Expansion noise - capillary tube into evaporator

[desA]

I have a small heat-pump on test, in my lab. The unit has a capillary tube installed, together with inlet filter/strainer.

At various points during the heat-up phase, noises like machine-gun fire, can be heard - from time to time. I traced the main noise to the junction between capillary tube exit, & evaporator inlet. Checking the sight glass shows a fair amount of small bubbles coming up from the condenser coil.

At times, the system goes completely quiet again. The noise seems to be associated with the bubbles, as is to be expected.

This is the first time I've worked with a capillary tube expansion device, on a heat-pump.

Does anyone have any thoughts on how to reduce the expansion noise?

[chemi-cool]

Noise from cup tube expansion are normal, they are far from a machine gun noise

What refrigerant are you using?
Is it a small 10 to 30 grams drier filter?

How can you tell the amount of refrigerant? Is there a sight glass?

It sounds to me that you don't have enough refrigerant or condenser is too small.

[Mark II]

Des,

I always tell my customers .."When it stops making a noise , then you know you have a problem."

[lowcool]

maybe go up a size or two

[desA]

Capillary tube setup.



I'd value comments & suggestions for improvement.

[desA]

Noise from cup tube expansion are normal, they are far from a machine gun noise

Sound-bite: sssshhhh....ssshhh... ka-ping... ping... ping... ka-ping... sssshhhh.... ka-ping...

What refrigerant are you using?

R-134a

Is it a small 10 to 30 grams drier filter?

See copper drier in pic below, at entrance to capillary. A larger filter-drier exists upstream.

How can you tell the amount of refrigerant? Is there a sight glass?

There is a sight glass. At the moment, a fair amount of bubbles seem to come through - mainly clouds of small size, during operation - the clouds come & go.

It sounds to me that you don't have enough refrigerant or condenser is too small.

I'll run SH trials today to check refrigerant mass (roughly). I do think that the condenser will be under-size, based on the system design. I'll check the performance theoretically, then report back on this.

[desA]

Des,

I always tell my customers .."When it stops making a noise , then you know you have a problem."

Hahaha... one sure check, it seems.

Why is a domestic refrigerator so quiet then?

[Gary]

The drier outlet should be pointed downward.

[desA]

maybe go up a size or two

This is what I am thinking, I must admit.

To my 'ear' (big screwdriver on cap tube region), it sounds like the bubbles make their way to the capillary tube entrance, then form liquid-vapour bubbles-slugs, which then accelerate through the capillary tube - at high velocity.

[desA]

The drier outlet should be pointed downward.

Thanks so much, Gary.

Which drier? The small one?

[Gary]

Both... the strainer and the drier.

[desA]

Both... the strainer and the drier.

Thanks, Gary.

So, basically, to form a liquid seal at the bottom, to ensure continuous liquid feed into cap tube?

Would this be enough to catch all/most of the bubble clouds, would you think? Remember, I can see clouds of vapour passing through the vertical sight glass, which sits not too far from condenser exit - before both filter & strainer.

[Gary]

Thanks, Gary.

So, basically, to form a liquid seal at the bottom, to ensure continuous liquid feed into cap tube?

Would this be enough to catch all/most of the bubble clouds, would you think? Remember, I can see clouds of vapour passing through the vertical sight glass, which sits not too far from condenser exit - before both filter & strainer.

Usually cap tube systems do not have a sight glass installed... because it tends to confuse people.

[desA]

Usually cap tube systems do not have a sight glass installed... because it tends to confuse people.

Can you explain further?

[Gary]

Unlike a TXV, the cap tube doesn't necessarily have solid liquid at it's inlet. Depending on the cap tube sizing and system balance, some work with lots of subcooling and some very little... and even this changes with conditions.

[desA]

Unlike a TXV, the cap tube doesn't necessarily have solid liquid at it's inlet. Depending on the cap tube sizing and system balance, some work with lots of subcooling and some very little.

Interesting.

With this kind of condenser - external coil around water storage tank - if the coil were of correct size (surface area), I would expect some level of sub-cooling to be present, at least. The reason for this would be that the temperature gradient in the storage tank (cold base, hot top), would tend to allow some level of liquid sub-cooling.

