the way?

You ever think of using a different refrigerant? R22 or R401A?

zolar1
03-05-2002, 02:49 PM
"The latent heat of vapourisation of R134a is 217kJ/kg of refrigerant."

R414B has a better rating I believe.

DaBit
03-05-2002, 03:07 PM
I too like experimenting with different configurations. I made a simple heat exchanger for a domestic refrigerator. I varied the size, length, some of the shape, and connection locations before I got it to work satisfactorily.

So, what were your observations? Any recommendations for my new upcoming heat exchanger?


Also, you might want to move your heat exchanger away from the condensor and fan. Blowing heat on the heat exchanger offsets the gains you were trying to get. Why did you wrap the cap tubing around the dryer? Was it to keep the tuping from getting in the way?

Ah, you are looking at the old setup :)
Well, of course I insulated the old heat exchanger with a thick layer of neoprene. This is sufficient to keep the liquid in the exchanger cold for many hours after the compressor has turned off.

The newer heatex of which I posted pics in this forum is wrapped into >5cm thick polystyrene foam. This does a really excellent job on thermal insulation. It takes more than 30 hours for the liquid to rise 10 &deg;C

About the wrapping of the captube around the filter: this was done by the refrigerator repair technician which also evacuated and filled my system. No bad words about that guy, but he does not know much about the theory of operation of a fridge. When a fridge is broken he fixes the leak, fills the system, and everybody is happy.

Gary
03-05-2002, 11:12 PM
Just a bit offtopic, but can someone explain me how the superheat is generated? Say that I am evaporating at -20 C in liquid which is -16 C, then the gaseous refrigerant entering the suction line will have a temperature between -20 C and -16 C. So, how can superheat be larger than 4K in this fictive situation?


I am referring to the compressor inlet superheat, not the coil outlet superheat. Measure the line temp about 6 inches from the compressor inlet. At your coldest evaporating temp (SST), the line temp should be 10-15K higher. In this case with a -20C SST (saturated suction temperature), the line temp near the compressor inlet should be -10C to -5C. Less than this endangers the compressor.

Your heat exchanger seems to be transferring heat very well. You could improve the performance by going to a TEV instead of the cap tube, but to get much more you will probably need a bigger condensing unit. What is the condensing temp (aka SCT)? And at what condenser air in temp?

If you can provide these, we can tell a lot more about the system:

Low side:

evap water in temp
evap water out temp
SST (saturated suction temp)
suction line temp at compressor inlet

High side:

cond air in temp
cond air out temp
SCT (saturated condensing temp)
liquid line temp at condenser outlet

Preferably, all of these should be measured when the water is at it's lowest temp.

Dan
04-05-2002, 12:43 AM
Great pics. How cold do you want the processor to be? Oil return is less as suction pressures are lowered.

I would extend the water inlet to the bottom of the evaporator and also put the water inlet and outlet inside and outside the evaporator so the water flows through the greatest surface area of the evaporator.

I am assuming that you are working with home or camper fridge condensing units, hence the 134A. A higher pressure refrigerant would make it easier to achieve lower temperatures, but you are looking at extra cost and further design considerations.

What is the wattage of the device you wish to make cold? There was another thread in this group that I lost track of where I asked the same question.

Gary
04-05-2002, 01:06 AM
How cold do you want the processor to be?

This is exactly the right question, Dan. After all, the bottom line is processor temp, not fluid temp.

We have been considering the system as it cools the fluid, not as it cools the processor. A much closer measure of this would be the temps of the water going to and returning from the processor.

zolar1
04-05-2002, 06:25 AM
So, what were your observations? Any recommendations for my new upcoming heat exchanger?

From my observations, a typical domestic refrigerator needs a small H.E. - it seems the optimal size is between 4-5 inches (10-18cm?) long, 1/2" in diameter.

I was mostly concerned with lowering the amp draw on the compressor. Most of the compressors I use draw between 1.25amps and 2.65 amps. Again, small stuff. I managed to be able to reduce the amp draw between 15% and 30% without any reduction in capacity. n fact, I believe the capacity went up. Regarding temperatures and such, I measured them with my hand, If it felt hot and the suction side pressure was low enough, I could add more refrigerant as long as I didn't see frost anywhere. I tried the H.E. across the compressor and across the evaporator, both configurations did work, but across the compressor resulted in a hotter compressor. Both designs used the cross flow pattern. The H.E. were a very simple design.

