Cap tube sizing rule

[paul007]

now Im not very smart on the matter at hand, but would tuning a system for say 200w with r404 say 6ft now and if the system was not loaded with a good 180w 24/7 wouldnt the gas enter the compressor without fully boiling off requiring a acculmator??

[Gary Lloyd]

Assuming proper refrigerant charge, given the small amount of refrigerant in these systems, that would be unlikely, although possible. More likely the difference would cause the liquid to move closer to the compressor, but not to the point of floodback.

[paul007]

thanks..

[chilly1]

Originally posted by Gary Lloyd
See... The old man isn't as dumb as he looks. (Actually, nobody is.)



That's .028, not .28, and it depends on what refrigerant we are talking about.

And the compressor size!

[Gary Lloyd]

And the compressor size!

Not really. The compressor size will make some small difference, increasing or decreasing the dP, but not that much.

[Russell_hq]

I had a go with some capillary tube models. Running the following parameters:

Capillary Tube Diameter: 0.028" ID
Refrigerant: R404A
Degree of Subcooling: 5K
Degree of Superheat: 5K
Condensing Temperature: 30ºC
Evaporator Pressure: 1atm

These are the results:

Tube Length (m)	Mass Flow (kg/s)	Cooling capacity (watts)	
1.5		0.001447438		156.3	146.6
1.75		0.001343377		145.1	137.6
2		0.001259301		136.0	130.2
2.25		0.00118952		128.5	124.1
2.5		0.001130382		122.1	118.8
2.75		0.001079422		116.6	114.2
3		0.001034907		111.8	110.2
3.25		0.000995581		107.5	106.6
3.5		0.000960504		103.7	103.4
3.75		0.00092896		100.3	100.5
4		0.000900391		97.2	97.9
4.25		0.000874355		94.4	95.5
4.5		0.000850498		91.9	93.2
4.75		0.000828531		89.5	91.2
5		0.000808215		87.3	89.3

The two cooling capacity figures are the results of 2 different models (Wolf et al. and Choi et al.) as you can see the capacities are lower than what we are seeing here in the forum. Someone want to give some input?

[chilly1]

Originally posted by Gary Lloyd
Not really. The compressor size will make some small difference, increasing or decreasing the dP, but not that much.

Displacment of the compressor(Kg/S) will affect your load temps
and capacity of the system. Sizing the cap tube for sizing the cap tube will get you no where.

Take for instance the 2.75m capillary tube and the massflow of 0.001079422 kg/s what happenes if we increase the size of the pump on this system? and at the same time enlarge the condenser?

[Russell_hq]

If we enlarge the condensor, then we should have a larger degree of subcooling so increasing mass flow rate and hence capacity. The high side pressure and degree of subcooling are the only significant factors for cap tube sizing. Increaseing the size of compressor would only be required if the current compressor could not handle the volumetric flowrate required at the evaporating conditions and would only add extra heat to the system IMO. So I agree with Gary when he says, increasing compressor size will make small difference.

[Gary Lloyd]

The two cooling capacity figures are the results of 2 different models (Wolf et al. and Choi et al.) as you can see the capacities are lower than what we are seeing here in the forum. Someone want to give some input?

What is also of interest is the amount of capacity difference per meter of length adjustment. I would assume that your chart is far more accurate in that regard than my simplistic 15W/ft. If we can work out why we are seeing actual watt measurements for a given size that are different from your chart, we could probably work out a graph that is very precise for our purposes. In other words, find an accurate center point (say 150 watts), and use the adjustment curve your chart suggests. I think we are progessing here.

[Epsilon]

I hope this will get somewhere
It would be a tremendous help for the beginning PC'ers

[Gary Lloyd]

Russell, I haven't seen your models, but experience tells me that the primary factor is condensing temperature. You might try generating a chart using say 35C for condensing temperature, and see how that shifts the numbers.

[Russell_hq]

Ok, did this and this is the result.

Tube Length (m)	Mass Flow (kg/s)	Cooling capacity (watts)	
1.5		0.001542142		166.6	156.0
1.75		0.001431272		154.6	146.4
2		0.001341695		144.9	138.6
2.25		0.001267349		136.9	132.0
2.5		0.001204341		130.1	126.4
2.75		0.001150047		124.2	121.5
3		0.00110262		119.1	117.2
3.25		0.00106072		114.6	113.4
3.5		0.001023348		110.5	110.0
3.75		0.00098974		106.9	106.9
4		0.000959302		103.6	104.1
4.25		0.000931563		100.6	101.6
4.5		0.000906145		97.9	99.2
4.75		0.00088274		95.3	97.0
5		0.000861095		93.0	95.0

This is 35ºC condensing all else equal

[Gary Lloyd]

I hope this will get somewhere

Actually, we have already gotten somewhere. The method I presented will get you in the ballpark and will work quite well. We are now closing in on home base. This is both good news and bad news. The good news is that with precision we can give you a length that will result in the absolute lowest temp. The bad news is that, given such precision, if your load is a few extra watts your temps fall off a cliff. We will probably want to go just a little shorter than the precision length so that you can have some room for extra wattage.