[lowcool]

when i was taught about charging with a capillary and sightglass fitted it was normal practice to have a small stream of bubbles,no bubbles=system was overcharged

[desA]

when i was taught about charging with a capillary and sightglass fitted it was normal practice to have a small stream of bubbles,no bubbles=system was overcharged

Thanks, lowcool.

What size bubbles are we talking about - roughly?
Do these bubbles come as a continuous flow, or are there spurts (clouds)of bubbles, followed by quieter patches, then spurts again?
What 'should' the system sound/feel like?

[mad fridgie]

Firstly sorry for not replying earlier.
First fix. Place ear muffs over ears, I think that should solve the problem.
Second drier upside down excess vapour entering. (well done gary, have a banana)
you can not use a slight glass for charge analise (only moisture indication) Reason charge migrates from evap to cond and back again.
A cap system is basically a flooded system.

[desA]

Firstly sorry for not replying earlier.
First fix. Place ear muffs over ears, I think that should solve the problem.

Engineers & our ears... small noises like that bother me.

Second drier upside down excess vapour entering.

Let's expand on this, after piping corrected to set inlet strainer downwards. Does the length of the cap inlet filter make a difference?

you can not use a slight glass for charge analise (only moisture indication) Reason charge migrates from evap to cond and back again.
A cap system is basically a flooded system.

Now, would that apply if the system were only operating in 'batch' mode, where it cycles cold to hot, stops, repeats - or. in 'semi-continuous/quasi-steady' mode, where the heating cycle can be reasonably consistent over long periods of water usage?

In the first case, I see migration inducing a transient operating state. In the second, the charge will tend to balance at the operating point, during the quasi-steady operation.

[mad fridgie]

I would have just poking in your are attemting to achieve a liquid seal.
One would expect some sort of equalibrium to reached, When and where it sits ?????????? I would say you just about always have some liquid in the evap, what level comes out/or is in the condensor changes. is more likely to be variable.

[mad fridgie]

PS, I do not like capillaries on forced draft evaps, Acceptable on static coils.

[desA]

I would have just poking in your are attemting to achieve a liquid seal.

One would expect some sort of equalibrium to reached, When and where it sits ?????????? I would say you just about always have some liquid in the evap, what level comes out/or is in the condensor changes. is more likely to be variable.

Fair comments. Thanks very much for that.

I figured this much, & will be watching the SH ultra-closely.

[Chef]

desA - Whats the conditions at the condenser and evaporator, temperatures and pressures whilst in steady state and what are the dimensions of the cap tube.
The capacity of the compressor is also very important, the volumetric flowrate at rated RPM is the best number to have.

Once the operating point of your tube is established the cause of the noise can be determined.

Chef

[desA]

PS, I do not like capillaries on forced draft evaps, Acceptable on static coils.

Me neither - they scare the pants off me, to be honest.

I far prefer to see a TXV in place.

I suspect for this unit, cost is a big factor. That said, these capillary tubes don't seem to be too cheap. I'd rather save elsewhere & go to a small TXV.

[mad fridgie]

desA - Whats the conditions at the condenser and evaporator, temperatures and pressures whilst in steady state and what are the dimensions of the cap tube.
The capacity of the compressor is also very important, the volumetric flowrate at rated RPM is the best number to have.

Once the operating point of your tube is established the cause of the noise can be determined.

Chef

Well here is the man to talk to, I bow my head, and leave the room backwards on my knees

[desA]

desA - Whats the conditions at the condenser and evaporator, temperatures and pressures whilst in steady state and what are the dimensions of the cap tube.
The capacity of the compressor is also very important, the volumetric flowrate at rated RPM is the best number to have.

Once the operating point of your tube is established the cause of the noise can be determined.

Chef

Hi Chef,
Thanks for weighing in here. You are our resident 'cap tube fundi'.
Once I've finished a few more runs, to settle the system, I'm going to have to hack into the lines to install a hp service port. After this, I'll be able to get definitive data, for us to work with.

One difficulty with a heat-pump is that evap & condenser operating points move over the course of the cycle. This will be interesting.

[Chef]

One difficulty with a heat-pump is that evap & condenser operating points move over the course of the cycle. This will be interesting.

