Puzzle

[Beemer01]

I was riding yesterday in a Suburban area on a winding road. Not pressing it, just loafing along.

The engine in my 85 K100 died, but in a manner that permitted me to safely coast to the side of the road.

It would crank, and almost catch, but in the end I sat there enjoying the weather.

20 minutes later, she started right up.

I drove for a couple of miles, and she died again. Several hours later it restarted just fine. By this time I had arranged a tow, so had it brought back to my home.

This morning I replaced the fuel filter - always do the easy stuff first - and it started up. But with the gas cap off, the 'loose change' sound I'd been hearing is quite audible and seems to originate from the fuel pump.

The bike started, ran for a few minutes on the centerstand, and has again died in my driveway.

Would a failing fuel pump evidence this type of sound?

[deilenberger]

Classic symptoms of a failing Hall-Effect-Sensor. Do a bit of research on the K tech pages on the IBMWR website.

The test for this is:

1. Take the cover off the HES (T-shaped cover at the end of the crankshaft on the front of the engine.)

2. Using a hair-dryer - with the engine running - heat up the HES. If it fails when it get hot - it's kaput. Then turn the heat off and use the dryer to cool it off - see if it restarts.

This is a betch'a a beer one for me. Symptoms are really classic.

[Beemer01]

The fuel pump gets power ignition on according to my voltmeter. The pump however does not work in place. The pump when removed from the tank and hooked up to a 12 V power source works fine.

I'm getting lost here.

[deilenberger]

The fuel pump gets power ignition on according to my voltmeter. The pump however does not work in place. The pump when removed from the tank and hooked up to a 12 V power source works fine.

I'm getting lost here.Where did you measure the voltage on the pump in place? There have been instances of the feed-through for the wiring on the tank going bad (rare - but it's happened..)

And under what conditions did you measure the voltage on the pump in place?

[Beemer01]

Removed positive and negative leads, lifted out of the tank, placed on an insulated surface. Attached the leads from a voltmeter and turned on the ignition.

12V indicated on the voltmeter key on.

[deilenberger]

Removed positive and negative leads, lifted out of the tank, placed on an insulated surface. Attached the leads from a voltmeter and turned on the ignition.

12V indicated on the voltmeter key on.With the pump attached - or just on the leads?

If just on the leads - if you have a high-resistance fault in the wiring - it may well read 12V (should actually be around 12.6V) but not run the pump, which draws the voltage down due to the load it places on the high-resistance fault. Bet if you do this with the pump attached - you'll measure considerably less than 12V.

You can make this measurement with the pump in place and inside the tank. The gasoline vapors inside the tank are too high a concentration (not enough oxygen) for combustion to occur. Even so - I'd do it carefully - and avoid sparking.

People have done it with the pump submerged in gasoline..

YMMV - and please be careful doing anything around the fuel tank.. do it with the bike OUTSIDE and with a fire extinguisher handy.

[Beemer01]

I am supposed to be heading out in a few weeks for a weekend motorcycle trip - Yikes! I'm starting to get worried.

Yes, the measurement was taken at the leads, disconnected from the pump. It may be significant that now I get no readings off the leads either attached or unattached (key on and dash lit)

When reconnected to the pump as well I get no reading. The pump removed from the tank and connected to my old analog 12v battery charger - runs just fine.

So the problem seems to be in the routing of power to the pump? I moved up the circuit and checked the power at the connector which comes out of the tank and got 12V (approx) of the two forward pins.

This would seem to indicate a fault with the in-tank wires or the connector under the tank?

[deilenberger]

This would seem to indicate a fault with the in-tank wires or the connector under the tank?If you have power at one end of the circuit, and don't at the other - the area between is obviously suspect. I'd guess you have a bad tank feed-through and wiring (it comes as a unit..)

http://www.realoem.com/bmw/showparts.do?model=0573&mospid=47906&btnr=16_0242&hg=16&fg=05 - Part #10.. Check the year/model or VIN number of your bike since it appears the PN changed around '93..

One point - two of the pins on the connector are used for the fuel level circuit - and you may see voltage on those with the ignition on. Read your circuit diagram in your manual to make certain you're measuring the voltage at the correct pins on the connector. (Brown is always ground on BMWs - so that eliminates one variable..)

[Beemer01]

I'll check this today - thanks Don!

