PDA

View Full Version : Non-condensable's and Sight Glass Bubbles



Saturatedpsi
29-08-2014, 01:56 PM
Howdy to everyone...

I'm trying to understand the effects of NCG's in high temp systems. Hours of research haven't as yet answered my question(s). I staged an "experiment" with an R22 split heat pump system (running in the cool cycle), beginning with decent performance numbers, then added a small amount of nitrogen to the system. TXV fed coil.

Beginning numbers:

Suction/SST...71 psig/41.6˚F
Superheat...13˚F
Liquid/SCT...213 psig/106˚F
Subcooling...11˚F

Sight glass full, located 12" downstream of liquid service valve

After adding nitrogen:

Suction/SST...73 psig/43˚F
Superheat...12˚F
Liquid/SCT...225 psig/110˚F
Subcooling...17˚F

Sight glass gave appearance of undercharge...less than full.

Anyone know what causes the bubbles/voids in the sight glass, and what they are? Nitrogen or *****? :confused:

Rob White
29-08-2014, 04:03 PM
.

Non condensables inside a system tend to end up in the hottest and highest
part, which normally mean the condenser. Because they can't be condensed they
just sit there taking up space. If they take up enough space they ultimately
reduce the capacity of the condenser.

If the efficiency and capacity of the condenser are reduced sufficiently then
it will stop condensing and vapour (saturated gas) will pass through the liquid line
and give the appearance of bubbles in the liquid, hence short of refrigerant.

In your experiment your liquid pressure went up in the second experiment
and the subcooling increased. Normally subcooling would decrease until the liquid
temperature met the same temperature as the condensing temp. With R22 I
can't explane why your system showed bubbles yet subcooled liquid, unless
that was just a quirk of the system and / or the condenser is over sized?

As an experiment try it again (as in the first option) but this time block the condenser
so the liquid temps meet the conditions in your second experiment. If you have
bubbles in the sight glass that might explane a lot?

Regards

Rob

.

Saturatedpsi
30-08-2014, 01:33 PM
Thank you for the input...I was slow to reply, just waiting on someone else to weigh in.


.
If the efficiency and capacity of the condenser are reduced sufficiently then
it will stop condensing and vapour (saturated gas) will pass through the liquid line
and give the appearance of bubbles in the liquid, hence short of refrigerant.

With the relatively small increase in head pressure, I wouldn't think the limits of the condenser capacity were reached. And, after the NC generated the increased head and subcooling, I was able to remove charge and lower the values back to the non-contaminated system numbers...if that means anything.


With R22 I can't explane why your system showed bubbles yet subcooled liquid, unless that was just a quirk of the system and / or the condenser is over sized?


I've read other posts in other forums, where bubbles were observed in a NCG contaminated system sight glass, which suggest it's not a system quirk. The condensing unit is "factory" design, so I wouldn't think condenser oversize is a factor.


As an experiment try it again (as in the first option) but this time block the condenser so the liquid temps meet the conditions in your second experiment. If you have bubbles in the sight glass that might explane a lot?

Based on what I've seen, substantially restricted condenser air generates an increase in subcooling and liquid temp increase....but with a mere 12 psig increase in head pressure, not sure what the liquid would do.

That's one point of interest with very small amounts of NCG's. They don't affect the numbers significantly. And without the option of observing "bubbles", in the absence of a sight glass, you'd never suspect contamination.

Thanks again.

Rob White
30-08-2014, 04:51 PM
.

Good post and it raises some valid points.

In truth I'm not sure if I can explain it fully to you or myself and
it seems to be a condition that could cause hugh debate, just like
the debate about the state of the refrigerant passing through a
liquid receiver (is it subcooled or saturated?).

One obvious point is the discharge pressure / temperature. As it goes up
due to the Non-condensable the subcooling will increase because the liquid
has heat removed. The subcooling will increase until the point when the system
fails and vapour passes through the condenser.

Interesting............ :)

Rob

.

Saturatedpsi
31-08-2014, 05:48 AM
.

Good post and it raises some valid points.

In truth I'm not sure if I can explain it fully to you or myself and
it seems to be a condition that could cause hugh debate...

