Normal "Leak-Down" During Carbonation... Or leak?

[Nagorg]

Hey HBT!

I'm trying to figure out if I have a CO2 leak or if what I'm seeing is normal for a fresh keg during early carbonation. I'd like to avoid loosing another 10# bottle of CO2.

For a leak-down test, I've been doing this:
- Isolate a single keg's gas line by closing the manifold valves everywhere else. **The line IS attached to the keg since I'm wanting to carbonate.
- Pressurize the line.
- Close the CO2 bottle valve.
- Observe the regulator gauges looking from drops over time.

For a new keg of flat beer, I get believe that I should expect this test to show drops as the CO2 gets absorbed into solution. But what rate of drop would be considered normal -vs- a leak?

If I do this test and wait for about an hour or two, I do notice that the pressure will drop for the tank pressure gauge with maybe a slight drop for the line pressure gauge. But I just did an overnight test and all gauges were completely bottomed out when I looked this morning.

When open the CO2 bottle again, I can hear the keg fill as the pressure is returned to the lines.

Is that rate of drop normal for a keg that's being carbonated? Or do I have a leak to find?

 

[WayFrae]

I wouldn’t really expect the lines to lose pressure over night if they aren’t hooked up to anything.

 

[Nagorg]

I wouldn’t really expect the lines to lose pressure over night if they aren’t hooked up to anything.

Maybe I didn't clarify, the line IS hooked up to the keg of beer since I'm trying to carbonate it and observing this loss. I get no pressure loss if testing the line when its not attached to the keg.

 

[WayFrae]

It sounds like you might have a leak in your keg then. I’d take off the o-ring gasket and give it a good wash and inspect it for damage. Also wipe around the lip under the keg to make sure where the gasket seats is clean. Then, wet the gasket a bit (or use keg lube) and pressure test the keg again.

 

[Broken Crow]

I get that I should expect this test to show drops as the CO2 gets absorbed into solution.

not really, I'm pretty sure there's some myth in there. Equilibrium does mean that CO2 is absorbed into the beer, but that does not change the mass inside the keg under pressure and the moment it is opened (via the tap) to lower pressure, the CO2 will begin to re-emerge from the beer. If you leave it at pressure and actually notice the pressure drop, then you have a leak.

NEXT DAY EDIT: I am WRONG. I Apologize.

 

[Nagorg]

you might have a leak in your keg then

This is what I'm trying to figure out. Like I said, the beer is flat and I'm trying to carbonate it. So since its being carbonated, I would expect that I'd loose pressure until the beer became carbonated based on the PSI I've set.

But is this rate of loss normal until the beer is carbonated or is it excessive? If its normal, I'd expect the loss to stop once the beer became carbonated.

 

[Nagorg]

If you leave it at pressure and actually notice the pressure drop

That's just it, I'm closing the valve after pressurizing the line. In this scenario, I'd expect that line pressure would drop as CO2 gets absorbed into the beer since the line no longer has a supply of pressurized CO2.

If the beer was fully carbonated, I'd expect there to be no loss. But since its flat, it makes sense to me that there would be loss. But is this rate of loss normal or not given that the beer is flat?

 

[Broken Crow]

But since its flat, it makes sense to me that there would be loss. But is this rate of loss normal or not given that the beer is flat?

Sorry, that's the part I'm having difficulty being clear on...wether the CO2 is in the beer or above it, not yet absorbed, the pressure will remain. If you vent a carbonated keg down to 2psi and wait, the beer will go flat down to 2psi over time, but the pressure in the keg will be unchanged at 2psi. It's not really different than pressurizing an empty keg... if it drops without a downward temp change, it leaks.

 

[Nagorg]

If there's only the small amount of CO2 in the line, even at PSI, then what happens to pressure as it gets absorbed into the beer? Wouldn't that cause a drop in PSI as that small supply of CO2 became absorbed?

I do not see an immediate drop in PSI. I have to let it sit for a while and then check the gauges later to see the drop. This seems to me that this could be "normal" for a beer that's not yet reached the target carbonation level. But if it dropped too fast, it could be indicative of a leak.

