Bad fuel tank reading

[Heyblue]

1984
Ran out of gas with fuel gauge reading just below 1/4. Bought and installed new fuel gauge sender. Tested out of tank and reads full with arm up and empty when down. Reading between one bar and no bars and the reserve light on. Noticed you have to wait for gauge to get a proper reading, i.e. the gauge is delayed from the arm position change. The tank is almost out of gas, just a few inches form bottom of tank. Put everything together and went to drive to gas station to fill up. Just after leaving the drive way noticed the gauge reading just under 1/4 level. Tried different things and discover the gauge reads correctly (one bar and reserve light on) when I turn on the ignition and wait a few minutes. I start the engine and after a few minutes the gauge reads just under 1/4 tank again. The only difference I can conclude is the battery voltage is higher with the engine running since the alternator is charging. Tried turning off the engine and then turning ignition on and the gauge goes back to one bar (since the tank is almost empty. I assume the dash cluster has a regulator to keep the voltage in it at a constant value, I thought 5VDC. I am guessing that power supply may not be working properly. Does anyone know if I am on the right track? One of the issues is I recently paid for a rebuilt dash cluster and it is still under warranty.

 

[Toms007]

It does seem to point toward the gauge cluster......but I guess it could be the new sender too.

 

[Heyblue]

It does seem to point toward the gauge cluster......but I guess it could be the new sender too.

The reason for replacing the sending unit is because the old one was causing a gauge reading if just under 1/4 tank and yet I ran out of gas. Though I didn't do as much testing on the old one, the results are the same. Anyone want a possible good used sending unit

 

[Heyblue]

Still hoping someone might be able to shed light on my problem of fuel gauge reading problem. Thanks

 

[Heyblue]

I also posted this on another similar site with little response also. One person there suggested a bad ground. Not sure how a bad ground could cause the different readings when engine is running verses only the ignition on. But maybe it is ground. The ground is common for both the fuel sending unit and the fuel pump. So maybe it is more the difference with the pump running or not, being it would run while the engine does.
MY QUESTION how do I fool the system to have the pump always running even though the engine is not, but the ignition is on?

 

[Heyblue]

I also posted this on another similar site with little response also. One person there suggested a bad ground. Not sure how a bad ground could cause the different readings when engine is running verses only the ignition on. But maybe it is ground. The ground is common for both the fuel sending unit and the fuel pump. So maybe it is more the difference with the pump running or not, being it would run while the engine does.
MY QUESTION how do I fool the system to have the pump always running even though the engine is not, but the ignition is on?

For those who may have a similar problem, I removed the pump relay and jumper-ed between the hot line and the line going to the pump so with the ignition on but engine not running the pump would run continually. Observed the gas gauge reading with the pump running and not running. The reading changed as it did in the above posts. Soo my conclusion is that maybe the ground from the sending/pump unit is not solid thous maybe changing the reference point for the sending unit. I will try to run a separate ground and test this theory. I will report back when I do that.

 

[Heyblue]

Closer

So had some time this afternoon. I ran a 12 Gauge stranded wire from the sending unit/fuel pump to the chassis and engine close to where the battery grounds are. My gas gauge now reads the same if the engine is running or not and it is reading as close to accurate as I can calculate. So the theory that the ground going to the sending/pump assembly was "soft" so that when the pump is running the ground reference moves enough to give poor readings. Hope this helps anyone in the future with a similar problem.