How to Test a Fuel Sending Unit on a Classic Car (Gas Gauge Stuck on Empty or Full)

The fuel gauge in my old Chevelle used to tell me I had a quarter tank right up until the engine coughed, sputtered, and coasted onto the shoulder of a two-lane highway about four miles short of a gas station. I walked it with a borrowed can. That was the afternoon I quit trusting the needle and started testing the circuit behind it.

I’m Dash, and if you own anything built before the computers took over, sooner or later the gas gauge is going to lie to you. It reads full after you have driven 200 miles. It sits on empty with a fresh tank. It works fine on the highway and drops to nothing every time you hit a bump. The good news is that a classic fuel gauge circuit is about as simple as automotive electrical gets. There is a sender in the tank, a wire, a gauge, and a ground. With a cheap multimeter and an hour, you can find out which one is guilty before you drop a tank you did not need to drop.

This guide covers how to test a fuel sending unit on a classic car, and it covers the three big American families: GM, Ford, and Chrysler. They do not behave the same way. A test that pegs a Chevelle gauge at empty will peg a Mustang gauge at full, and that one detail confuses more people on the forums than anything else. Get the logic straight first and the rest falls into place.

How a Classic Fuel Gauge Actually Works

The sending unit is a float on an arm. The arm pivots at the top of the sender and drags a little contact, called the wiper, across a coil of fine resistance wire. As the fuel level changes, the wiper slides to a different spot on the winding, and the resistance between the sender terminal and ground changes. That is the whole trick. The sender is a variable resistor to ground. It does not make voltage and it does not send a signal in any modern sense.

The gauge on the dash reads that resistance. More current flows through the gauge when the sender resistance is low and less flows when it is high, and the needle moves accordingly. The direction depends on who built the car. GM put low resistance at empty. Ford and Chrysler put low resistance at full. Remember that, because it decides what every test in this article means.

One more thing worth knowing before you start chasing ghosts. The float measures depth, not gallons. A tank that is taller in one spot or rounded at the bottom will make the needle move quickly through part of its range and slowly through another. A gauge that seems to fall from full to three-quarters in fifty miles and then hang at half forever may not be broken at all. It may just be reading the shape of the tank.

Know Your Ohm Range Before You Touch Anything

Every sender is designed to work with a gauge that expects a specific resistance at empty and at full. People write it empty-first, so a GM sender is called a 0-90 ohm sender and an old Ford sender is a 73-10 ohm sender. Mix a gauge and a sender from different families and nothing you do with a float arm will make it read right.

Gauge / sender family	Ohms at empty	Ohms at full	How the gauge is powered	Sender wire grounded (key on)	Sender wire unplugged (key on)
GM 1964 and earlier	0	30	Two-coil magnetic gauge, no separate regulator	Reads empty	Reads past full
GM 1965-1997 (most cars; some charts start at 1966)	0	90	Two-coil magnetic gauge, no separate regulator	Reads empty	Reads past full
Ford through about 1986-1988	73	10	Instrument voltage regulator, pulsed, about 5 V average	Climbs to full (test briefly)	Reads at or below empty
Chrysler / Mopar through 1986 (most)	73	10	Voltage limiter, pulsed, about 5 V	Climbs to full (test briefly)	Reads at or below empty
AMC 1950-1977	73	10	Check your shop manual	Should read full (low ohms = full)	Should read empty (open = high ohms)
Auto Meter, Stewart-Warner, Classic Instruments universal	240	33	Per gauge maker	Reads full	Reads empty

Those numbers come from the parts makers that live with this every day. Tanks Inc lists GM up to 1964 at 0-30, GM 1965-1997 at 0-90, Ford up to 1986 at 73-10, Mopar up to 1986 at 73-10, and AMC 1950-1977 at 73-10. Classic Instruments shows the same GM and Ford/AMC ranges, though it starts the GM 0-90 era at 1966 instead of 1965, so if you have a 1965 or 1966 car, measure what you have instead of trusting a chart. On the Chevelle side, the old Team Chevelle tech crowd has long said 1965-1972 cars use the 0-90 sender while the 1964 is different.