Cont...

zolar1
04-05-2002, 06:28 AM
On another note: I purchased a 'store bought' heat exchanger. It was rated up to 10 ton capacity and installed it on my personal refrigerator. When it runs, you hear the evaporator making noise, but not a boiling sound. More like the wind blowing across the window on a cold winter day. I was using a pretty big compressor (I go by LRA number, not HP rating). The LRA number is a 26. My amp draw is approximately 2 amps. The refrigerator uses static condensing coils up the back side.The refrigerator size is 21 cubic feet with built in icemaker (which with the heat exchanger, it makes ice at a faster rate). The freezer section is always around zero deg F +/- 3 deg F. The choice in refrigerant used was Hot Shot (R414B). The company claims it has a capacity of 92but/hr.

My deduction is this: too big of a heat exchanger causes evaporator 'groaning' and is overkill for the application.

Continued....

zolar1
04-05-2002, 06:32 AM
I am considering purchasing an electronic sight glass to further understand the physical characteristics of refrigerant in any given application, and to facilitate the charging of most systems. I feel it would be a good diasnistic tool and wouldn't have to open the system to check pressures. Installing a real sight glass is not cost effective on a domestic refrigerator or window a/c.

Any comments/experience/suggestions on the electronic sight glass?


PS about your heat exchanger design - If you alter your design slightly by attaching 2 tubes together then wrap them into a coil, leaving 1 tube open to the chamber, you might gain more efficiency.

Since your compressor appears to be the same type/kind as used in a soda cooler, why couldn't you use R22 instead?

DaBit
05-05-2002, 10:23 PM
Hmm, I got MySQL errors yesterday? Well, luckily I saved the text into a file.



I am referring to the compressor inlet superheat, not the coil outlet superheat. Measure the line temp about 6 inches from the compressor inlet. At your coldest evaporating temp (SST), the line temp should be 10-15K higher. In this case with a -20C SST (saturated suction temperature), the line temp near the compressor inlet should be -10C to -5C. Less than this endangers the compressor.


Where is the extra temperature rise coming from? How can there be a 10-15K differential over about 30cm/1ft of piping? I will do the measurements, and currently I drained a lot of refrigerant which I am adding little step by little step again.




Low side:

evap water in temp
evap water out temp
SST (saturated suction temp)
suction line temp at compressor inlet


Water in: -15.4 &deg;C
Water out: -15.6 &deg;C

This measurement is inaccurate since it falls within the 2% accuracy of the temp measuring devices I use. Flow is high, so the water temp differentials are quite low.

SST: -16.2 &deg;C
Suction line temp at compressor inlet: -8.2 &deg;C

cond air in temp : 18 &deg;C
cond air out temp : 23 &deg;C (inaccurate measurement)
SCT (saturated condensing temp) : ehm? If this is the temp of the gas going into the condenser, this is 54 &deg;C
liquid line temp at condenser outlet: 27 &deg;C

Please note thet the fans on the condenser are speed-controlled in such as way that if the liquid leaving the condenser is hotter than 32 &deg;C, the fans start blowing more air. Currrently the fans are operating at a very slow, almost silent speed.


Great pics. How cold do you want the processor to be? Oil return is less as suction pressures are lowered.

I would extend the water inlet to the bottom of the evaporator and also put the water inlet and outlet inside and outside the evaporator so the water flows through the greatest surface area of the evaporator.

How cold? Hmm, -80 &deg;C would be nice :)

You are right about the extension of the water inlet. But what you cannot see on the pics (they are already old again :) ), is that there is a pump mounted at the bottom of the exchanger.


I am assuming that you are working with home or camper fridge condensing units, hence the 134A. A higher pressure refrigerant would make it easier to achieve lower temperatures, but you are looking at extra cost and further design considerations.

I am using a condensing unit from a very old, very large R12 fridge. The reason I do not use R22, R404/407 or other higher-pressure refrigerants is because I simply cannot obtain them, or the parts required to work with them. Here in Holland you are not allowed to buy anything AC related if you don't have the required license. Not even a capillary tube, let alone refrigerants. Thus, a fridge repair shop helped me a bit, and they use solely R600a and R134a. Hence my choice for R134a.

My current goal is liquid @ -20 &deg;C, which is within the operating area of R134a. The compressor should be able to handle it.



What is the wattage of the device you wish to make cold? There was another thread in this group that I lost track of where I asked the same question.


In will be multiple devices:
- Pump: 10W
- Overclocked processor: 110W (yes, that much! No typo.)
- Gfx card: 30W
- chipset: 5W
- memory: 3W
- thermal losses: 30W

So, it is safe to assume a 200W load on the system.



We have been considering the system as it cools the fluid, not as it cools the processor. A much closer measure of this would be the temps of the water going to and returning from the processor.