The keg is literally brand new. It held pressure before I filled it with beer. New lid and post o-rings, again the keg is new. I can replace the post o-ring but I'm really just trying to understand if this rate of loss is normal or not. From what you're telling me, I should expect no loss as CO2 gets absorbed during carbonation and I just don't understand that. Maybe that's my problem but it just doesn't make sense...

Maybe I just need to let it ride. In ~8 more days I'll either have a fully carbonated keg of beer that no longer looses pressure during a leak-down test or I'll be headed to Airgas to swap out another empty bottle of CO2.

 

[Broken Crow]

This part:

But I just did an overnight test and all gauges were completely bottomed out when I looked this morning.

..should not happen. If the beer absorbed the CO2 then it would have pressure within the keg. Adding CO2 to beer is not like adding liquid. Though it comes out of liquid slower, the very head on your glass of beer is the pressure of CO2 leaving. I regularly fill kegs to set aside at serving pressure and when I tap them weeks later they are still at pressure, but now at equilibrium with fully carbonated beer.
I think you should be looking for the leak. Post some pics of your lines and disconnects.

 

[WayFrae]

pressurize the keg and spray starsan or soapy water on the seals, you will see bubbles if you have a leak there.

 

[Nagorg]

pressurize the keg and spray starsan or soapy water on the seals, you will see bubbles if you have a leak there.

I regularly see this mentioned but have yet to have this help me find a leak. Seems others report the same.

 

[Broken Crow]

If it helps, here's where I've failed in my process and found leaks a few times in new kegs:
-Second use of a cheap import 1.6G mini-keg; The gas post had somehow loosened and I had to tighten it back down.
-Twice, with new 5G Kegland kegs; on removing the disconnect, the poppet came up very slightly off-center and I had to give it a quick poke down and back up to seal.
-Once, a very elusive leak that totally confounded me and forced me to remove the lid..at which point I found that I had a hair wedged in behind the lid o-ring (My hair is down to my butt and gets everywhere! )
-Once; (and I'm still unsure of this one) I could find no problem at all save having forgotten to lube the lid o-ring, which I did on reassembly and then it held pressure just fine.
Hope it helps.

 

[Nagorg]

These (I actually have 2 doing this) are new Torpedo kegs. They do have the o-rings that they came with. I'm about to remove the disconnects and perform another leak-down test on the disconnected lines as a sanity check. My first leak was found to be a disconnect that needed more tightening. I'll make extra sure that these lines don't have a leak on their own.

I'll also then replace those stock o-rings with new ones and see if that helps. I did spray star-san around the lid but I see no bubbling. But I'm not confident that that method would be very effective for a slow leak.

And to be clear about what you say regarding that there should be no pressure loss if there are no leaks, not even with a flat beer... You are saying that I should be able to pressurize a keg of flat beer to the desired PSI, completely disconnect the keg from the CO2 line and as long as the keg is kept cold (38 degrees in my case) it should become fully carbonated in ~10 days with zero change in PSI?

 

[Broken Crow]

And to be clear about what you say regarding that there should be no pressure loss if there are no leaks, not even with a flat beer... You are saying that I should be able to pressurize a keg of flat beer to the desired PSI, completely disconnect the keg from the CO2 line and as long as the keg is kept cold (38 degrees in my case) it should become fully carbonated in ~10 days with zero change in PSI?

That's where I can't give a certain answer..Full disclosure: Since a child, I've percieved everything through physics, but in 2014 I was injured, including my brain and I've lost the math and science and frustratingly the entire world and every system in it is difficult for me now...my very comprehension of numbers and comparative quantities volume and measures is fundamentally broken...even my 'certainty' of any and everthing...so I have to rely entirely on the emperical and carry on in manners dictated by what remains of my memory of having dealt with before. I'm on this site largely for re-learning, especially the math and science, so at this point I'll tag my adopted mentor @doug293cz and hope he can wiegh in with the science in which I suspect there is some loss of pressure, but certainly nowhere near enough to drop the gauge overnight or even more that 1 or 2 psi over a couple weeks.
(I hate having to include the disclaimer and don't want to derail the thread...I'm learning here to, so thanks for demanding a full accounting.)

