Hey Guys,
     I'm having an issue that's driving me crazy trying to diagnose and decided it's time to bring in the big guns: You guys.

     My 1985 K100RS has been faultlessly reliable for the last 15,000 miles. Solid as a rock. I just finished giving it its every-other-year tune-up: Brake flush, fork oil, engine oil, transmission/final drive oil, air filter, spark plugs, battery, final drive splines, new coolant. After doing all that I took the bike out for a test ride/ warm up. As expected, it ran flawlessly. Did about 20 miles. Once the bike was hot, I came home to check my throttle body sync. It was perfect, so I didn't do any adjustment.

    Then I rode about a mile to the nearest gas station, filled up, and set off for a good ride. 20 seconds later the engine died with no warning. One second from perfect to dead. Dash lights still on. I coasted into a parking lot to assess. My first thought was that when I moved the tank to change the coolant (just scootched enough to reveal the coolant filler, not removed) I may have dislodged power into the tank. So I checked that connection. It seemed fine, but I gave it a good wiggle. Tried starting the bike and it jumped right to life. Great! I turned it off, put my gear back on and started it again. It started perfectly, ran for about one second then died. Gear back off. Tried to start again. Perfect start, ran for one second, died. But I could hear the fuel pump running. At that point my guess was that there was something in that tank of gas I bought that clogged the filter. So I pushed the bike home (not fun in 97 degree heat)

   Back home I pulled the filter. I tried connecting the fuel pump outlet directly to where the filter usually connects. The bike started right up and ran fine. I took that as confirmation of clogged filter. So I bought and installed a filter. The bike started, ran for one second, and died. Now my thought was bad fuel pump. I could hear it running, but I wondered if the pump part had come unfixed from the motor shaft. Maybe? Maybe the motor still could impart a little bit of motion to the pump, but not enough to push gas through the filter? So I ordered a new fuel pump. With the new pump installed I tried to start the bike. Perfect start, one second, dead. Sooooo discouraging.

   My next though was fuel pressure regulator. Maybe it had failed and was not holding any pressure in the injector rail. So I ordered a new regulator. While waiting for that I had the thought that perhaps I'd just gotten a tank of bad gas. So I drained the tank and replaced it with fresh gas (from a different station). No difference. When the new regulator showed up it didn't make any difference at all either. 

    I next thought that maybe I had a clog in that internal pipe in the fuel tank that goes from the filter to the output. I confirmed that my old fuel pump was working and tried using that in an external jar of gas, connected with new hose directly to the injector rail. No difference. Bike starts perfectly, runs for one second, then dies. Oh, I forgot to mention: With the 'choke' on it's worse. Choke on it runs for only half a second. With the choke off it'll run for about a second, sometimes 2-3 seconds.

    Having confirmed to myself that the problem is not within the gas tank (right?), I reconnected the new fuel pump in the tank. 

    Even though I strongly feel this is a fuel issue, I decided to check the spark plugs (all brand new and gap-checked). I pulled a couple and found them completely dry. I rechecked torque on all of them and then started checking that I had good spark. I found good, strong spark on number 1, same on 2. While checking spark on number 3 I found two things: That plug has good spark.... and the bike was running. Down a cylinder, of course, but running. What?!?

   So I reconnected spark plug 3 and the bike started right up and ran perfectly. Like absolutely perfect. Like nothing had ever been wrong. I put all the fairings back on, starting the bike every couple minutes to make sure I hadn't been dreaming. Perfect every time. That was yesterday. I didn't have time to ride, but I kept sneaking out into the garage to start the bike and marvel at the perfect running.

   Today I had time for a test ride. So I fired it up. The bike started perfectly and settled into a perfect idle. I let it warm up for a couple minutes while watching it suspiciously and putting on gear. It ran perfectly. Revved up with no trouble. I hopped on and took off for a ride. It ran beautifully for about 600 feet and then died instantly. 

   So that's where I am. I have nothing now. I thought of replacing injectors, but I just don't see how injectors could be the issue. Honestly, I don't know what else to try right now. 

   Any thoughts?
   Thanks!
    -Jon

Instant death... two things to start with:

- are you sure the FICU connector is seated properly?
- check the wiring to the coils.

Thanks for the reply, Dai!
   I would agree about the ignition coils, except that I found I'm getting a good, strong spark on all cylinders.

   The FICU, though... I think that's a solid possibility. I have checked the tightness of the connector, but I think I'll try cleaning the connection. I just had a long walk thinking about the issue and I think the key is that the spark plugs were bone dry when I checked them. I'm pretty sure now I have gas on the outside of the injectors and none in the cylinders. I could buy one injector failing on me, but all at the same time?  I'm very much leaning toward the idea that the injectors are not getting the signals they need. Could be the connection to the FICU. Could be the FICU itself has crapped out. I had a failure in my flasher unit once that was intermittent like this. I actually opened it up and found a cracked solder joint. Resoldered it and it's been perfect ever since. A cracked solder within the FICU could totally explain all of this. Anybody ever open one of those things up? I am NOT an electronics guy, but if it's openable and I can spot a crack I can resolder it. If I have to replace it anyway it might be worth a try.

