The Stovebolt Site Forums The Shop Area The Electrical Bay Fuel gauge diagnosis

Fuel gauge has stopped reading correctly. Showing 1/4 tank instead of being quite full. Tank and sending unit were replaced a few years back when I first got the truck. Gauge worked fine last I knew.

This is a 51 truck that was long ago converted to 12v. Big ugly resistor under the dash.

I have multimeter and test light. Haven’t pulled the sending unit to check ohms or anything. Hoping there’s something easier I could check first.

There is really no easy way to test but following a test procedure will help you pinpoint the problem quickly.

Seeing how many ‘bolters have issues with misreading gas gauges, I wrote the procedure about a year ago. I want to submit it as a “Tech Tip”, as soon as new submissions start to reappear. It was reviewed by another ‘bolter Jon G whom is a frequent poster on electrical issues.

Testing a Gas Gauge on a Stovebolt

Written for a ‘47 thru ‘55 first series, but may work for other years

The steps are as follows:


At the sending unit, remove the wire that goes to the gauge. The gauge should go to full. Now touch the wire to a known good ground. (Run a jumper to the battery negative if unsure). The gauge should go to empty. If you pass both of these tests, that tells you that the sending wire and the gauge are OK and the sending unit and/or the ground at the sending unit is the issue. Check sending unit grounds. Clean contact surfaces and/or replace wire and connectors.
If your gauge still does not work, remove the sending unit from the tank. Be prepared to buy or fabricate a new gasket, as it’s usually stuck to the tank. First, shake the float to see if it has been leaking. If there is any gas in it, it’s shot and will need to be replaced. We’ll test the sending unit to see if we replace the whole unit or just the float. To test, use a multimeter and set it to the lowest ohms scale (mine was 0-200). Place the positive (red) lead on the terminal that goes to the gauge and the negative (black) lead on the sending unit body making sure you have good contact (sand/scrape if you have to. Now move the float. At empty (float all the way down) it should read 0 Ohms (or close to it). At full (float all the way up, it should read 30 Ohms (or again close to it). If the unit works, you can buy a new plastic float from eBay. Look for one that fits a Jeep, 1955-1983 CJ5, 1955-1975 CJ6, 1976-1986 CJ7, 1981-1986 CJ8
Test for 12 volt supply on the gauge (the one from the ignition switch) if you find no voltage, fix/replace the supply wire.
Do a continuity test on the sending unit wire to the gauge. If you have an open fix/run a new wire.
Run a jumper wire from a known good ground to the back of the gauge cluster to insure the cluster has a good ground to the dash. If you gauge works you’ll need to clean up the area where the cluster touches the dash.
Unfortunately the next step is to remove the gauge (Disconnect your negative battery terminal before you start).This requires you to remove (or drop) the gauge cluster which houses the fuel gauge. I was able to lower mine sufficiently in order to remove the gauge. The bezel is crimped in a few spots which holds the assembly together. Using a small screwdriver, carefully pry open the crimps. Once the bezel, glass and inner trim are removed the gauge is visible. To remove the gauge, the adjacent gauges need need be loosened. There are two tabs (visible in the picture) that are behind the adjacent gauges and will prevent you from removing the gauge. Once loosened, you can remove the gas gauge. There are a few things to check. At this point, use the battery (or an external battery such as a motorcycle, tractor, auto, etc) so we eliminate any bad ground issues.
First let’s check if the gauge functions properly. Connect a jumper from the negative side of the battery to the sending unit terminal (the one with the red warning label). Connect a second jumper from the negative side of the battery to the gauge ground (shown in the picture). Connect a jumper from the positive side of the battery to the power supply terminal. Your gauge should go to empty. If it does not, the gauge is bad.
If your gauge checks out good, look at the back of the gauge where the gauge ground makes contact with the gauge cluster frame and clean/sand the contact area, insure there is no paint, rust or anything else inhibiting a good ground contact. Reinstall the gauge back into the cluster frame and tighten the nuts. Re-perform the gauge test but use the gauge cluster frame as the gauge ground. If the gauge now test good it’s time to reassemble. Tighten the adjacent gauges to the cluster frame.
There’s one more thing we need to check before we reassemble. Some of the reproduction gauges are made slightly thicker and the needle gets caught on the inner trim piece inhibiting the needles movement, making the gauge look like it’s not functioning. If you find this condition, a gentle push with your fingers will give you the additional clearance you need to keep it from binding. Mind you I do mean gentle as it moves easily and you don’t want to distort the inner trim piece. Once you have determined the needle moves freely, reassemble the rest of the cluster. The first time I installed the bezel, I used a screwdriver and a small hammer to crimp the bezel to the cluster and I felt it jared the cluster too much. The second time (yes, it took two times to get it right) I was able to use a needle nose pliers to get the crimps in. Protect the chrome from getting marred. Maybe some other “Bolters” have some other ideas on how to crimp the bezel on to the cluster. Now reinstall the cluster and hopefully it works! (Don’t forget to reconnect the negative battery cable)
At this point, we have eliminated everything but the sending unit. If your gauge still does not work, remove the sending unit from the tank. Be prepared to buy or fabricate a new gasket (you can’t use Silicone Gasket Maker as the gas will soften it and it will fail), as it’s usually stuck to the tank. First, shake the float to see if it has been leaking. If there is any gas in it, it’s shot and will need to be replaced. We’ll test the sending unit to see if we replace the whole unit or just the float. To test, use a multimeter and set it to the lowest ohms scale (mine was 0-200). Place either lead on the terminal that goes to the gauge and the other lead on the sending unit body. Now move the float. At empty (float all the way down) it should read 0 Ohms (or close to it). At full (float all the way up, it should read 30 Ohms (or again close to it). If the unit works, you can buy a new plastic float from eBay. Look for one that fits a Jeep, 1955-1983 CJ5, 1955-1975 CJ6, 1976-1986 CJ7, 1981-1986 CJ8.


