Fuel Gauge Pinned, Sending Unit Good

I'm back again with another problem in my 2003 that I just can't figure out. My sending unit was bad for the longest time. I dealt with it because it worked ok to a degree. I changed it because of emissions testing. My gas cap was causing a EVAP leak code, and after I cleared the codes I couldn't get the EVAP readiness monitor to set. I read online somewhere that it required the tank to be at a certain level (between 35% and 80% or something), and the sending unit just wasn't giving the right reading. So I changed it.

But it didn't fix the gauge. In fact, it made it worse. When I started the car, the gauge shot up to full (which it wasn't) and stayed there for some 10 or so seconds and then pinned at empty and stayed there. I was worried that the monitor still wouldn't set because it would never be in that range. But it did set that day, and I passed emissions just fine.

But the fuel gauge still does this to this day. I tested the connections based on the specs I found and it tested ok. The sending unit must be sending the right signal to the ECM because it set that monitor. So what is wrong with the gauge? Bad grounding has been suggesting, but where do I look? Could the gauge have gone bad somehow? The rest of them work.

I'm less tired of playing what I call "gas tank roulette" than I am of seeing the "LOW FUEL" light every day. In fact, I think I'm burning that screen out. It seems dim.

You replaced the fuel sending unit only?

Normally, that takes care of the problem....

Now you need a couple resistors(can be had at Radio Shack)........the range is 40-240 ohms(if I remember right.....GMTech1?)......40 ohms should read "E".....140 ohms, should read "1/2"......240 ohms, reads "F"......you disconnect the harness to the tank, and then install each resistor one at a time, on the fuel level circuit ....turn key to just "on".......if the gauge reads ok with the resistors, then the problem is the sender or the wiring from the harness to the sender.....

If it doesn't read correctly, then you need a scan tool......that signal usually goes to the PCM, and then the PCM sends the signal to the cluster....so if the PCM read the resistor right, but the gauge doesn't, it's the cluster......if the PCM does not read the resistor correctly, and the gauge reads what the PCM sees, then it could be the wiring between the harness and PCM or the PCM itself.....

Yep, I only replaced the sending unit. I tested it as well to make sure it was sending the right signal by pulling it back out again (a pain when it's all rusted) and moving the arm myself and it was in spec. I like the resistor idea, I'll try that tomorrow hopefully. That range sounds familiar, but it's been a while and I can't remember where I got my info.

My scan tool is pretty limited in what it reads, and I'm pretty sure it doesn't display fuel level. I'm convinced that the PCM is getting the right info because it finally set the EVAP monitor once the new sending unit was in. We'll see what happens.

Yep, I only replaced the sending unit. I tested it as well to make sure it was sending the right signal by pulling it back out again (a pain when it's all rusted) and moving the arm myself and it was in spec. I like the resistor idea, I'll try that tomorrow hopefully. That range sounds familiar, but it's been a while and I can't remember where I got my info.

My scan tool is pretty limited in what it reads, and I'm pretty sure it doesn't display fuel level. I'm convinced that the PCM is getting the right info because it finally set the EVAP monitor once the new sending unit was in. We'll see what happens.

the evap PCM testing is not done unless the fuel is in the middle range . then you will get a test monitor not being completed. under the current emission rules 1 monitor not tested is allowed . my 2000 silverado I have this as well ..12 yrs now the sender is showing no fuel on cold start. then after the fuel moves around the fuel sender works but the PCM does the test only on a start cycle. so I get this 1 monitor not completed. never got a ck lt on and it has never failed the em test..

if your testing shows you can get the fuel gauge to move with the resistors then the float may be stuck or the sender is wired wrong or defective.

the evap PCM testing is not done unless the fuel is in the middle range . then you will get a test monitor not being completed.
Cool, I'm glad that information was correct. That is what is so strange, the monitor completed after I installed the new sending unit, but the gauge is worse than it was before.

I used a different scanner that showed a high fuel signal voltage code that my scanner would not show, so now I'm even more confused. Apparently that code will set the gauge to empty, but what does this mean? A short somewhere? I tested the incoming voltage and it read 5V.