   -Jon

I've opened some FICU it's easy, but leave it for last. I try to resist my natural way of trying a series of hunches. I think the proper way a mechanic would go is first diagnose the problem, then fix it. 

With the help of a screwdriver on an injector you should be able to hear it clicking on and off while cranking. If not, clean the pins and the plug with electrical contact cleaner.

If it's still not clicking, it's  probably not getting earth or it's not getting power, which is easy to test with a computer safe test light.

Try this keep the start button pressed after the engine has started. No harm will come because if the engine is running over 711rpm the starter will not run but the ECU will keep power to the pump so the bike should keep going. Try that.

Daveyson, Thanks for the reminder to keep cool and be methodical. If I'd been more careful from the start I could have saved myself the cost of a fuel pump and pressure regulator. I Oh, well. At least now I have spares.    I tried cleaning the FICU pins and plug. No improvement. I even took a look inside the FICU, but there was nothing obviously amiss, so if there's a problem with the unit it's beyond my ability to repair. I put everything back together and the bike behaved just the same. Started, then died. That was a couple hours ago.

Just now I read your idea of listening to the injectors to see if they're firing. Great idea! I decided to go try it right away. I have a shop stethoscope so I used that. I put the stethoscope probe against one of the injectors and hit the starter. I could absolutely hear the injector clicking! I was surprised for a second and then I realized I could STILL hear the injector clicking! Because I had stethoscope plugs in my ears it took me a while to realize the bike is running perfectly again   So I still don't know how the injectors behave when the bike isn't working. Oh man, what a process. 

Rick, the pump (or power to the pump) isn't the issue. When the bike starts and then dies the pump is running the whole time and even runs for a moment after the bike dies.

-Jon

It's running perfectly now. That might be because you cleaned the pins. If they weren't making good contact, that can be fixed when they are clean and dry. It takes time for them to dry out. If lucky, you've just fixed it. Maybe the plug for the engine harness to the chassis harness also needs a clean. Anyway look for stuff like loose pin connections or wires not in full contact, stuff like that. You might have a dirty loose earth connection near the fuel rail, just in front of injector number four (the FICU can't provide earth if it's not receiving earth) I think. 

Did you clean the temperature sensor socket too? Follow the injector loom round the front of the bike - the temp sensor is between the radiator and the plenum chamber.

daveyson wrote:You might have a dirty loose earth connection near the fuel rail, just in front of injector number four (the FICU can't provide earth if it's not receiving earth) I think.

Urgh. That's exactly why all my home-built wiring looms have a continuous earth loop, picking up the frame and the engine as the last two stops and that's mostly to ensure that there is no electrical potential between the loom and the metalwork.

All right, I have an update for you guys:

The perfectly running bike didn't last. The next morning it was back to messed up 
But that gave me an opportunity to listen to the injectors while the bike was dying. They were clicking perfectly in time with the engine RPM. While the starter was running, and for a split second after, the engine was at full idle speed. The injectors were clicking at full speed. As soon as the starter stops running the engine slows to a halt, like it simply doesn't have enough power to turn itself over. Sometimes one second, sometimes 2-3 seconds. Each time, the injectors click in time with the dying engine, coming to a stop as the engine does. Seems like just what I would expect them to do, right?

This morning I decided to check the compression, wondering if I might be looking at something like a blown head gasket. Compression was low (has been for years), but consistent between the cylinders. So that's ruled out. While I was doing that I cleaned the spark plugs, spark plug wire connections, ignition coil connections. I rechecked that I had solid spark on all plugs. I do. I did find that I had neglected to tightly screw on those little spark plug connector caps. I had a moment of wondering if all this trouble had been caused by that little oversight, especially when I put everything back together and the bike ran perfectly!

The bike sat there and idled happily. It sounded completely normal and perfect. I revved it up to 3000RPM a few times and it responded enthusiastically. Just completely normal... for about 5 minutes. Then it died instantly. I didn't do anything. It was sitting there idling perfectly, then it died. Tried to restart. It ran for two seconds and died.


So, here's my current train of thought about the situation:

A) What caused this problem?