With a fridge system the condensor stays roughly at the same temp and the evap slowly drops in temp. This is opposite in a heat pump where the evap is roughly constant and condenser increases in temp.

Sometimes this is a benefit to a cap tube as it stabilises the tube conditions over a wider range of operating conditions. It means once the tube is correct dimensions it will operate at best efficiency over a wide range.

The bubbles are quite normal and gives a very good indication of what state your tube entrance conditions are.

Chef

[desA]

Thanks, Chef. Excellent points.

With a fridge system the condensor stays roughly at the same temp and the evap slowly drops in temp.

Very true. An a/c follows a similar pattern.

This is opposite in a heat pump where the evap is roughly constant and condenser increases in temp.

For a heat-pump, the evaporator Te,sat does actually drift upwards slightly, over the course of a heating cycle, although this drift is small compared to the Tc,sat movement.

Sometimes this is a benefit to a cap tube as it stabilises the tube conditions over a wider range of operating conditions. It means once the tube is correct dimensions it will operate at best efficiency over a wide range.

Good point.

The bubbles are quite normal and gives a very good indication of what state your tube entrance conditions are.

Now, I'd like to get your thoughts on what the bubbles should look like. This will surely be an important system indicator? In other words, large lazy bubbles, or intermittent clouds of tiny bubbles.

[Chef]

Now, I'd like to get your thoughts on what the bubbles should look like. This will surely be an important system indicator? In other words, large lazy bubbles, or intermittent clouds of tiny bubbles.

Without the full system details it is not easy to predict and your systems geometry will play some part in it but in general the following would be a good guide.

A If you have SC of around 4 to 6C and above it will be all liquid with maybe a few very small bubbles.

B With an SC of 1 to 4C there will be more bubbles and if the piping were perfect it should be steady streams roughly constant.

C As the SC falls to 0C then the bubbles will be larger are more of them with even slugs of gas passing temporarily followed by bubbles entrained in liquid.

In both conditions B and C above when the more gas bubbles enter the tube its resistance increases, the flowrate falls a little and the condensor pressure rises - just a little - but enough to increase the SC by a fraction and thus produce a stream of enhanced liquid flow. This in turn eases the pressure drop along the tube, increases the flow and drops the condensor pressure just a fraction, lowers the SC and more bubbles pass. The cycle may repeat itself.

D The SC is at 0C and condensor pressure continues to rise. In this case more and more gas will pass through the tube to balance its flow rate to the higher condensor pressure, X values from 5% to 10% may occur and the sight glass if Horizontal will show a clear liquid but will not fill the sight glass ie it will have liquid on the bottom and gas on the top.

E The condensor pressure rises even further the value of X increases to up about 40% (max). From 10% up to the 40% the pressure drop is high in the tube and flow rate is low because of all the entrained gas, the outlet of the tube will likely be sonic and now the evap pressure will start to fall as the compressor sucks more gas than the tube and evap can provide, the evap pressure drops and the system finds a new balance point.

Note - there is only one set of evap and condenser conditions that provide the perfect cap tube conditions for its particular length.
When the condensor is at lower temps than perfect there will be more SC than desired and when the condensor is at higher temps than perfect the tube will operate in the X region with gas entrainement. There is nothing you can do about this except optimise the perfect point so there is some SC excess and some small X during the whole cycle.

Post a picture of the sight glass.

Chef

[desA]

Thanks so much for a stunning review. I've added in what I'd estimate the flow regimes to be, judging from your descriptions. If you could amend these as you see fit, it would be greatly appreciated.

Without the full system details it is not easy to predict and your systems geometry will play some part in it but in general the following would be a good guide.

Agree. Very true.

A If you have SC of around 4 to 6C and above it will be all liquid with maybe a few very small bubbles.

Ok. Single phase liquid.

B With an SC of 1 to 4C there will be more bubbles and if the piping were perfect it should be steady streams roughly constant.

Ok. Bubbly flow.
Can you expand on the point - "if the piping were perfect"?

C As the SC falls to 0C then the bubbles will be larger are more of them with even slugs of gas passing temporarily followed by bubbles entrained in liquid.

