View Full Version : Oil separator

05-05-2004, 09:41 AM
There was some time ago a thread about oil separators on packs which I can't find back with the search engine.

There seems to be problems with packs, working of R404a at low condensing pressure where the separator can't separate the oil.
It's just like if you lower too much the discharge gas temperatures,that the separator isn't working properly anymore.

A colleague of my had enormous troubles with it. As soon they increased HP, problem was gone. And as soon the decreased it again, problem came back.

Ever heard of it?
What is the technical explanation for this?
What can be done to solve the problem so that can be worked with low high pressures without the oil problems?

05-05-2004, 02:25 PM
hi peter,

at what pressure does the problem start?


06-05-2004, 05:24 PM
Hi peter :)

Yes ive had this on a few supermarket packs :)

Best regards Mark

06-05-2004, 06:14 PM
hi mark

and....... whats your verdict? :confused:


06-05-2004, 07:53 PM
Hi Peter,
Like Mark, i have had it on some supermarket packs, Normally it has been down to the condenser fans coming in too early. Adjusting the cut in pressures has cured it.

Hope this helps



06-05-2004, 08:14 PM
hi guys,

found something that explains why does it happend,

at the bottom of page 2.


06-05-2004, 08:29 PM
Thanks Chemi,
Looks like an interesting link, it's printing off as i type.
I have had to strip qiute a few of these down recently, all on "NEW" R404 packs. It is surprising how little amount of dirt in the bottom of it it takes to block them up. The ones i have have done are the type with 8 bolts holtding the bottom onto it, Its a bit of a messy job but since done no more probs.



06-05-2004, 08:57 PM
Hi :)
not seen this problem to any great degree, but I am told it is due to low velocity, basically the gas is going to slow to allow the oil to be knocked out of it by impingement. Also to do with the low discharge temperatures associated with 404a, raise the head and the oil is warmer and less viscous. Raising the head will reduce the amount a compressor will pump, so the velocity will go down in the separator, this would seem to be the reverse of what we need, so I would say the oil temperature is the important issue here.
Kind Regards. Andy. :)

07-05-2004, 06:18 PM
Hi everyone :)
Maintaining high discharge pressures and temperatures is common on "supermarket packs".
The increased discharge set point keeps a constant duty on the pack improving oil momentum or velocity in normal operation and times of low thermal load.
When you talk of lowering discharge/ condensing pressures to "save energy" oil return is compromised.
The raise in discharge pressure , equally raises the suction pressure in turn running more compressors in order to maintain suction set point.This can help momentum of oil in pipework,traps,cases condensers return to its origin,as not all oil which leaves the compressor is returned via the seperator.
A common problem associated with oil seperators on packs can be down to poor maintenance procedures and contractors after a "quick fix".
Set points are kept reasonably high on the smaller/moderate packs much to the" energy saving" experts dismay.
This situation is not true in every pack installation as few stores are uniform .
Best regards Mark :)

07-05-2004, 08:16 PM
I don't work on supermarket packs but this does ring a bell. With the VRV air conditioners that operate with inverter compressors this was a problem that was identified with low load conditions. To overcome the oil return problems associated with low load/low velocities the systems have an inbuilt oil recovery programme. For every 8 hours of running the system will switch off the indoor fans, fully open the EEV's and run the compressors at full capacity for a short period. When the oil is recovered then the system reverts back to where it was.

Could this type of logic control be incorporated into Packs? Maybe not with TEV's?

08-05-2004, 07:03 AM
Hi Frank,

Just a quick reply. Has for the moment almost no time anymore to sleep.

Indeed, they do this (Mitsubishi VRF, DaikinVRV, Hitachi SetFree) - I thought that the PCB counts the running hours from the indoor units or so, not important - to have an oil flow back at high speed through the lines. AC suction lines can't mostly not be placed inclined towards the compressors or the techs aren't aware of the need for that. That's also the reason why they maintain and recommend (very) high speeds in the suction lines: to have a proper oil floodback.


I know it's common in supermarkets. But if your raiser line to the main suction line is sized for the correct speed, then this line will always have the correct speed (capacity in a cabinet stays +/- the same) In the main suction line, there yo have variations in speed due to the fact that some cabinets will shutt off if they reached preset temperature.

If this main suction line is properly installed, inclined towards the compressors, then it should come back, even at lower speeds.

But what is the reason that it can pass trough the oilseparator?

We had once a Profroid pack, 5 Bitzer compressors, each has his own separator (coalescent), calculated for his own max capacity and a main oilseparator for the total capacity (5 compressors) but problems remained the same.

Normally, they say that a coalescent separates 99.9%. :confused:


I thought the slower it goes, the better it separates. I thought they need to reduce the speed in the separator.
But on the other hand, the link Chemi provided states that mass flow will increase with lowering the head pressures. :confused: :confused:

08-05-2004, 08:05 AM
hi peter,

you got it wrong, you must be tired.

