Fuel Sending Unit Repair

[Keyav8r]

My 91 (HO) MJ fuel gauge has not worked correctly since I have owned it. I had previously tried to check the ohm output by reaching in through the filler neck with a bent coat hanger (the universal tool) and moving the float arm through the range. I got low and high level readings that seemed to be good and some intermediate readings that looked close. I had previously checked out the gauge and wiring from the tank to the gauge by unplugging and grounding the connector and by rigging a connector that would allow me to add resistance in steps to check the full range of the gauge. Gauge and wiring checked out okay.

Next step - I dropped the tank and removed the fuel pump/sending unit assembly.

Here it is on the workbench.

With full access to the sending unit I found that the ohm output was very spotty because the pivot point on the sending unit was loose due to wear and allowed the float arm to "flop" around, which caused the movable contact (part of the float arm attachment) to move away from the stationary part.

This is the point where pure dumb luck comes into play. While looking for a fuel pump connector to rig to test the gauge and wiring, I crawled under a 91 XJ at a P&P and found the fuel pump/sending unit laying on the ground, less the fuel pump. I took the sending unit off the bracket with the thought I might be able to "rig" it to repair mine.

This is my MJ sending unit on the left and the 91 XJ unit on the right. The obvious physical differences are the longer float arm and the longer blue wire on the MJ unit. The black wire is the ground to the bracket and is the same on both.

I hooked each one to my multimeter and checked out the high and low readings.

MJ unit high reading.

MJ reading low reading (had to hold the arm in place to get an ohm reading).

XJ unit reading high reading.

XJ reading low reading.

The XJ sending unit ohm range is pretty close to the MJ range and the polarity is the same, so it should work in my MJ.

Next, to address the float arm length and the blue wire length.

First, I carefully pried the float arms out of the MJ unit and the XJ unit and swapped them.

That last photo is the XJ sending unit with the MJ float arm snapped into place. Be careful in removing and replacing the float arm in the sending units - 25 year old plastic pieces.

Next I used a soldering iron to remove the blue wires from both units and put the longer wire on the XJ unit that now has the MJ float arm.

Last step - held the sending unit vertical and moved the arm throughout its range to assure I still had good ohm readings at all points. I checked the ohms with both digital and analog meters (per Don's recommendation).

In the next few days I'll sandblast and paint the bracket with Klass Kote epoxy (thanks to Akula's post for the coating info)

As soon as my new fuel pump arrives from Amazon (thanks again, John), I'll reassemble the pump and sending unit onto the bracket, install it in the tank, and re-install the tank.

Hopefully, this will give me a working fuel gauge.

I'll post a final report as soon as that is done.

 

 

Sent from my iPad using Tapatalk

[Hippie66]

Very good write-up. Just curious as to why you dropped the tank to pull the sending unit?

[Zambeezy]

My 91 (HO) MJ fuel gauge has not worked correctly since I have owned it.

What was your fuel gauge doing? Mine just shows empty all the time, no matter how full the tank is. It's next on the list for troubleshooting but I'm not sure where to start

[Keyav8r]

Hippie66 - I dropped the tank because I wanted to get a good look inside and also wasn't sure how much fuel was in it. Turned out it was a lot more than I thought. And, the sock came off the pump when I pulled the assembly. Could have dropped the driveshaft, but decided to drop the tank instead. Also gave me a chance to recheck the vents. When I put them in they were a little loose but they've tightened up pretty well. I guess the gas fumes swelled the rubber grommets a little.

Zambeezy - The fuel gauge would not go to full with a full tank. Sitting still and engine running it would show 1/2 to 5/8 tank then drop to around 1/4 tank when I started moving. Never went close to empty, but the low fuel light would come on sometimes. i just fueled up at 150 to 200 miles.

[Hippie66]

The grommets are cheap from Summit. I replaced mine when I put a new tank in.

[cruiser54]

Nice write-up!!

[dinghyboy]

Zambeezy - The fuel gauge would not go to full with a full tank. Sitting still and engine running it would show 1/2 to 5/8 tank then drop to around 1/4 tank when I started moving. Never went close to empty, but the low fuel light would come on sometimes. i just fueled up at 150 to 200 miles.

Keyav8r:  Thanks for the write up.

