Fuel Gauge Issue - 454 TBI

[Evil Trailblazer SS]

Yesterday 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.

According to the diagram of my fuel pump/switch/relay/valve system, the gauge gets it's signal from the fuel tank selector valve (Pink wire), is my problem a failing selector valve? How do I test this out to confirm?


Posted via Mobile Device

[blakeduren]

When the gauge pegs out past full, that is a grounding issue with the sending unit. Maybe check near the selector valve?

[Evil Trailblazer SS]

Since the pumps operate, it would not be a ground issue with the selector valve...

[hatzie]

Originally Posted by Evil Trailblazer SS

Since the pumps operate, it would not be a ground issue with the selector valve...

Not entirely true. If the tank ground is not completely bad the higher load component will complete the circuit anywhere it can. The fuel pumps draw way more current than the gauge. Guess who looses.

I've had all kinds of weird behavior from bad grounds. The headlights that didn't work with the switch but flashed with the turn signals on my 79 K20 (RIP) are a good example.

[Evil Trailblazer SS]

I agree with you there, the grounding issues only got worse on the new vehicles...my only concern is that both tanks (one at 3/8 and the other Full) both read Dead Empty when the fuel pumps are not running (Ignition On) and when the pumps are running, they peg out...the function of the fuel gauge is directly related to when the pump relay kicks on and off.
Posted via Mobile Device

[hatzie]

Clip a temporary frame to tank ground onto the pinch weld and frame with vise grips and see what happens.

[Evil Trailblazer SS]

Didn't make a bit of difference...tried it on both tanks. And I also tried a different fuel gauge, it acts exactly the same.
Posted via Mobile Device

[motornut]

in mine open sending unit wire goes full,grounded goes empty

[hatzie]

Originally Posted by Evil Trailblazer SS

Didn't make a bit of difference...tried it on both tanks. And I also tried a different fuel gauge, it acts exactly the same.
Posted via Mobile Device

E = 0 ohms to ground on the signal wire (gauge signal is grounded)
F = 90 ohms to ground on the signal wire
Nailed = Infinite ohms to ground on the signal wire. Open somewhere in the signal circuit.

This happened suddenly. The gauge goes to E with the ignition on and no fuel pumps. Which points to a short that opens up completely when the truck starts.
Turn on the ignition and jumper the fuel pump test lead on the ALDL plug to 12v. Does this NAIL the gauge?

Check the gauge switch inside the selector valve...
Get some Male WeatherPak pins and alligator clip test leads.

Switch to the LH tank then pull the 6 position WeatherPak plug off the selector. The truck will start and run without the selector valve plugged in you just can't change tanks.
Jumper the PINK wire to PINK/WHITE (Left Tank) by plugging in the Male WeatherPack pins and clipping the test lead in place.

Gauge should read 3/8 or F. If it reads "E" STOP you have a short in Circuit 30.

Plug the selector back in and switch to the RH tank then pull the 6 position WeatherPak plug back off the selector.
Jumper the PINK wire to PINK/BLACK (Right Tank) the same way. check the gauge. Should read 3/8 or F.

If the gauge works when you manually switch the senders the micro-switch inside the selector valve is bad.
If it doesn't start checking the pink wires and tank grounds for damage.
Make sure the Circuit 30 pins are fully seated in the bulkhead plug.

[Evil Trailblazer SS]

Okay, my ALDL does not have the fuel pump test lead built in, but I do have the test terminal at the fuel pump relay on the firewall. Connecting this to 12v+ yields no results. But with a test light, it does light up and the fuel pumps do turn on properly. Left when switched to left and right when switched to right.

Now, with the fuel tank/pump selector switch, since my truck is TBI, it will not run with the 6-way weatherpack in the selector disconnected. I was a bit confused on what exactly you wanted me to jumper, but I jumpered the pink wire in the weatherpack to the pink/white wire in the weatherpack as well and it showed no change on the fuel gauge, same with jumpering the pink to the pink/black wire.

This tells us I have a short in the pink wire (circuit 30) from the fuel tank/pump selector valve up to the gauge itself, correct? Now I just have to trace that all the way up and figure it out? Circuit 30 should travel up to the main harness bulkhead and through the firewall, correct?
Posted via Mobile Device

[hatzie]

Originally Posted by Evil Trailblazer SS

Okay, my ALDL does not have the fuel pump test lead built in, but I do have the test terminal at the fuel pump relay on the firewall. Connecting this to 12v+ yields no results. But with a test light, it does light up and the fuel pumps do turn on properly. Left when switched to left and right when switched to right. Now, with the fuel tank/pump selector switch, since my truck is TBI, it will not run with the 6-way weatherpack in the selector disconnected.