Later cars changed again. GM went to roughly 40-250 ohms around 1998, and Ford moved to a 16-158 ohm sender in the late 1980s (Tanks lists 1987, while FORDification and JEGS say 1989). If your car has been modernized with aftermarket gauges, the gauge maker sets the range. Auto Meter, Stewart-Warner, and Classic Instruments universal gauges are commonly 240-33, VDO is 10-180, Dolphin uses 0-90, and Dakota Digital systems are programmable to most ranges.

There is one oddball worth naming. The 1963-1967 Corvette did not use the simple one-wire GM sender. John Hinckley’s well-known write-up for the vette club describes a powered sending unit fed from the ignition switch and working as a voltage divider with the gauge. If you have a midyear, the ground tests below do not apply the same way, and you should follow a midyear-specific procedure.

Half-Tank Numbers and Why the Middle Can Be Wrong

Full and empty are the easy part. The middle is where a lot of replacement senders fall down.

On a GM 0-90 system, the late Fred Aldrich, who wrote as The Chevelle Engineer, gave the in-car targets as about 84-88 ohms full, around 40 ohms at half, and 0-2 ohms empty. That is close to linear.

Ford is different. Auto Meter’s tech notes, which FORDification reprints, explain that the original Ford factory sender was non-linear: about 10 ohms full, 25 ohms at half, 50 ohms at empty, and the remaining stretch up to 73 ohms sits a needle width below E. Some aftermarket 73-10 senders are linear instead, reading about 41 ohms at half. Install one of those in a 1967 Mustang and the gauge will be dead-on at full and empty and wildly optimistic in between. That is not a bad gauge. That is a mismatched sender.

Mopar gauge rebuilders on For B Bodies Only bench test the gauges at 5 volts with about 9.6 ohms for full, 23 ohms for half, and 73 ohms for empty, which tells you the Chrysler curve also runs low ohms at half.

Tools and Safety

You do not need much. A digital multimeter with a 200 ohm range is the main tool. FORDification makes a good point here: a meter with only a coarse ohms scale will barely move across a 10-73 ohm sender, so use the low range or an auto-ranging meter. Before you measure anything, touch the probes together and note what the meter reads. Most leads show a few tenths of an ohm. On a GM sender at empty, that matters.

Also gather a jumper wire with alligator clips on both ends, a 12-volt test light for Ford and Mopar cars, a wiring diagram for your exact year, wire brush and sandpaper for grounds, and a fire extinguisher within reach. If the sender has to come out, add a brass drift or a lock-ring tool, a new sender gasket, a siphon or transfer pump, and approved gas cans.

Gasoline does not forgive shortcuts. Work outside or with the garage door wide open. Keep drop lights and anything that sparks away from the tank and the filler. Disconnect the battery before you pull a sender or drop a tank. When you make a jumper connection near the tank, clip the ground end first on clean metal away from the filler neck and any wet spot, then touch the other end to the sender wire. And never run straight battery voltage to a Ford or Mopar sender. FORDification warns that the sender is built to run on the regulated pulse, and a 12-volt hot wire will fry it.

If your car has an ammeter, a temperature gauge, and an oil gauge that all work, that tells you something too. On Ford and Mopar cars, fuel and temperature usually share the same voltage regulator, so a temp gauge that reads normally means the regulator is at least alive.

Step 1: The Unplug and Ground Test at the Tank

This is the single most useful test in the whole job, and it takes five minutes. It splits the circuit in half: everything from the sender wire forward to the dash, and everything in the tank.

Find the sender wire where it connects to the top of the sender or the tank. On a GM car of the Chevelle era it is a single tan wire, with a separate black ground wire from the sender to the body. Turn the key to ON with the engine off.

For GM cars, pull the tan wire off the sender. With nothing connected, the circuit is open, which looks like infinite resistance, and the gauge should swing past full. Now touch that wire to a clean chassis ground with your jumper. Zero ohms means empty, so the needle should drop to E. Give it a little time. Aldrich noted that some GM gauges take a long time to respond.