[Gary Lloyd]

Note that increasing the condensing temperature had a significant impact on the centerpoint, but very little effect on our span (change per length adjustment). I suspect that the condensing temp primarily changes the centerpoint, while subcooling primarily changes the span.

We could test this by taking the condensing temp back to 30C and changing the subcooling to say 10K, and see what effect that has on the chart.

[Russell_hq]

Here we go:

Tube Length (m)	Mass Flow (kg/s)	Cooling capacity (watts)	
1.5		0.001638956		177.0	164.5
1.75		0.001521125		164.3	154.4
2		0.001425926		154.0	146.2
2.25		0.001346911		145.5	139.2
2.5		0.001279948		138.2	133.3
2.75		0.001222245		132.0	128.2
3		0.001171841		126.6	123.7
3.25		0.001127311		121.7	119.7
3.5		0.001087593		117.5	116.1
3.75		0.001051875		113.6	112.8
4		0.001019526		110.1	109.8
4.25		0.000990045		106.9	107.1
4.5		0.000963032		104.0	104.6
4.75		0.000938157		101.3	102.3
5		0.000915153		98.8	100.2

30ºC Condensing, 10K subcooling

Numbers changed a little after edit. Forgot to change a parameter in the model

[Gary Lloyd]

Hmmmmm... interesting.

I stand corrected. Both condensing temp and subcooling dramatically effect the centerpoint, while neither does much to the span. My faith in the span is growing.

[Gary Lloyd]

What is needed here is some real world testing. Conventional refrigeration is concerned with maximizing capacity for a given temp, while our concern is lowert temp for a given capacity. Therefore we need to know the 'falling off a cliff' point for our cap tube sizes. As I recall, Bowman has done such real world tests, and his input here would be extremely helpful.

With a few 'cliff' points, we could see if the disagreement with the models is in the centerpoint or the span (or both).

[Gary Lloyd]

As much as I hate to argue with the geniuses who give us the models... NOT... here is the bottom line:

We have seen a great many systems with longer cap tubes and higher heat loads working very well. According to the models, the temperatures should have taken a nosedive from lack of refrigerant flow, but they did not. Therefore, the models are WRONG.

There is something in my nature that makes me want to celebrate when the powers that be are proven wrong. In fact, my books totally trash the long accepted orthodox trouble shooting methods. Must be part of the iNTj thing.

[Russell_hq]

LOL.

These models are based on correlations with experimental data of specific refrigerants with limited operating conditions. So unfortunatly they are limited in their abilities, this will be what we are seeing here.

The models are also based on adiabatic capillary tubes, and we run non-adiabatic systems which give us greater capacity. Im currently trying to work out a model from a paper for suction line<->capillary tube HX which is based on mass, momentum and energy conservation so should be able to cope with a wider range of conditions. The models dont account for oil in the refrigerant either, that may be something that can be factored in.

[Gary Lloyd]

These models are based on correlations with experimental data of specific refrigerants with limited operating conditions. So unfortunatly they are limited in their abilities, this will be what we are seeing here.

Translation: Nice theory, but it doesn't work that way... LOL

I spin the wrenches out there in the real world. If it doesn't work on real machines, in my oh-so-humble opinion, it is useless crap.

[DaBit]

Every physical phenomenon can be captured in numbers Gary. If we succeed in capturing our captubes in numbers, and we write a nice piece of code around it, it will be a useful tool for the clueless.

[Gary Lloyd]

I agree, although I'm not certain that being clueless is a prerequisite for it's usefulness.

[Russell_hq]

Just did some sums based in 150 watt load condensing temp of 30ºC, subcooling of 5ºC, superheat of 5ºC and evaporating pressure of 1atm.

With a mass flowrate of 0.001389kg/s of R404a I calculated that the 2 phase region occurs approximatly 0.7m into a 0.028" ID capillary tube. So if we want to increase the mass flow rate through the capillary we need to start heat exchange within 0.7m of the start if the capillary.

[Gary Lloyd]

Even if there is no heat exchange between the cap tube and suction line, we can assume that the refrigerant will quickly achieve room temperature, which might translate to something like 10K subcooling shortly into the cap tube. With suction/liquid heat exchange, there would be a progressively increasing subcooling effect.

[Gary Lloyd]

Let's assume suction/liquid heat exchange, with an average subcooling midway into the captube of 28K. What effect would 28K subcooling have on the models?