I have yet to do the resistor test, and the scanner did not show fuel level so I'll have to see if I can get a hold of one that will.

So I found two resistors I had that measured out to 160 and 232 and tried them. No response. Just to be sure, I retested the signal voltage and it was a cool 4.99V. I couldn't get the leads of the tester into the connector to test the ohms the sensor was producing, but I checked it before and it worked and changed with movement just the way it should, so I'm ruling out a bad sending unit as well.

So at this point, I think it's safe to put the carpet back down, nothing is wrong back there. So where do I look next? Should I check the connection at the PCM? There is a voltage being sent to the sensor, so I have to assume that connection is fine. Although, how does that trouble code get set? Ugh...this thing is ridiculous.

So I found two resistors I had that measured out to 160 and 232 and tried them. No response. Just to be sure, I retested the signal voltage and it was a cool 4.99V. I couldn't get the leads of the tester into the connector to test the ohms the sensor was producing, but I checked it before and it worked and changed with movement just the way it should, so I'm ruling out a bad sending unit as well.

So at this point, I think it's safe to put the carpet back down, nothing is wrong back there. So where do I look next? Should I check the connection at the PCM? There is a voltage being sent to the sensor, so I have to assume that connection is fine. Although, how does that trouble code get set? Ugh...this thing is ridiculous.

the 5 volts at the sender output wire means that the sender rheostat in the fuel tank is open...what happens is as the level of the tank drops then the resistance drops this drops the voltage down . the lower the volts the lower the gauge indication.

with a sender wiper contact open/damaged you will get a full pegged indication. 2.9 volts is the normal max voltage ..5 volts being that high will throw a code..

I COULD BE WRONG !

the 5 volts at the sender output wire means that the sender rheostat in the fuel tank is open...what happens is as the level of the tank drops then the resistance drops this drops the voltage down . the lower the volts the lower the gauge indication.

with a sender wiper contact open/damaged you will get a full pegged indication. 2.9 volts is the normal max voltage ..5 volts being that high will throw a code..

I COULD BE WRONG !

No I meant the 5 volts coming in to the sensor. I don't know what is coming out of it, but I'm pretty positive it's whatever it should be. After I noticed the strange gauge behavior when I first turned the key on with the new sending unit, I immediately took the sending unit back out and tested its operation and it worked fine. I was merely saying the sensor is being supplied the proper 5V from the computer, so there is definitely a complete circuit there.

No I meant the 5 volts coming in to the sensor. I don't know what is coming out of it, but I'm pretty positive it's whatever it should be. After I noticed the strange gauge behavior when I first turned the key on with the new sending unit, I immediately took the sending unit back out and tested its operation and it worked fine. I was merely saying the sensor is being supplied the proper 5V from the computer, so there is definitely a complete circuit there.

the voltage 5 volts is when the sender is open .. the voltage at the sender should be max 2.9 volts. when the tank is empty the voltage will drop to 0 volts.

measure across the sender connections. I bet its a high /open resistance. KEY OFF !

place a 20ohm resister across these connections see if the fuel gauge drops..also measure the volts across the sender.

If it doesn't read correctly, then you need a scan tool


Mattyp6 TechII means a type of scan tool used by GM techs called a "TECH 2 or TechII or the new MDI , Not a code scanner.

If the gauge is reading empty, just disconnect the harness, and the gauge should go to Full...if it does that, the problem is between the connector at the tank and the sending unit....

something that just happened to me, that I can probably say is not wrong with yours, i just want to mention this incase someone reading this it might help. My fuel guage became erratic . I put in a 16oz bottle of SEAFOM which I had planned to do anyway and it restored my fuel guage to working perfect. I ran the tank down to about a half gallon went to the gas station , poured in the bottle then got a full tank.

something that just happened to me, that I can probably say is not wrong with yours, i just want to mention this incase someone reading this it might help. My fuel guage became erratic . I put in a 16oz bottle of SEAFOM which I had planned to do anyway and it restored my fuel guage to working perfect. I ran the tank down to about a half gallon went to the gas station , poured in the bottle then got a full tank.

