KTM 640 - engine randomly dies

No sidestand (or clutch) switch on the bike. All that was removed before I bought it 4 years ago.

Was on a trip, a buddies lc4 did something very similiar. Turned out to be a very small pebble in the float bowl. Periodically it would clog the intake, bike would stall. Then it would run great for x amount of time.

Good luck

I also got a 98 manual when I bought my 99?

I'd check voltage to the coil.

If you have an inductive timing light, hook it up to the plug wire and see if it's got a reading. Can also remove the spark plug cap and lay it against the engine. The spark should have no problem jumping from the cap to the motor.

Might want to check the plug. I found this on my Vstrom a few years back.

<iframe frameborder="0" scrolling="no" width="640" height="480" src="https://api.smugmug.com/services/embed/821494259_MJsqLX7?width=640&height=480&albumId=2675309&albumKey=q7mmS6"></iframe>

...dual plugged heads meant the bike ran OK, just sounded funny.

Can someone confirm socket size needed for the plug?

Braving the cold. She started right up. Letting it warm while idling.

Intermittent electrical problems are a pain in the ass.

Have multimeter and manual opened to section on testing ignition coil. Manual provides values for two resistance tests - blue/white to ground, and blue/white to plug wire.

At the top of the ignition coil test section the manual says that I need to "disconnect all wires" before using the ohmmeter to test for the given values.

The specified values KTM provides are

Below this table of values, the manual then says

"Before checking the ignition system check..." and goes on to list whether the ignition is on, the kill switch is on, the neutral light is on, motorcycle can be started with clutch in and that the battery is charged.

What is confusing to me is the seeming incongruence between the introductory paragraph, and the second paragraph. From reading other sites, resistance testing should be done with no voltage present. Is the KTM manual saying the same thing, but to first make sure these other things are OK -- before disconnecting everything?

Assuming I should have the battery disconnected, I'm showing "1. " on the multimeter readout on both tests, using the battery NEG as ground. If I tough the leads together I'm showing ~0.00. The MM is set to Ohms at 20k (the lowest value).

Please advise..

The book is telling you to check the easy stuff first. Should also ask is there gas in the tank and is the fuel petcock on.

To resistance test the coil, it has to be disconnected.

But if the bike is running OK, and you're making these checks than it probably won't show what the issue is.

I was suggesting that you make sure the coil is getting voltage. That would be the wires to the coil and should be 12volts.

But again, any checks really need to be made when the bike won't run.


So for me, the first thing I would do is wait til it stalls out again and can't be restarted, then verify if you are getting spark. You can do this several ways: pull the cap and let the plug hole rest against the motor. Crank and look for a spark (don't touch the cap while cranking unless you like 10,000 volt zaps)

Also pull the plug and re-install in the cap. Ground the plug against the motor (again don't touch while cranking) and verify spark.

If you're getting spark then the intermittent problem may have gone away while you were unhooking things, or it may not be a spark problem at all. (pain in the ass intermittent electrical issues)


Good Luck.


BTW, 20k ohms shouldn't be the lowest setting, usually there's a 200 ohms setting, though I don't know anything about your multi-meter.


BTWBTW, I'm not convinced it's an electrical issue. A fuelling issue can shut the bike down immediately. When people think of fuel issues, they thing of running out of gas where the bowl steadily leans out, you get some sputters as fuel sloshes and gets sucked up then none then sucked up. But, like what someone said earlier about a pebble in the float bowl, if something cuts fuel to the jets immediately, clogs it immediately, there will be no sputtering. It will shut off just immediately like you've cut the ignition, especially on a thumper.

Quick reply -- I used a timing light to confirm there was spark a bit ago -- seconds after it turned over and fired. It was getting spark - the timing light flashed. I rolled the bike out and crunk it up, let it idle. Ran 2-3 minutes then shut off as above. I put the timing light back on, and this time when I turned it over I get no light. No spark.

I'll double check the multimeter reading -- there is a 200 value on the dial.

Back in a flash.

Boon is right, manual is telling you to check the basics in a logical manner and not stuff around testing things willy nilly.

So, you have confirmed its electrical, theres no spark when it plays up. So now why is there no spark, no power to coil most likely?

Start with power, then move on to testing coil, cdi pickup etc.

Your not making much progress, I can suggest a few things...