For Ford and Chrysler cars, the logic flips. Unplug the sender wire and the gauge should sit at or below empty, because open means high resistance. Ground the wire and the needle should climb toward full. Do this briefly. CJ Pony Parts and plenty of Mustang owners warn not to leave a sender wire grounded for long on these cars, because you are putting the full regulated feed through the gauge with nothing limiting it. On the older Ford thermal gauges, the needle rises slowly, so watch for a minute rather than a second.

If the gauge does what it should in both directions, the gauge, the dash wiring, the body harness, and on Ford and Mopar cars the regulator are all doing their jobs. Your problem is the sender, its ground, or something inside the tank. If the gauge does not respond, the problem is between the tank and the dash, and dropping the tank would have been a waste of a Saturday.

Step 2: Measure the Sender Resistance

With the sender wire still disconnected and the key OFF, put one meter probe on the sender terminal and the other on a clean chassis ground. Not the sender body, not yet. You want to measure the sender and its ground path together, because that is what the gauge sees.

Compare the reading against what is actually in the tank. If you just filled up, a GM sender should read near 85-90 ohms. A Ford or Mopar sender should read near 10. If you are running on fumes, flip those. A reading that makes sense for your fuel level means the sender and its ground are probably fine, and you go back to the dash side.

A reading of OL, or infinite, means the circuit is open somewhere. That can be a broken winding, a wiper that has lost contact, or a dead ground. A reading of zero on a Ford or Mopar car, or a reading that never changes no matter how much fuel you add, points at a sender stuck at one end, a sunk float, or a sender wire shorted to the body.

You can also do this measurement from the front of the car. Aldrich’s GM procedure separates the Fisher connector, the body harness connector just outboard of the fuse block on Chevelles, and measures the tan wire on the body side. If the reading there matches what you saw at the tank, the body wiring is good. Then pull the connector off the back of the gauge and measure again. If the numbers check at the gauge and the gauge still reads wrong, the gauge is your problem. That front-to-back approach is how you find a connector that is green inside.

Step 3: Check the Sender Ground

The sender only works if its ground is solid, because the gauge is literally reading resistance to ground. A corroded ground adds ohms of its own. On a GM car that makes the gauge read higher than the fuel level. On a Ford or Mopar car it makes the gauge read lower. Same fault, opposite symptom, and it fools people every time.

Each family grounded the sender differently. On GM cars, Aldrich describes a black wire welded to the sender and screwed to the body with a sheet metal screw. Those screws rust, the wire breaks right at the weld, and undercoating gets sprayed over all of it. On Mopar cars, the original setup used a special clip on the sender tube that jumped over the rubber fuel hose to the steel line, according to a long-time poster on Mopar Forums. Lose that clip during a fuel line job and the sender may be floating electrically. Classic Mustangs and most older Fords ground the sender through the tank to the body.

The fast test is a temporary jumper. Clip one end to the sender body or its mounting ring and the other to clean chassis metal, reconnect the sender wire, and look at the gauge. If it suddenly reads correctly, you found it. Make the fix permanent with a real ground wire and a ring terminal under a clean, bright screw. If you have already chased charging problems with the method in our guide to testing the alternator and grounds on a classic muscle car, this is the same idea on a smaller wire.

Step 4: Ford and Mopar Cars Need a Working Voltage Regulator

GM gauges from this period use two coils that pull the needle against each other, with the sender tapped in between them, so normal changes in system voltage largely cancel out. They do not need a separate regulator. Ford and Chrysler did it differently, and that adds one more part to test.

Ford’s instrument voltage regulator, often called the IVR or CVR, takes 12 volts and turns it into a pulse that averages about 5 volts at the gauges. A 1977 Ford service description puts it plainly: it maintains an average pulsating value of five volts at the gauge. Chrysler used a similar voltage limiter. On some Mopar clusters it is a separate plug-in unit, and on others it lives inside a gauge. A cluster rebuilder on For B Bodies Only says the 1966-1967 Charger has the 5-volt limiter built into the fuel gauge, and that most of the clusters he sees have cooked the gauge’s heater wire because the limiter points stuck.