Since the coolant flow is relatively high, this falls within the inacuuracy of my measuring devices. Doing the math (heat capacity of the liquid, heat load, flow) shows about 0.3 &deg;C temp difference, maximal.

I also know that I have a temperature difference between the processor silicon and the liquid of about 10-12 &deg;C. This is unavoidable since most of this dT is generated in the processor's package. But I cannot give you processor temps right now, since the measuring devices used on mainboards are not calibrated for such low temperatures. It doesn't matter whether my liquid is 8 &deg;C or -15 &deg;C, the temperature monitor would display 19 &deg;C.



My deduction is this: too big of a heat exchanger causes evaporator 'groaning' and is overkill for the application.

Why is this? What causes this?



PS about your heat exchanger design - If you alter your design slightly by attaching 2 tubes together then wrap them into a coil, leaving 1 tube open to the chamber, you might gain more efficiency.

Hey, almost the coaxial type of heat exchanger ;)


Since your compressor appears to be the same type/kind as used in a soda cooler, why couldn't you use R22 instead?

I live in Holland where we have a government that sucks (so obtaining R22 is simply impossible), and beside that, R22 has a too high pressure. The compressor would simply not start.

Gary
06-05-2002, 02:30 AM
Your superheat at the compressor inlet is 8K. This is slightly low, but reasonable.

Dan is correct. The water flow through your heat exchanger is not efficient. The coil of copper tubing should separate the water in from the water out.

Looking at your picture, I would suggest soldering a copper plate to the bottom of the coil. Then extend either the water in tube or the water out tube through the plate, depending on the direction of refrigerant flow.

If the refrigerant flows through the coil from top to bottom, then the water should flow from bottom to top, thus the water in should be extended through the plate.

If the refrigerant flows from bottom to top, then the water out should extend through the plate.

The idea is to force the water to flow through the tube coil in a direction that is counter to the refrigerant flow.

Just so that we are using the same terminology, the change in temperature of a single medium is called a delta-T or dT. A comparison of temperatures of two mediums is called a temperature difference or TD.

For example, the difference between water in and water out is a delta-T. The difference between the processor silicon and the liquid is a TD.

Gary
06-05-2002, 02:51 AM
SCT is similar to SST. Just as the SST is the saturation temperature corresponding to the low side pressure, the SCT is the saturation temperature corresponding to the high side pressure. Both can be found using a pressure/temperature chart.

If you don't have a high side pressure tap, an acceptable substitute would be the temperature on a return bend 1/3 from the top of the condenser. The return bends are the copper bends along the ends of the condenser.

Gary
06-05-2002, 03:21 AM
Without SCT, I cannot properly calculate the subcooling, but given the liquid line temp compared to ambient temp (9K TD), I suspect subcooling to be essentially non-existant. This would indicate that the cap tube is too large.

We should keep in mind, however, that a more efficient (counterflow) evaporator design would need a larger cap tube, because of it's increased heat exchange.

Perhaps Prof Sporlan could give us a better evaluation of this.

DaBit
06-05-2002, 11:05 AM
Your superheat at the compressor inlet is 8K. This is slightly low, but reasonable.

I changed the refrigerant charge yesterday (I am doing this very slowly, giving the system 30 minutes after each change to stabilise), now my superheat is 12K.


If the refrigerant flows through the coil from top to bottom, then the water should flow from bottom to top, thus the water in should be extended through the plate.

Refrigerant flows from top to bottom. Someone told me this would give better results than bottom to top. My thoughts were: 'if it flows from bottom to top, then the liquid would boil and the compressor always gets gaseous refrigerant, so that is the more optimal situation'. Now, who is right?

I am going to change the heat exchanger anyway. I have too much trouble with the PVC, the cold and the methanol. Also, the liquid flow over the evaporator pipes is too slow.

Maybe I can get a second hand SWEP plate heat exchanger, maybe I will make a full copper coaxial heat exchanger.


Without SCT, I cannot properly calculate the subcooling, but given the liquid line temp compared to ambient temp (9K TD), I suspect subcooling to be essentially non-existant. This would indicate that the cap tube is too large.

I think the only substantial subcool comes from the capillary<->suction gas heat exchange.
How does this heat exchange influence the length of the capillary anyway?

For calculating the capillary length I used the compressor selection program from L'Unite Hermetique (Tecumseh, www.tecumseh.com). This program allows you to enter design parameters, then it comes up with a suitable compressor and capillary tube dimensions (I.D. x L). I picked the ir compressor (AEZ1380Y) which was closest to the Danfoss NL11F.