Hopefully, at this point, you issue is solved!

Perfect. Thank you.

During your troubleshooting, if you happen to zero-in on the fuel lever sensor, here is something I ran into.

Here are pictures of a new tank lever sensor compared to one that is less than 10 years old.
As you can see, the parts that were not coated already have corrosion and rust.
Why? Old Ethanol (E10) gasoline.
The sensor itself is basically a variable resistor that makes physical sliding contact corresponding to the level of the gasoline in the tank.
The sensor could have rust or corrosion which could cause the fuel level gauge to read incorrectly.

I've mentioned before, but the best 0-30 ohm sender unit I've found is one that was created in Taiwan and is sold on eBay by a seller whose id is renbusan. It is accurate and clever. The wiper never touches the resistance material but runs across precision metal fingers instead. And it wipes in a flat plane and not in an inverted arc, so it remains accurate. Body is all nylon. About $30 delivered last I looked.

This one?

https://www.ebay.com/itm/Fuel-level....m2548.l6249&mkrid=711-127632-2357-0

Also available from Amazon. Price is the same but the shipping may be cheaper.

https://www.amazon.com/Level-Sender-Sending-ranges-Adjustable/dp/B079LRRJXN?th=1

That is a really cool unit. I wish I'd seen this before I bought the standard one from NPD that works but isn't accurate as this one appears to be.

Yes, that is the one. The person who designed this had his mental ducks in a row. The black arc above the wiping area is the resistance medium (poured in place). Then below that there is the wiping arc. What looks like individual wires is really individual tiny pieces of plated metal equidistantly spaced. Those were bonded into the nylon, laser cut and extend into the resistance medium so all the resistance measurement is both accurate and won't change over time. Plus it wipes across a flat surface (unlike the upturned arc resistance wipers so there are no thin wires to wear through and pressure on the wiping surface is more uniform). I spent over 6 months on the fuel gauge sender project but stopped dead when I studied this sender. Sorry if anyone missed it, but I believe I posted that I couldn't improve on this scheme and suggested everyone consider this above anything else out there.

