Bench testing two cheap eBay fuel sending units

Hi there:

For a few years I've been dealing with the frustration of an inaccurate gas gauge on my '68 Valiant after I replaced the tank and sending unit with moderately-priced aftermarket items. I've had the same issue of "drops from Full like a rock within 30 miles" that several other people have dealt with. That wasn't all of it, though:

Issue #2: leaked past the sending unit seal even though everything was new. Tipped off that the replacement lock rings are too thin. Replaced with original, leak stopped.

Issue #3 leaked past the filler tube seal even though it was replaced when the tank was(2 years prior). Pulled the tank again today to inspect and found the seal was beginning to shrink and split even though it had the part # and DPCD symbol made into it JUST LIKE THE ORIGINALS.
Issue #4 observed today that the dust seal(the one around the filler tube at the trunk floor) that I replaced almost exactly one year ago with one from DMT, is beginning to dry and crack. He claims they are EPDM just like the originals...but the factory didn't use THIS version of EPDM because theirs lasted 40+ years. I may request a refund or replacement.

So, I've had too many issues. Last week I ordered another sending unit and DMT's supposedly better filler tube seal. Pulled the tank today after pumping out 12 gallons of gas...the gauge indicated 1/4 tank btw. I decided to test the new sending unit on the car before installing it. Guess what....it's readings matched the one I took out. Both were WAY off.

Tonight, I realized I had everything needed to make some scientific tests to determine of the sending units are all defective the way people claim they are. I was already convinced they were. I have the dash out of my 68,000 mile Dart sitting waiting to go back in, so I used it and a jumper pack to power things. I powered the gauge regulator with the jump pack and then created a ground loop through the sending unit and gauge back to the jump pack.

The test results were informative. I marked each sending unit body at the point where the gauge read Full, 3/4, 1/2, and Empty. The first surprise that neither sender displayed the sudden "drop" from Full. While the points that each indicated the chosen level on the gauge were slightly different between the two, again, neither displayed erratic readings. Based on the tight grouping of F, 3/4, 1/2 and 1/4...then the spread of space down to E, I would say that the sending units are designed closely enough to original that they would work correctly. Both sending units took a relatively significant amount of travel to reach and cross through the E mark and both did somewhat before the float arm reached the bottom stop. I also discovered that the replacement Spectre tank I bought back then has a reinforcing rib made into it that the original doesn't have in that location and as a consequence the float arm hits it before reaching the bottom stop.

So, this has helped convince me that the issue is elsewhere. Mopar only gives specs on the resistance at Full and Empty and the spec at empty is plus or minus 12 ohms, so it's pretty 'loose". I checked the ground path through the metal fuel line on my car at it passed with flying colors. Here's my opinion: I think the problem is in the fuel gauges. That Dart and my '66 Chrysler both spent their lives garaged out of the sun and heat and both gauges read accurately..for the '60's. On the other hand, the Plymouth baked outside it's entire life and I've had other problems with the cluster being warped and dried out. Someone on a board also claimed that the gauges could be damaged over time and they were pretty quickly ignored. Anyway, I'll post some pictures in the next post. Thanks for reading.

Sender on the left is new, the one on the right is 4 years old. Look the same otherwise. I also forgot to mention that I still have the tank assembly from my old '73 and it also was garaged and also had an accurate gas gauge.

Lots of threads on here about this. Attached are some charts I copied from other members. Hope it helps.

Good, scientific information. Subscribed.

Due to the tanks shape, the swing of the float lever is irrelevant to the reading on the gauge to gallons in the tank.
Product Review: A100 Fuel Sender For 67 Dart

Fuel Gauge Sending Units compared

Close as you will come to getting an OEM fuel sender..

The only way to test the senders is to put them in the tank and fill the tank 1 gallon at a time.

The bottom of the tank is rectangular but with a slope on 2/3s of it. Then the top half has the spair tire cut out

Then there is the guages itsself . On one of my charts you can see the restance needed to make the guage read E 1/4 1/2 F and it is curved. And it's curve closely matches the later mopar sender, but the other senders I tested were much more linear.

Good luck to you.

BTW... If anyone has an 18 gallon Dart tank they want to part with, (can have a few holes.) I will cut an access hole in the top And video and measure the inches up from the bottom per gallon.

That is interesting information, thank you.

What intrigued me was that neither of the sending units I tested behaved the same tested on the bench as they did on the car. On the car they drop off from Full much too quickly and in fact read 1/4 tank with approximately 12 gallons in the tank. On the bench, they REACT smoothly to downward changes in the arm.

Not only that, but these gauges react so SLOWLY to changes in the float position that I'm not convinced that the discrepancies at different points in the chart matter. That float is moving all the time and that combined with whatever buffer feature the gauge or reg has makes it even less of an issue.

What also leads me to conclude that the issue lies in the car's wiring or gauge is that several people state that the sender measures out fine, but they can only get 5/8 on the gauge. Those gauges read lower the more resistance there is in the circuit, so that tells me the issue is internal to the car. I bet if you took a known good gauge and ran it through an IVR but right at the tank, the mysterious 5/8 reading would miraculously turn into FULL lol.

They will act differently on the bench than in the car. The issue is the shape of the fuel tank. The fuel tank has a big chunk of it missing near the top for the spare tire. Since the top part of the tank doesn’t hold near as much gas as the bottom part, you don’t need to use much gas to get the float to drop a lot when it is full. You can’t really duplicate this just moving the float on the bench.

The thermal gage design takes awhile (several seconds) to respond to changes in the float position. This effectively “averages” the voltage coming into the gauge over the period of time it is on. This means that any temporary change, like the gas sloshing around for a few seconds won’t show up in the gauge reading very much. The gauge takes a finite period of time to heat up or cool down.

The threads Dana67 posted have a lot of good info. There can be all kinds of problems with your car, the gauge, the wiring, the IVR, the sender, grounding. Maybe one of them, maybe all of them combined.

I made a resistor box with switches to simulate the correct resistances for empty, half & full and connected it back at the fuel tank. That way I could verify that the wiring, gauge & IVR together were working properly. I got the right readings with the simulator box, so I knew it was my sender causing the errors. I used a device called a Meter Match to fix it. It is not perfect, but works pretty good.

I recently took a box of gauges from 50 years of 70 Darts (non-rallye) to a guy that does nothing but restore gauge clusters for antique cars. He showed me how cheaply made the originals were and identified some that were simply dead (both fuel and temp gauges). I decided to buy new fuel and temp gauges that are available from several sources (I got mine from Herbs). I looked at them carefully and they were much different from the originals and much more substantial in terms of the actual mechanism. I'm not to the point where I've put all this to use and it will be a few months, but this might be part of the solution.

TylerW said:

On the car they drop off from Full much too quickly and in fact read 1/4 tank with approximately 12 gallons in the tank. On the bench, they REACT smoothly to downward changes in the arm

Click to expand...

That is the shape of the tank. Let's say the upper half of the tank can hold 6 gallons (due to its shape)

The fuel level falls faster for the first 6 gallons then for the last 12.

But if the sender is not calibrated to the tank it will read incorrectly

The sender wire also has resistance that adds to the system so full might show less than full, 10 on the bench plus 2 to 3 for the wire and connections now you are getting 12 or so ohms and the guage will read less than full. Same on the low end will read 80 or so before it really is empty.