To check it, unplug the sender wire at the tank, turn the key ON, and put a test light between the wire and ground. A working regulator makes the light pulse on and off, about once a second on Mopar systems according to that same forum. A digital meter will show the voltage jumping around instead of holding steady. A steady light or a steady 12 volts means the regulator is stuck closed, and both fuel and temperature gauges will read high. No light at all means the regulator, a cluster connection, or the wire is open.

The same Ford service text gives one firm warning: do not spark or ground either terminal of the IVR. It can burn the harness, the regulator, or both. Test at the sender wire, not at the regulator spades.

If the regulator is bad, Ford and Mopar solid-state replacements are easy to find. Real Time Engineering’s IVR-3 gets mentioned often by Mopar cluster people as a drop-in fix for points-type limiters. Replace a bad mechanical regulator and you protect the gauges you still have.

Step 5: Bench Test the Sender Out of the Tank

Once the tests say the sender is the problem, it has to come out. On many GM cars, including the Chevelle, that means dropping the tank. On a 1965-1970 Mustang, the tank is the trunk floor and comes out from above once the filler neck, fuel line, sender wire, and perimeter bolts are off, which makes the job a lot friendlier. Drain or siphon as much fuel as you can first. A full tank is heavy, sloshes, and makes a mess you do not want on a creeper.

Most senders are held by a lock ring with tabs. Tap it around with a brass drift or use the proper lock-ring tool. Brass because it does not spark. Lift the sender out at an angle so you do not bend the float arm on the tank opening, and look at what comes out with it.

Shake the float. A brass float with a pinhole will slosh with fuel inside and sit low, which makes the gauge read low. A swollen or fuel-logged plastic float does the same. Look at the float arm for bends from a previous installer. Look at the strainer sock on the pickup while you are there, because a sender job is the cheapest time to replace it.

Now clamp the sender so you can work the float. Black meter lead on the sender body or ground terminal, red lead on the gauge terminal. Move the float from the empty stop to the full stop slowly. Tanks Inc and the parts houses all describe the same test: record empty and full, and those two numbers tell you the range. What you are really looking for is a smooth sweep. If the reading jumps to OL in one spot and comes back, the wiper is skipping across a broken or worn section of winding. That is the classic gauge that works at full and dies at half.

FORDification shows how the older Ford sender head opens by straightening the metal tabs and separating the halves. Inside are the wound resistance wire and the wiper. A wiper can be gently bent back into contact, and a wiper that fell off the shaft can sometimes be soldered back on. Light cleaning with fine steel wool is possible, but go easy, because those wires are very thin. On Mustangs, more than one experienced owner has said an original Ford sender with a new float will outlive several reproduction senders. When you have a good original, it is worth saving.

Reading the Symptoms

Here is how the common failures usually look, so you can match yours.

GM gauge stuck past full. Open circuit. A broken tan wire, a disconnected rear harness plug, a sender ground that rusted off, or a winding that broke completely. One Chevelle owner on Team Chevelle found 127 ohms at what should have been about two-thirds of a tank, well above the 90 ohm maximum, and the gauge rode high until it was fixed.

GM gauge stuck on empty. Something is grounding the sender wire. Look for the tan wire pinched by the tank strap, chafed against the trunk floor, or crushed by the tank when it was last installed. A sunk float or a wiper stuck at the empty end will do it too.

Ford or Mopar gauge stuck on empty or not moving. Open circuit, a missing sender ground, a dead regulator, or a gauge that is not making contact in the cluster.

Ford or Mopar gauge pegged full. The sender wire is grounded somewhere. A poster on Mopar Forums pointed at the wire passing through a grommet in the trunk floor of his Satellite as a common place for that to happen. CJ Pony also notes that on classic Mustangs a gauge that is not centered in the cluster housing can let the sending-unit post touch metal, which sends the needle to full as soon as you turn the key.

Correct at full and empty, wrong in the middle. Mismatched curve, usually a linear aftermarket sender in a car built for a non-linear factory sender, or a bent float arm.

