Fuel-sender problems (long)
 

After two years of an off-and-on fuel gauge not working correctly, and an entire afternoon today of installing a brand new, Porsche OEM fuel sender, I'm appealing to my dorki friends for help.

Pertinent info: 1993 964 C2, five-speed, <80k, owned 14 years and have put fewer than 30k on it. Has been very well maintained. An incorrect reading fuel gauge has been problematic almost from Day One, but just one of those things I would get around to fixing some day.

Sometimes, the gauge would go many, many miles while reading correctly. Then, in May 2018, it consistently would read only 3/4 when I had just filled it up. (Yes, the tank was full.). Prior to this, sometimes the gauge would get stuck at, say, the half-full mark. Tapping the gauge never did anything. All my research on various forums point to faulty fuel sender.

Finally, I installed a new, non-OEM sender on Sept. 26, 2016 (ordered it in May, 2016) at 77,370 miles. A little over two years go by, significant because the two-year warranty on the new, non-OEM sender had expired. It at least used to work sporadically; now, nothing. I could hear the float going up and down the sender but hesitant to take the unit apart to see if the wires inside were either corroded or broken.

Two weeks ago, I ordered an OEM sender from Pelican (their last one in stock) for $123.75 (twice the non-OEM unit). It arrived today, and I have just spent a good half-day trying to get it to work. Installation, of course, couldn't have been easier.

While bench testing the newest sender by hooking up the two-wire connector and tilting the sender so that the float is obviously at the top, which should indicate a full tank, the gauge either registers empty or full (when the needle goes quickly to the top; it's only done this two or three times). When tilting the sender downward so that the float is at the bottom, I get an empty reading. I tried the two-year-old sender, and it would work sporadically also. The very last time I tried this, when not in the tank, and the float at the top of the sending, I got a full reading (which is correct). But when I switched to the new sender, and tried this, I got "empty" both times.

There seems to be no rhyme, no reason to why either sender works or doesn't work. I've checked for broken wires as best I can, and the fuel/oil level gauge and found nothing obviously amiss. I even dismantled the gauge and couldn't spot any cold solder joints, etc.

Can I assume the fuel sender is not the problem (similar, though inconsistent results on two senders, one of them OEM and brand-new).

Should I be getting a voltage or resistance reading at the pins on the sender? Would I be looking for voltage or ohms? (I am electricity ignorant!) Should I be checking something else?

Any help, suggestions, remarks welcome. Thanks for reading so far.

--richard

Fuel sending units are just variable resistors. Put an ohm meter across the terminals and test it. I dont have a spec on hand but there should be a spec for full and a spec for empty and a half tank should be read somewhere near the middle.
On the other hand your gauge is just an ohm meter that measures the resistance.
Easy.

Dave, thanks for the response. Could you give me a quick lesson on using an ohm meter? Should the ignition switch be "on" for this testing? I just now tried to do this (ohms should be between 70.1 and 7.5, I've read on Rennlist) but got no readings either switch-on or -off other than OL, which is what it was when I first switched on the meter.


Attachment 58184

To use the ohm meter on the sender, disconnect the sender from the car. Put the ohm meter in resistance mode (capital omega - Ω). Hook up one ohmmeter lead to one side of the sender and the other lead to the other side of the sender.

Then, move the sender through its range of motion slowly. You should see the resistance value change slowly from a low value to a high value if the sender is not broken. If it stays at one value then your sender is busted.

This is probably easiest to do with the sender out of the car entirely. If the sender is hooked up to the car, then you’re measuring the resistance of a combination of the sender and the car’s instrument cluster, not the sender alone.

Ed

Seems like a bad wire/intermittent connection between the fuel sending unit and the gauge.

Thanks Ed. That makes more sense than the way I was trying. Will test tomorrow.


Thanks, Peter. I'll try to find a way to test for intermittent connection.

I agree, it sounds more like a bad wire (or possibly a bad gauge) than a bad sender.

You could test it by making a temporary test wire that connects the gauge directly to the sender, completely bypassing the wiring harness, and see if that makes any difference.

Another way to test is would be to disconnect both ends of the existing wire and hook the ohmmeter to both ends. Then, move the wire around and see if the reading changes. Although, since most of the existing wire is embedded in the harness, the first test is probably more conclusive.

If neither test makes any difference, take a look at the gauge. If it isn't the sender or the wiring, then that's the only thing left (occam's razor).

Test order for a stand alone issue. If there are other aspects of the dash acting up I would be inclined to start at the dash and work back to the sender. don't assume the sender is bad. The testing might have left some aspect out such as a poor connection or the like. My shop is a war zone right now but this is a park and lean in kind of fix. Feel free to come over some time and we can take a look.

Sender-wire-clean 12+Volts and good ground to gauge-gauge.

What Ed said :)

Latest development: Cleaned all senders electrical connections and at the gauge with spray. Gauge now shows slightly more than quarter-tank (tank is full, by the way), which it has shown occasionally before. I hooked up the DVM, turned the DVM to (capital omega - Ω), as Ed said, and ran both senders through the test, both out of the car. On both the new sender and the old; the DVM doesn't show anything but "OL." Does this show that BOTH senders are bad, or -- more likely -- that I don't know what I'm doing (highly likely).