 

[WayFrae]

These (I actually have 2 doing this) are new Torpedo kegs. They do have the o-rings that they came with. I'm about to remove the disconnects and perform another leak-down test on the disconnected lines as a sanity check. My first leak was found to be a disconnect that needed more tightening. I'll make extra sure that these lines don't have a leak on their own.

I'll also then replace those stock o-rings with new ones and see if that helps. I did spray star-san around the lid but I see no bubbling. But I'm not confident that that method would be very effective for a slow leak.

And to be clear about what you say regarding that there should be no pressure loss if there are no leaks, not even with a flat beer... You are saying that I should be able to pressurize a keg of flat beer to the desired PSI, completely disconnect the keg from the CO2 line and as long as the keg is kept cold (38 degrees in my case) it should become fully carbonated in ~10 days with zero change in PSI?

The PSI will go down if you disconnect it as the CO2 gets absorbed but at some point it will equalize. I’d expect there to still be pressure even after 10 days but it won’t be fully carbed and it won’t be at the pressure you set because you didn’t keep it connected.

You can pressurize the keg and shake it and hear the CO2 coming out of the tank.

 

[WayFrae]

I regularly see this mentioned but have yet to have this help me find a leak. Seems others report the same.

You need to use more soap or starsan in the mix then. You can also make it sudsy and put the bubbles on there. If it’s a super slow leak then it will be hard to catch. You could also submerge it in water

 

[doug293cz]

You can't leak test with uncarbonated beer or water in a keg. The beer/water will absorb CO2 from the headspace, and the pressure will drop, if the CO2 supply is shut off. If you want to leak test, use a keg with no liquid in it.

Brew on

 

[doug293cz]

Sorry, that's the part I'm having difficulty being clear on...wether the CO2 is in the beer or above it, not yet absorbed, the pressure will remain. If you vent a carbonated keg down to 2psi and wait, the beer will go flat down to 2psi over time, but the pressure in the keg will be unchanged at 2psi. It's not really different than pressurizing an empty keg... if it drops without a downward temp change, it leaks.

This is all wrong.

If the CO2 partial pressure in the headspace is higher than that which would be in equilibrium with the current carbonation level of the beer, then the beer will absorb more CO2 from the headspace. If the headspace is not fed by a pressure controlled CO2 source, then the headspace pressure will drop until the CO2 partial pressure and the carbonation level are in equilibrium.

If the beer is carbonated, and the headspace vented to lower the CO2 partial pressure, and no CO2 source is connected. Then the beer will release CO2 to the headspace, and the headspace CO2 partial pressure in the headspace will increase. This will continue until the headspace CO2 partial pressure and beer carbonation level are in equilibrium.

Brew on

 

[hotbeer]

When a certain volume of CO2 is dissolved in water or beer, there isn't a equal increase in the liquids volume. In fact I am not sure the chemical bonds made will not be tighter and there is almost no increase in the liquid. So pressure would fall in the headspace.

I don't know how significant the change is though for the parameters given.

 

[Nagorg]

Thanks @doug293cz for confirming my thoughts on this. So what I'm observing in this leak-down test with flat beer is expected. Which means that its going to be hard to determine if I have a leak until the beer is carbonated.

And I get that but does what I describe (the amount of time it take for pressure to drop) sound somewhat normal or is it too fast?

I may be asking for the impossible and may just have to give it the full amount of time and hope that its not leaking. I've replaced the o-rings now even though they were new. Its a cheap thing to try so why not.

 

[day_trippr]

fwiw, over the last many years 100% of my (highly infrequent) keg leaks have been through gas poppets - and that can only happen when the gas posts don't have QDs on them. So it never happens while conditioning or dispensing kegs - just conditioned kegs sitting in one of the fridges waiting to be tapped. And I have in fact tapped what turned out to be a flattened keg a couple of times - could have been the same keg...

Cheers!

