Mitsubishi Gas leak? PURY P250 YGMA

[marc5180]

I got a call 2 weeks ago to a site where one of their VRF's wasn't working. I found the system short of refrigerant got a total of 11kg out instead of the 28.5kg there should have been in.
I put the system under pressure and found a leak on the outdoor condensor suction shutoff valve.
I checked all the indoor flare nuts and the BC flare nuts and couldn't find another leak.
I told the guy that i had found a leak and that i would like to leave it under pressure over the weekend (or that night) to be sure there were no more leaks.
He insisted that we run the system up rather than leave it under pressure again as it fed a medical store, i did believe that was the leak sorted as the pressure hadn't dropped since we sorted the leak but i decided to add some uv leak dye incase of any furthur leaks in the future.I then vacced out and recharged with 28.5 kg of R10a.

However i got a call yesterday saying it has gone off again. I went down and took out 4kg of r410a from the system meaning that it had lost 24.5 kg in 13 days, thats nearly 2kg leakage per day!!

It must have a major leak somewhere, however i pressure tested and leak checked all indoor units again and bc box and checked it for any UV dye leakage but nothing. Everything was dry and no signs of a leak. I stripped the condensor down but again couldn't find any leaks.
I pressure tested it at 9am this morning and i left at 4pm and it hadn't lost anything, the same as when i carried it out last time. I've left it on over the weekend but i am going to call in tomorrow just to check if it has lost anything but i don't think it will have.

It's as though it only leaks when the system is running, has anyone ever come across something like this because im lost

[eggs]

I've had it in the past.
Try testing at a lower pressure. It turned out that the nitrogen test pressure had closed up the leak on the pipe braze and the leak only showed up at about 10 bar.

It's worth a try.

eggs

[marc5180]

I tried it at 15 bar and couldn't find a leak then raised it to 30 bar.

It lost 24.5 kg + 17.5kg which is over 42kg, but there is no sign of a leak, it's as though there is refrigerant trapped in the pipework surely if it was a leak there would be some oil around or uv dye.

I turned power off to all of the indoor units to be sure the LEV's were open and weren't trapping any refrigerant in any of the evaporators. Not too sure about the solonoids in the BC box but i think these are open.

I'm expecting to see a big drop in pressure when i go back but if not then it must be leaking when the system runs up but how?

[Brian_UK]

Any underground or cavity pipe runs ?

[puddleboy3]

Drop the fan impellers out of the indoors and check for oil in the drip tray. I had same type of problem on a R407c multi system and found that one of the indoor unit heat exchangers was leaking at the header.

[marc5180]

There's no access to a lot of the pipework because it runs across ceilings and through rooms where they have been plasterboarded and painted.

None of the indoors are the 4 way blow cassette type they are either ducted (again with no access due to them being above a ceiling) or they are the single blow underceiling units.

I went back to the site at lunch time and found the pressures hadn't dropped at all but i don't see how that can happen when it is losing nearly 2kg of refrigerant a day when it is charged.

[raz5]

There's no access to a lot of the pipework because it runs across ceilings and through rooms where they have been plasterboarded and painted.

None of the indoors are the 4 way blow cassette type they are either ducted (again with no access due to them being above a ceiling) or they are the single blow underceiling units.

I went back to the site at lunch time and found the pressures hadn't dropped at all but i don't see how that can happen when it is losing nearly 2kg of refrigerant a day when it is charged.

Have you thought of using balloon gas for testing , it may be a molecule size problem and the ofn is too big, I hear what your saying about 2kg per day and i may be off mark but i have found 410a to be prone to this type of leak problem

[marc5180]

Have you thought of using balloon gas for testing , it may be a molecule size problem and the ofn is too big, I hear what your saying about 2kg per day and i may be off mark but i have found 410a to be prone to this type of leak problem

Yes i think if i go back on monday and still no leak then i will try helium. I have never had to use it before as i have always found the leak with OFN.

[Thermatech]

Try seperating the system up into sections.

If you have access to the BC box then you can pressure test each set of pipes to each indoor unit seperately & also test the BC box seperately & gas/liquid pipe to the outdoor unit seperately & the outdoor unit seperately.

I know its alot of work to set this test up but 9/10 it will prove which pipe has the leak & all you need is good access to the BC.

[marc5180]

That's a good idea, the BC box is probably the only bit that we do have full access to all of the flare nuts and pipework.
I'm just worried that this is going to be a very time consuming job, the customer wants it on asap but he will just have to wait.