I'm guessing the Seafoam cleaned off the contacts on the sender, which are exposed to the gas in the tank. Mine were beyond dirty, almost worn off, so it's a good thing I replaced the sender.

Just to clarify, the 5V I keep referring to is measured across the harness, not factoring in the sensor at all. Not really sure how to test the voltage coming from the sender, but I did test the ohm range before installing it and it was in the proper range, and using known resistors did nothing to the gauge. It acted the same as it does when plugged into the sender, horribly wrong.

I just picked up a new gauge cluster from the junkyard, but it will be a little while before I have the time to get in there so I thought I'd reopen this discussion for more speculation. What I'm pondering now is the dash light and the corresponding warning dings. What triggers the "low fuel" display and the dings? Is it triggered by a low reading at the gauge or from the computer?

If the gauge is reading empty, just disconnect the harness, and the gauge should go to Full...if it does that, the problem is between the connector at the tank and the sending unit....

It does not. It pins itself at E (after sitting at F for a few seconds) no matter what.

first thing is the harness has 5 volts from the PCM. the sender is nothing but a variable resister. as this resistor moves the resistance changes.

so you installed a new fuel pump assy and the sender unit. problem is the same ? if so then my guess the wiring for the sender is bad. you will have to follow these sender wires back to find the defect.

if the fuel gauge jumps from full to empty then it is an intermittent wiring / connector issue I bet..

Yeah with the new sender the gauge pins at E whether it's plugged in, with a resistor connected, or everything disconnected. I'm thinking though since the EVAP readiness monitor was satisfied, that it has to be the gauge. That's why I'm wondering what triggers the low fuel light, the gauge or the computer?

Yeah with the new sender the gauge pins at E whether it's plugged in, with a resistor connected, or everything disconnected. I'm thinking though since the EVAP readiness monitor was satisfied, that it has to be the gauge. That's why I'm wondering what triggers the low fuel light, the gauge or the computer?

the computer .......gauge is a slave .. pcm is the boss !

the evap system does not use the sender. the in tank pressure sensor does the evap reporting to the computer.

Thanks J cAT and Tech II. Good info on testing. But it just struck me funny when Mattyp6 said " I just picked up a new gauge cluster from the junkyard".

the computer .......gauge is a slave .. pcm is the boss !

the evap system does not use the sender. the in tank pressure sensor does the evap reporting to the computer.
The readiness monitor supposedly requires the fuel to be at a certain level, or it won't set. I could not get it to set with the old worn out sender that would never stay in the middle range (I tried for days), but with a new sender it was set and ready in just a few miles. This is what leads me to believe the computer (PCM, ECM, BCM, whichever it is) is getting a correct reading from the sender, and the problem lies beyond the computer (wiring to the cluster, or the cluster itself). Is there any way to see what the computer is reading for fuel level to verify any of this?
Thanks J cAT and Tech II. Good info on testing. But it just struck me funny when Mattyp6 said " I just picked up a new gauge cluster from the junkyard".

Hah yeah, good point. It was so cheap compared to what I found online, that I had to grab it whether or not I need it. I'm sure it's just fine.

The readiness monitor supposedly requires the fuel to be at a certain level, or it won't set. I could not get it to set with the old worn out sender that would never stay in the middle range (I tried for days), but with a new sender it was set and ready in just a few miles. This is what leads me to believe the computer (PCM, ECM, BCM, whichever it is) is getting a correct reading from the sender, and the problem lies beyond the computer (wiring to the cluster, or the cluster itself). Is there any way to see what the computer is reading for fuel level to verify any of this?


Hah yeah, good point. It was so cheap compared to what I found online, that I had to grab it whether or not I need it. I'm sure it's just fine.

the EVAP testing uses the pressure sensor in the tank. as far as the PCM completing this as a correct result that would be dependent on the fuel level. too low/high the resulting test would not be stored. so since the level is being reported as you say the pcm did store the EVAP system test. next would be the dash circuit .. so I would change that and see if with a different dash panel the gauge work

2003 was a bad year for dash indicators from GM vehicles.