#1- I would recommend that you pop the outer flywheel side cover off and have a look at the stator and pickup coil... Look for any rub spots or shorts in the wires running in and inspect the faces on both the pickup and stator where they read the flywheel for debris scoring...

#2- Sometimes the issues that you describe can be caused by something as simple as a faulty spark plug cap or even the spark plug itself...

#3- When I had similar intermittent problems with my 640 A the problem ended up being some broken wires inside the ignition key switch... With the engine running try gently yanking and wiggling the wires running into the switches base cup to see if it kills the engine...

#4- Make sure you have good routing and slack in the manual decomp cable, I have seen it happen to where this cable gets hung up to where a small movement of the bars lifts the lever arm enough to kill the engine...

#5- If you haven't cleaned the carb, it wouldn't be a bad idea to take it apart, clean and inspect... Look for a visual wear witness ring on the rubber tip of the float needle... Some times a worn float needle can cause erratic running/stalling issues but it usually is not instant like an electrical issue...

Good luck...

Negligible voltage coming off of the blue wire from the part where the red MM lead is touching (solenoid?)



In the picture I am checking voltage off of one of the big wires (this left goes to the POS battery term, the right I think goes to the starter). I'm showing 12 volts with on the left term and none on the right. When starter is engaged the voltage drops to about 9 for both big terminals. The blue wire continually shows negligible voltage ~.05 .10v.

I tested resistance on the coil wires using the 200 setting. This time I get ~.20 from blue-white to ground, and still 1._____ from plug wire to blue-white.

Anyone know what voltage I should expect to see on the blue wire coming off the relay? Seems like I should see more than .05-.10v To be clear - the voltage on that blue wire (which routes to the coil) does not change when the starter is engaged.

The blue wire on the starter solenoid shouldn't have anything to do with the coil. Thos two small wires energize the starter solenoid. Maybe blue is the ground on the bike? I haven't looked at a wiring diagram.

Either way, if you push the button and he starter cranks the wires at the solenoid are doing what they are supposed to be doing.

No idea, yours being a 99 will be different to mine and probably to most peoples as I think that was a 99 or 2000 was a funny year which had a lot of differences.

Refer back to your wiring diagrams and manual, but I would be looking elsewhere first as you've probably gone off track.

Ok, disappointing weekend. Thought the loose connection I fixed a few days ago would surely fix the problem.

I'll work at this more next weekend. Sorry if I sound like a retard - never done any electrical troubleshooting.. all new to me.

Gotta be systematic. Like Auzzie Steve said earlier. You gotta make sure all components of the ignition are working. Make sure coil is getting voltage, make sure pulser/trigger is sending a signal etc. Systematic process of elimination.

And still you've gotta check that plug and cap. If there's no place for the spark to go, either due to bad cap or internally shorted plug, or the high voltage wire is shorting to ground before the inductive clamp, there won't be a charge for the inductive timing light to read..

Should voltage only goto the coil when the engine is being turned over? If I understand these systems, I think that's right.. then I'd need to hook up everything, and slide my POS lead into the coil blue/white connector, and to the battery ground for the other. Please correct if I have this wrong.

Usually there will be a constant 12 volts to the coil. The coil is fired by cutting voltage, usually by controlling the ground side of the coil.


EDIT: Looking at some of the online wiring diagrams, something doesn't look right. Shows blue/white wire from the tach to the coil to the ECU. Either they are supplying power to the coil from the tach (?) or the diagrams aren't correct... or the coil is running on voltage from the ECU and the coil is frame ground?? If it's the last then you may not have a constant 12volt at the coil. It may be pulsed. I'm not familiar enough with my 640 yet to tell you the correct answer.


EDIT EDIT.. Your post answered my question. The check from BL/W to ground means the ECU is supplying power to the coil, but I still don't know if it's pulsed or constant.

Your other check, BL/W to Ignition wire has to be incorrect. That check should be Ignition Wire to Ground = 10,80 &#8211; 16,20 k&#937;. (you'll need to change the meter back to 20k ohm for this check)

Yeah, the wiring diagram seems off. My bike doesn't have a tachometer - only a mechanical speedometer and indicator lamps in the dash. I mistakenly followed the blue/wt wire back to the relay starter - need to double check that and find where my coil voltage originates.

Edit -- the "k" from the chart blended with the ohm symbol, missed that. I'll run the test again in the next couple days and post findings.