Well, kudos to you for the acquiescence... :)

I don't lose sleep over the "science"...My objective is just being able to make a quick and accurate diagnosis.

What got me on the "bubbles" initially was the thought I had discovered a simple way to differentiate between NC's and the other stuff that causes high head/subcooling. I was gonna' try to find an "electronic sight glass" that TIF used to sell, confirm the bubbles, and make the call. But I think they (electronic sight glass) must have gone out along with propane leak detectors. :rolleyes:

It's only been recently that I came across some high NC content systems, and did a lot of wheel spinning before zeroing in on the true problem.

One was a TXV system that had to be overcharged to 25˚F subcooling just to get 20˚F+ superheat. I kept thinking in terms of liquid restriction (that I couldn't find) or TXV issues. Replaced two valves before someone suggested NCG's.

Did a condenser "pump down" and the indicated saturated temp was 16˚F above ambient. :eek: Removed and replaced the charge, and all the numbers fell into place.

Another thing I learned with that one, and high NCG content: you can't reduce the subcooling to low single digits by removing refrigerant, as you can with a simple overcharge situation. I had dumped liquid till the recovery drum pressure was equal to the system head, and still had 12˚F subcooling.

But, after all that I started trying to understand the "bubbles"...should have left well enough alone. :confused:

nike123
31-08-2014, 09:04 AM
What are your temperature and temperature intervals (superheat and subcooling) measuring units?

Saturatedpsi
31-08-2014, 01:31 PM
What are your temperature and temperature intervals (superheat and subcooling) measuring units?

˚F...apologies for not noting units. Will edit the posts.

Saturatedpsi
03-09-2014, 01:51 PM
Well, that horse didn't get very far down the track. :)

Does anyone think Henry's Law has something to do with it? :confused:

Henry's Law (http://en.wikipedia.org/wiki/Henry's_law)

Rob White
03-09-2014, 03:31 PM
.

I think Henry's Law relates to different gas and liquid combinations.

It states that the solubility of a gas in a liquid is directly proportional
to the partial pressure of the gas above the liquid.

In refrigeration the gas is the refrigerant.

My interpretation of gas is it is heated above the vapour (and well above the saturation temp) temperature (superheated)

Refrigerant can exist in liquid, vapour and gas forms. Gas is heated above the saturation
temperature but vapour can almost touch the saturation temperature.

Henry's law in my opinion is relating to different products, water and CO2 for example
makes soda water or carbonated water and that is in complete agreement with Henry's
Law, but we're dealing with refrigerant and air (OFN in your experiment), which are different
(like CO2 and water) but the pressures required to dissolve air (which is Oxygen and Nitrogen)
into the refrigerant are impossible in a normal refrigeration vapour compression cycle.

I don't know what causes the bubbles you see in the sight glass but I do know it is not
air (or OFN) that you see, it is an interruption of how the condensing process works and is
disrupted with Non condensables.

Bubbles can be seen in Zeotropic refrigerants because
of glide even if the liquid is subcooled,I can't explain that either :D

Regards

Rob

.

Saturatedpsi
03-09-2014, 03:56 PM
Fair enough...thanks. :)

mad fridgie
03-09-2014, 08:54 PM
Lest start with the most likely, incorrect measurement of liquid pressure and temp (as much to do with accuracy of measuring device and influence of ambient temp)
Velocity entrapment, where the speed of the liquid, draws non condensable bubbles think of a piece of wood in a fast moving river, even though it should float, it actually bibs up and down below its floating level)
Bubble production due to low pressure areas within the piping, especially R4***refrigerants, and it does take time for them to re-condense.

sterl
05-09-2014, 10:19 PM
2 cents worth....
Industrial systems have purgers to remove non condensables. Many operate in a vacuum on the low side and air gets in....Lots of ways. When you don't purge, the air in the condenser represents an over pressure (partial pressure) and the condensible medium (refrigerant) vapor pressure essentially makes up the rest of the total. If a condenser is entirely full of refrigerant liquid and gas, and you valve it off and let it sit for a minute or 2, its pressure will drop to that corresponding to its surroundings. Do this with a condenser that contains some NCG and the pressure will stay higher than that corresponding to its surroundings.

