Need electrical help, brick runs without fuel pump, whats going on?

[daveson]

Although I have had this problem, and fixed it, I'm looking for help to understand what happened.

When I first got the brick it was over fueling to the point of fuel leaking out between the headers and muffler. With the start button pushed there was a constant flow of fuel from all injectors. After cranking with the pumps electrical plug disconnected to dry it out, I was surprised that I could get it to run constantly. Connecting the pump would then result in a stall.

After checking things activated with the start button pushed, I got to replacing the fuel injection relay, which fixed the problem.

My wild, way out there guess, with more turns than the magic bullet ........ a faulty resistor in the fuel injection relay was causing a constant earth all the way back to the injectors. Connecting the pump with the injectors constantly open resulted in flooding and a stall.

Although gravity fed fuel with the help of manifold vacuum wouldn't be enough, the added help of injectors constantly open, was enough, but not ideal, for it to run with the pump disconnected.

Here is someone else with the problem at "K75s idle with fuel pump d/c" on youtube.

An attempt to attach.

Thoughts anyone?

[DJEwen]

What sorcery is this?

[Laitch]

[daveson]

Before I posted this thread I thought of removing the resistor from a relay to see if I could reproduce the problem, but I wasn't sure if that would result in damage to the computers. Would it?

The time I had it running with the pump disconnected, I rode it around the block,  about one kilometer, which seems to have done no damage.

However if I remove the wire from pin 12 of the fuel injection unit (earth to injectors) and earth it,  would that be a way to test it without risking damage to the computers,  or anything else?

Or leaving it connected and earthing the injectors some other way?

[daveson]

OK,  might cop some flak but anyway...

Another beautiful day... another beautiful ride... an historic day too... riding along... all of a sudden a thought bubble from nowhere... disconnect the injector earth's and earth them, then earth the injectors.

That should be a way to provide constant earth to the injectors without risking damage to the computers.

Yes?

[rbm]

Before I posted this thread I thought of removing the resistor from a relay to see if I could reproduce the problem, but I wasn't sure if that would result in damage to the computers. Would it?

It might damage the ICU although I think that risk is very low.  The resistor is for damping the voltage spike that occurs when power is removed from a relay coil.  The risk is that removing this damping resistor will possibly take our a transistor driver in the ICU from the resulting voltage spike that occur when the bike is powered off.  As I said, this is a low risk. You probably could do it and see if it changes the nature of the failure.

The time I had it running with the pump disconnected, I rode it around the block,  about one kilometer, which seems to have done no damage.

However if I remove the wire from pin 12 of the fuel injection unit (earth to injectors) and earth it,  would that be a way to test it without risking damage to the computers,  or anything else?

Or leaving it connected and earthing the injectors some other way?

Removing the wire from Pin 12 of the ECU plug would not be catastrophic.  What you’d be doing is removing the ability of the ECU to control the injectors.  If you grounded this free floating wire, then all the injectors would open up fully for the duration of time that the lead was grounded.

OK,  might cop some flak but anyway...

Another beautiful day... another beautiful ride... an historic day too... riding along... all of a sudden a thought bubble from nowhere... disconnect the injector earth's and earth them, then earth the injectors.

That should be a way to provide constant earth to the injectors without risking damage to the computers.

Yes?

Yeah.  As I mentioned above, grounding the Yellow/Grey wire from Pin 12 will open all the injectors fully for the duration of time you ground the wire.


I suspect that the engine vacuum combined with the injectors being open 100% of the time would permit enough fuel to enter the cylinders to support combustion.  If the fuel pump was to be started, the additional pressure would flood the engine and stall it.  I think this is what you saw originally.

It’s perplexing to me how this problem was solved with replacing the FI relay.  The FI relay supplies power to the injectors but doesn’t control their dwell time; that’s done by the ECU.  The injector dwell time is influenced by the TPS switch and engine temperature.  If the TPS is at WOT, then the ECU will set the injector dwell time to almost 100%.  Did you test the condition of the TPS?

Another condition that could cause fully open injectors is if they are dirty.  I guess your injectors are clean because you claim to have a perfectly running engine at the moment and you didn’t touch the injectors.  So, that’s not a possible cause.

I’ve gone through the startup sequence from Bert’s page, thinking “what would happen of the FI relay was defective?”.  Here’s a normal sequence.

1. The starter relay and Fuel Injection relay close, Hall sensors get power.
2. The FI relay energizes (+) the fuel pump, the idle switch, the injectors and the air flow meter.
3. The Jetronic receives (pin #4) the signal that the engine is starting and it sets itself on starting mode (enriching the air/fuel mixture)
4. ECU pin # 9-10-14 send power to the ignition coil
5. ECU pin #8 send engine rpm info to pin #1 of FI computer
6. ECU pin #7 sends ground to the coil of the FI relay as soon as the starter switch is depressed. The ground will stay on after the starter switch is released as long as the ECU receives a signal from the Hall sensors.

