Intermittent power loss diagnostics (1987 K75C)

My intermittent power loss has returned after an absence of 2500 miles. (Original thread: http://forums.bmwmoa.org/archive/index.php?t-65698.html)

Short version: Occasionally (3-4 times in 60 miles, in this case), the engine loses power. After a second, or three, it comes back on.

I previously had a bad kill switch that would shut everything down when it jiggled out of place. This doesn't feel like that. Not catastrophically jerky. The engine is just not there.

It will restart with the clutch held in, so it is not bump-starting itself. It has not (yet) died for a long enough period for me to get off the road.

All that, along with normal spark plug appearance, makes me think I have a fuel supply (or maybe air-related) issue rather than an ignition issue.

I've replaced the usual obvious suspects - fuel filter (several), spark plugs, crankcase vent hose, all fuel hoses, big rubber elbow after the air meter, gas tank electrical connector (but not the whole sending unit - yet). The gas and the tank are clean. Injectors just cleaned by Mr. Injector after the first failure 2500 miles ago.

I suspect an electrical fault in the fuel supply, either killing the fuel pump or the injectors. A failure of the pump or pressure regulator would yield the slightly smoother shutdown (as compared to the total electrical failure of the bike). Here are my (current) diagnostic questions:

1. Does it sound like that's the right direction, or am I barking up the wrong tree?

2. Because it's so intermittent, and brief, I have near zero ability to observe what's happening with the bike when the failure occurs. So my thought is to tap critical fuel-related electrical points (like fuse 6, and the coil side of the fuel pump relay) to power LEDs mounted next to the cluster. When the issue occurs, if current is lost at one of those points, its LED will go out. Assuming I don't have too many of those LED's, a quick glance will tell me if any of them is out, and I can follow up with focused examination of the indicated area. Does this seem sane, and at least plausibly productive?

3. Assuming it's not insane, the hardest point to tap (yet possibly the most relevant) is the two connectors on the fuel pump itself. (Either power or ground could fail, of course.) Any suggestions on how to temporarily route wires into the tank?

4. To test a ground connection, like, say, the ground on the fuel pump itself, I'd need to run power to an LED from a known good source, and then ground it through the suspect ground wire. Does that potentially introduce grounding problems for the component (e.g. fuel pump)? I don't think so, but best to ask; I don't want the diagnostic process to cause new failures.

Oh, and there's also no throttle position switch on the bike. Removed it last time and haven't put it back on. (Which doesn't rule out a problem on its connecting wire, of course.) But the switch itself and its adjustment are ruled out as causes.

Quick diagnostic test..

Ignition on, engine not running. Wiggle the key. If the dash lights flicker - the ignition switch needs disassembly and cleaning. Not at all an uncommon problem as these bikes get up in years. There is info on the IBMWR K-tech pages on doing it, and I have some photos on my website if you need them (search for Ignition on the K75S page..)

My symptoms were identical to yours.

Best,

Thanks. It passes the wiggle test, but I have enough "missing voltage" here and there on the bike that cleaning the ignition is high on the list.

Question on that process (and yes, I've read the ibmwr page on it). When I attempted it on the ignition switch from the parts bike, I had trouble (read: I wasn't able) reassembling the rearmost white piece onto the dark gray piece it clips to. There are two little wings on the gray piece that are meant to snap into cavities on the white piece, and that fitment is necessary to compress the springs in the white piece, or else the switch just won't function. One of those two just wouldn't go back in, no matter how many times I carefully realigned all the bits. Any tips on getting it apart and back together nondestructively? That unsuccessful attempt on the parts bike is really why I haven't done it already.

I've also had this problem

on an 85 K100 - engine briefly cuts out - will restart easily. I traced my problem back to that pesky undertank fuel pump connector. (It will just get worse)

I'd spend a bit more time there - other threads here talk about permanent solutions to fix - from my experience just cleaning and reseating it won't really fix the problem

James. I think maybe I had the same kind of problem. I have done this maintainence a number of times with no hitch and was lucky enough to not mess up the layer of the assembly below the switch layer. There are springs and balls that can get out of place. they also can get away from you if not careful. Finding them in the grass is just short of a miracle. DAHIK. If that level is not correct the switch will not turn far enough to turn on the power. Hope this helps. Harold in Kansas

Beemer01 said:

on an 85 K100 - engine briefly cuts out - will restart easily. I traced my problem back to that pesky undertank fuel pump connector. (It will just get worse)

I'd spend a bit more time there - other threads here talk about permanent solutions to fix - from my experience just cleaning and reseating it won't really fix the problem

Click to expand...