       1 - I messed something up while tuning up the bike:
                * Fork oil, brake flush, transmission/final drive oil, final drive splines - No way those could be related.
                * Engine oil, air filter, coolant change, new battery, new spark plugs - Possibly related. But I've done those things many times, and all went smoothly. I feel very confident I did them correctly. And the bike ran perfectly for 20 miles after doing all of these.
                * Check throttle body sync. This was the last service I did before the problem (After the 20 mile warm up), so maybe the most suspicious. But I've done it dozens of times, and I didn't even do an adjustment this time. It was already spot on.
       2 - Something happened when I filled up with gas
                  Timing makes this really suspicious. The bike ran for not more than 30 seconds after I filled the tank. Totally perfect and reliable for 15,000 miles and then 30 seconds after a fill-up the bike dies? It feels right, which is why I have spent so much time looking at the fuel system. But feeling right doesn't always mean it is right. The gas is not necessarily the issue here.
       3 - The timing is a complete coincidence
                  This issue might not have anything to do with what I did to the bike. Some component could have been RIGHT on the edge of failure for some time, and that was just the moment it let go.

B) What clues are there as to the nature of the problem?

         1 - The problem is intermittent
                    I feel like this is a huge clue. There are plenty of things that can go wrong and keep an engine from running, but 99% of them will incapacitate the engine until they are fixed. Period. If, for instance, my timing chain broke the engine wouldn't run. But it would NOT RUN. It wouldn't work (perfectly) intermittently. Intermittent issue says electrical to me, a sitaution where I have contact sometimes, sometimes not. But I'm still stuck on the thought of fuel. I still wonder if I have some small chunk of crap in the fuel system that is floating around, sometimes blocking fuel flow, sometimes not. I realize that's not likely, but maybe?
         2 - The problem is completely binary
                     There's not been a moment of the bike running sub-optimally. When it's working it's the bike I've known for years. Runs and sounds exactly right. When it's not working it just doesn't run at all. Leads me to think I'm looking for something broken, rather than mis-adjusted. Something broken in such a way that it sometimes allows perfect function, then becomes misaligned or loses electrical contact or something and the bike just dies.
         3 - I can't find a single thing WRONG
                     * I have compression. Even if it's low, it's the same as it's been through thousands of miles of reliable riding. 
                     * I have air. Kind of a no-brainer. Even if I had a super dirty air filter (I don't) the bike should still at least idle.
                     * Spark. I can't see them sparking in the cylinder, of course. But when I pull the plugs test them against ground I see a solid spark on each.
                     * Fuel. I can't see gas being squirted in the cylinder directly, but I have new fuel pump (which I can hear running), new pressure regulator. I can hear the injectors clicking. They might be dirty, but that seems like it would make the bike run rough rather than alternate between perfect and dead.
                      So what more do I need to make the bike run?



Finally, there's Dai's question about the temp sensor. That stirs up a half memory from when I was first getting this bike running. Isn't there a thing where if that temp sensor reads high enough it will just completely shut the bike down? Could that be what's going on? The bike is definitely not overheating. Most of the time it's dying it's completely cold. But could that sensor be sending a signal that the whole bike is about to melt? Would that let the bike start and run for a second before dying?

   -Jon

Have you checked the Hall Effect Sensors (HES)?

Oh, the HES! That's a possibility. I've read about those buggers giving some really wacky symptoms. Is there a good way of testing it? My service manual just says "Bring it to a BMW dealer"

-Jon

Another update:
   I remembered that I had a spare temperature sensor, so I installed that. No improvement. While I was that far into the bike I also cleaned the connection at the ignition control unit and the connection for the hall unit. So those are clean and firmly connected.

   Speaking of the hall sensor, anybody else have thoughts about Dai's idea of that as a culprit?

   -Jon

Try keeping  the button pressed like Rick mentions, I think that was his point.

Bikesmith wrote:Oh, the HES! . . .  Is there a good way of testing it?

Try the Hall sensor test found in the Troubleshooting the EFI guide.

Follow these instructions to get there:
Click the Portal tab at the top of this page, then click on the words Tech Page in its center, then be agreeable, then click on the words Electrical Diagrams under the category Electrical. Also in that category is the Troubleshooting the EFI guide and the Cleaning the ignition switch guide.

I'm having a brain freeze right now(I blame it on coming of age in the mid '60's), but it seems to me that when the engine is cranking below 700 rpm the fuel injection or ignition is controlled differently than after the engine catches and the rpm goes above 700.  For the life of me, I can't recall what the function is that is handed off, or what is the control factor that is being introduced at that time. 

Maybe someone else knows.  Whatever it is is electrical and as you mentioned intermittent.

Is your barn door sensor(MAF) plugged in tightly?

A weak hall sensor often fails when it gets hot, does that seem to match your symptoms? For example if it always works normal first start in the morning. I'm thinking when you're riding, then stop for fuel it gets a little hotter cause the coolant circulation stops. If it doesn't go back to normal until it cools down, that's another hint. When it plays up and goes back to normal quickly after spraying water on the hall sensors cover, that's a big hint.