 

My '92 Gauge never goes above 3/4 and never goes blow 1/4. It seems to operate normally between those two ranges lowering relatively steadily. I've always filled up when it hits 1/4 (not knowing how low it could go) and it's never taken more than 12 gallons. I didn't even know i had a low fuel indicator.

 

Judging from your resistance readings; If I disconnect the connector back at the fuel tank, I take it the gauge should got to full? If I then ground the gauge wire to body would it go to empty?  Is the gauge wire easy to identify at the connector?

 

Thanks.

[Keyav8r]

dinghyboy - IIRC, the center (purple or blue) is the gauge wire. The black is the ground and the brown or orange is the 12V. I would go look at mine to confirm but it's at the office because I plan to use the shop blast cabinet to clean it prior to coating.

I believe you're right on the testing by disconnecting and grounding. You can probably do the grounding by jumpering across the center pin (or socket) and the ground pin (or socket) on the body half of the connector. You can also add resistors in the circuit (if you can find them, Radio Shack is almost extinct in our area) to get intermediate readings.

[Keyav8r]

Thanks, Cruiser!

[cruiser54]

Get it all sorted out and it will be under "Helpful links" on my website if that is okay with you. 

[64 Cheyenne]

You can also add resistors in the circuit to get intermediate readings. Huh?

 

Could somebody expand on this? If my understanding is correct, your using a resistor with a set known value to check against the output value of the sender, to more or less calibrate the sender? Where and how is the resistor wired to test? And once satisfied with the results, does a person ever need to bend the float arm or otherwise change the orientation level of the float in the tank to read more or less true readings?

[johnj92131]

 

You can also add resistors in the circuit to get intermediate readings. Huh?

 

Could somebody expand on this? If my understanding is correct, your using a resistor with a set known value to check against the output value of the sender, to more or less calibrate the sender? Where and how is the resistor wired to test? And once satisfied with the results, does a person ever need to bend the float arm or otherwise change the orientation level of the float in the tank to read more or less true readings?

 

 

You can bend the arm to change the low fuel point on the fuel gauge.  But it is a whole lot of work to try and get right.  I did that with my Passat, but found it much more work than I gained in accuracy. With the last 4 to 6 gallons in the tank, fuel sloshed all over the place, a small incline would change the gauge reading.

 

With the Comanche, I took a different approach.  Every time I saw the low fuel light go on, I wrote down how many miles I had on that tank right then.  When I filled the tank, I computed the MPG for the entire tank.  With a little math, I was able to calculate how many gallons of gas was left in the tank when the low fuel light went on.  For my truck, I usually have 6.5 to 7 gallons left when my low fuel light goes on. 

 

I have done the same excercise with half full, quarter full and right on the "empty" line.  So I have a good idea of how much fuel is left most of the time.  I have put as much as 24.2 gallons of fuel in the Comanche with out running out of gas.  But, I do not suggest anyone try this unless your are willing to get stuck on the side of the road.

 

Here is an example of the math I used.  At 340 miles, the low fuel light goes on.  I fill the truck up at 420 miles with 21.00 gallons of gas. (420/21= 20 mpg), then 330/20mpg = 17 gallons when the low fuel light goes on.  Full tank = 23.5 gallons, subtract 17 gallons when low fuel light goes on = 6.5 gallons left when low fuel light goes on. 

 

I worked this out last summer over the course of two Pacific Ocean to Atlantic Ocean back to Pacific Ocean trips and it worked out very well.

[HOrnbrod]

 

You can also add resistors in the circuit to get intermediate readings. Huh?

 

Could somebody expand on this? If my understanding is correct, your using a resistor with a set known value to check against the output value of the sender, to more or less calibrate the sender? Where and how is the resistor wired to test?

 

The best way to "calibrate" the sender is to check the potentiometer resistance with an analog ohmmeter. The range of the fuel sender potentiometer (HO in this case) is from 0 - 105 ohms. The - meter lead is connected to the grounded end of the pot's winding, and the + lead goes to the pot slider terminal. The pot slider is what moves when the float arm rises and falls and sends a varying resistance to the fuel gauge. What you are looking for is 0 ohms at one extreme (full indication), and ~105 ohms at the other (empty tank) with a smooth transition, i.e. no dead spots along the path.

