This is one I posted on the iATN and thought I'd share it with some of my experts here.

It's a weird one folks. This is a dual fuel sending unit type fuel tank. Fuel gauge slowly raises to full after the 2 new fuel senders were installed. It is verified there is 1/2 tank of fuel. The resistance has been confirmed at both fuel senders, and at the receiver. The cluster has been removed and checked. Spot on with RM readings. The gauge raises and lowers perfectly with the senders laying on top of the tank. No obstructions in the tank. At this point we have it narrowed down to this. If I simply remove the bolts at the sub sender (pass side), and leave it there, the gauge will now go back to half where it should be. Obviously removing a ground because the bolt is creating a ground to the tank body. (metal tank) What's your thoughts on this?

The ref voltage going into my main sender is 12v. It goes through the resistor and the variable resister (float) of the main and comes out at 2.5V (1/2 tank).
I get the same reading (2.5V) at my sub fuel sender going in and coming out (unbolted to the tank)... but if I insert a bolt, then my ref voltage drops 50% or more at the sub. And The gauge raises to full.
I say there's a short to ground. Thoughts? Happy 4th.

Here's my take, without any experience in this situation:

You nailed it--that's the only thing that makes sense, to me. Since there is a sender for each half of the weird shaped tank (upside down U), the circuit needs to stay in series (additive), reflecting the gas in each half. By shorting the sub sender unit to ground, I expect the signal voltage doubles, with the fuel gauge (volt meter) then indicating the tank is full. The resistance in the sub sender unit is eliminated from the series circuit and the signal voltage to the fuel gauge is overstated.

kitch428, does this accurately reflect the picture, in your eyes?

What's the rationale for designing a fuel tank like this for a RAV4? It seems to be asking for trouble, to me.

The rationale for designing a fuel tank like this is because of the 4WD model which the drive shaft takes up the middle hump of the tank design. It's the same tank for the 2WD model to keep expenses down.
In the sub sensor side, there is a "jet pump" that works off a siphon if that side should get higher fuel level than the main side from splashing over the 'hump'.
You're absolutely correct in your thoughts. They are in series. If either one is disconnected, the fuel gauge goes dead. They both rely on a common ground which of course has been checked and good.
The only thing that makes sense here is an internal failure of the current flow in the sub sensor bracket itself.
One has been ordered and awaiting parts to arrive. We'll see I guess, huh tex?

Let us know the final solution.

Thanks for the replies. It's fixed. Somehow, in the metal bracket that holds the sub fuel sender, the insulation between the top and bottom wiring is interrupted and shorting to ground. Bolting it down creates another ground. Ref voltage goes down, resistance goes up, and gauge pegs. Go figure. Replaced bracket, problem solved. I told him, 4 hrs labor on this one, because this ones for the books. Hope it helps someone in the future.