I'm also with point seven five that the two second thing is a hint, the start button is the signal to the computer that the engine is turning, as long as the button is pressed, plus two seconds, then it's the hall sensors that tells the computer the engine is running, so the computer says OK I'll allow fuel (by providing earth to pin 85 of the fuel relay) as long as the hall sensors tell me that the crankshaft is turning.

Okay! A sudden flurry of things to respond to! Thanks you guys so much for all your help here!

I'm starting to think that this from Point-seven-five is the key: "but it seems to me that when the engine is cranking below 700 rpm the fuel injection or ignition is controlled differently than after the engine catches and the rpm goes above 700"


This is exactly what seems to be happening! The bike starts well every time, but then as soon as the starter turns off (700rpm) The bike just sputters out. I tried holding the starter button and it confirms this thought. When I hold the button, the bike starts fine, then at 700rpm the starter turns off, everything goes weird and the engine starts dying, then the starter comes back on and the bike comes back to life. It just repeats that cycle for as long as I hold the button.

On this page (https://www.kforum-tech.com/electrical/EFI/bike-wont-start-EN.htm) I found this passage:

Bikesmith wrote:

 Is this maybe a bad ignition module, not doing the things it's supposed to do to keep the bike running after it starts?

Tests for the ignition control unit are in the EFI troubleshooting guide. Give it a go.

Hopefully the ignition module and it's plug just needs some electrical contact cleaner. (Edit, but I like Laitch's idea better of first diagnosing the problem)

The 700rpm thing in Bert's guide refers to the starter relay, that's not your problem I think cause your brick has been cutting out while riding.

I have experienced this on a few Ks, will run ok if you keep your finger on the button but cuts out as soon as you release it.

To clarify the fuel pump operation: Fuel pump gets signal from 2 sources, first is the start button, but when the engine fires its coming from the Hall Sensors instead of the start button. 

The fact that you have fuel and firing when you are pressing the start button tells you all circuits are good at that point. But the fuel pump is not receiving the signal from the Hall sensors. 

My good friend, a local ex BMW tech guru, who has since passed away almost always found electrical connections to be the issue instead of the Hall sensor itself. The big plug into the box under the seat is the prime suspect.

Doh, I missed it that you cleaned the ignition module plug. Since the problem is intermittent, maybe wiggle the wires while doing the hall sensor tests (including continuity from the hall sensors plug to the ignition module plug while wiggling the wires)

Okay, now I am REALLY thoroughly baffled 

Here's the update: It's not the ignition control unit. In my typical putting-the-cart-before-the-horse fashion I replaced it. I found a super cheap one locally and decided to just try it. Not difference at all.

I also tested the hall sensor again, in situ this time. Gave me an opportunity to make sure my timing was set perfectly. All good there.

Point-seven-five asked about the barn door. I had a thought that a faulty air flow sensor could be telling the engine that it's getting a tiny amount of air and it's responding with only a tiny amount of gas. Long shot, I know. But it's easy to get at, so I pulled it and tested it. It's fine. And I cleaned the connection while I was there.

So, just now, I did what testing I could of the ignition computer (even though I already replaced it). Here's what I was looking for:

[size=15][b][size=15][b][size=15][b][size=15][b][size=15][b]Pin #[/b][/size][/b][/size][/b][/size][/b][/size][/b][/size] [size=15][b][size=15][b][size=15][b][size=15][b][size=15][b]Function[/b][/size][/b][/size][/b][/size][/b][/size][/b][/size] [size=15][b][size=15][b][size=15][b][size=15][b][size=15][b]Testing mode[/b][/size][/b][/size][/b][/size][/b][/size][/b][/size] [size=15][b][size=15][b][size=15][b][size=15][b][size=15][b]type[/b][/size][/b][/size][/b][/size][/b][/size][/b][/size] [size=15][b][size=15][b][size=15][b][size=15][b][size=15][b]Expected value[/b][/size][/b][/size][/b][/size][/b][/size][/b][/size] 1 Direct ground continuity with ground at the battery Ω .5 max. 2 power to Hall sensors control with ignition ON Volts 12v 3 Ground to Hall sensors continuity with ground at the battery Ω 0 4 Shielding of Hall sensors wires continuity with ground at the battery Ω 0 5-13 signal from Hall sensors Cannot check with voltmeter/Ohmmeter (see below for testing)   6 Signal from starter switch When starter is running Volts 12 7 Ground to FI relay Engine not running ignition ON Volts 12 7 Ground to FI relay When starter is running Volts 0 to 1 8 Signal Engine RPM When starter is running Volts ~ 10 to 20 mV 8 Signal Engine RPM When engine is running Volts ~ 3 to 4 v 9-10-14 Ground to primary of ignition coils When starter is running Volts ~ 1.5 to 4v 10 +12v  of Electronic ignition module Ignition ON Volts 12v 11 Control ground to starter relay When starter is running Volts 0 to 1 11 Control ground to starter relay Engine running, starter switch depressed, engine rpm < 710 Volts 12