So the liquid exiting a saturated (industrial) condenser that contains some Non Condensibles will be COLDER than the measured pressure of the condenser would chart to, because the pressure in the condenser is falsely high....and the condensing is happening at the partial pressure of the refrigerant, not the measured pressure in the condenser. If you then pass that liquid through a high pressure receiver your receiver pressure will be considerably lower than condensing....Which can really mess up some styles of head pressure controls.

So you don't have 17-deg of subcooling. The nitrogen is occupying some of your condenser volume and fouling portion of its heat transfer surface. So you have your original refrig charge but effectively a smaller condenser...You would expect less subcooling.

sterl
05-09-2014, 10:19 PM
Oh: And its Dalton's law of partial pressures.

Saturatedpsi
07-09-2014, 12:12 AM
Oh: And its Dalton's law of partial pressures.

I understand the implications of Dalton's Law. The additive pressure of the nitrogen creates an apparent (though false) increase in SCT for the given system refrigerant, resulting in an increased "calculated" subcooling value.

But my question was with the "bubbles"...How/why they form, and what are they. :confused:

Another "case study" I forgot to mention, from another forum:

407C chiller, with high head, high subcooling, high superheat, bubbles in the sight glass. The tech had an identical system to compare it with, and early on, suspected the ill-running system had suffered some NCG intake, not finding any other reasons to cause the symptoms.

However, he didn't have the option of simply following his instincts, and decided to get some lab results before starting an expensive project. He took a refrigerant sample from the liquid line of the equipment. The vapor phase of the sample was analyzed, and the results confirmed a high (near 20%) NCG content.

So, he started with an empty (well vacuumed) sample container, filled 60-80% of the container capacity with liquid, and ended up with significant NCG's in the sample.

That seems to confirm the NCG's are in the liquid phase of the refrigerant...well, at least in the liquid "side" of the system. And if the sight glass shows bubbles, my simpleton thought process puts 2 and 2 together, concluding the bubbles are NCG's.

I'm not taking an argumentative position with any particular point of view... ;) Just trying to learn something. :confused:

nike123
07-09-2014, 03:53 PM
I think, since you are measuring pressure of refrigerant to obtain saturation condensation temperature of refrigerant, that any addition of nitrogen or other non condensible gases makes your calculation of subcooling wrong, since measured pressure does not correspond any more with actual saturation temperature.
Therefore, bubbles in sight glass, when "subcooling" is high "enough" by your calculation, are not nitrogen! They are simple not enough subcooled refrigerant.
Some refrigerants stop bubbling in sight glass at 5K and some, like R407C, at even 9K of actual subcooling.

I think in 101 thread Gary has explained when sight glass is starting to appear full regarding subcooling.
Ad to that some pressure change from different friction and change in shape in sight glass and preceding pipe/drier (or what ever else is before sight glass) and you got your picture.

Regarding experiment of your colleague i tend to suspect in way how sample is obtained. If we talk about non condensible, than they cannot be in liquid phase. That is what their name is for. Therefore, some mixing is happened when sample is taken. Why and how, we will probably never know.

Experiment should be repeatable at controlled conditions in order to became fact.

But, again, I could be totally wrong!:D
I thinking here in statical conditions. What is happening when molecules of gases are mixed by their flow, is mystery to me!

Saturatedpsi
08-09-2014, 03:05 AM
Lest start with the most likely, incorrect measurement of liquid pressure and temp (as much to do with accuracy of measuring device and influence of ambient temp)


I would think the measurements were as accurate as you can get in the field. Digi-Cool AK900, and tubing temps confirmed with a second TC thermometer. ;)



Regarding experiment of your colleague...some mixing is happened when sample is taken.

The colleague's case wasn't an experiment...actual system that got contaminated through sloppy service procedures. And, he was very meticulous with the sampling procedure, making certain the samples were accurately representative of the stuff he was extracting from the liquid line.

:)

nike123
08-09-2014, 04:58 AM
So, it looks like you are suggesting that Nitrogen, which is inert gas, has chemically reacted with refrigerant in system, formed compound and condensed as such to liquid state at system operating temperatures, and, in such liquid state, transferred as sample to bottle for examination, where latter in that bottle it broke its chemical connection and appeared again as separate inert gas mixed (not in chemical connection) with evaporated refrigerant.