 

[doug293cz]

When a certain volume of CO2 is dissolved in water or beer, there isn't a equal increase in the liquids volume. In fact I am not sure the chemical bonds made will not be tighter and there is almost no increase in the liquid. So pressure would fall in the headspace.

I don't know how significant the change is though for the parameters given.

The pressure drop can be calculated if you know:

The beer volume​

The headspace volume​

The starting carbonation level of the beer​

The CO2 partial pressure in the headspace​

The temperature​

​

Brew on

 

[doug293cz]

Thanks @doug293cz for confirming my thoughts on this. So what I'm observing in this leak-down test with flat beer is expected. Which means that its going to be hard to determine if I have a leak until the beer is carbonated.

And I get that but does what I describe (the amount of time it take for pressure to drop) sound somewhat normal or is it too fast?

I may be asking for the impossible and may just have to give it the full amount of time and hope that its not leaking. I've replaced the o-rings now even though they were new. Its a cheap thing to try so why not.

The rate of pressure drop is going to depend on the headspace volume to beer volume ratio, the temperature, the starting CO2 partial pressure, and the starting carbonation level. The further you are from equilibrium, and the smaller the headspace relative to the beer volume, the faster the pressure will drop.

You can do the test with an empty keg. Start with the keg full of air, and then pressurize to ~15 psi. This will only require about 40 g (less than 2 oz) of CO2. The shut off the CO2 as you did before. If the pressure drops, then you have a leak. If the pressure holds, then you are good.

Brew on

 

[tchampagne]

This isn't exactly what the OP asked but I've lost my share of CO2/tanks over the years. I was aggravated enough to try and figure out a way to monitor CO2. It took me longer than I'd like to admit but finally came up with a solution that works quite well for me. I did consider using weight as a way to monitor (this is how I remotely monitor my keg levels) but during force carbonation it is valid (i.e, not a leak) to use CO2 and thus have the weight decrease.

I have a three tap keezer and two tap kegerator and in both cases I keep my CO2 tank inside the keezer/kegerator. This allows me to monitor the CO2 levels in the 'air' inside each keezer/kegerator. If there happens to be a leak whether it be a hose, poppet, valve, etc. the CO2 level in the air will change. Heck, just purging the keg triggers my alarm/alert to occur (it might be a it too sensitive).

I use a simple Raspberry Pi and this CO2 monitor (https://www.amazon.com/JESSINIE-Monitoring-Concentration-4-5～5-5VDC-0-10000ppm/dp/B0B7MBM7RN/)

Here is a screen shot of my CO2 dashboard.

 

[Broken Crow]

This is all wrong.

If the CO2 partial pressure in the headspace is higher than that which would be in equilibrium with the current carbonation level of the beer, then the beer will absorb more CO2 from the headspace. If the headspace is not fed by a pressure controlled CO2 source, then the headspace pressure will drop until the CO2 partial pressure and the carbonation level are in equilibrium.

If the beer is carbonated, and the headspace vented to lower the CO2 partial pressure, and no CO2 source is connected. Then the beer will release CO2 to the headspace, and the headspace CO2 partial pressure in the headspace will increase. This will continue until the headspace CO2 partial pressure and beer carbonation level are in equilibrium.

Brew on

Thanks! That's another incorrect preconcieved notion of mine you've busted...( I'm gonna have to find some way to pay you as an online tutor)
I'm trying to do the math because your description for calculating the drop reads like a visual diagram in a text-book, but my math is broken... Is it possible for the gauge to bottom out overnight? Part of brain keeps insisting to me that the drop wouldn't be so steep and I'm truly trying to understand.

 

[NoBigDeal]

Hey HBT!

I'm trying to figure out if I have a CO2 leak or if what I'm seeing is normal for a fresh keg during early carbonation. I'd like to avoid loosing another 10# bottle of CO2.

For a leak-down test, I've been doing this:
- Isolate a single keg's gas line by closing the manifold valves everywhere else. **The line IS attached to the keg since I'm wanting to carbonate.
- Pressurize the line.
- Close the CO2 bottle valve.
- Observe the regulator gauges looking from drops over time.