So here's what I found: 1 is supposed to be ground. It is.
2 is power to the hall sensors. Supposed to be 12v. I was getting about 10.9 volts there.
3 and 4 are supposed to be ground for the hall sensor. They are both good ground.
6 is supposed to be 12v when the starter is running. I'm getting about 8 volts when the starter is running.
7 should be 12v with ignition on and 0-1v when the starter is running, which are the results I got.
8 Should be 10-20mv when the starter is running. I was getting about 5 volts. It's 3-4 when the engine is running. I couldn't test that. Partly because my bike doesn't run, but also because I had to pull the gas tank to reach this plug.
9 and 14  are supposed to be 1.5-4v when starter is running. I got 12v as soon as I turned the ignition on, which dropped to 9-10 volts  when the starter was running.
10 is supposed to be 12v and it is.
11 should be 0-1 when starter is running. It's .85. It's supposed to go to 12 volts when the engine is running, but I can't test that.

So, yeah, I'm baffled.   Unless someone has a better suggestion, I think starting tomorrow I'm just going to trace through the whole electrical system, cleaning and checking everything. Battery to kill switch to ignition switch, etc. Just follow the flow of the power, making sure I all is as it should be.


-Jon

I spose I should stop posting on this thread cause it's getting over my head, but that hasn't stopped me in the past.

That's not too bad, some of the readings are a bit low, but while cranking it will normally drop a bit under 12Volts. Pin 8 (rpm output signal to the fuel injection computer) is way out. It's the hall sensor outputs (to ignition module pins 5 and 13) that tell the ignition module what the rpm is,  so maybe revisit the hall sensors and wires from hall sensors to ignition module.

It would be interesting what readings you get with the previous ignition module.

As @daveyson points out above, the fact you are gettinging 5V on pin 8 while cranking instead of 10-20mV is indicative of a problem with the ICU or the Hall Sensors or wiring. Pin 8 is a signal to the ECU to say that the engine is turning over and it should work the injectors.

Earlier you said you bench tested the hall sensors so we can assume that they are working.

You also said earlier that you swapped out the ICU with no change in symptoms. So we can assume that the ICU is operational.

1. Check the pins in both the male and female plug housings where the Hall Sensor cable plugs into the harness under the tank.  I'm suspecting that there might be a poor connection there causing the hall sensor signals to intermittently reach the ICU. Although the parts may fit tightly together, it's possible that the contacts are not making when assembled. Clean the contacts well. The HES signals are very low level.

While out for a walk (I do my best thinking while walking my dog) I had a thought. I bench-tested the hall sensor and it tests as just fine. But what if the signal from one of the sensors is not making it through the harness to the ICU? There's one gap in that metal cup thingy and two sensors, so the ICU should be receiving two hall signals per revolution. If only one of those is getting through, would the ICU be sort of half running the engine? Half the required number of injectors pulses and sparks?

See, the engine trouble definitely begins as soon as the starter turns off, but it isn't instantaneous death. It tries to keep running for a second or two. As I said earlier in this thread, it's like the engine just doesn't have enough power to keep itself running. So, half the gas, half the sparks? Maybe?

I'm going to check that out later today. The hall test (https://www.kforum-tech.com/electrical/Hall_sensors/Hall_sensors.htm) done at the ICU plug should do it, right?

I'll post again after I do the test, but I wanted to post the idea right now. Partly to see if my thinking makes sense. Partly because I find it helps clarify my thinking to put it in writing.

-Jon

Bikesmith wrote:While out for a walk (I do my best thinking while walking my dog) I had a thought. I bench-tested the hall sensor and it tests as just fine. But what if the signal from one of the sensors is not making it through the harness to the ICU?

It seems like that is what robmack in his last post is suggesting that you check, starting with continuity at the sensor's harness plug connection.

Thanks @Laitch . I though I might have been writing in Esperanto or something and no-one was understanding what I wrote.

Laitch wrote:

Bikesmith wrote:While out for a walk (I do my best thinking while walking my dog) I had a thought. I bench-tested the hall sensor and it tests as just fine. But what if the signal from one of the sensors is not making it through the harness to the ICU?

It seems like that is what robmack in his last post is suggesting that you check, starting with continuity at the sensor's harness plug connection.

That HES signal to the ICU is for ignition and you know thats working because it is firing correctly as long as you are holding the start button.

But the fuel pump also relies on this signal. Its getting the signal from the start button but not the signal from the HES. Connectors, cable breaks. . . . . . . .

Robmack,
I'm sorry if it seemed like I was ignoring your post. I wasn't. Because of your post I was planning to check into the hall sensor more thoroughly. I was only just running with your thought and speculating about whether a faulty connection between hall and ICU could explain my symptoms.