 

The only use for a fixed resistor I can see is to test the circuit for fuel meter deflection based on overall circuit resistance. This is done by substituting a known fixed value resistor for the potentiometer, say 33 ohms, which should indicate a 1/2 full tank on the gauge according to the FSM. The resistance picked off the pot is non-linear due to the irregular arc of the float arm. If you get a half full indication on the fuel gauge at 33 ohms, the circuit is good. But every solder joint, connector, and corrosion point in the circuit will add resistance, so the fuel gauge will normally indicate something less than 1/2 full because on an HO sender, more resistance equals less fuel gauge needle deflection. This is opposite from the Renix fuel gauge, as more circuit resistance results in more meter deflection. Thus the HO and Renix fuel gauges are of opposite polarities. All the dash gauges read varying resistance from their respective senders to vary the meter movement coil and indicate and none are known for their high degree of accuracy. It is what it is........

[Keyav8r]

About adding resistors to get additional readings - while testing my fuel sender I opened the circuit to get the empty reading and grounded the lead to the gauge to get the full reading. As an experiment I picked up a fuel pump connector at P&P and wired the ends together (with the exception of the lead to the gauge) so that I could splice it into the circuit.

Then I bought some resistors (10 ohm and 100 ohm) that I could jumper into the circuit to add resistance and drive the gauge to read other fuel levels.

I had an approximate 30 ohm resistance from the sender which gave a gauge reading of about 5/8 full.

Adding one 10 ohm resistor gave an approximate 40 ohm resistance. Adding two in series gave an approximate 50 ohms. Putting two 100 ohm resistors in parallel in the circuit gave an approximate 80 ohm reading. Et cetera, et cetera. The ohms are approximate as the resistor values approximate (10 ohm resistor metered at around 9.6 ohms, 100 ohm metered at around 96-97 ohms).

Adding the resistance in the circuit drove the gauge to corresponding fuel level readings back toward empty, confirming my assumption that the gauge was good throughout the empty to full range. That could have been assumed from the empty and full readings, but I just wanted to do the experiment.

 

 

Sent from my iPad using Tapatalk

[64 Cheyenne]

Ahhhhhhh I see....

Thanks for the clarification

[Keyav8r]

Continuing my sending unit repair - I sandblasted the fuel pump/sending unit bracket and painted it with gasoline resistant paint.

When my new Bosch fuel pump arrived, I installed on the bracket.

The Bosch pump is shorter than the old pump but it comes with a white spacer and two rubber washers to adapt it to the bracket. You'll need to open up the inside diameter of the top clamp and trim 3/16" - 1/4" off the "fins" on the white spacer to get a good fit on the bracket. Sorry, I didn't make photos of those steps but I believe Knucklehead did a write-up with photos on this.

Next, I cut the top connector off of the fuel pump wires and spliced them into the original wiring.

Then I mounted the "new" sending unit on the bracket (yes, I went back and cleaned up the fuel pump wiring before everything was installed in the tank.

Tested the sending unit again just to make sure it still worked.

Installed the finished product in the tank with a new gasket and lock ring.

After the tank was installed in the truck, this is what my fuel gauge looked like with no fuel in the tank.

I poured around 4-1/2 gallons in the tank and tried to start it up but no joy. Couldn't hear the pump running and no fuel to the rail (checked at the Schraeder valve). So, I pulled the pump back out and discovered that I had not completely seated the wiring connector on the pump and it had come loose. Made sure it connected well and tested the pump with a couple of jumpers from the battery. Re-installed it in the tank, cranked the engine for about 5 seconds and it fired up.

I poured the rest of the fuel I had drained from the tank (about another 4-1/2 gallons) into the tank, giving me approximately 9 gallons of fuel and this is what my gauge looked like.

The photo quality isn't really good, but it's reading a little under 1/2 tank. With this same amount of fuel in the tank, the gauge was reading about 5/8 full before this repair.

Hope this helps someone else with a sending unit problem.

 

 

Sent from my iPad using Tapatalk

[HOrnbrod]

Nice work Fred-man. If and when my sender dies, I'll stick my NOS one in, then send you my broken one for repair. Deal?

[Keyav8r]

I'm sure we can work something out, based on your apparently bottomless supply of MJ goodies!

[texasjack]

My fuel gage goes way past full and stays there. I wonder if it has lost a ground?