Ok. Slug flow.

In both conditions B and C above when the more gas bubbles enter the tube its resistance increases, the flowrate falls a little and the condensor pressure rises - just a little - but enough to increase the SC by a fraction and thus produce a stream of enhanced liquid flow. This in turn eases the pressure drop along the tube, increases the flow and drops the condensor pressure just a fraction, lowers the SC and more bubbles pass. The cycle may repeat itself.

Ok.

D The SC is at 0C and condensor pressure continues to rise. In this case more and more gas will pass through the tube to balance its flow rate to the higher condensor pressure, X values from 5% to 10% may occur and the sight glass if Horizontal will show a clear liquid but will not fill the sight glass ie it will have liquid on the bottom and gas on the top.

Ok. Annular flow.

E The condensor pressure rises even further the value of X increases to up about 40% (max).

Ok. Annular - semi-annular.

From 10% up to the 40% the pressure drop is high in the tube and flow rate is low because of all the entrained gas, the outlet of the tube will likely be sonic and now the evap pressure will start to fall as the compressor sucks more gas than the tube and evap can provide, the evap pressure drops and the system finds a new balance point.

Ok. Sonic choke.

Note - there is only one set of evap and condenser conditions that provide the perfect cap tube conditions for its particular length.
When the condensor is at lower temps than perfect there will be more SC than desired and when the condensor is at higher temps than perfect the tube will operate in the X region with gas entrainement. There is nothing you can do about this except optimise the perfect point so there is some SC excess and some small X during the whole cycle.

This is the crux of the design, isn't it? Where would you aim to set this balance point in a heat-pump cycle - cold start / mid-range / hot? Bearing in mind that most of the time a heat-pump of this nature will be operating in the mid-to-hot range as it 'tops up' heat in the storage tank.

Thanks so much, Chef for an absolutely stunning post. It is incredibly helpful & very much appreciated.

Post a picture of the sight glass.

I've been trying to snap decent pics. With my camera, it is easier said than done... Will need to practice.

[mad fridgie]

I have had a closer look at your pic, (but without knowing all components)
I suspect that your present cap, is too long and too small diameter. You may struggle to even get a descent starting point.

[desA]

I have had a closer look at your pic, (but without knowing all components)
I suspect that your present cap, is too long and too small diameter. You may struggle to even get a descent starting point.

Thanks for that.

Where would you go, from the present cap tube, and why?

[Chef]

Can you expand on the point - "if the piping were perfect"?

First MAD has a point, it does look like it is somewhere too long but only your data will allow this to be confirmed.

Having the 2 driers feeding from the bottom to the top might allow them to act as accumulators and upset the natural balance of charge dissipation as the system heats up.

When it is subcooled these will be full of liquid and so deplete the evap from what could be more liquid and more performance.

When it is in X regime these will be part gas and part liquid and act like a bubble pipe - no reason for that but means your charge will not be perfect.

If they are small compared to the evap size and condenser size dont worry but they may have affect on the last 10 to 15% of the performance?

Chef

[desA]

First MAD has a point, it does look like it is somewhere too long but only your data will allow this to be confirmed.

This machine has been designed & built by others. I bartered it for design time on a project. The plan is to debug it as far as possible & at the same time trim costs all the way down - if possible. So, it is a fantastic learning tool.
I will get into the specific retrofitting side once we get some clear direction from this thread. I have the Danfoss cap tube program & can purchase a second alternative. I'd then like to design/select the optimum cap tube to suit the system. This seems like it will be a challenge.

Filter-drier & inline strainer
On the current machine the liquid line goes - tank condenser - sight glass (vertical upflow) - filter-drier (horizontal) - inline strainer (vertical upflow).

Having the 2 driers feeding from the bottom to the top might allow them to act as accumulators and upset the natural balance of charge dissipation as the system heats up.

Did you mean bottom-to-top, or top-to-bottom?

When it is subcooled these will be full of liquid and so deplete the evap from what could be more liquid and more performance.

Mmhhh... good point. Would this require a tad more refrigerant to make up for the inventory hold-up?

When it is in X regime these will be part gas and part liquid and act like a bubble pipe - no reason for that but means your charge will not be perfect.

