Keg Leak Troubleshooting

[klangers]

HOLY THREAD RESURRECTION BATMAN

I think I have got maybe a keg out of my last two co2 bottles. I might've got 12 kegs from the one before that.

I am having a ******* frustrating time finding the leak. I have a metal manifold from the gents at Gryphon that I believe has one-way valves in each tap and I have another one way valve before the manifold

despite this I come home today to three empty kegs and a very empty c02 bottle. After a ******* long day this has done little to improve my mood - NO ******* BEER AND NO ******* GAS MAKES KAHLERISMS SOMETHING SOMETHING

I'm having to drink the second best beer in Australia out of stubbies instead of my own draught beer.
WHAT IS THIS, A THIRD WORLD COUNTRY?

Anyway, I'm frustrated and needed to vent. For some reason even though I use the soapy water trick on LPG all the time I have never done so on my keg system. The tips in this thread about disconnecting stuff and seeing if pressures drop are great, too. Will report back. I have three fermenters full of beer that need to get into kegs and I can't pressureise any of them so they're all sitting near zero for a day or a week or something until I get around to getting my CO2 bottle swapped.

This is also how I feel at the moment.

• 24 hour Pressure test on bottle and regulator - pass
• 24 hour pressure test on lines - all isolated at manifold - pass
• Soap test on all joins; isolated at manifold - pass
• Soaped 2 kegs (the only kegs I have full ATM) - found one leak in lid. Re-seated lid O ring and applied keg lube. No more bubbles - pass
• Soap tested all connections - relief valve, disconnect. Pass
• 24 hour pressure test on disconnected disconnects - fail. But no bubbles in soap test or full immersion test. WTF!!!??

Can anyone help me as to why the hell I have a slow leak when my disconnects aren't attached, yet both water immersion and soap bubble tests say otherwise?

 

[Grott]

mmmmm.... the disconnects and their connectors show no leaks, there is no leaks where these lines connect to manifold so it has to be the gas line (tubing) itself. A split? hole? You say disconnects as in plural therefore you need to pressure test each line on there own by turning off the others at the manifold. Then you will know which line is the bugger.

 

[Dave70]

After re connecting my gas line, first time after shifting house, I dumped a bottle of CO2 in three days. The o'l spray test reveled nothing - until I grabbed the tube and moved it from side to side near the barbed nipple on the reg. Hissssss..bubble bubble.. A teeny split that must opened up once the bottle was back in position. Cut about 20mm of the end of the tube and re connected. No worries.

I'd also suggest fitting the hose to the fittings using a little soapy water and finger strength, not softening it up with a lighter or hot water. Pretty sure that was the culprit in the first place.

 

[goatchop41]

This is also how I feel at the moment.

• 24 hour Pressure test on bottle and regulator - pass
• 24 hour pressure test on lines - all isolated at manifold - pass
• Soap test on all joins; isolated at manifold - pass
• Soaped 2 kegs (the only kegs I have full ATM) - found one leak in lid. Re-seated lid O ring and applied keg lube. No more bubbles - pass
• Soap tested all connections - relief valve, disconnect. Pass
• 24 hour pressure test on disconnected disconnects - fail. But no bubbles in soap test or full immersion test. WTF!!!??

Can anyone help me as to why the hell I have a slow leak when my disconnects aren't attached, yet both water immersion and soap bubble tests say otherwise?

I had this issue (2.6kg bottle disappeared after about 3 kegs, I was disconnecting it when not using, but forgetting to turn the bottle itself off!), and it was the inside of one of my disconnects that was the issue. I didn't diagnose it until I flipped the disconnect upside down (so that I could see the little sticky-out bit that seals the disconnect until it is pushed down in to the post) and half filled the inverted disconnect with star san. Pumped the reg up to 40-45 psi then, lo and behold, I see a few little bubbles start to come up from around the little thing that sticks out and pushes the poppet down!

I promptly turned off pressure, disassembled the disconnect, lubed the tiny o-rings in it and reassembled - no more leaks since!

