'got my backwater device installed, used a 3/4" gray plastic threaded plumbing nipple into the tank hole, but did not want to make it permanent, or damage the hole with widening it, so I had to sand down the threads that would go into the tank. Threaded the other end of the nipple to a 2" ~ 3/4" PVC reducing adapter, and then the rubber union collar and backwater device float valve.
To keep the ball from air sealing the lower fitting by sitting on the hole, I placed a piece of round plastic with holes punched in it for the ball to sit on, above the pipe, and air to pass thru around the ball. Otherwise, the suction could shut the valve the wrong way.
The valve mounted this way is almost at the top of the dam height, so it's too high for the water to float the ball when the boat is level, but when heeled, I'm sure the water will fill the column and float the ball.
Not trusting the pressed-in / partially screwed plastic to fiberglass fit to be stong enough to withstand the 500 # or so of water weight when heeled over, even with the heavy application of marine silicon caulk, I rigged up a retainer rod across the top of the valve's large plastic top castle nut, attaching the ends of the rod to the top lip of the dam, 180 degrees across. This rod keeps a mechanical retention force to keep the valve and nipple in the hole. Made the rod out of flex copper 1/4" pipe, and flattened the ends, then 7/32" drilled holes in them to bolt down over the dam and valve. I left the pipe/rod in the round to pass thru the slots in the valve's castle nut.
So now trusting the valve assembly will stay in place, from its tight friction fit in the hole, generous silicone marine caulk, tightly screwed together at its rubber union clamps, and retained with the copper flex pipe rod like a car battery holder, even if it gets a small amount of water escaping the ball to gasket seal. The valve will never close when on the level, but then its seal is not needed. Any water that goes up the column should shut it with pressure.
I was concerned about the passive hose arrangements which vented thru the sink's thru-hull, allowing a slow blug blug leak of ballast water to escape the tank when heeled to port, and suck replacement air when it righted, to repeat next heel, and eventually accellerate as the air over the ballast water increased, to produce a partial ballast situation.
Even if a bit of water escpapes the valve when it first seals, and maybe it won't even allow that, it should remain in the dam or cup, and indicate how much is splashing past.
Maybe another column of rubber stacked over the valve to hold the water until the ball drops and the water falls back into the tank could be rigged, but right now that would interfere with my copper retainer rod.
As of now, if the set-up isn't satisfactory, I can still fairly easily remove it altogether and put the plug back in the unchanged hole.
If it works really well, I will probably epoxy it in place, and then plumb a pipe over the valve to carry the stale tank air over to the sink thru-hull, but leave the ball valve for ballast retention. Then it would be truly maintenance free. An audible whistle as the air is pushed out would be great, or maybe the ball will oscillate between the gasket and plastic perforated floor, from the air pressure varying. Reasoning that if the escaping air column forces the ball up against the rubber seat, the water coming in thru the open gate valve would reach a point where it stopped, then back out a bit and allow the ball to fall free, and the air to again escape.
The thing is... obviously I have not yet tested this in the water, - 'boat is on the trailer - so when I have time for launching, rigging and sailing again, we will find out how it all works and the actual results not just my wishful thinking and theories.
