Troubleshooting my fuel gauge issues.

[timk225]

Last autumn, I bought a complete standard A body instrument cluster for my 1973 Duster (not a rallye dash). I took it apart, cleaned it, painted it, and got one of those $80 electronic voltage regulators to replace the factory bimetallic regulator, for a steady 5 volts.

I installed the new instrument cluster this weekend, and the new gauge didn't move any differently than the one that was in the car before.

Over the past 8 months, I've replaced the entire fuel system. Lines, tank, sender, pump, everything. Even the 1920 Holley got rebuilt. Before I assembled the new tank and sender, I tested its resistance with a multimeter through its range of motion. I was getting the correct 10-73 ohm range.

So right now, my fuel gauge is sitting right on the E line. I got under the back of the car today and tested the resistance of the sender itself, from its output lug to its fuel pipe. It was showing 50.9 ohms, so it has about 1/3 of a tank of gas still in it.

With the tank full the old gauge showed full, but it dropped fairly quickly, and I think this other gauge will do the same.

Although the factory is happy with letting the fuel sender get its ground through the fuel line mounting clamps, that's not good enough for me. I tested from the sender output to the fuel line, going through the clip-on strap, and it was getting its 50.9 ohms.

So to supplement the grounding, I took some stranded wire, stripped off a good 4-5 inches on each end, and wrapped it multiple times around the fuel senders output pipe, and to the newly installed copper 1/4" vent line. Then up in the engine compartment, at the other end of the 1/4" copper vent line, I ran another piece of wire from it to the radiator support, so it has more grounding.

This didn't change the gauge indication at all.

Where to look next for a connection with extra resistance? Firewall connector? I haven't touched that yet.

I could put in an AutoMeter fuel gauge, but I'd like to have the stock cluster fully operational.

 

[RustyRatRod]

Have you grounded the wire to the sender to see if the gauge goes to full?

 

[Dana67Dart]

I installed the new instrument cluster this weekend, and the new gauge didn't move any differently than the one that was in the car before

What are you trying to achieve?

At 50 ohms from the sender you should have had between 2 and 4 gallons in the tank (1/3 tank as you mentioned)

Product Review: A100 Fuel Sender For 67 Dart

I guess I'm not sure what you are seeing.

Or what is not right with your gauge right now.

 

[67Dart273]

I agree with Dana. Is there a problem, and if so, what?

Are you aware of the metermatch product, and the look alikes from china and ebay?

 

[timk225]

I've seen the videos for the meter match, I could try it although I think there's a non-gauge issue right now. I think my sending unit is a Spectra. I don't remember. The arm the float swings on is about 3 inches shorter than the OEM sending unit and its arm.

 

[harrisonm]

I will repeat what @RustyRatRod said. "Have you grounded the wire to the sender to see if the gauge goes to full?"
That is one of the first things I would have done. You will need a helper. With the ignition to the full ON position, have your helper sit in the driver's seat and watch the fuel gauge. Meanwhile, you crawl under the car and pull the plug off of the fuel sending unit on the gas tank. Note: Sending units in our Mopars do not actually send any signals to the gauges. They simply allow a variable amount of current to flow to a ground. When the gas tank is full, the float is all the way up, and the rheostat allows more current to flow through to the ground. That current flows through the bimetal coil in the gas gauge, that coil is hooked to the needle, and the current will push the needle to full. When the tank is empty, the float is all the way down, and the rheostat allows no current to pass through to the ground. That means that there is no current flowing through the bimetal coil, and the needle does not move. So when you pull the plug off of the fuel sending unit, no current is passing through to the ground, and the gauge should be pegged at "E". Tell the helper the plug has been pulled, and they should say the needle is on empty. Then tell your helper you are going to ground the plug to a known good ground and have them tell you what the needle is doing. Then, using a jumper wire ground the plug to a known good ground. The needle should go to "F" fairly quickly. As soon as the helper says it at "F", I would remove the ground wire. If it does not go to "F", I would say you have a problem with the gauge or a bad wire connection somewhere.

