View Full Version : Capillary Tube Characteristics

21-01-2009, 09:21 AM
Capillary Tube Characteristics

The capillary tube is generally envisaged as a control element for the refrigeration cycle but it should be thought of more as a fixed function element and it is the other components, the compressor, the condenser and to some extent the evaporator that adapt to balance the system.

The condenser
The condenser not only provides a liquid reservoir for the refrigerant not able to pass through the tube at the current system pressures but more importantly acts as a pressure regulator.

When the compressor mass flow is greater than the tubes mass flow, refrigerant accumulates in the condenser and reduces the surface area for condensation. This reduction in area causes a rise in the delta T and a corresponding rise in the saturation pressure until the pressure is sufficient to pass enough mass flow down the tube to reach equilibrium.

The condenser also provides what can be described as an ‘end stop’ to the discharge pressure by limiting it to the saturation temperature based on the current ambient conditions. For example if the ambient is taken as 32 deg C then it is unlikely the condenser will allow the fluid to fall below, say, 36 degrees C. The actual difference being based on the efficiency of the condenser but there is another controlling factor at this stage, this can be termed as bubble enrichment and is where the liquid form of the refrigerant contains a small amount of refrigerant in gaseous form. Again this becomes very important later in the tubes performance.

The evaporator
This unit has less effect on the tubes performance but is closely associated with the compressors function by helping to control the suction conditions and so the systems mass flow rate. It can play a control function in extreme conditions when evaporator starvation becomes the controlling element to reduce the mass flow rate and hence obtain equilibrium.

The Compressor
The main controlling element though is the compressor. It reacts to the tubes mass flow characteristics at the prevailing ambient and evaporator temperatures and continues to change its performance till it matches the tubes requirements.

Generally the suction pressure is controlled nicely by the evaporator and follows a classic path, reducing mass flow through the compressor as the temperature of the evaporator falls. So in order to match the tubes mass flow capabilities the compressor discharge pressure varies until there is sufficient head pressure for system equilibrium to exist.

Typical pulldown of a refrigeration system.
To understand how the capillary tube acts it is convenient to simulate a simple refrigerator from ambient startup to final stable operation. There are five separate regimes that capillary tube goes through during the pulldown cycle and each of these phases is controlled by a different interaction between the other main system components. The discharge pressure curve has been annotated with the time since startup and is of major interest up to the time=30 minutes.

Startup Phase Time 0 to 1 minute.
A short period where the refrigerant is being pulled out from the oil and the tube is trying to pass uncondensed refrigerant as a gas which would only represent about 0.5% of the mass flow rate. It is a transition state not involving the other elements of the system but does help to give rise to the very recognisable loop of the curve. It is also useful to determine the system charge, a long saught after parameter.

Build up period Time 1 to 5 minutes
During this period the compressor continues to suck refrigerant from the oil and more liquid builds up in the condenser. Most of the gas is thus being condensed and starts to fill up condenser with liquid until the area left available for actual proper condensation is sufficiently reduced that the head pressure keeps rising. When the pressure reaches a level where the tubes flow rate begins to equal the compressors the flow through the tube is fully established.

During the early part of this phase liquid interspersed with gas and then liquid again passes through the tube and is easily noted by listening to the evaporator plate. This emits a shush shush sound. Eventually at 5 minutes the tube is passing liquid at the start of the tube and flashing to gas as it passes along the length and the system starts to establish a stable feedback.

Liquid Phase Time 5 to 10 minutes
This phase is generally characterised by a high suction pressure and hence a high mass flow rate. To enable this high mass flow, in this instance over 1g/s to pass through the tube it needs excess pressure and a high subcooling, in the region of about 14 degrees C. The capillary tube at this point has liquid flowing through at least half or more of its length and as the pressure reduces along the tube, gas flashes off and increases the two phase flow velocity to over sonic speed. In this regime the evaporator pressure has no influence on the tubes flow rate. It is also characterised by vibration and is clearly audible.

Classic Phase Time 10 to 30 Minutes
The system through out this phase is stable and the capillary tube is operating in what can be termed as its classical regime. At the beginning of this phase the liquid is subcooled by about 8 deg C and loses its subcooling excess pressure by friction losses down the first part of the tube. This is typically about 1/3 of the length (but this varies from system to system) at this point and then it starts to flash into gas and as the refrigerant proceeds along the tube it loses more pressure and so more gas flashes off. Now about 25% or 30% of the refrigerant is in a gaseous state by the end of the tube and the volumetric flow rate is significant with internal velocities ranging from 50 to 100 m/s. This high velocity causes realistic pressure losses from frictional flow and now the tube follows both the suction and discharge from the compressor.

By the end the end of this phase the subcooling is down to about 3 or 4 degrees C and the refrigerant starts to flash to gas almost immediately it enters the tube, the whole length of the tube is now effectively filled with a gas/liquid mixture and as the pressure drop is dominated by the gaseous phase, significant pressure drops are seen, so the mass flow reduces in step.

The lower the mass flow, the lower the suction pressure becomes and so the compressor passes less mass, again balancing the tubes requirements.

By the end of the classic phase the condenser has no liquid reservoir and all the liquid is now within the evaporator. This classic regime has a very small operational envelope.

Limiting Phase 30 to 90 Minutes
As the system approaches final equilibrium there is a large change in the control mechanism. The liquid has a specific volume of about 100 to 200 less than the gases specific volume at the typical temperatures found in a refrigerator at operating conditions. Obviously the system has to continue to pass mainly liquid through the tube to achieve the mass flow required and so the tube enters a rather strange phase of bubble enrichment.

