I have a 1991 Trans am convertible with TPI. I am at a loss on what to do or try next. My car runs fine but the gas gauge does not work (reads empty). Unplug the unit and it reads full. I replaced the sending unit and still does not work in the tank (reads empty). If I remove the sending unit from the tank and just plug it into the car and move the float the gauge reads just fine. Just to make sure, I rotated the gas tank with the sending unit inside and the ohm meter reads that it changes so it’s not stuck. Any ideas?

 

Fuel gauge goes to empty when the circuit is 0 ohms. You've got a short to ground somewhere.

 

If the brand new sending unit sends perfect signal to the guage out of the tank, but in the tank sends an empty signal but the ohm meter registers it as about half when flipped on its side, where do you think my ground issue is? I have never had this problem before. Any help is appreciated.

 

Maybe verify the float arm geometry is correct, and that the float actually floats? Perhaps the float cracked and sinks.

These are situations where you might have to start pulling and tugging on things while measuring actual resistance across the terminals until it acts goofy and you can isolate the problem.

 

Yes it’s all new, ohms verified fully working as well as the gas guage working. I just don’t understand what would cause the sending unit to work outside the tank but not inside the tank? To test I had the ohm meter verifying that the arm goes up when I rotate the tank empty to simulate fuel levels. Also no issues with the pump itself works in or out.

 

Tell me where I may have missed something:

1. The fuel level sensor measures the expected resistance from end to end of its sweep, with no intermittent opens or shorts;
2. If the fuel level sensor is connected to the wire harness and operated manually, the fuel gauge responds normally;
3. The new level sensor gets installed into the tank, and the complete tank assembly gets mounted to the vehicle, whereupon the fuel gauge no longer operates.

Is all of that correct?

The process of Installing the tank includes providing a solid ground for the tank to the vehicle chassis, and may include moving, pulling, or compressing the wire harness.

If the sensor is connected to the harness in a manner that operates the gauge normally, does that change if a separate/additional ground wire is connected from the tank unit mounting flange and a good chassis ground?

 

That is correct. So it seems that when I put the sending unit in the tank it provides additional ground which causes it not to read on the gauge? Like I said I never seen anything like this. It’s so confusing why would it work outside the tank with no hound besides the 3 wire harness. Then in the tank doesn’t work? I’ve done fuel pump/ sending unit swaps about a dozen times over the years and never had this problem before.

I always test to make sure the sending unit arm moves freely and test with a meter to make sure I’m getting a proper reading on these new sending units. Also I have the same exact result with different tanks and sending units which lead me to believe it’s the car. Could I have a rubbing wire on the body somewhere?

 

I'd disconnect battery and use a multimeter to check continuity between high side of sender (pin B of C313 body pass-thru connector) and pretty much every else it shouldn't be grounded to. Keep in mind the fuel sender and fuel pump share the same ground going back into car (pin A of C313 connector).

Interior Side (hard wired):
BLK: Chassis ground (14 AWG)
PNK: Fuel gauge
TAN / WHT: Fuel pump power (14 AWG)

Exterior Side with Weather-Pack connector:
Pin A: BLK, Fuel Pump & Fuel Gauge ground (14 AWG)
Pin B: PNK, Fuel Gauge
Pin C: TAN, Fuel Pump power (14 AWG)

Be careful, there might be an intermittent short to ground inside that fuel tank, and there are gas fumes in that tank that will go up in a ball of fire if you cause any sparks in the air pocket. A good reason to always disconnect power from car battery.