Olaf,
  I know that I'm getting some spark. I confirmed that. What I don't know is whether it's the correct frequency of sparks. If I were getting, for example, half the number of sparks that I should and half the number of injector pulses that I should, maybe the the bike would try to run but not have enough power to keep itself going, which is just the behavior I'm seeing.

The behavior I get when I hold the starter button down is that the bike starts, then the starter turns off, then the engine starts to die, then the starter comes back on and the bike gains new life, then the starter turns off, then the bike starts to die, then the starter comes back on, and so on for as long as I hold the button. My impression is that it's all about that 710rpm switch. Below 710 it seems like the bike is doing what it's supposed to. Once I hit 710, the bike switches from start mode to run mode and everything goes to heck.

Oh, and the fuel pump runs the whole time. Every time I try to start the bike, the fuel pump runs from the moment I push the starter button until a short time (half second?) after the bike has completely died.

The fuel pump can't run at anything lower than full power, right? It's either fully on or fully off, depending on the relay? The thought has crossed my mind to get a manometer to check the fuel pressure, but I suspect that would be wasted effort. It's a new pump and new pressure regulator. I think I can assume fuel pressure is fine.

-Jon

Some will say that I am crazy as my name stand for, but years ago I did have a similar problem on my K100.
Bike started well but always died on idle. This was a sudden fault. I rode the bike about 50kms without problems, stopped for 5 min. and when I re-started the bike, I experienced this failure.
Found out that the battery was bad. This was weird because the starter had plenty of power. It was certainly a bad cell in the battery and unless you demanded a lot of power, the bad cell provided a cut in the circuit.
This could also explain why your bike runs as long as you push the starter button (demanding more power to the battery)
Remember, the alternator doesn't provide power under 1,500rpm. Try starting the bike and keep the rev over 1,500rpm.
Also try to replace the battery or a least jump it when starting the bike. See if it makes a difference.

Bikesmith, good point, that would be a problem if you were only getting half the number of sparks, but I don't think that's a fit with your problem. 

When it's in no go condition, I think you had spark at all four cylinders. Let's say you weren't getting a signal from the top sensor, then you would have no spark at cylinders one and four ( bottom sensor allows spark for two and three) I spose that eliminates my guess as well. 

(Edit: but I spose if there is a signal, but not a correct one, the injectors might not be open for the correct amount of time)

By the way it's great to see a post by Crazy Frog, beer O'clock.

Bikesmith wrote:Every time I try to start the bike, the fuel pump runs from the moment I push the starter button until a short time (half second?) after the bike has completely died.

That's what a functioning pump does.

It might be time to follow up on robmack's suggestion before speculating into the weeds.

How do you know your good deal ignition control unit is any better than the one that was in the Brick first? Your test data of the original shows a problem, providing you've tested correctly. What's the data on its replacement—exactly the same?

daveyson wrote:By the way it's great to see a post by Crazy Frog, beer O'clock.

Yes CF is still alive and now has a busy life. The last 4 years have been like hell for me.
I am now exclusively riding (a lot) the K12LT with the sidecar. However, I don't forget my old K buddies.
The admins are doing a great job and I let them manage the forum. New blood is always nice.
Cheers to all.

See, I am still crazy as ever. This was last week end at a motorcycle event with 30 friends. Lots of secondary roads and twisties on the [url=https://www.cbisland.com/tourist-attractions/cabot-trail/?gclid=EAIaIQobChMIhYjOg_CK-QIVYwiICR0ZrwsTEAAYAyAAEgKLOfD_BwE#modal-The Cabot Trail]
[/url]

All right, I tested the hall sensors right at the ICU plug. Both sensors worked perfectly. It sure is looking like they're not the problem.

Daveyson, Thanks for the clarification on how the ICU uses the hall input. I guess that rules out an avenue of investigation.

Laitch, You're absolutely right that one used ICU is no more likely to be functional than another used ICU. But both malfunctioning in exactly the same way? Hmmm. Worth testing the other one,  though. Every additional piece of data is a step closer to the solution, right? I'll go do that right now and report back.

Crazy Frog, battery is an interesting thought. This one is brand new, but it sounds like the one in your experience was too. Hmmm. I'll give it a go. The thing that makes it tricky from a jump start perspective is that this is my only 12-volt vehicle  . My other two motorcycles and my truck are all 6-volt.

Bikesmith wrote:Crazy Frog, battery is an interesting thought. This one is brand new, but it sounds like the one in your experience was too. Hmmm. I'll give it a go. The thing that makes it tricky from a jump start perspective is that this is my only 12-volt vehicle  . My other two motorcycles and my truck are all 6-volt.

At least you can monitor the battery voltage.  