 

[Dana67Dart]

I've seen the videos for the meter match, I could try it although I think there's a non-gauge issue right now. I think my sending unit is a Spectra. I don't remember. The arm the float swings on is about 3 inches shorter than the OEM sending unit and its arm

ALL of the aftermarket senders are NOT calibrated correctly to our gauges and tanks.

See the graph and chart in post 3.

I have done extensive testing with several aftermarket and an OEM sender.

You still have not described the problem you are having.

 

[70SwingerGuy]

ALL of the aftermarket senders are NOT calibrated correctly to our gauges and tanks.

Thanks for that info Dana, I had been wondering why my gauge was a bit off, a previous owner replaced the fuel tank and sending unit, so that explains it

 

[Dana67Dart]

Thanks for that info Dana, I had been wondering why my gauge was a bit off, a previous owner replaced the fuel tank and sending unit, so that explains

If you look at the graphs the 9 gallon in the tank should be 23 ohms but are typically almost 50 ohms and that's closer to 1/3 tank or 3 -3.5 gallons

 

[70SwingerGuy]

If you look at the graphs the 9 gallon in the tank should be 23 ohms but are typically almost 50 ohms and that's closer to 1/3 tank or 3 -3.5 gallons

And Im assuming that there is no way to fix that on our end?

 

[timk225]

I checked my Year One order history, here is the fuel sender unit I have: https://www.yearone.com/Product/Chrysler-A-Body/BG914#prettyPhoto

Dana67Dart - the problem I am having is that the gauge reads Full when the tank is Full, but the gauge goes to E before the tank is E. The tank is still 1/3 full, more or less.

 

[timk225]

I thought of a troubleshooting step to do. I'll have to look in the 1972 shop manual I have, (anyone got a download link for the 1973 manual?) and identify which wire at the firewall connector is the fuel gauge sending unit wire. Then, check the ohms at the sender, and then come up to the engine bay, and check the ohms to ground at the firewall connector. If I get the same reading there, I'll know the resistance is accurate to that point. I don't know if I'll be able to get up under the dash enough to get an ohms reading right off the back of the circuit board.

I've looked at the AutoMeter Z series 2-1/8" fuel gauges. For the 73-10 ohm range, they have one that is linear, and one that is non-linear. The 1/4, 1/2, and 3/4 lines are in different positions. I wonder which gauge would be right, or closer to right, for the sender I have. I guess I'd have to meter match it. I see AutoMeter has a fuel bridge for their gauges that does the meter match job.

i wanted to reuse the original sender out of the old gas tank, but it was in really bad shape.

Is an OEM sender with the correct resistance windings available?

 

[Dana67Dart]

And Im assuming that there is no way to fix that on our end

Only with a meter match.

There is no amount of arm bending that will fix it

The tank is not rectangular.

The first 4 gallons will go down really fast as most of the top of the tank is the spare tire.

The next 8 gallons will be fairly constant

Then that last 4 will be fast but not as fast as the first.

Just the shape of the tank.


THEN....

The gauge wants 10 ohms for full
And 73 for empty.

41.5 is the average or 1/2 way between the two numbers, the gauge wants 23 for 1/2 scale.

So the sender needs to be calibrated for the tank AND the gauge.

 

[timk225]

I remember the good old days of '93-'04, when all I had were a Volare, an Aspen, Valiants, a Dart, a Duster, and a Satellite. Standard procedure at the gas pump was to fill until the pump kicked off, push on the quarter panel a few times to shake the air out of the tank, fill some more, shake, fill, shake, fill, shake. I could get another 1-1/2 gallons in that way, depending on the car. I wonder what sort of an idiot I looked like doing that. Only when the filler tube was full and I couldn't shake any more air out, was the gas tank considered full. Depending on where I worked I could run 5 days on one tank of gas.

 

[timk225]

The gauge wants 10 ohms for full
And 73 for empty.

41.5 is the average or 1/2 way between the two numbers, the gauge wants 23 for 1/2 scale.

So the sender needs to be calibrated for the tank AND the gauge.