I will qoute for you: "if velocity is too low there will be in sufficient forces to cause the smaller droplets to collide and drop to the sump"

which means that to have a good oil separation, you need high head pressure.

chemi :)

08-05-2004, 12:38 PM
Hi Frank,


I thought the slower it goes, the better it separates. I thought they need to reduce the speed in the separator.
But on the other hand, the link Chemi provided states that mass flow will increase with lowering the head pressures. :confused: :confused:

Two differing points. Low velocity will result in drops of oil just passing through the oil sep, a lazy flow. Very high velocity and the oil sep will not have the ability to slow the gas down enough for thr oil to be knocked out.
As an exmple say their was only one compressor, fixed displacement. Raising the head will result in reduced mass flow and reduced velocities, your compressor would simply not push as much refrigerant around the system. But in a mutiple compressor pack reducing the head causes the suction to drop and compressors to off cycle, reducing the velocity in the pipe work due to reduction in capacity (number of compressors running) so high heads are employed to ensure oil return, when in fact all that is required is more load, but due to the low ambient the load is not available.
Hope this helps.
Kind Regards. Andy. :)

08-05-2004, 06:28 PM
Yes Andy ;) :D

Dave Goodings
08-05-2004, 07:53 PM
Hello Frank Regarding VRV systems that force the eev open to assist in oil return do you ever have any problems with this on systems, :confused: because on a recent thread where Marc was having pipework installed in a riser by abseilers I mentioned about oil return ie traps where normally on a larger commercial/industrial sytem you would have a double riser for low load conditions to return oil, this is an interesting if not dangerous solution to this problem!!!

09-05-2004, 07:23 PM
Hi Dave

Daikin insist that oil traps are not installed on VRV systems. On a recent training course we were shown examples (photo's) where just by installing pipework that drops down to pass under a structural beam and then rises back up again is classed as an oil trap and invalidates the warranty. Pipework on a VRV (&VRF) system must be installed with long slow bends, Refnet joints on the horizontal (within 30deg) as gravity will effect liquid flow and always without oil traps. It is vitally important to nitrogen purge due to the large amount of brazed joints. As with other manufacturers, Daikin are getting quite hot on Warranty claims and investigate any compressor failure to determine the cause - not only by looking at the compressor but by investigating the whole system to ensure that it is not an installation defect that caused the failure.

Not many engineers look further than the end of the job and what they call "commissioning"

Just as an aside - how many on the forum bother to employ the manufacturers or distributors to come out and commission? I expect it is a minority :eek:

09-05-2004, 07:35 PM
Just as an aside - how many on the forum bother to employ the manufacturers or distributors to come out and commission? I expect it is a minority :eek:

hi frank,

in most cases, the tech in the field nows more then the distributor!

chemi :)

09-05-2004, 09:40 PM
Hi Dave

Pipework on a VRV (&VRF) system must be installed with long slow bends, Refnet joints on the horizontal (within 30deg) as gravity will effect liquid flow and always without oil traps.

We never install the Refnet joints in VRF's (Mitsubishi), always simple T' s and that's because 1. The Belgium importer gave us permission for this after many field tests 2. they are real techs and understand that a VRF is not more then some sort of a pack with many evaporators connected to it.

We did at least 15 VRf's like this.

I'm sure if you follow the same rules as with a central compressor like in a supermarket, you will never encounter problems.

But's it's because a lot of the Daikin installers - at least in Belgium - are central heating techs who don't know nothing of refrigeration and the special piping for it but followed a one day steam course 'How to install a VRV'.
So they have to be very carefully with them.

I saw Daikin installations VRV, where each bend was soldered with a solder turn and soldered with tin, and in mm copper like they use for drinking lines.

And to be honest, the Refent joints are not more then expensive streamlined T's or better, Y's.

We even did it at least 5 times wiith VRV's without any problem.

If they - Daikin - should say that an omega-turn aroudn a beam is an oiltrap, then they should explain me why and how this could be a danger for the compressor and what you should do then in such a cases.

I agree like Chemi that the techs in the field know mostly more then the so called experts.

10-05-2004, 09:03 PM
I'm sure if you follow the same rules as with a central compressor like in a supermarket, you will never encounter problems.

Hi Peter

Like my previous post about automatic oil return cycles - maybe one of the reasons that supermarket packs do not employ this kind of device is because of the restictions in supermarket designs. EPR's come to mind as well as other things.

It has been well documented by Daikin and others the many reasons why compressors fail. We at the moment are investigating an inverter compressor failure on a VRV system installed by others (16HP) - we have been invited to replace the compressor but as the system is only 18 months old we are not happy just replacing the pot - we need to know why it has failed as we don't want to be asked the same question in a few months time.

Any restriction in a system that uses automatic oil return cycles will effect oil return ;)

10-05-2004, 09:16 PM
Hi frank :)
Yes,These automatic oil return cycles on a modern supermarket pack installation would not be practical :(
Too many variables to consider .
Good theory though for A/C.
Best regards
Mark :)

10-05-2004, 09:33 PM
Can you post what you have found on that VRV?
Because these are situations you can learn from a lot.

