Load tests would have yielded resistance at the connector. Good to hit all the connectors on the under side and under hood.

Earlier test with a 150 ohm resistor in the connector/plug (at the fuel pump) showed 150 ohms at the instrument cluster terminal positions B7 and B17. So, if there was some contaminants on the connectors in the plug, it would seem that the resistance reading would have been higher then. I think the fuel pump is original to the car and after 18 years and 171,000 miles there may have been some corrosion on the fuel pump side of the connections.

Terminal resistance may not show up with a simple resistance test. The resistance doesn't show itself until current starts flowing. .

After a couple of weeks with a functioning (sort of) fuel gauge, I've found that the accuracy isn't real good. At one point the gauge was indicating 3/4 full and the trip odometer was showing 165 miles. A the cheap gas station I was able to get 8.5 in the tank, which is about half empty. Maybe the aftermarket sending units aren't as precise as OEM.

I think your 1997 PA has an external fuel filter.

I take that to mean it had something happen to it on December 7, 2007:



. . . unless somebody put 68,312 miles on it in a year:



68,312 miles divided by seven years elapsed equals ~9,759 miles per year, so completely reasonable.

Could be. They are usually fairly close though. The big questions are:

Does it move all over the place as originally reported? . . . or is that resolved?

Does the computer still have a code of "bad fuel tank sending unit"? . . . or is that resolved?

The speedometer shop in Atlanta thought the cluster may have been opened previously, but couldn't find any obvious signs that anything had been replaced. Their take was that the service date was Dec 7, 2007 with 68,312 miles on odometer

Fuel gauge needle behaves correctly now, but I'm thinking that it doesn't accurately reflect actual quantity in the tank. I haven't had anybody hook up a good scanner, so don't know if it is still showing a code for bad sending unit

For what it'* worth, it hasn't been until the last ten years or so that manufacturers have focused on the accuracy of fuel gauges, generally to support a reliable "DTE" feature. Prior to that, the fuel level represented on the fuel gauge was designed to give a general idea of the approximate fuel level. Is that what we're talking about here?