If I understand you correctly the fuel pump will not run with the test lead powered and the valve unplugged? That's your problem.

The Fuel Pump Relay feeds voltage to the TAN/WHT wire on the dash switch. The dash switch also has a BLACK ground wire tied to the ground buss bar under the dash.

The dash switch is a polarity reversing DPDT switch. (For an explanation of how this switch works, go to Basic Electricity and scroll down to the 'Double Pole Double Throw' section). When the LH tank is selected TAN 920 is hot and GREY 921 is ground. This powers the LH pump while the RH pump has two grounds... When you flip the switch the opposite is true.
The selector valve actually has a DC motor fed by 920 & 921. It has its' own internal polarity reversing overtravel switch. Once the valve shuttle has switched to LH or RH it shuts off its' motor and doesn't move till you flip the dash switch.

According to the wiring diagram the Fuel Pump feed wires 920 & 921 are spliced into the harness upstream from the valve. So if you start the truck, unplug the valve, and don't change the dash switch position the correct fuel pump should continue to run because you haven't interrupted the current flow.

The selector valve also has a SPDT switch to control which sender feeds the gauge. When you unplug the valve the gauge circuit is interrupted. Circuit 30 to the dash gauge is connected to one of the senders 930 & 931 through the switch. It really looks like the fuel pumps are grounding through the gauge circuit.

Unplug the instrument cluster plug with the fuel pump running. Does the fuel pump die?
If it does try this;

• Ground the tanks by making a paint and rust free spot on the frame and on the tank pinch weld with steel wool or emery cloth.
• Clamp temporary 16ga ground wires with vise grips to all 4 bright spots you've made.
• Jumper the fuel pump test lead so that the pump runs with the valve plugged in.
• Unplug the valve and the fuel pump should continue to run.
• Unclip the test grounds and the pump will die.
  

Originally Posted by Evil Trailblazer SS

I was a bit confused on what exactly you wanted me to jumper, but I jumpered the pink wire in the weatherpack to the pink/white wire in the weatherpack as well and it showed no change on the fuel gauge, same with jumpering the pink to the pink/black wire.

Backfeeding the fuel pump ground through a sender might actually damage it. I'm not sure. If there's no change with a temporary ground...

Originally Posted by Evil Trailblazer SS

This tells us I have a short in the pink wire (circuit 30) from the fuel tank/pump selector valve up to the gauge itself, correct? Now I just have to trace that all the way up and figure it out? Circuit 30 should travel up to the main harness bulkhead and through the firewall, correct?
Posted via Mobile Device

Unfortunately no.

[Evil Trailblazer SS]

Originally Posted by hatzie

If I understand you correctly the fuel pump will not run with the test lead powered and the valve unplugged? That's your problem.

Okay, I misunderstood you, I did not keep the fuel pump test circuit connected when I followed the rest of your directions. I just did and the gauge had no reaction to jumpering the pink/white or pink/black directly to the pink (Circuit 30) in the 6 position weatherpack connector. The gauge pegged past full (3 o'clock position) just like before.

Originally Posted by hatzie

The Fuel Pump Relay feeds voltage to the TAN/WHT wire on the dash switch. The dash switch also has a BLACK ground wire tied to the ground buss bar under the dash.

The dash switch is a polarity reversing DPDT switch. (For an explanation of how this switch works, go to Basic Electricity and scroll down to the 'Double Pole Double Throw' section). When the LH tank is selected TAN 920 is hot and GREY 921 is ground. This powers the LH pump while the RH pump has two grounds... When you flip the switch the opposite is true.
The selector valve actually has a DC motor fed by 920 & 921. It has its' own internal polarity reversing overtravel switch. Once the valve shuttle has switched to LH or RH it shuts off its' motor and doesn't move till you flip the dash switch.

According to the wiring diagram the Fuel Pump feed wires 920 & 921 are spliced into the harness upstream from the valve. So if you start the truck, unplug the valve, and don't change the dash switch position the correct fuel pump should continue to run because you haven't interrupted the current flow.

The selector valve also has a SPDT switch to control which sender feeds the gauge. When you unplug the valve the gauge circuit is interrupted. Circuit 30 to the dash gauge is connected to one of the senders 930 & 931 through the switch. It really looks like the fuel pumps are grounding through the gauge circuit.

Unplug the instrument cluster plug with the fuel pump running. Does the fuel pump die?

