Intermittant fuel pump.

Joe 1320

Torqueus Maximus
Joined
Feb 28, 2002
The car: 1986 GN
Casper's hotwire kit, new fuel pump ( old one too weak), new hanger (old and sender was fubar) , new fuel tank (old one leaked) , new relay (thought that would have cured the issue).

There is an intermittant condition where the fuel pump will not activate on start up. As a example, drove the car on a trip yesterday. Drove 120 miles to a business meeting the car ran flawless. At end of the day, drove another 50 miles and stopped to get gas. After filling up, the car wouldn't fire as the fuel pump wouldn't activate. I messed with the connections at the harness going to the tank to no avail. Opened the hood and connected the hot wire from the underhood lamp to the fuel pump test connection behind the alternator and it fired up. I left it hooked up and drove the rest of the way home. I pulled into the driveway, left the car running and disconnected the jumper expecting the pump to stop and it didn't. I used the key to shut it off.

This isn't the first time it's done this, it's really starting to piss me off. Oil pressure shows fine, there doesn't seem to be any pattern to this. Sometimes it operates perfectly, other times it doesn't. Doesn't seem to matter whether it's a cold or hot start. Ironically, once started it will never cut off while in operation. Fuel pressure and volume are fine, the performance is just great. It's the non priming issue that's getting to me.

Any thoughts on where to check?
 
Some insight to how the fuel pump wiring is. This should help in troubleshooting the issue.

The likely reason the pump keeps running with the test plug wire removed is the oil pressure switch. This switch is in parallel with the fuel pump relay. Either one will activate the pump. This switch closes once the oil pressure is above 4 psi (or so).

Since this is an intermittent issue, it may also stay running since it is possible that the pump relay finally activated. The way to test which one is running the pump is to remove the test plug wire, and disconnect the oil pressure switch connector. If the pump turns off then the pump relay is not active.

The ECM drives the fuel pump relay by providing +12 volts to the coil (dk grn/wht). This is from pin A1 of the ECM. The other side of the relay coil is run to ground (blk/wht). Either one of these may be intermittent (check the wiring). A leaky windshield will allow water to run into the ECM connectors and corrode them.

The relay socket may also be corroded.

As for the ECM control of the relay, at key-on, engine-off the ECM will power the relay for 2 seconds. Then when the ECM receives reference pulses from the ignition control module (ICM) it will again power the relay. Note that the ICM won't send reference pulses until it sees the cam sensor signal toggle.

So it can take up to two engine revolutions before the ECM again powers the pump relay.

Power for the pump is from the INJ/FP fuse. This wire (pnk/blk) is run to both the oil pressure switch and the pump relay.

The power wire to the pump itself is gray. This wire has a splice in it and is the same electrically from the oil pressure switch, pump relay, and test connector. They are tied together (splice) and then run to the fuel pump as a tan/wht, then tan, then back to tan/wht.

RemoveBeforeFlight
 
Great post. I could be wrong but I want to say the ecm is providing the ground to pull in the relays not 12vdc.
 
The ECM does send +12 volts (ignition voltage level) to the fuel pump relay. There is a small transistor (PNP) that provides the voltage. These pop easily, but since the issue is intermittent it is likely OK. For one that popped, there are two small transistors on the lower board about in the center. The taller of the two is the one that goes bad.

Can use a PN2907 to replace it. Easy & a lot less $'s then another ECM. Sometimes while troubleshooting, or if the pump relay socket wire insulation goes bad and the ECM voltage drive get shorted to ground the transistor is toast.

RemoveBeforeFlight
 
Thanks for the replies. I went though everything and found an intermittant contact in the relay supplied in the casper's harness. Works great now. :D Thanks guys!
 
Top