For a new keg of flat beer, I get believe that I should expect this test to show drops as the CO2 gets absorbed into solution. But what rate of drop would be considered normal -vs- a leak?

If I do this test and wait for about an hour or two, I do notice that the pressure will drop for the tank pressure gauge with maybe a slight drop for the line pressure gauge. But I just did an overnight test and all gauges were completely bottomed out when I looked this morning.

When open the CO2 bottle again, I can hear the keg fill as the pressure is returned to the lines.

Is that rate of drop normal for a keg that's being carbonated? Or do I have a leak to find?

 

[NoBigDeal]

Hey HBT!

I'm trying to figure out if I have a CO2 leak or if what I'm seeing is normal for a fresh keg during early carbonation. I'd like to avoid loosing another 10# bottle of CO2.

For a leak-down test, I've been doing this:
- Isolate a single keg's gas line by closing the manifold valves everywhere else. **The line IS attached to the keg since I'm wanting to carbonate.
- Pressurize the line.
- Close the CO2 bottle valve.
- Observe the regulator gauges looking from drops over time.

For a new keg of flat beer, I get believe that I should expect this test to show drops as the CO2 gets absorbed into solution. But what rate of drop would be considered normal -vs- a leak?

If I do this test and wait for about an hour or two, I do notice that the pressure will drop for the tank pressure gauge with maybe a slight drop for the line pressure gauge. But I just did an overnight test and all gauges were completely bottomed out when I looked this morning.

When open the CO2 bottle again, I can hear the keg fill as the pressure is returned to the lines.

Is that rate of drop normal for a keg that's being carbonated? Or do I have a leak to find?

There are many comments in this thread, but I haven't seen a mention of a soap test. This test uses a liquid that bubbles when you spray it on a gas leak. There are commercial leak detector products, but you can mix a small amount of liquid soap with water in a spray bottle and spray the connections in your gas system from the CO2 bottle to the keg lid. Don't forget the keg lid pressure relief valve and the gas and liquid posts on the keg. Even a very tiny leak should result in bubbles. If you use Star San as a sanitizer, your Star San spray bottle will work.

In my case, I found leaks at the union between the CO2 bottle and the regulator, in the threaded connections on the regulator body, in hose connections (both barb and flare) from the regulator to the gas connector, and from the keg lid. I also found leaks in the gas connectors when I didn't assemble them correctly after cleaning or if there was damage to the poppet valve.

If this doesn't locate the leak, I remove the gas connector from the keg. With the CO2 bottle valve shut, a constant pressure in the remaining portion of the gas system tells you the leak is from the keg. If you still see the pressure drop and the outlet of your regulator has a shutoff valve, shut it. A constant pressure means the leak is from your gas hose or its connections. If you still have a pressure drop, the leak either is in your gas hose or is an internal leak in the regulator.

If you haven’t found and corrected your CO2 by now, there are commercial CO2 leak detectors that could detect the CO2 at the leak location.

Although it won’t help your current problem, I’ll offer something that might help when you’re ready to fill your next keg. I perform a drop test on each keg after I clean it. After I assemble the keg with new O-rings and keg lube, I hold the lid in position while I add enough CO2 pressure to hold the lid in place. Since I don’t close the lever on the lid, even the smallest leak will let the lid fall to tell you that there is a problem that you can correct before the keg is full of beer. This doesn’t require anything more than a little patience since a slow leak may take a few days before the lid drops.

Good luck with your soap test and corrective actions.

 

[mac_1103]

There are many comments in this thread, but I haven't seen a mention of a soap test

Post #11.

 

[Nagorg]

Post #11.

And post #12

 

[doug293cz]

Ok, let's look at the OP's situation:

A corny keg has a total volume of about 5.3 - 5.35 gal, so let's assume 18.9 L (5 gal) of beer, and 1.5 L (0.4 gal) of headspace volume (which includes the CO2 tubing and internal regulator volume.) If the beer was fermented at 68°F (20°C) then it will start with a carbonation level of 0.84 volumes, and the total CO2 mass in the beer will be 31.34 g. If we purge the headspace of air, and pressurize to 12 psig (26.7 psia CO2 partial pressure), chill to 4°C (~40°F), and shut off the CO2 tank, then we will have 5.39 g of CO2 in the headspace.