Circling back to this topic, the float wasn’t doing float things, so I replaced it with the Jeep one that Phak1 suggested (nice tip, btw). All seemed to be fine. I go to top off the truck last night because I wanted to see where “full” really was. And…now gauge is reading half-ish with a full tank

Should there be one or two wires at the sending unit? My truck has two. When I replaced the tank 7 or so years ago I just wired it up the same way it was.

Anyway, I guess I will go back to the drawing board and do some of the other tests mentioned in Phak1’s excellent write up.

The second wire grounds out the sending unit. Just a pigtail from one of the mounting bolts to the cab. You should also check to see if the cab is properly grounded.

I assumed it was a ground. But wiring diagram seemed a little unclear to me on the subject.

Thanks!

There is a ground wire and the center post is a resistance wire (resistance back to ground). Using your multimeter with the function set to resistance (ohms) and set to the lowest reading, connect one lead to the center post and the other to ground. With your tank full, the resistance should be 30 ohms. Make certain you have a good ground connection, too. If you get something like 14~18 ohms resistance with a full tank, your sender isn't doing the right things. Let us know, please.

Will try to do that later this evening while it’s fresh on my mind.

For what it’s worth, the 5 screw setup on top doesn’t align with holes to give tons of options on which way the sender lever points. I’m not 100% confident it has full range of motion.

That is common. Sometimes (not often) the sender unit will allow you to loosen/rotate/adjust but usually there is a notch cut which prevents that. I had one that I used two adjustable wrenches to twist once.

Somehow or another, it worked in the past. But I didn’t pay enough attention to how it was clocked before I removed it this time around.

As long as the part of the shaft next to the sender and the sender itself are parallel to the front or rear side of the tank, it should all be Jake.

So having issues with my 1950 Chevy 3100. Got it running (mostly), and last nagging issue is the fuel gauge already read EMPTY. Reviewing this thread I did some testing at the sending unit, underneath the cab. After cleaning grounds, and cleaning the sending unit connector... now the fuel gauge reads FULL (I mean slammed completely to the right). Disconnecting the sending unit wire, nor grounding the gauge wire at the sending unit connector has any effect on the fuel gauge reading FULL (doesn't even flick when connecting or grounding). At one point I accidentally connected the sending unit wire with the ignition switch on... which also seems to be when the fuel gauge went from reading EMPTY to now reading FULL. Switch on or off makes no difference in the results though (before or now).

QUESTION: Does this mean I somehow burnt out the gauge ? I would expect a burn-out gauge to go to its resting state, not full power state (although, see below, my thinking on operation may be backwards).


I have determined that at least the sending unit itself seems to be bad as it still reads ZERO Ohms with a half-full tank. That will be replaced tomorrow probably, but Phak1's trouble-shooting steps in this thread do not seem correct to me looking at the electric schematic for the 3100. If others have found this to be correct, then obviously I'm wrong, but maybe someone could point out my error to help me finalize my trouble-shooting.

1) 12V from the battery (6v in my case via an inline resistor at some point) goes via ignition switch to one side of the fuel gauge.
2) Other side of the fuel gauge goes to the sending unit
3) Other side of the sending unit goes to ground.
4) I assume the fuel gauge reads EMPTY when no power, meaning that is its resting state (i.e. NO POWER = EMPTY on gauge)
5) That would mean maximum voltage to the gauge would translate into maximum deflection of the gauge (i.e FULL on gauge).
6) An empty tank however would provide maximum resistance translating to minimum voltage to the gauge (i.e. EMPTY on gauge).

The trouble-shooting write-up however has the gauge indicating FULL with no power (i.e. "At the sending unit, remove the wire that goes to the gauge"). And likewise when you ground the wire feeding the gauge at the sending unit... it is suppose to read empty (although that is completing the power circuit and so should be providing maximum voltage and deflection of the gauge (i.e. FULL).