If it hasn't lost anything on monday then i will have to try the helium but like i said above, i've never come across a leak that i havn't found with OFN.

[Brian_UK]

As you have ducted units try using a leak detector in the downstream side of the units.

Run them, put the detector in and then turn them off; this reduces the airflow but allows any refrigerant to concentrate.

[AbsoluteWDJ]

Try seperating the system up into sections.

If you have access to the BC box then you can pressure test each set of pipes to each indoor unit seperately & also test the BC box seperately & gas/liquid pipe to the outdoor unit seperately & the outdoor unit seperately.

I know its alot of work to set this test up but 9/10 it will prove which pipe has the leak & all you need is good access to the BC.

I agree with this marc. On the majority of pressure tests i carry out i'll shut service valves so that outdoor and interconnecting/indoor pipes are divided. Pressure test both sides at equal pressure. Leave it for a weekend and hopefully it narrow down where to start leak checking. It sounds that your situation could take a bit of time. Hope your customer understands this!! Oh yeah be aware of tiny cracks in flare nuts. Good luck

[Brian_UK]

I'm just worried that this is going to be a very time consuming job, the customer wants it on asap but he will just have to wait.

They always want it yesterday don't they

[andy c]

I have been to two of these units very recently, which have held a pressure test, but lose refrigerant over a period of about a month. Both leaks were found on the internal pipework of the bc box, at the point where the pipe splits from one to two. Its very easy to re-braze these, once the bottom cover, electric box and drip pan are removed. I found them by pulling a wire of the compressor and switching the power back on, to energise the solenoids in the bc box. Hope this is of use. Regards, Andy C

[airconadam]

Is their no way to open all the valves on the system from the outdoor board? I have had a similar problem with daikin and you can open all the valves in bx boxes and the outdoor unit to ensure all the pipework is open,this also ensures a full recovery and a correct vac.

adam

we also spoke to mitsi and they said power down everything and that will open all the valves???????

[marc5180]

The power to the condensor has been turned off as well as power to all indoor units on this system. I haven't tried turning power on and disconnecting the compressor yet.

Went back to site today and the needle hadn't moved at all indicating no leak...
I went out and picked up some 99% helium and cut the pipework coming out of the condensor and braized in schraeder stubs so both the condensor and pipework could be tested independantly.
Pressure tested both to 25bar with the helium and waited for 2 hours with no drop again.

Theres no signs of oil anywhere or uv dye, it's as if there is no leak. I have never come across a leak like this before.

All i can think now is that it only happens when the compressor starts and begins to vibrate something in the condensor, hopefully when i go back tomorrow the condensor will have lost some pressure or im off on long term sick

[Brian_UK]

I know that you said you can't see any dye but is it possible that the leak is from the access port fittings; they're tight when your manifold lines are connected but leak with a cap on...

[marc5180]

No we have checked and double checked them, 100% sure it isn't them.

[andy c]

When i carried out the pressure test to both systems, (0n two different sites) they held a nitrogen pressure test of 24 Bar for the whole weekend. This had me confused at first, because the systems lost 18Kg of refrigerant in a month or so. They had both also held under a vacuum, which also suggested no leaks present. The leak wasn't identified on both occasions, under normal procedures. But after disconnecting the CN Plug from the board for the LP device, to stop the compressor from being able to run, I powered up the system whilst under nitrogen pressure, and within a few minuites, the leak could be heard within the bc box. I don't know whether the solenoids are energised with power or de-energised, but on both sites it worked for me. The second site i attended, was after two of my colleages had also carried out the pressure test, and been unable to find a leak. A possible tell tale sign is to have a feel in the bottom of the drip tray of the bc box for oil. Regards, Andy C

[marc5180]

But after disconnecting the CN Plug from the board for the LP device, to stop the compressor from being able to run,

I'll try that tomorrow but i did remove the bottom cover of the BC box and i could see through to the solonoids but couldn't see any oil or dye.

It's something else for me to try though so thanks.

[laf100]

I too recommend dropping the bottom of the BC box to check for leaks more closely, I have had leaks here twice. I also swear by my Fluke digital pressure gauge, it will show you to .1 psi resolution, and more often than not you can see it dropping. Neither of these explain your 2kg per day loss though!!

If you dont have full access to all your fan coils or pipework, you could be waisting your time!