Could you perhaps expand this a little further?

If they are small compared to the evap size and condenser size dont worry but they may have affect on the last 10 to 15% of the performance?

Very interesting viewpoint. In this particular machine, I consider the evap performance to be marginal. Could you expand this point a little further?

Thanks again... excellent food for thought.

[mad fridgie]

The problem you are going to have is choosing adesign point for your cap. the nature of the hot water heating beast is that it is always in a wide range of flux (compared to say a fridge which generally is more stable once the cabinet is at the required temperature)
maybe you should look at a single pass water heating system, controled by head pressure, this will give you a constant to work with, so then you only have to optimised around SST.
I do not think you should use your sight glass at all to judge charge. The only benefit (excluding moisture) is to show that you have flow. (Gauges would show the same)

[Chef]

When the system has too much accumulator volume after the condenser it will steal liquid from the evap.

When the system goes in to X mode then the evap will be full of liquid (as full as it can be without flood back) so this is the max charge you should use.

But when in SC mode there is more liquid in the condenser/accumulator than you need so less in the evap and lower performance.

The cap tube is just a simple pipe and has no control function whatsoever - it is the changes in condenser pressure and SC and the changes in evap density from pressure and temp that provide the system balance and so it these items you need to be watching. They will change all the time as the system parameters alter and will work to match the compressor flow so its a combination of four parts all the time to get an understanding of whats going on.

Without real data from the system it is difficult to run a simulation and see whats happening.

Oh and by the way - as discussed on many threads before be very careful using Dancap as its results are 'unique'.

Chef

[Chef]

I do not think you should use your sight glass at all to judge charge. The only benefit (excluding moisture) is to show that you have flow. (Gauges would show the same)

Your quite right that selecting charge on a sight glass in this system would be almost useless. But it does show what condition the refrigerant is at as it enters the tube and for research and cap size changes it can be very useful.

It would be great if one could buy a glass tube - say 300mm long - about the same diameter as the compressor discharge tube to see the refrigerant condition. What a tool that would be.

Chef

[desA]

The problem you are going to have is choosing adesign point for your cap. the nature of the hot water heating beast is that it is always in a wide range of flux (compared to say a fridge which generally is more stable once the cabinet is at the required temperature)

I think that this is the key, here. The varying heat-loads, on both the evaporator side & condenser side, keep the system in relative dis-equilibrium.

maybe you should look at a single pass water heating system, controled by head pressure, this will give you a constant to work with, so then you only have to optimised around SST.

A very good idea. This kind of device ends up with far less in the way of variables. It's a way of thinking, more than anything else.

I do not think you should use your sight glass at all to judge charge. The only benefit (excluding moisture) is to show that you have flow. (Gauges would show the same)

Not to judge charge, but more of an overall indication of the instantaneous state of the system. Charge optimisation has always to be done together with a number of other parameters - even then, things change on the day...
These RHVAC systems are a lot more sensitive than many folks would have us believe.

[nike123]

Maybe this article could have some interesting information.

[desA]

Some further feedback from today's runs:

1. Leveled the heat-pump mounting plate - evap more level;
2. Adjusted system to various quasi-steady settings:
2.1 Full sight-glass
2.2 Capillary noise minimal
3. Change load to next quasi-steady point:
3.1 Bubble appeared until steady - then disappeared;
3.2 Capillary noise minimal.

Moved right up the heating range to ~ 57'C.

4. Quasi-steady at 57'C
4.1 Capillary sounds like fast flow;
4.2 No intermittent noise.

5. Set up to tank heat-up cycle
5.1 Capillary noise - pop, ping, but more constant

6. Added mastik dampers to cap ends
6.1 Quietened noise a lot - more constant - sounds like plug flow.

An interesting day.

[desA]

Maybe this article could have some interesting information.

Interesting article. Does anyone have access to this paper through a University library?

ScienceDirect, as usual, want an enormous amount (USD 31.50) for a paper - an incredibly greedy company is Elsevier, considering that the authors pay to have their paper published.

[chemi-cool]

Very interesting discussion.

I have learnt a few new things today.

I may add one thing concerning the noise.

1. would it be quieter if a few smaller diameter cup tubes will be instead of one large diameter?