[cjack]

If you have power at one end of the circuit, and don't at the other - the area between is obviously suspect. I'd guess you have a bad tank feed-through and wiring (it comes as a unit..)

http://www.realoem.com/bmw/showparts.do?model=0573&mospid=47906&btnr=16_0242&hg=16&fg=05 - Part #10.. Check the year/model or VIN number of your bike since it appears the PN changed around '93..

One point - two of the pins on the connector are used for the fuel level circuit - and you may see voltage on those with the ignition on. Read your circuit diagram in your manual to make certain you're measuring the voltage at the correct pins on the connector. (Brown is always ground on BMWs - so that eliminates one variable..)

This is an '85 I guess. The ETK fiche on the '85 fuel tank and attaching parts shows the connector on this one with the fuel gauge sender. I don't see that fiche panel in OEM parts. The connector is on the front left side of the tank and the fuel level sender is an upright tube which showed two different remaining fuel levels on the instrument gauge. Entirely different from the '86 and newer. This connector mounted right to the tank has a history of getting intermittent from the pins getting loose (opened up a bit...maybe tank vibration, etc. since it is hard mounted to the tank).
Comments about the four pins are well taken for either vintage. Check for power between the brown and green/white on the harness plug. These connector pins on the tank side become grey and violet inside the tank. This may be as simple as carefully closing up the female connector pins so that the male pins make better contact if you have power outside the tank on the harness plug and no power inside on the grey and violet.

[deilenberger]

Jack...

Thanks for the brain-nudge.

Indeed - the '85 had a very bad habit of the connector loosening up. Pinching the female bits a tiny bit on the connector that goes to the wiring harness normally clears up the problem just fine. That was another reason BMW went to the other design starting in '86 (aside from the fuel gauge being basically useless..) People have stuffed the aluminum foil from a cigarette pack into the female connectors and then pushed it on in a pinch.. I guess that worked also.

Damn, have to remember to check the year/model people ask about. That one should have popped right to the surface since I owned an '85..

[BuddingGeezer]

I have read where metallic anti-sieze instead of Dilectic grease being used to increase the conductivity between the connections.

Ralph Sims

[Beemer01]

I tested the female connector under the tank - key in. One pin has power, two pins ground out properly and one pin has no power. The wire colors are very difficult to make out, but since there is no power to the pump from the in-tank grey and violet wire, it would seem that there is no power to the pump positive wire.

I've checked the wiring circuit diagram in Clymers and it looks like that wire goes to a 7.5 ampfuse, which is fine.

Any thoughts?

[deilenberger]

I think you have to be a bit more agressive in determining the color of the wires going into the plug. If necessary - carefully slit back the insulation for the wire bundle a bit and remove some so you can see the colors.

We have no reasonable answer for you until we know if power is getting to the undertank female connector. If it is - and there is no power on the corresponding wire inside the tank, then the through tank connector is suspect - or the plug is loose.

Have you tried - as others have suggested - very slightly (jeweler's screwdriver is good) tightening up the female connectors in the connector under the tank? That may very well fix your problem.

[Beemer01]

OK, I'll remove the tank, as the wire loom supporting the female connector lacks sufficient length for me to pull it out any further.

The insulation on the wires is pretty faded, so as I indicated it's difficult to tell the color.

I used a jewelers screwdriver to adjust the connectors inside the female connector, but last night tested the connector directly. The results were one lead live, two which had ground and one which was dead. Tweaking the connectors had no effect on this.

Either the fuel indicator circuit is dead or the pump power wire is dead. Since the pump is not working, I'm betting that this is the offending circuit.

If I'm reading the Clymer wiring diagram correctly, the dead connector(and corresponding wire) is the one that goes to the 7.5 amp fuse. The fuse is OK.

This is a constant pressure pump for the voltage delivered s/b around 12 volts.

[breyfogle]

Beemero1: I had a somewhat similar thing happen to my K75. The bike would not start and the only obvious symptom was that the fuel pump would not run. I could, however, force the pump to run by hot wiring the the pump directly off the battery.

Turns out that the fuel pump only gets power if the FI module energizes one of the relays in the relay box and the FI module only does this if it thinks the HES is generating an output signal. In my case, the HES was fine, the FI module had partially failed, and was not providing power to close the relay. It took a good shop tech several days of trial and error futzing because the faulty FI would sometimes start working (which was interpreted as a false "fix" for whatever the tech had just tried) only to fail again after a few minutes of run time.