Fuel Pump / Wiring Harness Connector Problem - SOLVED

[houstontaylor]

I have a 2003 Chevy Tahoe with about 360,000 miles. A few years back I had a dealer replace the fuel pump as a precaution because I was traveling long distances and it had high mileage. On a long trip soon after that the car shut down on the highway and the problem turned out to be a short in the wiring harness connector, apparently, which shorted out the pump. The dealer in the town near where it broke down installed a new pump and wiring harness adapter with the original style square connector. At some point after that the fuel guage began to give suspect readings and has worsened so that it often gives an empty or full readings when the fuel level is known to be not near those levels at all. The fuel pump (again), instrument cluster, and ECM have been replaced. The sending unit wiring runs through the computer on this vehicle. Since the problems have been occurring more frequently now, I was able the disconnect the pump sending unit wiring harness under the fuse box and the resistance in the rear half of the circuit to the tank showed open circuit when the fuel guage showed empty. I suspect a bad fuel tank wiring harness adapter splice but haven't taken it to a shop yet to have that redone and see if that solves the problem. Before I go to that expense I would like to ask if anyone else has seen problems with fuel pump wiring harness adapter or if there could be another reason for this problem. Thanks.

 

[OBSandaNNBS]

It's always a freaking wire with a Chevy...I can't really help but hope you got this solved!

 

[strutaeng]

If you have a bi-directional scanner, you can monitor the voltage the fuel sender is putting out. There's a table of voltages correlating to fuel tank levels somewhere I've seen posted, maybe by @swathdiver ?

If not, you could back probe the ECU pins for the sender (look up a wiring diagram) and check with a DVOM. I did that once on my 04.

But your pump isn't causing problems, right? Just the fuel tank level is the issue?

360k?! Pretty impressive TBH. Is that the original engine?

 

[swathdiver]

I have a 2003 Chevy Tahoe with about 360,000 miles. A few years back I had a dealer replace the fuel pump as a precaution because I was traveling long distances and it had high mileage. They also installed the standard wiring harness adapter for the newer pump style. On a long trip soon after that the car shut down on the highway and the problem turned out to be a short in the wiring harness adapter which caught fire and shorted out the pump. The dealer in the town near where it broke down installed a new pump and wiring harness adapter. At some point after that the fuel guage began to give suspect readings and has worsened so that it often gives an empty or full readings when the fuel level is known to be not near those levels at all. The fuel pump, instrument cluster, and ECM have been replaced since the sending unit wiring runs through the computer. Since the problems have been occurring more frequently now, I was able the disconnect the pump sending unit wiring harness under the fuse box and the resistance in the rear half of the circuit to the tank showed open circuit when the fuel guage showed empty. I suspect another bad fuel tank wiring harness adapter splice but haven't taken it to a shop yet to have that redone and see if that solves the problem. Before I go to that expense I would like to ask if anyone else has seen problems with fuel pump wiring harness adapter or if there could be another reason for this problem. Thanks.

If you're still working on this, sign up for the shop manual for a month at ALLDATADIY. See if the truck is throwing codes for this and then follow the troubleshooting guide(s).

My father-in-law is experiencing similar issues right now, the gauge on the cluster is not moving but has been tested to be good and so are the sending units in his fuel tank (Cadillac). The issue is probably a poor wiring repair near the fuse box under the hood where the car was in an accident. It has not been confirmed yet by more abler hands.

 

[houstontaylor]

If you have a bi-directional scanner, you can monitor the voltage the fuel sender is putting out. There's a table of voltages correlating to fuel tank levels somewhere I've seen posted, maybe by @swathdiver ?

If not, you could back probe the ECU pins for the sender (look up a wiring diagram) and check with a DVOM. I did that once on my 04.

But your pump isn't causing problems, right? Just the fuel tank level is the issue?

360k?! Pretty impressive TBH. Is that the original engine?

Not original engine. Replaced with a remanufactured one at 290,000 miles.

Not original engine. Replaced with a remanufactured one at 290,000 miles.

At one point I thought that the problem might be the little corrosion on some of the ECM pins due to a pinched rubber seal but that wasn't it. And the 5 volt reference line out of the computer showed 5 volts when the circuit was disconnected under the engine compartment fuse box.