 

[Grott]

mmmmm.... the disconnects and their connectors show no leaks, there is no leaks where these lines connect to manifold so it has to be the gas line (tubing) itself. A split? hole? You say disconnects as in plural therefore you need to pressure test each line on there own by turning off the others at the manifold. Then you will know which line is the bugger.

klangers have you done the checks yet? We'd all like you to find this f*^n leak. Best of luck.

 

[klangers]

klangers have you done the checks yet? We'd all like you to find this f*^n leak. Best of luck.

Haha thanks Grott.

I've completed one 24 hr pressure test on line "C". There are 7 other lines, so this will take some time

 

[Grott]

Oh, that's a lot. If able you could speed up process as only a few hours should give you a result but if working then 2 a day? Set one up before off to work and check/set up next when you get home etc.

 

[peteru]

A binary search algorithm will get you there much quicker than a linear search.

Turn half ON and half OFF.
Test.
If no leak, then mark the ON lines as GOOD.
Repeat the above steps on a set of unmarked lines, but each time half the number of tested lines until you get to individual ones.

Assuming you have a single leak somewhere among 8 connections, you should find the culprit in 3 test. It'll take a few more goes if you have multiple leaks.

 

[klangers]

Clever idea, but I don't know how many leaks I have. I know it's definitely more than 1. They're very, very slow indeed. For all I know it could be a systematic issue with all the hose clamps or disconnects.

 

[GalBrew]

I have just undergone fixing my leaking keg fridge after coming back from holidays to find the gas bottle empty. I had 3 leaking kegs, which required replacing the gas post seal. The only foolproof way of working where your leaks are (if you can't see them after spraying starsan everywhere) is to isolate one part of your system at a time, set the regulator, turn your bottle off and wait to see if the high pressure gauge drops. With a slow leak this can be an overnight wait! Took me two weeks to fix the problem, which was compounded by undercarbed kegs.

 

[klangers]

So after doing 24 hr tests on each of my lines (minus one which has no disconnect), all but 2 have slow leaks. These all take a few hours to notice any change whatsoever. We're talking very slow indeed. Splits or pinholes would be far faster.

I have 4 SS disconnects on the way, so I'll swap in them and see what happens.

If it's not the disconnects, then I'm going to throw the whole fecking manifold in the bin and do it with Festo pneumatic components. The ball valves could even be leaking through the spindle packing, and I don't have time for that crap.

 

[Grott]

The SS disconnects are value as they fit and seal well. I found a lot of the cheap plastic ones would often leak if any sideway pressure was applied which is very hard to detect. I suggest when you get the disconnects you replace the keg post seals as well.

I think you may being fixing your problems and all will be good.

 

[klangers]

More results.

So I took off a plastic disconnect and attached a barbed stainless one. I ran out of hose clamps so a press fit over the barb was holding it on.

The stainless disconnect appeared to have solved the problem as there was no HP needle movement over 24 hours. However, after 36 hours it had moved. F%^k me. Here we go again. This bloody bush league hose-barb and ****** ball valve manifold is next on my hit list. Leaks through valve stems I've seen before.

Hell, the leak's slow enough that we could even be talking diffusion through the vinyl hose.

I am going to run a parallel line which is 100% Festo pneumatic components into a SS disconnect which is pushfit >> FFL thread. If this leaks then I'll give up and resign myself to the fact that excruciatingly slow leaks are reality in a non-welded system.

 

[Grott]

Klangers, you shouldn't be looking at the high pressure gauge, this with will change with temperature changes in the room. You should be looking at the pressure you have set in the gas lines. So if you set say 30psi and turn the tank off with gas disconnects off the kegs then if no leaks it will remain at 30psi.

 

[GalBrew]

Klangers, you shouldn't be looking at the high pressure gauge, this with will change with temperature changes in the room. You should be looking at the pressure you have set in the gas lines. So if you set say 30psi and turn the tank off with gas disconnects off the kegs then if no leaks it will remain at 30psi.

That is partially true, if you set your reg to 30psi and turn the bottle off if you have no leak it will stay at 30psi. Also the high pressure gauge will fluctuate with room temp. However, If you have a very slow leak the high pressure gauge will drop off first, followed by the low pressure gauge. Depending on how slow your leak is, this could take a couple of days.