So you're saying the gauge reads 10 ohms at full, 23 ohms at 1/2, and 73 for empty? Chrysler could have done a lot better with designing the gauge and sender.

 

[Dana67Dart]

So you're saying the gauge reads 10 ohms at full, 23 ohms at 1/2, and 73 for empty? Chrysler could have done a lot better with designing the gauge and sender.

they and ford all did just fine. the gauges are thermo-electric so they do not react fast and are stable in various air temperatures. I have had a car with a fully electric fuel gauge and it would wiggle constantly when driving.

If you want a precise fuel level then put a flow meter on the fuel inlet and on the fuel line to the carb. AND if you have a fuel return line one there too. then measure the in and the out and the return and have a computer do the math then display the results.

watch and learn

 

[67Dart273]

So you're saying the gauge reads 10 ohms at full, 23 ohms at 1/2, and 73 for empty? Chrysler could have done a lot better with designing the gauge and sender.

There are lots of things than are not linear. Back in the day a couple of us scored some meters out of civil defense radiation survey devices. Basically, "rough" geiger counters, which measured on much higher scales. The measurement is ridiculously non linear. In order to make the gauges look a little bit more "even" the factory made the meter movements themselves non linear, so that the SCALE that your read would be closer to being so. These were "on the face of them" 50 ua meters, nice and sensitive, VERY attractive. Only we could not get the damn things to act like we thought they should!!! Full scale was 50ua, OK, but 1/2 scale was way off the mark

Finally, a retired USAF Colonel, who on a side story had test flown some captured German stuff in WWII for evaluation, and who was incredibly well versed, set us straight.

Similar to one of these girls

 

[timk225]

I researched the wiring in my 1972 Chrysler shop manual. Good news, the fuel gauge wiring does not go through the firewall connectors. It is one of the wires in that bundle that travels through the passenger compartment, along the left side, under the carpet. There's an 8 pin plug and it's the dark blue wire. I found it on page 8-151 in the 1972 manual.

So this is where I can take an ohm reading and see if there's extra resistance there compared to what the fuel sender is putting out. Probably not, if it is just a solid wire with no connectors further back in the body of the car anywhere. I can also see which wire in the instrument panel wiring plug it is, I could take a reading there too if I pulled the dash loose again enough to unplug it.

The carpet is out of my car now, so I have easy access to it. I could run a whole new wire all the way to the gauge and bypass all the old wiring if I wanted to.

 

[Dana67Dart]

So this is where I can take an ohm reading and see if there's extra resistance there compared to what the fuel sender is putting out

I would expect 1 to 5 ohms additional for the wire, but that is insignificant compared to the aftermarket senders inaccuracy.

 

[timk225]

Some more ideas today.

I'm sure someone has thought of it before, but how about taking apart a new fuel sender and rewinding it with less wire, thereby lowering its overall resistance range. Might take a few tries, but I think it could be gotten closer to OEM values. Use fewer turns of wire through the mid range for closer to OEM resistance values where needed.

Or, something like this digital ohm meter.

https://www.aliexpress.us/item/2251832658801671.html?src=google&pdp_npi=4@dis!USD!9.00!7.20!!!!!@!65165642956!ppc!!!&src=google&albch=shopping&acnt=708-803-3821&isdl=y&slnk=&plac=&mtctp=&albbt=Google_7_shopping&aff_platform=google&aff_short_key=UneMJZVf&gclsrc=aw.ds&albagn=888888&ds_e_adid=&ds_e_matchtype=&ds_e_device=c&ds_e_network=x&ds_e_product_group_id=&ds_e_product_id=en2251832658801671&ds_e_product_merchant_id=106967304&ds_e_product_country=US&ds_e_product_language=en&ds_e_product_channel=online&ds_e_product_store_id=&ds_url_v=2&albcp=20268592310&albag=&isSmbAutoCall=false&needSmbHouyi=false&gad_source=1&gbraid=0AAAAAD6I-hEM8j5VmWw4_bda6mzJDFxu1&gclid=Cj0KCQjwh_i_BhCzARIsANimeoETjd5bvvHBZnAGEhbwinp2yOnlT4jfqsTqezFueO5txXgkLVAuFBYaAu_gEALw_wcB&gatewayAdapt=glo2usa

I don't know if it is ok to have power on it all the time, at least from a key switched power source, but you know what ohm number = full and what ohm number = empty, then just go off of that.

