Water tank conundrum

[Jolly Roger]

have a water gage problem which has been driving me nuts for months, and I wonder if anyone with experience with pneumatic gages can throw some light on it.
I have two 150 gallon watertanks either side of the boat. The tank gages are air pressure operated, like the Tank Tender system, but older, circa 1977. The makers are long gone.
A simple 3/8” straight plastic tube, the depth of the tank, bolts into the top of the tank with a gasket and six screws. A 1/8” thin plastic pipe pushes in the top and goes to the gage where it pushes into a tube on the gage. The gages are made like a pocket watch, having a diaphragm and gears which activates the gage indicator as the water pressure varies.
One side works perfectly—when I fill up the tank the gage moves to full and decreases as the water is used. When the tank is empty the gage shows zero.
But when I drain and fill the tank on the other side the gage slowly moves to full, then drops off to zero within about an hour, but that varies a bit.
"Air leak" you say! Either in the plastic pipe or gage! but wait. I’ve installed new connecting pipe, which I pre-tested for cracks or leaks. I’ve also sealed all the joints in the tank tube with epoxy, including the connecting pipe joint.
Here’s the rub. If I swap the gages over, the good gage drops off to zero and the faulty side gage stays at full. So it’s not a faulty gage, right?
Also, when I submerge the faulty tube in the bath-tub at my home it records full, (with either gage), and stays full, the last test being a week in duration. There is absolutely no sign of air leaks in either the tank tube or the connecting pipe to the gage.
But the moment I screw it back in the top of the full tank, bingo! it drops off to zero within an hour, with either gage fitted .
I don’t want to change the system for a sight-glass or lever gage, because it doesn’t use any power and is really simple. One side works perfectly, so why doesn’t the other?
These sort of conundrums are frustrating, like intermittent electrical problems, so any thoughts would be much appreciated.

[Auzzee]

The only thing I can think of is the possibility of a bent sensor pipe. If the pipe is bending slightly horizontal on, for instance, a baffle in the tank, the rocking of the boat may be allowing a bubble of air out with each movement.

This could explain the 'fault' being specific to the location rather than to the gauge equipment.

Or it could be bad fairies.........

Cheers.

[Jolly Roger]

That's an interesting thought Auzzee. But there are absolutely no impediments to the tank pipe which goes straight down in a corner of the tank. The connecting pipe bends of course, up to the gage, which is about two feet above, but not with sharp bends.
Nor can it be the boat rocking, because it just doesn't, being in a very locked in position in the marina. Also, when I test it in the bathtub the pipe is actually at an angle of about 45 degrees, because the water isn't deep enough in the bath, but there is no air leak.
I'm inclined to think it actually is fairies.

[Auzzee]

Bloody Tinkerbell!
I'm otherwise stumped. I hope someone else can offer a fix.

[Oldsquizzy]

Murphy...tis always murphy.