Loss of Pressure But No Leak

[FarmerJohn]

Hi,

I have done some copper pipework for the flow and returns runs for my central heating system. The overhead runs are 22mm and I drop-down in 10mm microbore to each of 5 radiators. The 22mm run is about 20m long and the first 10mm connection is about 15m along.

First I air pressure tested the return pipework which held 4 bar for over an hour. Then I air tested the flow pipework which looked like it was losing a small amout of pressure over quite a long time. So I left this on over night and found that the pressure dropped to 3 bar after a 10 hours. Then to try and find the leak I did an 8 bar water pressure test. Water was pumped into the high point of the pipework until there was a flow at the first radiator exit point which I then fitted with a push fit cap. Did the same for all remaining 10mm exit points. Finally pumped the water pressure to 8 bar and left overnight. In a period of about 20 hours the pressure dropped to 7 bar. Looked through all the pipework but could not find and sign of a leaking joint.

As a point of interest as I relieved the water pressure from 7 bar to 6 bar I noted how much water was released which was significant. Since I found no water leaking it must have been air escaping.

Can anyone suggest how I would go about finding this leak either with an air or water test? Is there a way to fill the pipework without it containing any air?

There is nothing in the system except the pipework, that is, it is not connected to the combi or any radiators etc.

Any help would be appreciated.

Regards
FarmerJo

 

[John D v2.0]

Air can leak through a smaller hole than water can.
If there's no (compressible) air in the system you would need to take into account temperature changes too.
Lots of heating systems have really tiny leaks that evaporate as soon as they form

 

[FarmerJohn]

Yes I did wonder about the temperature. In my case (very cold house at the moment), normal daytime temperature is quite low. When I started the tests it was probably about 5C and went below this overnight. Its about 3.2C now. During this cold snap I suppose the temperature could get to about 10C.

I was wondering weather or not to add some perfume into the water and see if I can smell it along the pipe runs. Any thoughts on this?

Regards
FarmerJo

 

[oldbutnotdead]

Pipework only you've not got a huge volume so your leak is going to be tiny.
2 options;
1 air pressure- 4 bar is quite high and dangerous (if a fitting comes loose it'll fire off like a bullet), 2 bar is less hazardous so pump it up to that then go and get some leak detector fluid (don't use fairy liquid and water, it can promote corrosion), start with compression fittings and move on to soldered joints.
2 Fit some rads, bleed it, fire it up, see what happens.

 

[FarmerJohn]

Yes I like the leak detector fluid idea. Going to get some now. Thanks.

 

[FarmerJohn]

Got some spray and applied it to the copper work and so far found no leaks. I have added some additional pipework and one of those runs are also leaking. I did discover that some of the 10mm end caps were leaking when tested in bucket of water.

Just out of interest, are pushfit endcaps a good idea when pressure testing, with air or water etc?

Regards
FarmerJo

 

[FarmerJohn]

Also just noticed that the pushfit end caps I am using (John Guest) have a minimal working temperature of 20C. All my testing has been done at temperatures less than 10C and even below 0C overnight. Would this cause the endcaps to slightly leak?

Regards
FarmerJo

 

[oldbutnotdead]

I've used 15mm and upwards pushfit endcaps for pressure tests with no problem but the o rings don't last for an infinite number of removals/reinstalls. The o rings will be less elastic at lower temperature so that could be a contribution but more likely there's a bit of debris or deformation going on (10mm tube is very soft, unlike standard 15mm- a pushfit won't deform 10mm back into shape where a compression fitting would).
Try warming the endcaps (bucket of warm water is good), also use warm water on your dip tests, spin the endcaps if they're still bubbling-might cure it

 

[FarmerJohn]

Yes that sounds like good information. I will try the warming the endcaps in warm water and spinning the endcaps before testing. Will report back with the results.

 

[oldbutnotdead]

A quick exercise for you to ponder. A metre of 22mm copper has a volume of 300 ml ish. A metre of 15mms holds about 130ml. A metre of 10mms holds about 50ml.
So 20m of 22mm is only about 6 litres. If you can't cure the bubbles then determine a leak rate (upside down measuring jug, stopwatch) and see how much of your pressure drop is due to not quite sealed stop ends.
Or (if you've been sensible and left slack on the end of the 10mms drops) then sacrifice some olives and put the valves on.
TBH if the setup is holding 2 plus bar and leak fluid has found nowt in any of the joints I think you'll be fairly safe.

 

[oilhead]

The effect of a small leak is much more noticeable at lower pressures than high pressures as the amount of the media leaking is a bigger percentage of the volume at low pressure.

 

[FarmerJohn]

Tried the endcaps in hot water and that seems to work with all the 10mm pipework. All four runs are now holding pressure. To be honest I don't totally trust the dry test gauge that I have, it could be leaking.

I completed my 22mm/15mm bathroom pipe runs to shower and sink etc. and just bit the bullet and switched it on to mains water pressure. No leaks at all.

I did do an experiment to water pressure test about 2m of 15mm pipe at about 6 bar. The pressure could easily be increased by warming up the pipe slightly with the soldering torch. What did surprise me a bit was that just moving the connecting pipe slightly also changed the pressure.

Many thanks for everyone that has helped me with this problem, I have learn't a lot about testing pipework ready for the next time.

Regards
Farmer Jo