Needle bounces over bumps. A wiper losing contact, a loose ground, or a connector that is barely hanging on.

Fuel and temperature both read wrong at the same time. On Ford and Mopar cars, suspect the shared voltage regulator first.

On Ford trucks with dual tanks, FORDification notes that if grounding each tank’s sender wire pegs the gauge, the selector switch is the likely problem, not both senders. That is a nice reminder that the cheap test saves you from replacing two good parts.

Buying a Replacement Sender That Actually Matches

Buy by year, model, and tank, not by price alone. The ohm range is the first filter, and it has to match your gauge. After that, check the fuel line size, because many GM senders were offered with 5/16 and 3/8 feed lines, and check whether your car needs a return or vapor line on the sender. Match the float arm length and shape to your tank. A new lock-ring gasket is not optional.

Test the new sender on the bench before it goes in. Run the float through its range and make sure the numbers match the gauge. It is not unusual to find a brand-new part that is off at one end. FORDification points out that many senders can be adjusted by carefully bending the float rod or the stop tangs, which is how you fine-tune a universal or close-enough sender. Bend in small steps and recheck.

If you install the sender, reconnect everything, and the gauge is close but not perfect, you have options short of pulling the tank again. Tanks Inc sells an inline calibrator called MeterMatch, and Dakota Digital gauges can be programmed to the sender you have. For a numbers-matching car, though, the right answer is a factory-range sender and a little patience with the float arm.

A Simple Resistor Check for the Gauge Itself

If you want proof that the gauge is good before you go any deeper, you can feed it known resistance in place of the sender. Disconnect the sender wire at the tank or the back of the gauge and connect a resistor between that wire and ground, key ON.

For a GM 0-90 gauge, a jumper to ground should give empty, something near 40-45 ohms should land around half, and 90 ohms should read full. For a Ford or Chrysler 73-10 gauge, about 10 ohms should be full, about 25 ohms should be near half on a factory Ford curve, and 73 ohms should be at or below empty. Give the Ford and Mopar gauges a minute to settle. I use 5-watt resistors from an electronics supplier so nothing gets hot, and I write the values on a piece of tape so I am not squinting at color bands under the dash.

If the gauge reads correctly with known resistors and wrong with the sender, the sender or its ground is the problem. If it reads wrong with known resistors on a GM car, the gauge needs attention. If it reads wrong on a Ford or Mopar car, check the regulator before you blame the gauge. This is the same mindset we use when we test a ballast resistor on a classic muscle car with a multimeter: know the expected number, measure, and let the number settle the argument.

Final Thoughts

A classic gas gauge is not a mystery. It is a float, a variable resistor, a wire, a ground, and on Ford and Mopar cars a little regulator. Know which family you are working on, because GM reads low ohms as empty and Ford and Chrysler read low ohms as full. Do the unplug and ground test first. Measure the sender from the tank and again from the dash. Clean the ground. On Ford and Mopar cars, check the regulator pulse. Only then drop the tank.

I still carry a gallon can in the trunk of the Chevelle, mostly out of habit now. The gauge has been honest for years since I replaced a sender with a burnt spot at about half and added a real ground wire where the factory one had rotted off. It took a meter, a jumper, and one afternoon. The walk to the gas station took longer.

When you finish, write the full and empty readings on a tag and stick it inside the trunk lid. Somebody down the road, maybe you, will be glad to have the numbers.

Featured image: early-1950s Chevrolet convertible at a Conoco station in Colorado Springs, photo by Jacob Moore via Pexels. Inline images: Ahnaf Piash via Pexels; Captainpisslord (CC0), CZmarlin / Christopher Ziemnowicz (CC BY-SA 4.0 and CC BY-SA 3.0) via Wikimedia Commons.

Note To Reader: Car Nut Chronicles is reader supported. We may post content which contains links to affiliate sites such as Amazon where we may earn a small commission. This helps us keep the lights on and the articles you enjoy being produced. Thank you for your readership, we appreciate you!