Thanks - I've replaced the connector with a new Delphi Weather Pack connector. That doesn't, of course, rule out a problem with the connection through the sending unit itself, which is also a weak point.

JamesInCA said:

Thanks. It passes the wiggle test, but I have enough "missing voltage" here and there on the bike that cleaning the ignition is high on the list.

Question on that process (and yes, I've read the ibmwr page on it). When I attempted it on the ignition switch from the parts bike, I had trouble (read: I wasn't able) reassembling the rearmost white piece onto the dark gray piece it clips to. There are two little wings on the gray piece that are meant to snap into cavities on the white piece, and that fitment is necessary to compress the springs in the white piece, or else the switch just won't function. One of those two just wouldn't go back in, no matter how many times I carefully realigned all the bits. Any tips on getting it apart and back together nondestructively? That unsuccessful attempt on the parts bike is really why I haven't done it already.

Click to expand...

It's actually fairly simple once you get the locking screw out..

http://www.eilenberger.net/K75S/IgnitionSwitch/index.html



See photo 4. That shows the orientation of the inner bits as assembled from the factory. You want the same orientation when reassembling it. I'd suggest assembling it so the back with the Y and the other bit facing up, aligned as shown in the photo. Then simply press the white part down on it (also aligned as shown in the photo.) It should just snap back together. Photo 4 is pre-cleaning. Didn't take one after cleaning the black plastic carrier, but you want to scrape the embedded contacts in it clean, and use some light grease on reassembly.

hhshort said:

James. I think maybe I had the same kind of problem. I have done this maintainence a number of times with no hitch and was lucky enough to not mess up the layer of the assembly below the switch layer. There are springs and balls that can get out of place. they also can get away from you if not careful. Finding them in the grass is just short of a miracle. DAHIK. If that level is not correct the switch will not turn far enough to turn on the power. Hope this helps. Harold in Kansas

Click to expand...

There are no balls that I recall (but it's been a long time since I've done one,) and the springs should be retained in the white plastic part.. I would do it in a contained area. Inside a shoebox is a good idea, keeps the parts from running into the corners of the room.

JamesInCA said:

Thanks - I've replaced the connector with a new Delphi Weather Pack connector. That doesn't, of course, rule out a problem with the connection through the sending unit itself, which is also a weak point.

Click to expand...

These are two different connectors. The '85 K100 was unique in it having a simple push in 4-pin connector that went into a recessed electrical fitting in the bottom of the tank. All that held it together was luck and a bit of friction. Over time - gravity overcomes luck and friction and it comes apart. It was a particularly poor design choice IMHO.

The one you replaced was on a pigtail emerging from the float assembly mounting plate. Different beast. These are known to fail at the feedthrough that takes the wires through and into the tank. Just replacing the actual plug connector doesn't address that problem. You might try wiggling the wires right where they enter the feedthrough.. it is not an uncommon failure point unfortunately, and the fix is expensive, a new float assembly is needed (to keep the gasoline inside the tank mostly..) It wouldn't hugely surprise me to see these failures increasing as ethanol continues it's relentless attack on older parts that weren't designed to be resistant to it.

springs and balls 1985 K100rt

There are no balls only springs on the switch level of the of the switch. The level below the switch provides the detents and stops for the switch. There are springs and balls. If you are careful this level is not disturbed. If you are doing this maintenence in a campground with out removing the assembly and pigtail from the bike things might get a little dicey. In my case something got out of place and the switch would not make its full turn, therefore never turning the power on. In an other case something could pile up and not allow the assembly to compress. Harold in Kansas

deilenberger said:

The one you replaced was on a pigtail emerging from the float assembly mounting plate. Different beast. These are known to fail at the feedthrough that takes the wires through and into the tank.

Click to expand...

Yes, I've had one of these fail previously, though in that case it just wouldn't work at all - nothing intermittent about it. Been hoping it isn't that again as, apart from a Hall effect sensor, that's the most expensive part that might be at issue here. I will try wiggling it as you suggest to see if I can make the connection through the plate falter.

Harold, thank you for the pointer to the spring and ball lurking behind the rotating piece in the bottom of the white cup. It's entirely possible that a misalignment there is what caused my reassembly of the test switch to fail. I just went and took it all apart, lined everything up correctly, and reassembled it without a hitch.