I would start with MiniFlyer's suggestion first about checking for an overly rich mixture on nose down with an almost full tank...
Then:
ORIGINAL: jetmech05
well don't forget there is pressure in the tank and centrifiugal force...
Unfortunately the "Centrifugal force" and "fuel stays in the back of the tank" is a misconception.
Posted videos back this up.
On even lightly angled downlines the clunk can and will become uncovered once the tank begins to empty.
As the tank approaches 1/2 or less the problem gets more and more dramatic as fuel is consumed.
Uncovering the clunk creates a large air bubble that can starve the engine of fuel.
Increasing the fuel tubing length doesn't help much, nor does wrapping the fuel line around the tank ( verified with videos we took btw... air still gets sucked into the line. ).
An engine that is able to run a bit richer when the nose is pointed down, and the clunk uncovered for a few seconds can sustain operation because it has two things going for it...
1 - the engine is unloaded and the airflow helps keep the engine running or idling
2 - the nose down attitude pushes a bit more fuel into the carb helping it to get through those "bubbles".
In an ideal situation the clunk WOULD travel forward and stay submerged in the fuel.
The problem with this is it can indeed get "stuck" up front.
Kinking, where the fuel is actually cut off, is really more problematic in the smaller tanks where the angle of the bend exceeds the tubing's strength/ability to keep the wall diameter somewhat consistent. In this case the wall collapses or kinks and the fuel flow is cut off.
The larger tanks ( assuming that the internal tank piping is set up to stay out of the clunk's way ) with medium tubing the wall thickness and strength, coupled with the unencumbered travel, permit you to set up an arrangement where the clunk may be permitted to travel forward.
MiniFlyer states: "The only way a clunk can suck air at half-tank is for the nose to be down at more than a 45 degree angle."
As you go beyond half a tank things get much worst rather quickly.
As fuel is consumed the required angle becomes less and less.
To solve all of this on several planes with 16oz+ fuel tanks I've set things up so that the clunk is permitted to travel forward.
To do this I have to be VERY careful with the vent tubing so that it stays close to the top wall of the tank. The intent being to eliminate any areas where the clunk and line may lodge or be caught into.
This has worked rather well for me, particularly in those cylindrical tanks with curved backs.