You mean unplug the instrument cluster plug that goes into the gauge case/cluster? I've had the entire dash ripped apart three times this week...I guess I can say I am getting good at it, but it sucks still!

I just tried this and the truck runs perfect with the gauge cluster plugged in or unplugged. The fuel level gauge does not flinch when I plug it in or unplug it.

Originally Posted by hatzie

If it does try this;

• Ground the tanks by making a paint and rust free spot on the frame and on the tank pinch weld with steel wool or emery cloth.
• Clamp temporary 16ga ground wires with vise grips to all 4 bright spots you've made.
• Jumper the fuel pump test lead so that the pump runs with the valve plugged in.
• Unplug the valve and the fuel pump should continue to run.
• Unclip the test grounds and the pump will die.
  

Backfeeding the fuel pump ground through a sender might actually damage it. I'm not sure. If there's no change with a temporary ground...

Originally Posted by hatzie

Unfortunately no.

Dammit! I hate when things are difficult!
Posted via Mobile Device

[Evil Trailblazer SS]

Last post updated with more information...

[hatzie]

Originally Posted by Evil Trailblazer SS

Last post updated with more information...

It runs with the gauge unplugged. Good. That narrows it down.
It should stay running with the selector valve unplugged too. If it doesn't then there's a ground issue.

You can test the senders with a DMM set the range close to 90 OHMS. Plug one end onto a good ground , preferably the tank itself, and the other onto one of the pink striped wires. The full tank should be somewhere near 90 ohms and the 3/8 tank somewhere between 20-40 ohms.

You're left with checking over the solid pink wire for damage between the selector and the dash. Nailing the gauge means it's an open rather than a short.

[Evil Trailblazer SS]

Originally Posted by hatzie

It runs with the gauge unplugged. Good. That narrows it down.
It should stay running with the selector valve unplugged too. If it doesn't then there's a ground issue.

You can test the senders with a DMM set the range close to 90 OHMS. Plug one end onto a good ground , preferably the tank itself, and the other onto one of the pink striped wires. The full tank should be somewhere near 90 ohms and the 3/8 tank somewhere between 20-40 ohms.

You're left with checking over the solid pink wire for damage between the selector and the dash. Nailing the gauge means it's an open rather than a short.

Okay, so I got the grounds tested and the gauge does not react to having alligator clip 10 gauge wires on sanded clean surfaces of the tank and frame rails. It still does the exact same thing...

So the truck will run with the selector unplugged ONLY if the test wire lead is connected.

Now, for your next recommendation. I will grab the DMM and check, but do you really think it is at all likely that both senders are failing when it seems that it should be something caused upstream of the tanks, like where circuits 920 and 921 tap betweenm the dash switch and the selector valve?

[hatzie]

Originally Posted by Evil Trailblazer SS

Okay, so I got the grounds tested and the gauge does not react to having alligator clip 10 gauge wires on sanded clean surfaces of the tank and frame rails. It still does the exact same thing...

So the truck will run with the selector unplugged ONLY if the test wire lead is connected.

You have two competing problems.

The gauge behavior doesn't change with the added tank ground. E till the pump is running then nailed. You have a short to ground between the gauge and the selector valve plug in circuit 30.

It ONLY runs with the gauge wire (circuit 30) plugged in either by jumper wire or the valve... You are ONLY getting fuel pump ground from the shorted Circuit 30 wire. You don't have a good frame to tank wire. Alligator clips may not carry enough current for the fuel pump. It should run with the selector unplugged if it has a solid tank ground.

Originally Posted by Evil Trailblazer SS

Now, for your next recommendation. I will grab the DMM and check, but do you really think it is at all likely that both senders are failing when it seems that it should be something caused upstream of the tanks, like where circuits 920 and 921 tap betweenm the dash switch and the selector valve?

I doubt it. Just trying to be thorough and definitely rule out the senders because it's easy to test right at the selector connection.

[Evil Trailblazer SS]

Okay, I built the jumper wires with big 50amp rated copper clips and 10 gauge wire. I doubt they weren't able to carry the load...

What do you suggest I do next, trace the pink (circuit 30) and find where it may be open and grounding?
Posted via Mobile Device

[hatzie]

Originally Posted by Evil Trailblazer SS

Okay, I built the jumper wires with big 50amp rated copper clips and 10 gauge wire. I doubt they weren't able to carry the load...

