Dual Tank 1987 TBI Fuel Gauge Issue - Please Help

[Evil Crew Cab]

So, for the last month and a half (or longer) I have been troubleshooting an issue that simply keeps evolving and getting more and more complex.

One night, the fuel gauge was working perfectly, I drive home and shut the truck down. When I went out a couple hours later, the fuel gauge pegs way past the full line. All the way to the 3 o'clock position.

Now, I figured out a weird thing is happening with it. When I turn the ignition on but don't start the engine, the gauge reads dead empty, whenever the pumps are running (fuel relay engaged) the gauge pegs out to the 3 o'clock position. This is happening with both fuel tanks, not just one of them.

Then it progressed to my ECM-B fuse blew for no apparent reason, replaced fuse and drove on for a couple weeks. Then it blew again. After that is became more frequent and the fuel gauge decided to start reading way past Full (3 o'clock position).

This is a diagram on the TBI dual tank set-up on the >8500gvwr trucks.

You must be registered for see images attach

When the truck was running, I would pull the plug on the tank selector and the truck would die. This isn't good because the pumps should run fine when the selector valve is disconnected, it just won't change over tanks/pumps.

We did a bunch of diagnosis and ended up finding that when I had pulled the bed to change the pumps/pick-up socks when I first got the truck, I had inadvertently smashed the pink fuel gauge wire (circuit 30) between the frame rail and the bed. If you look closely on the diagram next the the fuel gauge, it shows that the pink wire (circuit 30) is in the B location of a 2-wire weatherpack connector. It shows that the A location is "Not Used". But, my truck has a tan/white wire (circuit 120) that dead ends into that 2-wire weatherpack terminal.

So what we found is that the tan/white wire was sending the pink wire straight to ground causing the fuel pumps to back feed ground through the in-tank sending units and ground via the smashed section of the pink wire (circuit 30) robbing ground from the sending units.

Well, I got all of this squared away but ever since then, the fuel gauge only reads way past full (3 o'clock position) until I nearly get the fuel level to empty. Then it reads a tick above E.

We traced the pink wire (circuit 30) from the 2-wire weatherpack connector all the way to the firewall bulkhead and could not find any sort of cut or rub marks or anything suggesting a short to ground in the wire. Now the only thing I can think of is that when the pumps were grounding through the fuel level sending units, I somehow friend the sending units, is that possible or even likely. I cannot get a good reading to test Ohm's on the sending unit and gave up after frustrating myself for 4 hours last weekend. I am normally very good with electrical diagnosis but I am stumped on this one! I also pulled the plug off the tank selector valve and jumpered the pink/white and pink/black wires directly to the pink wire (circuit 30) to test if the sender micro-switch was faulty inside the selector valve and this yielded nothing...I have another gauge cluster (non-tach) with a functioning fuel gauge and it acts the same way so I know the gauge is not bad. It just doesn't make sense that both sending units would fry at the same time and cause the same problem.

Someone please help!

 

[HotRodPC]

I'd have to dig way deep into my memory, but IIRC, if you get an infinate FULL pegged, that means you have a dead short in the wire to the guage somewhere. Meaning the sending wire to the guage is grounding out on something.

So does it do this when switched on one tank? or does it do it on both tank??? If its on one tank, then follow the wire from that tank to the switch. If its on both tanks, then it should be between the switch and the guage.

Someone please correct me if wrong, cuz I am not 100% certain about this. Just trying to jog the old memory.

 

[Evil Crew Cab]

You are correct, but the weird thing is, when the tank gets low, the gauge rears a tick above E...in both tanks...

Plus I have traced the pink wire (circuit 30) which is the wire from the selector valve all the way to the gauge. No visible problems with it and it worked perfectly until the smashed wire incident...