Do you have a educated guess on the length of the cap. tube? I have heard everything from 3 meters 0.6mm I.D. up to 1m40 1mm I.D. Quite a large difference in flow resistance.

I will do a full measurement tonight of all the numbers again with the new refrigerant charge.

Gary
06-05-2002, 12:01 PM
Refrigerant flows from top to bottom. Someone told me this would give better results than bottom to top. My thoughts were: 'if it flows from bottom to top, then the liquid would boil and the compressor always gets gaseous refrigerant, so that is the more optimal situation'. Now, who is right?

The refrigerant will boil in either direction. Top to bottom is better for oil return.

It is essential that the water flows in the opposite direction (counterflow or crossflow, never parallel flow) and that the water flows over the tube coil, to provide maximum heat exchange between the water and the refrigerant.


Maybe I can get a second hand SWEP plate heat exchanger, maybe I will make a full copper coaxial heat exchanger.

The plate heat exchanger would be the better way to go, preferably with a TEV instead of a cap tube. Then make two small coaxial suction/liquid line heat exchangers with about six inches between them. Mount the TEV bulb on the suction line between them. The TEV bulb must be insulated and must not come into contact with the liquid line.


I think the only substantial subcool comes from the capillary<->suction gas heat exchange. How does this heat exchange influence the length of the capillary anyway?

The suction/liquid line heat exchange you are referring to does not not influence the length of the cap tube.

I am referring to subcooling in the liquid line before the cap tube. There must be some liquid line subcooling to ensure that liquid, not vapor, is entering the cap tube. It is not the subcooling which influences the cap tube size, but the reverse.

I would do everything I could to improve the efficiency of the evaporator first before judging whether the cap tube is properly matched to it. More refrigerant could be added to increase the subcooling, but you need a more efficient evaporator to boil that refrigerant off or the superheat will be too low, flooding the compressor with liquid. The cap tube selection program you used assumes an efficient evaporator. Everything works together. The evaporator is the weak link in your chain.

We could downsize the cap tube to match the efficiency of the evaporator, and thus improve performance, but better yet would be to increase the evaporator efficiency.

DaBit
06-05-2002, 12:54 PM
Improving the evaporator is indeed the way to go. As I said before, the current evap will be replaced by a plate heat exchanger (If I can obtain one. It is incredibly hard to get that stuff in The Netherlands without having a VAT number and a licence. Normal people should go to the normal store, but there are no stores selling this stuff since you need a licence before you can buy it.), or a coaxial heat exchanger with counterflow if I cannot obtain a plate heat exchanger.


About the TEV: I have been looking at the smallest Danfoss TEV's, type T, orifice 0 (not that they need to be Danfoss, but I just have data on them). Judging from the datasheets the refrigerant flow is too low for even the smallest TEV's. What is your opinion on this? Would a TEV equipped with Maximum Operating Pressure be more suitable?

About the mounting of the TEV bulb: I am not entirely sure what you mean. Is this drawing correct?
(dots = liquid line, | and - are suction line, BBB is TEV bulb)


| . | | . |
| . | | . |
| . | | . |
| . | | . |
| . | | . |
| ................ |
| | | |
| +----------+ |
| |
+------------------+
BBB

Prof Sporlan
06-05-2002, 02:38 PM
Refrigerant flows from top to bottom. Someone told me this would give better results than bottom to top.
You may get better improvement with bottom feeding the evaporator if your refrigerant velocities within the evaporator are low. Bottom feeding would promote trapping of the refrigerant in the coil, giving you something of a "semi-flooded" evaporator. Trapping liquid in the coil, however, can set up a TEV hunt, as liquid slugs can exit the coil affecting bulb temperature.


Judging from the datasheets the refrigerant flow is too low for even the smallest TEV's. What is your opinion on this? Would a TEV equipped with Maximum Operating Pressure be more suitable?
The Prof is reasonably sure no TEV manuafacturer makes a nominal 200 watt R-134a TEV :). This size unit would normally use a capillary tube. But one could expect a small capacity valve to work ok, e.g., a Sporlan EFJ-1/8-ZP, or a Danfoss cartridge TEV with a 0 cartidge, if using a TEV is desired.

An MOP-type TEV would be a good idea, to prevent a pulldown condition from overloading the compressor.

DaBit
06-05-2002, 03:46 PM
You may get better improvement with bottom feeding the evaporator if your refrigerant velocities within the evaporator are low.


Since I need 8.6mm I.D. for a 700ft/min flow (I did the math), I think my current speed is around 1300ft/min. This is an educated guess based on pipe size, and amount of refrigerant passing (about 4kg/hour)



But one could expect a small capacity valve to work ok, e.g., a Sporlan EFJ-1/8-ZP, or a Danfoss cartridge TEV with a 0 cartidge, if using a TEV is desired.