Let’s say one of the 87 contacts on the FI relay is defective but the other works.  That means injectors get power but the temp relay and fuel pump don’t get power.  When you crank, the ECU thinks the engine is cold and enriches the mixture, increasing the injector dwell time.  Since the injectors are open, cylinder vacuum draws fuel into the system and combustion happens (ICU tells the plugs to fire, AFM meters correct amount of air, etc.).  So your engine fires and keeps the RPMs up, drawing sustaining fuel from the fuel system.  But the engine temperature remains “cold” so the injectors remain open.  Replacing the FI relay corrected this broken contact situation and brought the system back to normal.

Well, that's my half-baked theory.  Maybe someone else has a better one.

[daveson]

Beauty.

Brilliant.

That makes sense.

I'll give it a go.

"...Did you test the condition of the TPS?" No,  I never thought about that.

I've got a few tests I can do now,  and be pretty relaxed about pressing the start button.

But it might be a couple of weeks before I try to cross this bridge again.

[daveson]

[daveson]

Quick update; removed relay pin to pump. Result; no go.

First,  thanks for the helpful PM's I have received.

Couldn't help myself as its a quick test. Instead of removing the pin, I tried the lazy mans way, replacing the FI relay with a horn relay, I think this reduces the resistor from 560 Ohms to 470, not a biggy. With the help of Roberts post I pressed the start button with confidence that I won't do damage. Result; idles for five seconds, then stops. I couldn't get it to rev. (note to self; some short cuts lead to taking the same path twice, stop it)

Next, I cut off the inner (pump) 87 pin of the FI relay. Result; idles for five seconds, then stops. I couldn't get it to rev.

I put a normal relay back in, and the bike ran normally, so I haven't cooked the computers.

About two years ago when I got this brick (a black '89 K100RT) and fixed it by replacing the relay, I thought "Oh, the old stuck relay story" and threw the relay in the bin. I regret that now.

About six months ago, I revisited the problem and realised that thought was obviously wrong. I tried multiple combinations of removing pins, and jumping pins, to reproduce the problem, without luck. This is what led me to guess it was the resistor, its the only thing left (except the coil)

Strangely, when I had this problem, it only ran well at revs. In the other cases I've heard about, they idle well.

Could the injectors be getting constant earth from an exposed wire somewhere, which was maybe disturbed while replacing the tank? ICU to EFI, EFI to injectors or somewhere. But I cant reproduce the problem with wiggling wires, and with a good relay it has always been good. And those wires look good.

For the next test I might simply earth the injector earth's, while the pump is disconnected.

 I tried to update my bike profile but deleted most of the info. I'll update it later, but briefly I have five bricks;
a black one I call blacky.
a blue one l call bluey.
a brown one l call brownie.
a red one l call reddy.
a purple one.

Please don't ask me what I call the purple one.

The quest continues...

[Soggz]

Go on then...’What do you call the purple one?’...

[daveson]

I asked you not to ask me that, 99.

Oh well OK, I call her purpilly, I think thats how you spell them.

Anyhow, I'm not doing too bad, six munths ago I couldnt even spell brick ryder and now I are one.

[Soggz]

B’dum...tsk!
You here all week?

[daveson]

Success, we have reproduced the problem, bricks can run without a pump.

Instead of disconnecting the injector earth wire from either end, I tried the lazy mans way and simply scotch locked a wire from the injector earth wire to the battery negative post, and disconnected the electrical plug to the pump.

Result; the brick runs well at 2000 RPM or higher, I soon got it to rev down to 1000, but it was rough as guts that low. There was no clicking from the injectors. I stopped it after five minutes, worried about the computer, I know, I shouldn't have been.

Then I removed the jumper wire, reconnected the pump, and the brick ran normally, so I did no damage.

I think the brick in the above video has a slightly different problem, as it idles well.

I think there are a few take home messages from this:

1; If your brick is over-fuelling to the billyo, the injectors aren't clicking (and the temperature sensor is good)  you may have constant earth to the injectors.

2; If your on a ride, your fuel pump packs up and you haven't got a spare, disconnect the pump, earth the injector earth wire, and you can ride home in limp mode (I didn't ride it today, its night, its got no light, but I did ride it last time) When I say limp mode, its actually OK at revs but you might have to ride the clutch a bit at intersections to keep the revs high, to avoid a stall (although I'm wondering about the fuel temperature) Also,  this risks overheating the injector coils.

3; As crazy as it sounds, if your brick goes when the pump stops, and stops when the pump goes, it possibly has constantly open injectors.

Thanks heaps Robert (once again) for your help, and for giving me the confidence to press the start button.

I'm bloody rapt (once again)

[daveson]

Soggz: "B'dum...tsk
               You here all week"

No, even better,  I've done another Monty Python.

[daveson]