Simple fix?

   
2000 GMC Safari Van

Whew.... Its been a tough week. But here it is 4:30 and Friday... I'm all caught up and if I play my cards right, I can sneak out the back door and start Monday with a clean slate. But no such luck... (bg) The dispatcher caught me as I was cleaning up and handed me a problem child.

The van has been in and out of the shop for the last couple of months and the customers concern has not been fixed. Intermittently the fuel gauge will do a dance and then either get stuck on full or on empty. The fuel pump assembly (which includes the level sensor) has been replaced. Twice the instrument cluster (IPC) has been replaced. This was all done by our 15a department which handles most of the body trim and electrical work.

My mission (whether or not I accept it or not) is to fix the concern. The shop is buying... Just get it fixed. Well, there is still a little time in the day. Friday night traffic is a bummer anyway, so lets take a look.

Currently the fuel gauge is reading full. Take a look at some of the scan data. I've only listed four PIDS in order to clean things up.

Click here for instructions on using the snapshot plotter below

I shook the van violently and waited for a moment.... the fuel gauge began its dance and then I saw this:

Now would be a good time for a schematic and a short description of how the fuel gauge works.

From the service manual:
The vehicle control module (VCM) requires an accurate indication of fuel level for evaporative emission (EVAP) system diagnosis. The fuel level in the fuel tank changes the rate of vacuum decay for the EVAP system leak DTCs.

The fuel level sensor changes resistance based on fuel level. The fuel level sensor has a signal circuit and a ground circuit. The VCM applies a voltage of about 5 volts on the signal circuit to the sensor. The VCM monitors changes in this voltage caused by changes in the resistance of the sensor to determine fuel level.

When the fuel tank is full, the sensor resistance is high. Therefore, the VCM will sense a high signal voltage. When the fuel tank is empty, the sensor resistance is low and the VCM senses a low signal voltage.

The VCM uses the input from the fuel level sensor to calculate the fuel level in the fuel tank. This information is then sent to the instrument panel (IP) cluster through serial data.

Ok.... Sounds straight forward enough.

With a full tank, the fuel level sensor will have a resistance of 250 ohms. With an empty tank the resistance drops to 40 ohms. There is a DTC P0463 ( Fuel Level Sensor Circuit High Voltage) stored but it may have been set on the last visit.

What would you do with this? Can we trust the scan tool data? Using a DMM to ohm out a circuit puts very little load on it. If I unplug the rear harness and ohm out the level sensor, can I trust the results?
      

 End part one

         
From part one:
What would you do with this?

If you thought to give it back to the guys who had been throwing parts at it before.... I tried. (bg) So it looks like I have to see what I can do.

From the forum:
You didn't complete the diagram but there are several scenarios that could play out.

Correct. But I did post the schematic that is shown in the manual. Here is how the manual describes it.

This would never work. No matter what the fuel level was, you would always have a 5v fuel level signal. Here is what we really have. Well, sortof, there is more later...

Now lets do one better. With the fuel tank on full, the level sensor resistance is 250 ohms. With an empty tank, the level sensor should read 40 ohms. A full tank for this vehicle should register 2.5 volts at the PCM. Knowing this we can calculate the PCMs input resistance.

Ok.... Now lets think for a moment. The fuel tank is either full or empty. We either have a false full or false empty value.

For a moment lets assume that the tank is empty and that our high voltage was due to a faulty level sensor. Does this fit?

Now assume that the tank is full and that our low voltage is due to ?? .... Does this fit?

To be honest, you could put up a good argument for either one.... I think the latter situation (full tank) is the more likely. What to do? Why call the customer up of course! Doing so confirmed my suspicions. The tank is full. So now we need to explain how the sensor voltage could drop like this.

Could it be the level sensor? It would have to be shorting out.... Not likely at all. But we could possibly have sensor wiring shorting to ground somewhere

Here the sense voltage could vary anywhere from 2.5v to 0v. Not very likely a scenario but it could happen. Now what would happen if inside the PCM the input resistance increased?

This would indeed drop our sense voltage.... Or how about this, what if the 5v applied to the circuit degraded to a lower value?

Yes, that could cause our problem...

If we can assume that there are no wires grounding out on the frame etc, can we say for sure that the PCM is faulty? Lets look further at this circuit... While there is no telling how many 5v regulators are in the PCM, those that are supposed to know tell me its like this...

Now, what if one of those circuits that use the 5v reference shorts it out? The regulator can only handle so much current until it can no longer supply 5v.

That could do it.... If one of the sensors was killing the 5v regulator you could see a low fuel level sensor voltage. However... You would almost definetly see a DTC for one or more of those sensor circuits.

The snapshot that I presented in part one had most of the PIDS removed (by me). In it you could see the MAP sensor voltage hanging out in the normal range. I think we have just proved that the PCM is faulty. Assuming of course that we don't have the level sensor signal wire grounding out somewhere....

I replaced the PCM and had the van driven 100 miles. So far so good... (bg)
  

 Copyright © 2007 Flatrater Diagnostics