 

[HotRodPC]

Now that would make me wonder if its a faulty sending unit in one of the tanks, but then why would it make both tanks read wrong??? Unless each tank used the other for a ground??? hmmm You're right, that is weird. But when the tank gets low then it reads low, that means its should be losing its grounding out which to me would mean the sending unit it grounding out at just above empty. Which tank gets low then the guage goes to E? That is likely the tank with a bad sender if that is indeed the issue. Just thinking outloud.

 

[Evil Crew Cab]

The weird thing is, both tanks have their own grounds. As you can see in this photo, the sending units simply have 12v+ feeds going into the bucket and the sending unit and pump are both grounded to the top of the bucket from the inside. You will see the ring terminal wire next to the weatherpack connector, that is the ground wire for both the sending unit and the pump in the bucket. This single wire is ground to the frame rail with it's own bolt. I sanded and cleaned both points on the frame rails and still nothing.

You must be registered for see images

Over on the **** site, one of the members was helping me through this and this is what he though, which I went through all possibilities and nothing fixed it...

You must be registered for see images attach

 

[HotRodPC]

hmmm, maybe the switch itself? But you can switch back and forth from one tank to another, and both tanks are working, just not reading the guage right?
For some reason, that would just seem to me the problem has to be at the switch or up to the guage since both tank act the same. If one tank was wrong and the other was right, then it would seem to me the problem is from that tank up to the switch. I dunno bud, hopefully someone else will chime in soon with a different outlook and another suggestion.

 

[Evil Crew Cab]

The dash switch is a brand new GM piece and the selector valve isn't the problem because even when I bypass the level sensor signal by pulling the 6-way connector off and jumpering the signal wire directly, it still does it...

 

[crazy4offroad]

Sounds like you need a diode (kind of like an electrical check valve) on your gas gauge sender wires or somewhere, one is overriding the other.

 

[Evil Crew Cab]

I even just went out and ran a completely new pink wire (circuit 30) from each sending unit up to the firewall bulkhead to find no difference. I really think it has to be the sending units in the tanks, but that just seems odd to me!

On a side note, the selector valve does not switch over when I flip the dash switch from tank to tank unless the truck is running, that doesn't seem right as all my previous trucks have done that...

 

[Evil Crew Cab]

Sounds like you need a diode (kind of like an electrical check valve) on your gas gauge sender wires or somewhere, one is overriding the other.

A diode will not help in this situation unless the micro-switch in the tank selector valve is bad. The whole point of the level sensor signal going throught the selector valve is that it will only feed one of the sending unit signals to the gauge at a time.

 

[crazy4offroad]

OK, I only have experience with the older ones where the sender wire goes to the selector switch. But it sounds like you might have zeroed in on the problem. Any troubleshooting or toll-free helpline to call?

 

[Evil Crew Cab]

...Any troubleshooting or toll-free helpline to call?

I wish...even just finding some of the wiring schematics proved to be a nightmare! The one I showed above with the Hot Fuel Module (>8500gvwr & 454 TBI) were so obscure that some of my GM tech friends had no clue about them!

 

[crazy4offroad]

Weird. I know GM fuel sending units are mostly supposed to be 90 Ohm but not sure if there are other ratings, is it possible 8500+gvwr trucks have a different rating or gauge? Are the sending units something like a poteniometer? I know they (pot switches) can fry easily if overvolted or grounded wrong. Otherwise I would start testing the transfer valve. Does it get ground from the mount? Just throwing things out to look at.

 

[Evil Crew Cab]

I already looked at that and tried eliminating the selector valve from the equation by manually jumpering a wire between each tank's sending unit feed directly to the pink wire (circuit 30) and found no change. So the selector valve cannot be the issue. At this point I have done days of diagnosis with no firm answer, so I just ordered brand new sending unit assemblies for both tanks. The stock ones are a little beat up and show their age...plus the ground was re-designed to a better set-up so maybe that will help with grounding...

And no, the >8500gvwr trucks all had the same Ohm relations for the gauge as every other truck.

 

[Evil Crew Cab]

Okay, apparently I did fry the sending units when the pink wire (circuit 30) got smashed and grounded between the frame and bed...