Is that what you are saying?
I tend to strongly suspect in that scenario.


I can only believe that there was no liquid column at exit of condenser, since there was not enough actual subcooling at that point of system.
If there is no liquid column at exit of condenser than it is possible for nitrogen gas and evaporated refrigerant mixture to occur at any point of system.

To measure your actual saturation temperature, you need portion of condenser with no change in temperature. And to confirm that you actually have subcooling, your condenser need to have at its exit temperature at least 5K or more lower than that zone of no temperature change.

With AK900, you can check your condenser temperatures at U turns to confirm that. Or you can use thermal vision camera for that purpose.

Saturatedpsi
08-09-2014, 06:04 AM
So, it looks like you are suggesting that Nitrogen, which is inert gas, has chemically reacted with refrigerant in system, formed compound and condensed as such to liquid state at system operating temperatures, and, in such liquid state, transferred as sample to bottle for examination, where latter in that bottle it broke its chemical connection and appeared again as separate inert gas mixed (not in chemical connection) with evaporated refrigerant.

Is that what you are saying?
I tend to strongly suspect in that scenario.


I can only believe that there was no liquid column at exit of condenser, since there was not enough actual subcooling at that point of system.
If there is no liquid column at exit of condenser than it is possible for nitrogen gas and evaporated refrigerant mixture to occur at any point of system.


I'm not suggesting anything...just presenting data. The OP was a question relative to the implications of the data, and the bubbles.

In the experiment, there was an initial subcooling of 11˚F, which would seem sufficient to provide a liquid seal, to keep the added nitrogen in the condenser section.

If you're suggesting the presence of any amount of NCG's can totally disrupt the refrigerant condensing conditions to the point of zero subcooling, fine. I can neither confirm nor disprove your theory.

And if it's the answer to the question, can you provide some data to support it?

nike123
08-09-2014, 08:02 AM
I only say that 11F subcooling you calculated is not actually 11°F real subcoolin if any amount of nitrogen is added to refrigerant.

Also, you have some pressure drop from friction from condenser to sight glass. And, you probably have some cross-section change from liquid pipe to filter to sight glass where pressure changes again.
All that summed, and you ended up with bubbles of evaporated refrigerant in sight glass, if you have some liquid column or, maybe it is nitrogen-refrigerant mixture if you don't have solid liquid column at exit of condenser.

Try to read 101 thread and when we can expect there is no bubbles at sight glass regarding ACTUAL subcooling.

Saturatedpsi
08-09-2014, 01:08 PM
OK...I appreciate your comments and participation in the discussion. :)

sterl
08-09-2014, 06:20 PM
Sat, this topic had more interest than I would have anticipated...And I decided maybe it wasn't closed?

So I did a little Goggle drilling.

http://www.nist.gov/data/PDFfiles/jpcrd249.pdf

In the chemical sense at least: Nitrogen is soluble in liquid R-22. See the attached, fluid 4.12e P 590 (29 of 38 pages in the PDF....) is R-22 and some of the surrounding fluids are refrigerants as well....So the point about Henry's law is well taken. I ran the equations through at the types of pressures and temperatures you posted and calculated molal concentrations at 5.13 * 10^-5. Or a Mass Concentration of 8.31 * 10^-6. Those numbers don't seem very impressive, but:

If you take that solution and reduce its pressure to 73 PSIG and 42 Deg in the evaporator: The Liquid Fraction will lose concentration of the dissolved N2 and thus your gas in the evaporator will have a partial pressure of R-22. To the tune of about 3.6 Psi partial pressure.....

So as the liquid solution moved from a zone with a total pressure of 225 Psig of which approximately 213 was the saturation pressure of the refrigerant, to a "nominally sat" vessel: the N2 in liquid solution would want to perk out. In a vessel, without the influences of pipe dynamics and the flowing liquid stream, the math would indicate that the new pressure would be 0.66 psi higher than saturation. In a pipe, if the gas had the opportunity to devolve, the same change in pressure would generate gas as N2 at nominally one-ten thousandth the volume of the passing liquid. A 10-Psi pressure change makes that about 2-ten thousandths.