I'd rather have a proper functioning OEM fuel gauge, or at least an auto meter gauge with the autometer fuel bridge (meter match equivalent but with 5 set points)

 

[Dana67Dart]

Man you are really trying to reinvent the wheel.

Put a wanted ad for a used OEM sender, send it out to a rebuilder.

Call it a day.


This is what the OEM resister board looks like.

Note the angle of the board and the pitch of the windings and the areas with the very close windings.

Every length of the resistance wire adds resistance so it is very Non linear.

The sweep makes contact to an area on the board to provide a variable resistance to ground

The length of the arm with the float on the end. The angle it is in the tank etc, ALL contribute to the calibration.

The top image is an aftermarket sender that looks VERY much like the OEM A body sender.

But the calibration is no better than the other aftermarket senders


The bottom photo is an actual OEM resister board

I thought about having the resister board made and putting it into the aftermarket sender, but it is cost prohibitive

Product Review: A100 Fuel Sender For 67 Dart

 

[timk225]

I was just thinking, if it would be possible to design a fully electronic solution, which is essentially a microchip controlled potentiometer. It could be programmed for every gallon. And custom set for whatever aftermarket sensor you had. 80 ohms in = 73 ohms out. 43 ohms in = 23 ohms out. 55 ohms in = 60 ohms out. Whatever numbers apply, I'm just making up numbers to show my idea.

Of course cost and practicality would make it not viable, but I'm just wondering if it is possible.

I should have saved my original fuel sender. Didn't think I'd ever touch it once I put the Year One sender in. Anyone know what brand (Spectra, etc) the Year One unit is? Link: https://redirect.viglink.com/?forma...com/Product/Chrysler-A-Body/BG914#prettyPhoto

 

[ESP47]

We all like to geek out on different stuff when it comes to the hobby. One guy might put a bunch of work into eeking out a few extra HP while another puts in a bunch of work to get a more accurate fuel gauge reading. It's just part of what makes hot rodding fun.

 

[Dana67Dart]

I was just thinking, if it would be possible to design a fully electronic solution, which is essentially a microchip controlled potentiometer

That would be very do-able.

All you need is a variable input that covers the entire tank height and a raspberry pi to interpret the input and convert it to the needed out put.


That is essentially what a Metermatch does.

You could even make a linier strip with 18 pickup points based on the gallons in the tank.

How I did all my checking was one gallon at a time with a jossle (sp?) between. Tank in car, car sitting on level ground.

Drained tank to empty, measured resistance, added 1 gallon, measured resistance, repete till full.

 

[timk225]

I got some data points today. A few days ago, with the sender reading 50.9 ohms, the gauge was on E. Today, after more driving, the sender was at 55 ohms and the gauge was still solidly on E. Then I put in 5.0 to 5.5 gallons from 3 gas cans, and the sender rose to 27.8 ohms, and the fuel gauge went to just slightly over 1/4 tank. And I drove it a couple miles to let it settle.

Then I got comfortable on the ground with the drivers door open, to check the wires at the A pillar, and find the fuel sender wire. But I wasn't finding the right wire to check. The 1972 manual I have says it is the dark blue wire. I checked the dark blue, the black, the brown, and even the dark green wire, and none were even close on the resistance reading, using several ground points under the dash. I'll have to find a 1973 or 1974 manual and see if there's a wiring difference.

I've done a lot of computer work, mostly hardware based, but not much programming. I'm familiar with the Raspberry Pi, I have one sitting here that was a controller for some old Avalon 741 bitcoin miners from several years ago. Pis can do a lot of things. I wonder what electronic device would measure the input resistance and provide the output resistance.