Ok yeah I get the ideal gas law and all, but I don't get how you can have a gas constricted by an air lock decrease in temperature but not pressure. If the temp of the gas in the headspace decreases it will decrease in volume AND pressure.
The volume in the fermenter is relatively fixed. As the pressure drops, air is pulled into the fermenter.
Also, I've never seen air sucked though the water of an air lock. If there would be a suction created, then wouldn't it suck the water back first (assuming that the water level covers the holes of the inner piece)?
This is what I saw. No speculation here. Those with better bottles and 3 piece air locks will tell you that if you squeeze and release, air can bubble both ways. Not all of the liquid in the airlock gets sucked into the fermenter. You can try it yourself. put an airlock on an empty beer bottle or flask. Put it in the fridge. You will see maybe some but not all of the liquid gets sucked in along with some air. Then take it out of the fridge and watch as it warms up. Since there is now more air in the bottle, as it expands it will bubble out again. You can keep repeating this as many times as you want and still have liquid in the airlock.
One last thing is in your OP you stated that co2 was still being produced, and pushing out through the lock.
Here is what I said:
I usually switch from blow off tube to 3 piece airlock after fermentation slows down.
Obviously the heating/cooling effect is enough to overcome the miniscule amount of C02 being produced. This is what I saw. You are right that it will only occur once fermentation is done and CO2 generation is very small. After noticing this, I moved the fermenter to a cooler with a small frozen ice bottle in it to keep the temperature relatively constant 68F for a long period of time. At that point the airlock was bubling about once every 120 sec.
If that's the case the I would think that the co2 generation would be vastly greater than any suction force created by a one degree differential.
there is much more than a one degree change in air temperature when the beer temperature is being controlled to within 1 degree. For example, when controlling the temp to 68F, just before the cold cycle kicks on, the air temperature is probably close to 68F or maybe a little higher. Then, when the thermostat trips the cold cycle on, cold air at about 32F comes out of the cold side until the beer tempeture drops back down enough to trip the thermostat off. So the air temperature differential between cycles is about 32F. A chest freezer may have a higher temperature variation since it is designed to go way below 32F. I don't have a chest freezer so I'm just speculating on that. The headspace temperature is probably not cycling that widely, but I'm sure it's significantly over 1 degree.
However, if you can document this phenomenon and prove it, then I would assume that data would drastically decrease 3 piece airlock production. They basically have one purpose and you would be proving it an ineffective design.
I wouldn't go that far. Everyone's setup is different and I only saw it happen on mine. In my OP I said I didn't know if this is a problem and YMMV. I just wanted to point out that it is possible and I think people should be cognizent of it. I didn't know it could even happen until I witnessed it for myself and gave it some thought. My solution for my system is to keep the blowoff tube on until I rack to secondary and then put the airlock on. Believe me I have nothing against airlocks.
Thanks to all that contributed to this thread, It helped me think through what was happening. Fun stuff.