Is the compressor damaged electrically or mechanically?

We always open the compressors (posted some pictures on this forum) to see what happened inside it. Otherwise it's difficult to predict possible causes.
We measure also the oil quantity.

But in your case - I think - compressor is still under warranty , so you probably can't open it.

What we often did in supermarkets was opening all the SV for some times (+/- 15 minutes) so that we had few times a day max velocity in the lines.
During night, we also closed them all, pumped down the whole system and then controlled after a delaytime the ***** level in the receiver.

11-05-2004, 09:45 PM
Hi Peter

We did not install this system but as it is now 18 months old the customer has asked us to take it on board our service and maintenance contract with him. The initial call out was due to high temperatures within some of the rooms. From the smart controller we could see that some rooms had set points of 22C but temperatures of 27/28C. Looking at the outdoor unit we could hear the contactors and sol valves clicking but no inverter compressor running. after a short period, when the system checked itself for temperature differnces and found none it applied the second on/off compressor. This led us to investigate the inverter compressor. Although the contactor was pulled in the pot was not running. We found the windings open circuit. From the fault code log on the controller it was apparent that at some stage the system had gone out on LP indicating a loss of gas. If you install these things correctly you never get a gas loss. Things are not quite adding up now and it is not just a case of replacing the pot. We need to investigate the pipework layout carefully and see if it falls within the design parameters. We need to recover the gas charge and check the quantity to compare with the correct amount of charge. We need to do an oil acid test to see if there have been comtaminants. We need to see if the oil is clear to see if the pipework was nitrogen purged during installation. After all this can we then decide what, if not a manufacturing fault, was the most probable cause for the failure and if the other compressors are at risk.

After this we have to do a written report for the customer so that, if it is a fault caused by incorrect pipework design or a dirty installation or something else that is not of the manufacturers making, the customer is willing to go after the company who installed the system for them to make the corrections at their cost. Daikin are willing to provide the compressor to us under warranty as we are Daikin dealers and trusted in our diagnosis but we do not want to get involved with bickering and in-fighting with the installers - we will just provide impartial advice to our customer. Obviously, we cannot accept the system on maintenance contract in it's present condition.

If the weather is good on Thursday this week we are due to de-gass the system so I will take my camera and record any findings that give rise to the fault and I will post them for you. In the past though I've had trouble posting photo's due to the restriction of the file size that can be attached to a post - am I missing something here or is there another way to post pictures with a larger file size?

12-05-2004, 06:37 AM
These are sometimes the worst or most dangerous jobs you can do: repair something someone else installed.
But it can also be a job where you easily can prove that you're much smarter then the one who initially installed the system. Good for the ego.

You know probably I'm a judicial expert in Belgium - don't know if this sort of job exist in the UK - so perhaps a few hints (maybe you knew this already but many makes a lot of mistakes against some basic rules here.

Take a lot of pictures - preferable date stamped- before you do anything and make also pictures while you're busy.
Each time you encounter specific problems, make a picture of it.

In Belgium - and I think also in the UK - your client can't use your expertise against your colleague unless he's (colleague) aware of the fact that you will do a repair/modifications and will try to find out what went wrong with is installation.
There is a legal term for this but I didn't found it in a dictionary. Your colleague - or his lawyer - must have the ability to be there when you're working on the installation and be able to deny officially or disprove (refute?) what you're saying or doing.

If you encounter big professional problems, then I should not hesitate to report this to your client. Are the faults more minor or little problems, things that are not quit wrong, more esthetically faults- but that you should have done otherwise, then I should not say this. It's always easy to break down another.

Try to find for each 'fault' a written prove in books or in Daikin manuals, otherwise it will be your word against his.
In a dispute where a judge has to judge a technical case, he's not a techician at all but if you can show him a book or an install note of teh manufacturer, then he would take this for true.

Try not to point personally to the installer. Don't mention his name in your report, never, except when you start your report and by saying 'According to the client or a sticker, the initial installer should be...' .
It's not because your client is saying this or there is a sticker somewhere on a unit that he's the one who initially installed it.
It can have been also a subcontractor.

Just describe the fault (don't get personally) and what in fact must have been done according to the recommendations of Daikin or recommendations of a good practice.

Don't say 'Jim's Airco Servie has soldered the bend the wrong way..'
Better should be 'I found that this specific bend - see picture X - was soldered the wrong way for the reason that can be found in ASHRAE section Y.... or in the Daikin Manual. Or it better had to be sodlerded like .... according to Daikin' Make a referral to Daikin or another wel known reputable name.

End your report with a legal sentence - don't know the correct English one - that you made the report to your best ability, that nothing you have written can be used against yourself, and that you behold all your legal rights despite what you wrote on you report.

12-05-2004, 04:01 PM
hi peter,

very good.

I think if someone is called to fix a problem on any unit,AC or refrigeration, this is an ethic code of behaviour among professional and respectfull people!