2. Because the refrigerant is R 134a, in a short time it most likely will block the cup tube, my suggestion is to use a small 1\4 drier, size 052 and braze the cup tube to the 1\4 tube on the outlet.

And Des, how about some work today?

[desA]

When the system has too much accumulator volume after the condenser it will steal liquid from the evap.

Fair comment.

When the system goes in to X mode then the evap will be full of liquid (as full as it can be without flood back) so this is the max charge you should use.

Where in a heating cycle would you expect 'X' mode to occur? Start-up, hot?

When you say 'full' of liquid - are you saying that all the charge will migrate to the evap? This unit has a fair vertical leg on the suction line, before entering the compressor. I'd assume that the maximum evap liquid level allowed would then correspond to a fair portion of this height? Where would you determine the 'critical charge'?

Surely, the tank condenser could be expected to retain some refrigerant?

But when in SC mode there is more liquid in the condenser/accumulator than you need so less in the evap and lower performance.

So, the whole balancing act is realistically determined by how much liquid that cab safely be 'buffered' in the evaporator?

The cap tube is just a simple pipe and has no control function whatsoever - it is the changes in condenser pressure and SC and the changes in evap density from pressure and temp that provide the system balance and so it these items you need to be watching.

Fair comment. Yes, correct.

They will change all the time as the system parameters alter and will work to match the compressor flow so its a combination of four parts all the time to get an understanding of whats going on.

Your paper from the Indian university explained some of this pretty well, I thought. Good stuff.

Without real data from the system it is difficult to run a simulation and see whats happening.

Agreed. I'm looking at tapping into an existing port on the high side, to at least get the initial data, so we can begin making sense of where we are. I'm probably going to have to order in a number of additional thermocouples, as my lot is still in my large test heat-pump.

Oh and by the way - as discussed on many threads before be very careful using Dancap as its results are 'unique'.

I did wonder about that. I have another program on hand, but will also write up a program to calculate this. You wouldn't just happen to have written one in EES, would you - one that can fit into the Academic version I happen to have on hand.

[desA]

I may add one thing concerning the noise.
1. would it be quieter if a few smaller diameter cup tubes will be instead of one large diameter?

Interesting thought. I wonder if anyone has thoughts on this?

2. Because the refrigerant is R 134a, in a short time it most likely will block the cup tube, my suggestion is to use a small 1\4 drier, size 052 and braze the cup tube to the 1\4 tube on the outlet.

Isn't that what we have at the start of this capillary? Is R-134a worse than other refrigerants for blockage?

And Des, how about some work today?

The R&D is part of my work...

[Gary]

When you say 'full' of liquid - are you saying that all the charge will migrate to the evap? This unit has a fair vertical leg on the suction line, before entering the compressor. I'd assume that the maximum evap liquid level allowed would then correspond to a fair portion of this height? Where would you determine the 'critical charge'?

Surely, the tank condenser could be expected to retain some refrigerant?

During the off cycle, virtually all of the refrigerant charge is going to migrate to the evaporator, because it is cooler than the condenser.

That vertical leg is a very good thing.

[desA]

A further thought.

What 'noises' could one expect if the refrigerant were (slightly) over-charged?

For instance, I noticed that on many occasions, with a full sight-glass, during quasi-steady operation, noises did, on occasion occur. It made me realise that the bubbles, when they showed up, were not the main cause of the capillary noise, but rather some other effect.

[desA]

During the off cycle, virtually all of the refrigerant charge is going to migrate to the evaporator, because it is cooler than the condenser.

True, you are correct - especially for this type of tank condenser.

That vertical leg is a very good thing.

What kind of vertical height would you recommend - relative to evaporator height? Is there some kind of rule-of-thumb?

[chemi-cool]

Hi Des,

I am changing on a weekly base blocked driers on small cup tube systems such as air driers, beverage cabinets, water coolers, domestic fridges and freezers.

This refrigerant is too good when it comes to cleaning.

By using regular 1\4 drier, the problem disappear.

[lowcool]

has anyone used that 1234 whatever refrigerant it is yet.
maybe you could step us into the future des by taking that path.
i know more r&d = more work