Hmmm. When I measure the battery voltage straight across the terminals it's 12.45 volts. But I am getting some wonky voltage levels here. 

For instance, that 5 volts coming off of ICU pin 8. It's supposed to 10-20mv when the starter is running, then jump to 3-4v when the engine is running. Does that indicate that the ICU is not aware that the starter is running? Maybe because the signal coming into pin 6 (signal from starter switch), which is supposed to be 12v, is only 7-8 volts?

Well, a friend is bringing his K75 by tomorrow for help doing his transmission spline lube. So, I'll have a spare battery here tomorrow and can test your idea, Crazy Frog.

Also, I tested the second ICU and got basically exactly the same results.

I wonder if the computer is tricked into believing it's running when it's actually only cranking cause the start button is continuously pressed. Hopefully the problem disappears with a new battery.

Pin 6 is a bit too low, but while cranking it will be below 12 Volts anyway. If you get the same low reading at pin 10 while cranking, that would be a pointer.

Best to leave your original ICU in I think, you know it's history better.

https://www.k100-forum.com/t17957-what-the-heck-is-going-on?utm_campaign=popular&utm_medium=topic&utm_source=newsletter


Its definitely electrical control problems over fuel supply to injectors.

I just read the the entire thread, I may have missed or misunderstood a few bits, but sounds like one of my recent problems. It turned out to be the load shedding relay. I went through pretty much all you have. I changed it and had a permanent fix.

I've got my friend's K75 here (just pulled the transmission), so I've been able to pull his battery and load shed relay to try them in my bike. Neither helped.

Barney Weasel, you said "Its definitely electrical control problems over fuel supply to injectors." Do you mean electrical control of the injectors themselves? Or are you talking about electrical control of the fuel pump? Through all of this work on the bike, I've always been able to hear the fuel pump working normally. The fuel pump starts running the instant I push the starter button and runs until half a second after the bike dies. I'm not 100% sure that it's providing the full needed pressure, but I am 100% sure that it's getting power the whole time the bike is trying to run.

Okay, here's something weird to think about. I just realized I could use my inductive timing light to check if spark plug rpm matches engine rpm. First time I tested it I had forgotten to put the load shed relay back in, so the fuel pump didn't come on, so I was just spinning from the starter. Spark plugs 1,2,3 were all firing at around 440-450 rpm, which I assume is starter speed. But cylinder 4 was half that, around 220. Then I realized I'd forgotten the load shed relay and put it back in. Retested and found cylinders 1,2,3 briefly hitting 900ish (varied some with the success of each starting attempt). Cylinder 4 was still about half that, around 450ish.

So, how do I have good spark on 1, but not 4? Don't they come from the same source? Is this a bad ignition coil? I know there's a way to test it, but I can't remember where I saw that. Anyone know where I find that? Or know the test procedure off the top of their head?

Yes, could be a bad coil, swap coils around and see if the problem moves. Could be a bad Spark plug or lead. 

Fuel pressure might be too low, probably wants to be tested, but can you see fuel return to the tank? (you might see it dribbling down from the top of the tank) If the relay has five pins, that's the fuel injection relay, that's why the pump cut out. As mentioned, fuel pressure might be too high. If for example fuel isn't returning to the tank, the fuel pressure regulator can't work to reduce the pressure from the pump from about 60 psi to about 30, but I think that's unlikely to be an intermittent problem.

I've removed the load shed relay from my brick, but it stared and ran normally except the unimportant stuff like the headlight didn't cut out while cranking, that's it's job. But yes a faulty fuel pressure relay might cause your symptoms. But I think you eliminated that possibility by swapping the five pin relays.

Wouldn't it be funny if simply fuse six was weak, or something like that. Have a good look at the wires and plugs from the fuel relay to the pump.

Probably the easiest test is the one to try first (apart from swapping fuse six)  I'd swap lead four with lead three, if no different put them back, then maybe swap one and two.

Edit: just noticed from your avatar, I think you have the earlier model, the coils are often a bit suspect on those.

Bikesmith wrote:So, how do I have good spark on 1, but not 4? Don't they come from the same source? Is this a bad ignition coil? I know there's a way to test it, but I can't remember where I saw that. 

Have you followed through with robmack's suggestion in post #26?

The coil tests are in the Troubleshooting the EFI document with the rest of the tests. Usually the coil just fails altogether but your Brick is a wellspring of dysfunctional possibility.    Back in post #3 you asserted there was a strong spark on all plugs. Check tightness of wire connections and remove oxidation from both plug wire and coil terminals, then follow robmack's suggestion. Spark plug terminal and plug wire connections must match or unreliable performance will follow. Plugs with terminal nuts seat in a wider cylindrical plug wire terminals than plugs with screw terminals. If a screw terminal plug is mated with the wider type terminal, it is likely to lose contact intermittently.