Most of the fluid entering the tube is in liquid form but it has entrapped within it a very small percentage of gas in the form of small bubbles. As soon as the pressure starts to drop as the mixture passes along the tube the gas expands and creates increasing velocities and hence increasing pressure drop and a lower mass flow rate. The lower the flow rate becomes the more bubbles are entrained in the liquid entering the tube until it could be up to 25% to 30% of the mixture. The main effect here is to move h4 away from the saturation line and reduce the refrigeration effect. As the temperature decreases further the refrigeration effect also continues to decrease rapidly and the COP changes respectively to point where inefficiencies dominate. Most system operate in this regime when they reach design running conditions.

Now the system is responding solely to the suction pressure and as the evaporator becomes colder the suction pressure falls and so does the compressors mass flow rate.

Looking at the delta pressure curve (red) in graph 1 shows that the pressure differential across the tube actually increases whilst the mass flow rate decreases – a most counter intuitive idea and not one expected in say a TX system.

This is caused by 2 possible operating conditions, the first is the bubble enrichment regime (as above) and the second may be due to the tube entering a phase where the suction pressure is so low the quality of the mixture in the tube reaches a point where the gas velocity is over 100m/s. This velocity causes such a pressure drop that the mass flow has to decrease to meet the available conditions.

The reason for all this strange behaviour is because the quality changes as the conditions at each end of the tube change. The lower the pressure in the evaporator the more the liquid/gas mixture becomes a gas and the higher velocity and so the pressure drop.
The PH diagram shows the systems performance at various times in the pull down and also the simulation allows the system heat loss to be plotted to determine the absolute minimum temperature that the system can reach.

Subcooling and superheat
Very little can be determined about the system condition from either subcooling or superheat. At every phase of the systems operation both of these factors are heavily involved in the control process and without knowledge of exactly where in the pulldown cycle the system is curently at any deductions are tentative at best. Basically SC and SH cannot be used to diagnose a capillary tube system.

The best way and probably the only way is to run a simulation of the system. The main controlling factors can then be studied and compared to actual values.

Key variables are the system charge and its relation to the volume of both the condenser and evaporator. A small condenser volume will induce large startup pressures and possibly overload the compressor during the first 5-10 minutes, an excessively large condenser will cause lower head pressures at critical points and thus reduce flow rates in the cap tube, the evaporator will become starved and control will be obtained from depression of the suction volume. A situation not desired for efficient operation.

Changing the diameter or the length of the capillary tube without matching it to the condenser, evaporator and charge is purely guesswork and does not allow the systems performance and startup properties to be optimised.

It is far better to size the compressor to the tube and relate the evaporator to the charge and adjust the condenser size to establish the desired performance.

There is much more to do yet!


21-01-2009, 10:36 AM
Thanks Chef, by first fast reading, this is very good article.
You should publish this.

22-01-2009, 08:42 AM
Thanks Chef, by first fast reading, this is very good article.
You should publish this.
Thanks Nike123 - it was something started last year to write a general program to predict the cap tube length but it never seemed to fit all conditions and so a simulation model with the tube program included in it showed why it was so complex. The results of the simulation program are a bit too big to include in an already oversized post. :D


26-01-2009, 06:40 PM
nice one Chef, i have made a copy for mi database so i can reed it whenever i want to.
keep going on,

26-01-2009, 07:40 PM
Really impressing Chef :)

You have clearly done a lot of work on this subject. I definitely learned something new here.

28-01-2009, 08:11 AM
Thanks icecube51 and SteinarN.

I hope it helps and maybe one day I will get round to posting the programs, they would be much more fun.;)


15-02-2009, 09:09 PM
On the P_H Diagram we see you have plotted the system load. We have been trying to plot this for our cabinet but dont see where you get the values for H from. Can you explain this to us as it helps to show our cabinets performance at changing temperature and load.

If we fill the freezer with produce how will this change the P_H Diagram?

16-02-2009, 06:21 AM
On the P_H Diagram we see you have plotted the system load. We have been trying to plot this for our cabinet but dont see where you get the values for H from. Can you explain this to us as it helps to show our cabinets performance at changing temperature and load.

If we fill the freezer with produce how will this change the P_H Diagram?

Hi mrfrostie.
The program actually calculates the system heat loss and not the total system load but the available cooling to cool down the produce can be found.

Using some simple relations:- Q=U*A*(Ta-Tc) is the heat lost from the cabinet so if the cabinet internal temperature Tc is say -10C the amount of heat loss can be obtained for known values of Ta, U and A.

Now Q=M1*(h1-he) where he is the enthalpy value that lies somewhere on the line h4 to h1. This amount of cooling is that required to overcome the heat loss from the cabinet.

For your application Qproduce=M1*(he-h4) and this is the amount of cooling left over to chill down the produce. Now you can use standard load calculations to get how long it will take to freeze the produce.

Now assume Tc is set to -20C the cabinet heat loss increases and the so therefore does the value of M2*(h1’-he’). The value of M2 will be less than M1 as the cabinet temperature is lower and so he’ moves to the left on your diagram.

So now you can construct 2 lines on your diagram, one at -10C and one at -20C (corrected for the actual evap temp of course) and then plot he on the -10C line and he’ on the -20C line to get the heat loss line.:D

If you dump a whole load of new produce in the freezer then the changes in cabinet temperature may mean you will need some extra lines on your diagram.

If you send me as many details as you have I can try and run it for you!


27-02-2009, 07:15 AM
This excellant post helps me realize just how much I appreciate TXV's on several fronts.
Only problem is the metric charts, difficult for us not into metric to follow, but when I get time I will change/convert the charts and re read your post.

After all these years in refrigeration this helps prove that cap tubes are best charged by weight, anything else is a guess.

Thanks for the great post

Thanks for the excellant post.

10-03-2009, 01:03 PM
really good one. i liked it very much.

many thanks