I just installed a through wall pipe on the wootank, it doesn't seem to leak. Much. (mental: well, at least it's not leaking a lot)
I drilled the hole while the tank is full, putting a block of wood on the inside to prevent a jet of cold water. The water is approx 6C, and keeping your arm in it for any period of time is not fun. Then I removed the block of wood and inserted the flanged pipe, with a bit of duct tape over the hole. The pressure alone was sufficient to keep the flange in place and almost no water leaked out. I then screwed on the outside flange and tightened, added a stopcock and removed the duct tape. No obvious drips, it appears not to leak... of course it could be leaking very slowly inside the plywood which we will find out about in a few months when it fails spectacularly. We'll see.
This means I can adjust the overflow level more precisely than cutting away the wall of the tank, and also hopefully will stop the water leaking between the liner and ply when it goes over the overflow (and we're talking many kl here over a period of months), which I suspect has happened at some point.
It also allows me to try out the super low energy flood and drain idea I've had rattling around for a while. The idea is to have the f&d bed at just above the tank's water level and use an axial flow pump to raise the water level the requisite 10cm. This would, for a 3.6x0.9 m bed, require at least 160J, which is 200* less than the current system.
(As an interesting aside, we might save even more by setting the flood and drain at slightly below the tank level, reversing the pump so as to keep the average at the neutral point. Extra points by having the system resonate as a wave guide so that most of the energy is conserved)
<insert pictures here>
No obvious leaking. Rate is below evaporation.