[nike123]

Check this manual:
http://www.refrig.com/LeakDetectionManual.pdf

[marc5180]

Thankd for that Nike

[Goober]

I have to agree with laf100, if you can't get to all the pipework you may be wasting your time. Also with all the pipework insulated the leak dye may not be visable. Tried cutting the insulation in random places to see if any oil/dye present?

[marc5180]

Went back to the unit today and found the needle on my gauges hadn't moved indicating no leak.

I cut open the ceiling to get full access to the bc box and i removed the electric box and tray so i could see inside.

I turned power on to the condensor and ran it up with the compressor disconnected. No signs of leaks at all in the Bc box or anywhere on the condensor.

[laf100]

I do feel your pain marc5180 !!!

As you have the "luxury" of good access to the BC, my next step would be to dis' all the branch pipes, and cap off the pipe connections of the BC using flare nuts & bonnets. You have then approximatley halved your system, and can p/test the CU, main pipework & BC box, without actually having to cut/re-braize any pipework.

If no joy, then test each branch pair & AHU individually from the BC access point. I have only had to do this once...and it ain't fun!

[airconadam]

I still cant see us finding a leak tho after 4 days of the system under pressure???? with helium and nitrogen! its as if the system only leaks when the unit is ran up?? maybe its a case of running the system up and using the electronic leak detector??

adam

[marc5180]

I do feel your pain marc5180 !!!

As you have the "luxury" of good access to the BC, my next step would be to dis' all the branch pipes, and cap off the pipe connections of the BC using flare nuts & bonnets. You have then approximatley halved your system, and can p/test the CU, main pipework & BC box, without actually having to cut/re-braize any pipework.

If no joy, then test each branch pair & AHU individually from the BC access point. I have only had to do this once...and it ain't fun!

If we had a drop in pressure then this is what i would have done but we have had no pressure drop in over 5 days.

[Goober]

Breath, and breath, and breath. Patience is going to win this one. you'll get there.

[AbsoluteWDJ]

Just out of curiosity. How old is the system? and has it had any previous faults of this kind. Might be clutching at straws but has there been any works carried out in or around where the pipework runs. I have had a couple of occasions where a builder has put a screw straight through refrigerant pipe!

[AbsoluteWDJ]

Just out of curiosity. How old is the system? and has it had any previous faults of this kind. Might be clutching at straws but has there been any works carried out in or around where the pipework runs. I have had a couple of occasions where a builder has put a screw straight through refrigerant pipe!

In saying that you would still expect to see a pressure drop....Stick with it my friend....Is the whole system under pressure?

[sinewave]

Does the system leak in cooling or heating?

Sometimes leaks only happen on a specific cycle.

[marc5180]

Update; got back to the system today and the pressure still hadn't dropped at all.
I decided to vac out and recharge with the intention of running the system up and checking for the leak with the sniffer.
Put system on vac and got it down to 1000 microns (1Torr) and turned the vac pump off and waited for it to rise but it didnt at all.

I recharged and ran it up. I put all 7 indoors on cooling and checked the condensor with the sniffer, then went to the BC box, then all indoor units.
All of them were cooling and there were no signs of a leak again, so i put them all into heating.

Again checked the condensor, the BC box and all indoors. If there is a leak then it has me beat because i have spent 5 solid days trying to find it.

I cant see it being on pipework because helium would find the leak over a weekend so it must be on something that is expanding and contracting like some sort of a seal but i can't think where.

I'm going back tomorrow to check it again and see how it is running, i'm even dreaming about it at night. I woke up in the middle of the night last night saying "i've found it" my girlfriend told me.

[marc5180]

AbsoluteWDJ,
the system is about 2 years old and no building work has been going on that i know of.

Sinewave,
I've tried on both heating and cooling mode now and still no signs of a leak.

[El Padre]

I have had this before and put it down to the BC box, changed the box and solved the problem, I think that a joint was leaking only when all the units are in heating mode, as you can exclude the condenser and pipework from being at fault that is the most likely culprit!

[AbsoluteWDJ]

I woke up in the middle of the night last night saying "i've found it" my girlfriend told me.[/quote]

What exactly did you find!!!
My fingers are crossed for you mate.

[marc5180]

I have had this before and put it down to the BC box, changed the box and solved the problem, I think that a joint was leaking only when all the units are in heating mode, as you can exclude the condenser and pipework from being at fault that is the most likely culprit!