Can you see that in a sight glass? That depends on orientation and actual velocities and the change in pressure between the sight glass and the liquid header on the condenser and the effect of an intercepting vessel and the lift in the liquid line and a bunch of other things....But if there was some gas devolving at the liquid filter-drier, eventually it would want to burp through.

I would say the influence on the low side pressure (long time later...) is more significant, and would be of greater concern for a circuit operating at a low suction pressure than it is on an AC circuit. Goes to prove: Multiple Vacs is a good concept.

Saturatedpsi
09-09-2014, 12:47 AM
Thanks for your invested time and effort... :)

Saturatedpsi
16-09-2014, 03:48 PM
Let me add another wrinkle...:D

I just discovered gas and vapor aren't the same thing, as I had thought for 30 years (learn something every day).

Gas vs Vapor (http://en.wikipedia.org/wiki/Vapor)

If I'm reading the info right, the vapor phase of the ***** we're calling a "gas" actually isn't, in which case, the Gas Laws won't necessarily apply...

If they don't apply, how would one then, predict the results of the addition of nitrogen or any NC, to the system? :confused:

sterl
18-09-2014, 05:24 PM
Origin of ideal gas laws were supercritical gases: Nitrogen, oxygen, methane, carbon dioxide that are always in the gaseous state at normal temperatures and pressures. The definition for vapor is the gaseous state of a fluid that can co-exist at normal temperatures and pressures with its liquid or solid state. A vapor will act like a gas if the "second state" is not present or otherwise not being approached. So for the 18th century scientists chasing these topics: Nitrogen was a gas. Liquifying it was a long ways off. But in a liquid nitrogen cylinder, gaseous nitrogen is present in saturation with the liquid...

And for the simplistic compounds considered "Vapors" the gaseous state, if it was removed from the solid and liquid, acted just like a gas once they were superheated past the prevailing saturation temperature. The extent of departure from ideal gas behavior for the fluids typically of our interest can be characterized on a PH chart. The change in slope of the isentropic lines as you approach saturation on the superheat side of the vapor dome reflects the departure from the Ideal Gas Laws. So if those isentropic lines are appreciably curved, some error is going to be introduced if one were to try to base property changes on the ideal gas laws....As example of where this leads: with a mix of fluids in a refrigerant, how would a compressor manufacturer predict discharge temperature and establish the extent of cooling required?
An industrial circuit tends to operate at very low superheats (say 3 to 4 degrees F.) so to those compressor manufacturer that departure of some significance. Much more involved for fluids that are in both conditions, supercritical gas and vapor in different portions of the same circuit. Transcritical CO2 circuits are the ready example and should one of those be operated at temperatures on the low side near the triple point: the departures from Ideal get very interesting and highly significant with all 3-phases and the supercritical gas in the same pressure envelope.
Note that the article link above deals with specifically nitrogen solubility in "vapor" liquids at a range of temperatures and total pressures. So the total pressure on that liquid would be the partial pressure of the volatile plus the PP of the N2....And the tiny dissolved fraction of N2 would want to come out of solution as the total pressure decreased. If there were no pressure differences prior to the expansion point in the fridge circuit, the released proportion of the N2 would blow through the evaporator and get slightly warmer but contribute almost nothing to the refrig effect....Your measured suction pressure would then be higher than the saturation corresponding to the refrigerant temperature. And should this be a TXV circuit, your superheat would be goofy and the evap a little starved.

Saturatedpsi
18-09-2014, 07:16 PM
Origin of ideal gas laws were supercritical gases......should this be a TXV circuit, your superheat would be goofy and the evap a little starved.

Again, I appreciate your input. :) ..and it's a lot of physics/chemistry for me to try to comprehend. :confused::D

I'm still fixated on the "bubbles" as a possible (and predictable) symptom of NC contamination, solely as an assist in diagnostics. I recently "found" my old TIF "electronic sight glass", that I no doubt wrote off years ago as less than useful, for whatever purpose I acquired it in the first place.

I've played around with it some, but not yet decided if it will consistently "hear" the bubbles...but I'm hoping for that result.:D

NewmanRef
18-09-2014, 08:17 PM
I reckon it's just short of gas guys! :p