Like I said, obviously I'm mistaken in my thinking... but the only way I can see trouble-shooting steps being correct is if the OFF resting state of the gauge shows FULL... which seems unlikely. Since it hasn't worked since I got the truck though, I really am not sure.

Kudos to the first one that shows me the error in my thinking :-)



Thanks... Joe

Joe,

Any chance you got the wires crossed on the gauge? There's a tag on it that says if you put battery voltage on the sender terminal, the sender will be damaged. I don't know if that would mess with the gauge itself, but it might be possible.

Please take a look at this explanation I wrote up a while ago: https://tinyurl.com/mr2xspy7

As Kevin pointed out, you may damage the sender (not the gauge) by reversing the wires and putting battery voltage on the sender terminal. The good news is that isn't a promise. Let's for right now hope that didn't happen. What I wrote should explain the working of the gauge, sender, etc. At ignition off, your needle will go to empty. Please also keep always in mind that your sender unit is a variable resistance thing. That's all it does...the unit provides between 0 and 30 ohms resistance to ground. Let's start with what I've written and go from there. If you need to catch me directly, use [email protected]

I should mention (as I've done many times here), if you need a new sending unit the most accurate and probably the most long-lived one is sold on eBay by a seller whose ID is renbusan. About $30. I think he may also be selling the sender on Amazon. Made in Taiwan and extremely well designed. I mentioned it on page 1 of this thread, but I mention it again in case anyone needs it.

Thanks Jon for the redirection to the other thread, I never considered a dual magnet scheme.... I guess this was done for less bouncing of the needle when driving.

As background, I am simply trying to get the fuel gauge to work, it has not worked since I purchased the truck. First action was getting it running, now fixing the "little" things. So I have not touched the gauge itself, I started on the sending unit as it was the more likely culprit. Initially it stayed on EMTPY. I cleaned the ground at the sender (under the body), no change. I pulled the sending unit connector, cleaned and added dielectric grease and replugged. At one point I did the reconnection at the sender with the Ignition ON, and the gauge went to FULL and has stayed there since.

As previously noted, I've unplugged the sender connector, I've grounded it, and I've removed the battery ground and the gauge remains stuck at FULL. I've even tapped (with a rubber mallet) the gauge and dash area when the batter was disconnected in case it was mechanically stuck.... still stuck at FULL.

So with no ground connection to the battery it seems to me that the gauge has to be mechanically broken... unless there is some other aspect I'm missing. I guess if mine is a spring-based return mechanism it could be the spring popped off when the gauge slammed to FULL, but how would I tell. The truck is a 1950 3100 with original parts as far as know in this area.

Hi Joe,
In 1950, the wiring scheme followed exactly the crude little drawing I made in the re-direct thread. And the gauge would go to E when the ignition was off. If you are stuck at FULL when the battery is not connected, then I would say something has gone wrong with your gauge. I have seen one of these get stuck on FULL, however. I never really understood why it happened, however once it was un-stuck it worked fine. I thought the needle must have become stuck against the gauge plate, but honestly by the time I got it disassembled, it had gotten freed so I never really knew for certain. Let's take a look at your gauge...image below. As you see there is only 1 moving part. That is the pivoted needle/permanent magnetic hourglass shaped piece of steel. With your ohmmeter, check the resistance between the two connections (wires to ignition and to sender not connected). You should see around 23~25 ohms if I recall correctly. Let's start there.

The needle in the gas gauge falls to empty with gravity when the power is removed
If the truck is upside down in a ditch the needle will go to full with ignition off
If it sticks to full without power there is a problem in the needle mechanism.
-s

Thanks guys... so seems I'm disassembling the dash... which I really am not looking forward to. I do have a new Classic Radio to install anyway, so at least I'll only need to the seats out once (to replace the broken sender, pull the gauge and check it, and do the radio). Timing is good, I'll do this before adding my final Antifreeze mix since the thermal sensor has to be removed to pull the dash gauges too I think. Sounds a party waiting to happen.......Anybody up for a Road Trip to NC :-)

Yes, a party sounds interesting. Anywhere but DFW should be cooler. Keep us posted....