Now the beer starts to absorb CO2 from the headspace, since at 4°C and 26.7 psia CO2 partial pressure (12 psi gauge pressure) the equilibrium CO2 volumes in the beer would be 2.52 volumes (which is way more than the initial 0.84 volumes.) As the beer absorbs CO2 the beer carbonation level increases, and the headspace pressure decreases. When equilibrium is reached, the beer carbonation level will be 0.93 volumes (34.76 g of dissolved CO2), and the headspace pressure will be -4.8 psig (9,9 psia CO2 partial pressure) and the mass of CO2 in the head space will have dropped from 5.39 g to 1.97 g, for a reduction of 3.42 g. For the gauge pressure to drop to 0 psi only requires the beer to absorb about 3.42 * 12 / (12 + 4.8) = 2.44 g of CO2 to be absorbed by the beer.

It's very plausible that the beer could absorb 2 - 2.5 g of CO2 overnight.

Brew on

 

[doug293cz]

Ok, let's look at the OP's situation:

A corny keg has a total volume of about 5.3 - 5.35 gal, so let's assume 18.9 L (5 gal) of beer, and 1.5 L (0.4 gal) of headspace volume (which includes the CO2 tubing and internal regulator volume.) If the beer was fermented at 68°F (20°C) then it will start with a carbonation level of 0.84 volumes, and the total CO2 mass in the beer will be 31.34 g. If we purge the headspace of air, and pressurize to 12 psig (26.7 psia CO2 partial pressure), chill to 4°C (~40°F), and shut off the CO2 tank, then we will have 5.39 g of CO2 in the headspace.

Now the beer starts to absorb CO2 from the headspace, since at 4°C and 26.7 psia CO2 partial pressure (12 psi gauge pressure) the equilibrium CO2 volumes in the beer would be 2.52 volumes (which is way more than the initial 0.84 volumes.) As the beer absorbs CO2 the beer carbonation level increases, and the headspace pressure decreases. When equilibrium is reached, the beer carbonation level will be 0.93 volumes (34.76 g of dissolved CO2), and the headspace pressure will be -4.8 psig (9,9 psia CO2 partial pressure) and the mass of CO2 in the head space will have dropped from 5.39 g to 1.97 g, for a reduction of 3.42 g. For the gauge pressure to drop to 0 psi only requires the beer to absorb about 3.42 * 12 / (12 + 4.8) = 2.44 g of CO2 to be absorbed by the beer.

It's very plausible that the beer could absorb 2 - 2.5 g of CO2 overnight.

Brew on

Nerdy math alert:

For those curious about how the calculations are done for the above, I submit the following derivation for determining the equilibrium pressure in a sealed system if you know the beer volume, headspace volume, temperature, and mass of CO2 in the system. The first equation allows you to calculate the starting mass of CO2, if you know the fermentation pressure (blue PA) and the current headspace pressure (green PA.) In OP's case they are not the same, but in other cases they may be.

The first thing we do is calculate the total CO2 mass. Blue PA = 14.7 psia, and green PA = 14.7 + 12 = 26.7 psia. We can then use the last equation below to calculate the new CO2 partial pressure when everything comes to equilibrium.

In the above equations, the meaning of the variables are:

CO2 is the mass of CO2 in grams​

VolB is the volume of beer in liters​

VolH is the headspace volume in liters​

PA is the CO2 partial pressure in psia​

TK is the absolute temperature in °K​

The blue part of the first equation comes from equation 2.1 in the attached .pdf converted for absolute (rather than gauge) pressure, and temp in °K (rather than °F.) 1.977 g/L is the value and units for 1"volume" of carbonation - i.e. the density of CO2 at STP. The green part of the first equation is just the ideal gas law.

Brew on

 

[Nagorg]

Thats awesome... Thanks @doug293cz !