I haven't found any leak on the BC box or anything to indicate that there is a problem with it, so i don't want to change it until i do. I'm going to go back tomorrow and put all indoor units on heat and stay with it for a few hours and monitor closely.

[marc5180]

I woke up in the middle of the night last night saying "i've found it" my girlfriend told me.

What exactly did you find!!!
My fingers are crossed for you mate.[/QUOTE]

Haha, i'd been dreaming that i had found a pin hole in the pipework, then woke up and realised i had been dreaming

[rude]

i hate waking up in the middle of the night thinking about work.

[nike123]

This is one of perfect examples why I hate even think about direct expansion systems in place of good old hydronic systems.
Another one is their unnecessary complexity.
Don't forget, everything leaks! So many joints, to many leak potential.

[El Padre]

Update; got back to the system today and the pressure still hadn't dropped at all.
I decided to vac out and recharge with the intention of running the system up and checking for the leak with the sniffer.
Put system on vac and got it down to 1000 microns (1Torr) and turned the vac pump off and waited for it to rise but it didnt at all.

I recharged and ran it up. I put all 7 indoors on cooling and checked the condensor with the sniffer, then went to the BC box, then all indoor units.
All of them were cooling and there were no signs of a leak again, so i put them all into heating.

Again checked the condensor, the BC box and all indoors. If there is a leak then it has me beat because i have spent 5 solid days trying to find it.

I cant see it being on pipework because helium would find the leak over a weekend so it must be on something that is expanding and contracting like some sort of a seal but i can't think where.

I'm going back tomorrow to check it again and see how it is running, i'm even dreaming about it at night. I woke up in the middle of the night last night saying "i've found it" my girlfriend told me.

At least she did'nt wake you up saying that she's found it!!

[dannycool]

A long shot, but another thing to consider is the presence of other engineers, i have experienced other engineers obviously being out of gas and stealing it from other systems, and a couple of times they have recovered from our equipment by mistake.
If it is a shopping centre or multi-occupancy office check the vistors log book around the time of your call outs

[airconadam]

I called into the site today tested every unit on heat and cool and all worked fine with no problems touch wood we wont get called back but surely its not fixed itself???

adam

[marc5180]

A long shot, but another thing to consider is the presence of other engineers, i have experienced other engineers obviously being out of gas and stealing it from other systems, and a couple of times they have recovered from our equipment by mistake.
If it is a shopping centre or multi-occupancy office check the vistors log book around the time of your call outs

We had thought that someone could be tampering with it so we had marked the caps on the condensor.

It took 2 weeks before we were called back to it and that time is up next thursday. I have my fingers crossed and everything else for that matter
Only time will tell now.

[Suffolk Fridge]

a mate of mine has just found a leak on a pury-p250ymf-c system that took him nearly a week to find. It was a hairline crack on a refnet in the bc box. Refnet only leaked on running, due to expansion/contraction. He only found it out of sheer desperation by dropping the bc box out of the ceiling (no mean feat with this particular box) capping all the connections and pressure testing it on the floor at 38 bar. even then it only showed up a tiny amount with leak spray.
In my experience i have found that putting dye in a city multi system is pretty ineffective as the oil separator is so good, it doesnt even leave the condensing unit. Fingers crossed for you mate i have been there many times.

[R1976]

This thread is great! It's like a gripping series for RAC engineers and I can't wait for the outcome!

Sometimes these series do go on though, I got bored of Lost on series 87.

Seriously though, Leaks like this are a nightmare and I wish you luck. I have had cracked flare nuts that only open on heat mode.

[jteixiera]

Just a quick question. are the indoor units powered? Hopefully yes as The indooor units have electronic LEVs

[marc5180]

Mitsubishi told us that the indoors had to be powered down, then we rang back up and the guy told us to leave them switched on which we did. Everthing had power apart from the condensor.

[airconadam]

all the indoor units were powered up also spoke with mitsi and they said powerdown to open the LEV??? we tried both!! no difference!

adam

[jteixiera]

I think you must cut the system into sections(i.e.) pressure test the condenser on its own, isolate each fan coil unit from the BC box and pressure test each unit with the associated flow and return pipe work and pressure test the bc box on its own. I have found a few leaks in the bc box solenoid valve heads. once you take the pc board off the bc box you can get to all the solenoid valves and really appreciate what the bc box does.