 

[houstontaylor]

Thanks for the ideas! It turns out that the original design fuel pump from AC Delco and the Delphi one before it apparently have design issues. Both pumps, one from the dealer, gave erratic readings and were new.

If you're still working on this, sign up for the shop manual for a month at ALLDATADIY. See if the truck is throwing codes for this and then follow the troubleshooting guide(s).

My father-in-law is experiencing similar issues right now, the gauge on the cluster is not moving but has been tested to be good and so are the sending units in his fuel tank (Cadillac). The issue is probably a poor wiring repair near the fuse box under the hood where the car was in an accident. It has not been confirmed yet by more abler hands.

The standard pump design with the square harness connector (2x2 they call it) at the tank has problems with the electrical connector, float location, and possibly more. It wasn't just the wiring harness connector. The splice from the dealer where it broke down was okay. When the shop I was working with installed a redesigned model that is on the market with a flat wiring harness connector (1x4 they call it) the problem was solved! The shop said that in addition to the new connector design, the sending unit float was closer to the center of the tank and can't rub on the walls of the tank and throw off the readings. They said the new design pump, which I found on the NAPA website, is part number MU1615. The new pump came with a new flat connector wiring harness adapter. There is a reason that someone redesigned the pump, apparently.

 

[houstontaylor]

Sorry for late response but was waiting to try another solution. I had thought the original square design electrical connector had been replaced already but it had not. The problem was SOLVED with a new design fuel pump/sending unit. I, and the shop I was working with, was having a hard time imagining that two brand new pumps with sending units could be bad but there was eventually enough evidence to try the resigned pump (part MU1615 per the shop) with a flat electrical connector (1x4 they call it) instead of the square connector (2x2 they call it) I found on the NAPA website (the shop said the new pump float was also closer to the center of the tank, as described in the comment above). The old float may have been hitting the side of the tank, causing more error in the readings. It seems there was a problem with the square harness electrical connector design on the originals. It was harder to find the problem when the gauge issue was more intermittent but it finally got so bad that the gauge needle spent most of its time on empty, fortunately. I had separated the pump wiring harness connector below the engine compartment fuse box and checked resistance on the small purple and black wires for for the sending unit. It showed open circuit on the rear part of the circuit to the tank when the gauge showed empty. Eventually I tried hooking up an ECT short circuit/open circuit finder from Power Probe and traced the wires down the frame rail to where the wires for the pump separate and go over the tank. It showed the open circuit to be over the tank (the connector it turns out). The shop found that when they wiggled the pump electrical connector, the gauge reading fluctuated. When the new harness connection adapter was installed, they soldered the connections and sealed them with heat shrink tubing with the internal sealant, instead of using the crimp butt connectors that came with the kit. The problem with faulty brand new parts is that we, including a dealer, initially ruled out the pump/sending unit as the issue and hence I ended up replacing the instrument cluster and ECM as part of the search earlier on. At least you all will know something else to check now that is not obvious it you have the erratic fuel gauge issue.

 

[houstontaylor]

Not original engine. Replaced with a remanufactured one at 290,000 miles.

At one point I thought that the problem might be the little corrosion on some of the ECM pins due to a pinched rubber seal but that wasn't it. And the 5 volt reference line out of the computer showed 5 volts when the circuit was disconnected under the engine compartment fuse box.

 

[houstontaylor]

Here is another point I forgot to make. Previously when filling the car I thought the gas pump was shutting off early but it turns out that the tank was actually full and the guage was showing 1/4 low. I found this out by running a video probe into the gas tank filler hose (the screen had been removed or opened up by a shop) and saw that it was full of fuel.

 

[West 1]

The GM Fuel Pump Connector is a known failure issue. Any quality Fuel Pump will come with a new improved connector for you to splice into the OEM harness.

 

[Fless]

The GM Fuel Pump Connector is a known failure issue. Any quality Fuel Pump will come with a new improved connector for you to splice into the OEM harness.

True. I'm a bit surprised that pumps with the old connector are even available. Thought they would all be updated by now.