crank position sensor errors

[DPrice]

Hello,
I'm having some issues getting our teams WR450 up and running on our Performance Electronics ECU. At the moment I am trying to get the ignition in order while running it on the stock carb. It starts about half the time, and the other half just kind of sputters and fails to fire for more than a few revolutions. I have been trying to verify timing via a timing light but am getting a very inconsistant reading which I am attributing to a funky crank position sensor signal.

We keep getting an intermittent crank position sensor error while cranking as well as while the engine is running. It comes up when cranking, which I understand is normal to some extent, but when the motor decides to fire, we still get a crank position error every few seconds that comes up and goes away.

I got a reading of what the output of the sensor is with an oscilloscope and it looks fairly consistent, but I don't really know what is acceptable and what is an inadequate signal. I also do not know how close the sensor supposed to be away from the trigger wheel? I know that the spacing is important, but is there a way that I can set it without repeatedly moving the sensor and trying to fire the engine?

If anyone is willing to offer some advise or insight on issues that they have encountered it would be much appreciated. Thanks

David

[Grant Mahler]

I have used a couple pieces of steel welded together with a nut to thread the CPS trigger into/out of. However, if you have it installed in the motor and are getting consistent readings via an oscilloscope, then its probably fine. Most of these sensors are very forgiving, anywhere from .040 to 1/4" seems to work.

As for the oscilloscope reading, take a picture and post it. That way we can see what it looks like.

[Steve O]

Can't offer any insight as to tips with your ECU; however, I can tell you a list of problems I had with mine and the solution ( we have a yfz 450 which I assume should have similar behavior as they are very similar motors). Before I get into that however, if your timing light is one that is meant for timing cars or cheap, it probably won't work as the frequency of the ignition is way to high for them... the motor fires every 360 deg.

As far as setup for your ECU is concerned, most ecu's need to be told this motor is a 2 cylinder motor not a one due to the above reason of a fire every 360. The motor does this because it has a crank sensor not a cam sensor and therefore does not know intake stroke vs the exhaust stroke. Second make sure your sensor type is set for internal reluctor. Your problem could lie with the reluctance filter your ECU has and it may need to be adjusted.

As far as the sensor is concerned. I think you should be seeing a couple hundred mv but not sure off the top of my head. However, Grant is correct in the fact that the sensors in these motors are very forgiving. It should be a biphasic pulse every 360 degrees. Two things to know with wiring the sensor, it may be an obvious statement, but make sure you have your polarity correct or you will be out of phase and you will be firing too late. Once you get all of that in order, there is one other thing that has to be done for certain setups with certain ecu's... you sometimes have to put a diode on the negative terminal of the trigger place the diode so it allows electricity to flow out of your ECU's negative terminal but not back in. With a previous setup we were using before I had done that, the engine would run great up to about 8 grand and the ecu would then read that it was at 16000RPM and rev limit the motor and generally stall it because the fueling and ignition was for a wrong RPM.

For troubleshooting purposes, you may want to hook up the ecu parallel to the stock setep and see what the ecu is doing with the signal of the running engine. Check if its getting a proper rpm reading and then throw your scope on the fuel injector and ignition and see if they are sending out the right signals. If they are then I would check to make sure your charge time and other settings are set right for your coil.

[murpia]

Are you using the stock crank sensor and trigger wheel or the PE supplied hardware?

Regards, Ian

[DPrice]

Steve O

Right now we have the ecu set to a single cyl, I will try switching it to a 2 cyl to see if that makes any difference. Our timing light is a Sears advance timing light, so I would assume its good enough for our application. Even if it could not charge quickly enough to fire every revolution, I would think that it still would fire accurately every 3rd or 4th which would be good enough to get a timing reading.

When you buy a PE ecu, you specify to them if it is a hall effect, or a variable reluctance type sensor, so I have to assume that they sent us the right one. There is no step in the installation manual to put a diode in the line, but it seems like an easy thing to add so that I can rule out more things that could be wrong.