Well, and IS using a TEV desired? Will it ease the setup of the system? Will it increase performance over the probably incorrect size captube I am using now?

I have only limited ability to experiment with different capillary sizes since I can only pass by the refrigerator repair guy once in 2 months or so. And I do not have an oxyacetylene torch nor a decent vacuum pump. So, if using a TEV eases optimizing the system, it's already worth it's weighth in gold (well, almost ;) ).


Also, the compressor I use now is a low starting torque compressor for use with a captube. Will I get starting problems when I mount a TEV?

fernbay
07-05-2002, 12:14 AM
The beauty of stuff like this is based in the KISS princpal.
Keep It Simple, Stupid!!!:cool:
how much has it cost you against just buying a faster mobo/cpu?
What happens in three years time when nanotube transistors give us cpu's running 30x faster? we'll be limited by hdd access speed then!!:p
might be better to spend money on standard upgrades.

anyway, you've already gone and spent money... so to late..
your commited now.... matter of pride, right?!!!
try this... improve the effiency of the heat exchanger.
create a vortex in there. get a flat peice of copper about 20 mm long to fit inside the end of the 1/4" water inlet.
put a half twist into it, poke it into the water in, and tack it in place with the oxy. The volume of water coming in contact with the coil will be greatly increased. :cool:
to take it further, you could put the twist into the end of elbow instead, and solder it to the water in and angle it towards the wall of the heat exchange. sorta like a double vortex.

hope this helps with ideas!! ( if only he would use his skills in persuit of cold beer :D )

zolar1
07-05-2002, 04:43 AM
You could use the compressor itself as a vacuum pump.

Have you ever looked inside the center tube of a gas hot water heater and saw the twisted metal?

DaBit
07-05-2002, 10:14 AM
The beauty of stuff like this is based in the KISS princpal.
Keep It Simple, Stupid!!!:cool:
how much has it cost you against just buying a faster mobo/cpu?


Ehm, the entire thing costed me around EUR 150 ($120 ??). This includes all parts and used silver solder (which IS expensive). I got most things for free.

So buying a faster processor is out of the question. Besides that, I don't think you can find a stock single processor PC which outruns mine. I am able to throw 400-500MHz above Athlon XP stock speed, or 1.5GHz above P4 stock speed, to give you some rough numbers.



What happens in three years time when nanotube transistors give us cpu's running 30x faster? we'll be limited by hdd access speed then!!:p


Then I am able to run those nanotube transistor CPU's 40x faster :P



try this... improve the effiency of the heat exchanger.
create a vortex in there. get a flat peice of copper about 20 mm long to fit inside the end of the 1/4" water inlet.
put a half twist into it, poke it into the water in, and tack it in place with the oxy. The volume of water coming in contact with the coil will be greatly increased. :cool:


Sounds OK. A good thing to do for my new coaxial heat exchanger if I cannot get a plate heat exchanger.



hope this helps with ideas!! ( if only he would use his skills in persuit of cold beer :D )
[/quote]

Oh, I did. put the bottle into the cold liquid and it is chilled in a minute or so :)


You could use the compressor itself as a vacuum pump.

Is an R134a compressor able to pull a deep enough vacuum? And how long does it last as a vacuum pump? I am used to leaving the vacuum pump on the system for 24 hours.
Hmm, I have the old NL7F laying around somewhere. If it is suitable as a vacuum pump it would help a lot. Oxyacetylene torches can be found in more places than vacuum pumps.

DaBit
07-05-2002, 11:28 AM
Message to the moderator: this topic is going incredibly offtopic.

Since the original question was answered, I will open another topic which is easier for future reference

zolar1
07-05-2002, 02:59 PM
What's the topic name so i can find it?

DaBit
07-05-2002, 03:23 PM
You found it already :)

zolar1
08-05-2002, 05:19 AM
I think I can improve on your design a little.

I need to know what size tubing you need to adapt to.

You mentioned 1/4" for the refrigerant, but what about the water inlet and outlet?

How cold do you want the water to be?
Does the physical size matter or does it have to be small?

The larger the size of the container holding the water mix, the greater the load can be placed on the whole system (intermittently)

You mentioned the PVC piping expanded at a different rate than the copper. I think you can still use the PVC, just use a bigger diameter and longer length. A baffle would be in order between the water inlet and outlet, much like a steam generator.

Still thinking about the design changes needed, but need more info please.
Can you use 1/4" inlet and 5/16" or 3/8" outlet for the refrigerant?