Well, I think I may have this resolved. Original issue was a float that wasn’t floating. I corrected that, but I think maybe the swing arm wasn’t contacting the tank wall.

Jon’s suggestion of a little twist/ correction with two adjustable wrenches did the truck. A full tank reads full at roughly 34 ohms. We will see how it does as this tank empties.

Have to give the gauge a tap to get it to return to empty when key is off. It might be a bit sticky, I guess.

Is your fuel gauge new? The needle may be touching the glass.

No, it’s old

Reviving my old thread because it contains some useful background on where I’ve been.

Short summary: new fuel tank and sender (when I bough the truck). Gauge worked fine until I began this thread. Discovered a leaking float. Once that was fixed I had a little alignment thing to solve to get full float arm travel. Then the gauge worked for a while…until it didn’t.

For the past _____ months, the gauge has been pegged at full any time the key is on. I’ve just lived with it, but finally got annoyed enough this weekend to attempt yet another diagnosis and fix.

I assumed I was dealing with a crummy sender ground. That is not the case. Test light shows the sender ground working properly. Gauge stays pegged full even if I directly ground the “center post” sender wire. Resistance from sender center post to ground is 18 ohms. We should be reading something other than totally full.

Working through the steps in the procedure posted on page 1, it would seem that my next moves are:

* Test for voltage at the sender. (Easy enough, I think it can do this by “feel” without disassembling gauge cluster).

* Check for continuity, sender wire to gauge. (Again, sounds easy enough)

Depending what I learn there, I need some guidance with the further instructions:

* Run a jumper wire for cluster to ground - Any place on the back of cluster? Or so I need to be more targeted to the fuel gauge?

* If I have to go as far as removing the fuel gauge - Am I removing the entire cluster with the nuts from behind the cluster? And then uncrimping the fuel gauge from the cluster?

Thank you for posting this! It will help me help another guy with his truck fuel gauge problem too.

That testing procedure “Testing a Gas Gauge on a Stovebolt”, is now a “Sticky Thread” located at the top of the Electrical Forum. Here is a [LINK]

Could you take a look at my questions on last post of the previous page?

If its full with the ignition on, what happens when you turn off the ignition?

Goes back down to zero after I tap the glass once or twice.

The circuit and needle movement requires electricity to pull the needle up
Gravity pulls the needle down
A sticking needle can’t move up without the movement actively driving the needle high

I would be suspecting an intermittent open connection between the sending unit and the gauge in the dash
And since tapping the dash seems to resolve your sticking high needle
The possible intermittent connection is in the dash end of this situation.

While the needle is high, measure the voltage on the sender at the gas tank
Then tap the gauge to fix the sticking needle and see what the voltage at the tank changes to
-s

Isn’t the garage tied to switched power? How would it have electricity keeping it “up” with the key “off.”

I’m assuming you mean gauge not garage.

I believe what 2manyTrucks means is the orientation of the gauge requires the ignition to be powered for the gauge to overcome gravity and register the level of the tank. Once the ignition is shut off gravity takes the pointer down to E.

If you short the wire at the sending unit with the ignition on and tap the gauge, will needle move down?

Yes, gauge. Autocorrect making things fun.

Pretty sure that no amount of tapping send the gauge to empty. It stays pegged full all the time if key is on. But I have not tried to intentionally short the wire at the sender and tap the gauge. I can try that.

Edit: I have grounded the sending wire with ignition on, but I have not tried to also tap the gauge whilst doing so.

Shorting the wire at the sending unit should send the gauge to empty. Tapping the gauge unsticks it. Give it a try and get back to us.