What do you suggest I do next, trace the pink (circuit 30) and find where it may be open and grounding?
Posted via Mobile Device

HMM. If the tank ground is good it absolutely should run with the valve unplugged. Unhooking the gauge circuit at the selector valve should not affect the fuel pumps. Yours shut off without the gauge circuit plugged in. That's not good.

The PINK wire should be a fairly short run from the selector to the bulkhead plug. The gauge short might be in the cab between the bulkhead and the instrument cluster. Not fun.

[Evil Trailblazer SS]

I don't get it, how would the pumps get their 12v feed if the selector valve is unplugged? The 12v feed to the pumps is fed through the selector valve according to the diagram I posted above...
Posted via Mobile Device

[hatzie]

Grounding the body of the fuel tank might actually do NOTHING for the TBI sender and pump circuit.
It looks like all of the TBI sender/pump wiring including the ground passes through a strain relief in the sender directly to the pump and sender in the tank. Regular senders ground to the tank body.
Dorman R/V TBI Sender picture
See if you can find a single wire with or without a ring terminal bundled into the sender pigtails along with the 2 position WeatherPak. That's your sender and pump ground.

Originally Posted by Evil Trailblazer SS

I don't get it, how would the pumps get their 12v feed if the selector valve is unplugged? The 12v feed to the pumps is fed through the selector valve according to the diagram I posted above...
Posted via Mobile Device

Look closer. The 12v fuel pump feed splices are between the selector valve and the dash switch. The grey and tan circuit 920 & 921 wires are a pair of "Y" wires. They actually have 3 terminations on one wire. 1 at the fuel pumps, 2 at the selector valve and 3 at the dash switch with a spliced connection between the valve and dash switch to make the "Y". If the pump power fed through the selector valve there would be 2 more connections at the valve and an extremely complicated microswitch assembly inside the valve for pump power.

A picture's worth a 1000 words. I marked up your diagram and colorized circuits 920, 921, & 30 to make the wires easier to follow. The PDF is more readable than the cut down JPG.

[Evil Trailblazer SS]

Okay, I got it now...thank you for spending so much time with me on this! I will pop the bed and check with the grounding and circuit 30...
Posted via Mobile Device

[Evil Trailblazer SS]

Okat Hatzie,

new problem same issue related. On my way to work Thursday night (booo! Night construction work) and the truck blew the ECM B Fuse. As I have been messing with the selector valve this morning trying to trace down the issue and go through all the diagnosis you talked me through before, the fuse blew again. Does this help in your diagnosis at all?
Posted via Mobile Device

[hatzie]

Originally Posted by Evil Trailblazer SS

Okat Hatzie,

new problem same issue related. On my way to work Thursday night (booo! Night construction work) and the truck blew the ECM B Fuse. As I have been messing with the selector valve this morning trying to trace down the issue and go through all the diagnosis you talked me through before, the fuse blew again. Does this help in your diagnosis at all?
Posted via Mobile Device

Either the fuel pump is drawing too much current now or there's an intermittent short somewhere in circuit #120.

[Evil Trailblazer SS]

Okay, so I got it figured out...Circuit 120 (white/tan) has a wire that runs all the way down to the top of the LH fuel tank. It shares a connector with the pink fuel level sending unit wire and basically dead ends into the 2-way weatherpack. When I pulled the bed to change the pumps I somehow squished both of them between the frame and the bed when I put everything back together. So I replaced the squished parts of both wires and the pumps work flawlessly and the ECM-B fuse is not blowing.

The problem now is that the fuel gauge is haywire. LH tank reads E and RH tank reads past F. As I drive, the gauge will bounce around randomly. Did I fry the sending units in the tanks?
Posted via Mobile Device

[rfmaster]

Originally Posted by Evil Trailblazer SS

Okay, so I got it figured out...Circuit 120 (white/tan) has a wire that runs all the way down to the top of the LH fuel tank. It shares a connector with the pink fuel level sending unit wire and basically dead ends into the 2-way weatherpack. When I pulled the bed to change the pumps I somehow squished both of them between the frame and the bed when I put everything back together. So I replaced the squished parts of both wires and the pumps work flawlessly and the ECM-B fuse is not blowing.

The problem now is that the fuel gauge is haywire. LH tank reads E and RH tank reads past F. As I drive, the gauge will bounce around randomly. Did I fry the sending units in the tanks?
Posted via Mobile Device

Disconnect Pink/white and Pink/black circuits at respective sending units (930/931) >> junction B. Measure resistance of the sending unit inside each tank. Should be between 90 and 0 Ohms depending on the fuel level. If it reads very high resistance or infinity that's the source of your problem.
//RF