Murpia,

I am using the stock sensor with a ring that we machined to have the 12-1 tooth pattern, that we then welded to the outside of the generator rotor where the stock timing tooth was.

Another idea that I had was that maybe the stock sensor does not have the resolution to be sending the sin wave 11 times every revolution as compared to only once with the stock system. When I read the signal with the oscilloscope it was only while cranking. In a day or so when I get the oscilloscope back I will take a reading while cranking the engine, as well as one while it is running. I will post pictures of the graphs as soon as I get my oscilloscope back.

Thanks,
David

[DPrice]

Ok, so quick update with pictures.

http://photos-635.ll.facebook.com/photos-ll-sf2p/v190/243/40/6420635/n6420635_34643855_5253.jpg

This first picture is of the crank position sensor signal taken from the CPS wires about 2 inches before they go into the ECU, so this is with any noise that it would be picking up by being in close proximity to sources like coil wires.

http://photos-635.ll.facebook.com/photos-ll-sf2p/v190/243/40/6420635/n6420635_34643857_7028.jpg

http://photos-635.ll.facebook.com/photos-ll-sf2p/v190/243/40/6420635/n6420635_34643858_7912.jpg
These are closer up views of the same signal. Both were with the engine running at about 5000 rpm. It sounded a little ragged at times, but close to constant rpm.

The only thing that I can see that may be giving an inadequate signal is the changes in amplitude from tooth to tooth. Does anyone with more experience looking at this kind of signal have any input about the quality, or consistency of it? Or maybe something that is obviously wrong with it that I don't see?

Thanks,
David

[screwdriver]

I agree with your conclusion. If you look at the top picture, notice how consistent the deviations in amplitude are.

I don't really know what could cause this. Possible causes:
* Loose sensor
* Sensor too far from tooth
* Excentrially mounted / loose tooth-wheel
* damaged tooth-wheel
* Waaaaay too long wiring
* Faulty wires or connectors

Cloud you post the post the osci-settings (ie div/V and div/time). It could well be that the signal you're measuring there is just too small, ie the Sensor is too far from the tooth.

[DPrice]

screwdriver,

O-Scope settings are ath the top of the pictures, for the bottom picture it is 500mV per division and 200mS per division

Of your possible reasons for the inconsistancies, Waaaaay to long wires is the first thing that stands out to me. The wires to the CPS are about 5 feet long. The wheel itself is not loose, and neither is the sensor itself, although I will double check tonight when I have a chance. I will also move the sensor a little closer while I am checking that everything is tight. I am also going to try to center the sensor on the teeth a little better than it is and see if that does anything.

Thanks a lot for everyones help,
David

[murpia]

The amplitude variation, which repeats from cycle to cycle, suggests some eccentricity in the wheel. But, nothing too serious really.

The signal amplitude of +/-1V (2V peak to peak) needs to be checked against the ECU spec's to be sure it'll trigger. The PE manual mentions minimum peak of 0.8V for cranking, but does not mention running specifically...

Also, I would expect the ECU to trigger on the zero-crossing of the signal, so amplitude variations with wheel speed don't have an effect. Whether the ECU triggers on the positive or negative crossing needs to be checked with PE, this can put your timing out a bit and usually the sensor performs better in one orientation and connected a certain way round. That will also depend on whether you have wide teeth and narrow slots or vice versa.

At least you have a scope and a running engine to check all this stuff, I'm sure you'll work it out.

Regards, Ian

[Grant Mahler]

Originally posted by murpia:
The amplitude variation, which repeats from cycle to cycle, suggests some eccentricity in the wheel. But, nothing too serious really.

The signal amplitude of +/-1V (2V peak to peak) needs to be checked against the ECU spec's to be sure it'll trigger. The PE manual mentions minimum peak of 0.8V for cranking, but does not mention running specifically...

Also, I would expect the ECU to trigger on the zero-crossing of the signal, so amplitude variations with wheel speed don't have an effect. Whether the ECU triggers on the positive or negative crossing needs to be checked with PE, this can put your timing out a bit and usually the sensor performs better in one orientation and connected a certain way round. That will also depend on whether you have wide teeth and narrow slots or vice versa.