Congratulations on not buying a load shed relay.  Your impulse control is improving.  Testing your friend's ignition control unit might have been interesting. Is there any fuel in the cylinders after try to start this beast?

Back in post #3, you indicated that the plugs were "bone dry." It's likely that the injectors aren't opening. What opens them are combined ICU/FICU signals based on Hall sensor signals. daveyson has run his Brick with no running fuel pump so it seems like an electronic problem and the search for it would likely start from ICU to fuel injection relay to FICU.

I have a good flow of fuel coming back into the tank. I could check the pressure, but I think that's down the list. Gas coming back into the tank seems a strong indication that pump and regulator are doing their job.

Daveyson, Swapping the coils is a good idea. I'll try that. I tried replacing fuse 6 . No change.

Laitch, I have cleaned the under-tank hall connection. And I've tried testing the hall unit from the ICU plug and it worked perfectly. I'll try cleaning that connection again. Can't hurt. I see in the EFI troubleshooting the part about the measured resistance at two points. Is that the extent of coil testing? If I get resistance measurements in that ballpark then the coils are good?
I thought about testing my bike with my friend's ICU or FICU, but his K75 parts won't work in my 100.

Bikesmith wrote:All right, I tested the hall sensors right at the ICU plug. Both sensors worked perfectly. It sure is looking like they're not the problem..

By doing this, you have validated the circuitry from the HES to the undertank plug. What it the state of the plug itself and the wiring from the harness to the ICU. Can you test the HES output at the ICU on pins 5 and 13 in the same way?

Here is a link to a BMW Diagnosis User Manual with a variety of tests to perform on the bike systems: https://drive.google.com/file/d/15l0WZVibIlK4gW9FEnDGdCsVz4d3d31M/view?usp=sharing

The sparking coil secondary is the same for Cyl. 4 as it is for cyl. 1. So how can you measure the same secondary for two different plugs with your timing light and get two different results? Would expect to see both 1 and 4 at the same frequency. There are 2 HES's, 2 coil outputs going to 2 coils with shared secondaries. Both 1 & 4 should fire at the same time, and both 2 & 3 fire at the same time (bu 1&4 fire at a different point in the engine cycle than does 2&3). So your result is indeed most strange. I don't know what to make of it other than possibly your test method is flawed.

Can you test the resistance of each spark plug lead. It should be 5KΩ with the lead completely removed from the coil and plug. Use the information in the manual above

Which plugs are you using? I can't find that in the thread,

Measure the primary resistance and secondary resistance for each coil. Use the information in the manual above to make sure they are in range.

 Hi all!
My rig (K100LT + EML) had some issues in september 2021, that seem to be quite similar. The bike always has been 'solid as a rock', but from one day to another it started to stutter when picking up gas. Or shut down completely at the strangest moments.
All dash lights stayed on etc and the bike always started immediately afterwards, but this behavior was anoying and in someway dangerous, as the shutdown sometimes happened when overtaking at higher speed...

I searched for a long time, replaced and checked almost everything mentioned in this post but with no result.
The solution appeared to be very simpel: one of the contacts on the back of the relay holder from the fuel injection relay had become loose. Because of the tied space the cable was in however, this connector 'leaned' into the relayfood. Most of the time: it was in contact and everything was fine... 
You can imagine that roadbumps etc cause the connector to slip out and there you are: engine 'dead'...

I'd say: give it a try, Bikesmith? There's nothing to loose...

https://i.servimg.com/u/f77/18/65/25/09/20210910.jpg
(Image showing the culprit in the relaybox under the tank...)

"daveyson has run his brick with no running fuel pump... "

When I first read that I thought, no I didn't, I ran it without the load shed relay. Then I thought to meself, that's strange, then the penny dropped, then I realised, ah yes, I did do that. 

The different readings with the timing, I thought to meself, that's so strange it must be the coil to be that strange. 

With the ICU I'm thinking it gets input from pin 5 to provide output at pin 14 (and input at 13 for output at 9) but I'm hoping it's a bad coil cause that's a simple fix. I think pin 14 passed your test, but maybe you didn't have your intermittent problem during your test. Intermittent problems are a problem.

Is this problem now permanent instead of intermittent, and how did you get your readings for pins 9 and 14?

Wasn’t your problem a won’t start problem and stops problem. These bricks start on three and even two cylinders. If yours doesn’t start or dies then it will be large loss off fuel or power to several cylinders. It’s puzzling

Puzzling it is. 

A couple of things. 

Thing one. Does the timing light clip have a directional arrow and does it always point towards the spark plug, including cylinder four? 

Thing two. Question: Let's say cylinder four is under compression, does that mean the circuit would go through cylinder one first, then reverse for the next spark? (I think so not, but I'm starting to go non linear here)