At least you have a scope and a running engine to check all this stuff, I'm sure you'll work it out.

Regards, Ian

Big thumbs up. I second all of these statements. Crossing pos/crossing neg should be able to be set (sometimes called rising/falling trigger).

[Pete M]

I'm not real familiar with the PE ECU, but if it were a motec i'd say a few things. Make sure you have it configured as a rising trigger, because that is the most consistent edge in your traces. If you can't change that on the PE, then flip the wires.

Otherwise, check that you're within spec amplitude wise. On the motec, you can set the minimum voltage as a function of RPM (for noise immunity) but other ECUs might have it fixed. Either way, there'll still be a minimum. Moving the sensor closer to the wheel (or using a bigger wheel) should make the voltage higher.

[B Lewis @ PE Engine Management]

Hi David,

Sorry for coming in late. I missed this thread earlier. If you are still having problems, give me a call at (513) 777-5322 ext 211 or drop me an email at brian@pe-ltd.com to discuss.

[DPrice]

Another update...

I moved the sensor a bit to center it in the middle of the teeth. When I rechecked the signal from the CPS with the o-scope the variations in amplitude were still there but the minimum amplitude was right around .8V while cranking as opposed to around .4V cranking before I moved the sensor.

I tried running the engine and it started and ran a whole lot better that it did before, but still not great. There were no crank position errors,(yay!) but spark was still not quite consistent. I checked with the timing gun again and every half second or so it would flash way off from the top dead center mark. So now I guess I have to track down some other source of error. I would assume that incorrect settings in the coil charge time could have some effect, so thats the first thing I am going to double check.

Thanks everyone for your help. I really appreciate the advise and support. I hope everyone's cars are coming along and I'm looking forward to seeing what teams come up with this year at comp.

David

[TMichaels]

Hi there,
just want to add my two cents.

Always use twisted pair wires for inductive sensors.

When using inductive sensors, the following is always true: The signal amplitude becomes higher, if the change in magnetic flux is higher-> centre the sensor as good as possible over the trigger wheel; move the sensor as close as possible to the trigger wheel; use trigger wheels with a big difference between tooth and not-tooth (sorry, not a native)

I am not familiar with the PE ECU, but you should also consider the compensation for the ignition coil charge time over battery voltage.

Regards,

Tobi

[Steve O]

Sorry I haven't had a chance to respond since way earlier in this post...our craftsman advanced light didn't do anything for us and just fire erratically most of the time. However, if you are getting it to work and it seems like the results make sense, then your flashing being way off could mean your trigger edge is set wrong for the motor (I would advise against switching your edge based solely on the signal looking slightly nicer, as suggested before, as this changes the timing of your engine completely.), and make sure the polarity, of the CPS wires, is correct. You can do this with your o-scope by seeing which one is naturally a sine wave and which is a cos wave. The sine should be your pos. signal. Also, keep in mind that you should also have another edge trigger setting specifically for your coil setup in your ECU. Changing that will vary when your coil is firing.

Steve

[Kirby]

Probably a bit late, I agree with what murpia said. Likley small eccentricities in your wheel manufacture are leading to the amplitude variation.

That said, I have seen signals with worse noise; where the engine has run fine.

Could be an issue with your trigger settings.

What is the distance between your missing tooth and TDC? Is there enough time in there for the ECU to do its calculations. Usually 35deg is a pretty good choice. Is the distance between the gap and tdc correctly set in the ecu?

[DPrice]

I wanted to thank everyone for their input. I finally got the problems figured out and now the engine runs pretty excellent. I redid some grounding points on the engine and ecu, and got some connections a little cleaner and that seemed to fix the ignition issues I was experiencing after I fixed the CPS. Thanks again.

David

[netogodoy]

have you guys figured out what is the maximum sensing frequency for the stock sensor ?

Could you guys point out wich sensors could replace the original ?