2009 Harley engine difficult to start

[tryskriel]

The engine is finally running but have a lot of difficult to start it.
It run fine when we start it but very difficult to start.

Tach signal seem a little bit confused when spark plus are on it

I can remove both spark plug and i run it at 362rpm when cranking very stable but i just install my spark plug and tach go from 0 to 450rpm and come back to 0.
I did try to play with Tach signal masking, next pulse tolerance, cranking rpm, rpm lag factor but nothing seem to work fine.

Those engine run really ackward and more when it's time to start.

Anyone would have an advice

[tryskriel]

We have burn another starter but we remove it and did try to put it on the chassis dyno with a car at side to make it run what we want without the starter.
Have play more with second trigger angle, have time it to 225 degree and with the timing light, both spark occur at right place, bike seem to run fine, good AFR but the rpm cut each 10 seconds

We did install some decompressor on the heads to help a bit with the starter but we would need a bigger starter

The tach is very stable but it always cut the signal each 10 seconds, dont know why

I did play with
-Tach signal masking
-Next Pulse tolerance
-Prediction setting
-Input smoothing factors
=Falling or rising edge

Nothing seem to be better, any idea would be great

[grippo]

That is the first time I have ever seen anything like this. I was expecting the usual loss of VR synch, but that was rock steady - no misses, no noise, trigger+/- was dead zero. So I am not sure what is causing the loss of synch (rpm goes to 0). The only correlation I can see is with the IAC stepper popping closed, then it resynchs and runs. I don't know if this is cause or effect of the problem. Can you disable the IAC both in the software and disconnect it, and see if it still does this ? Don't disconnect while the engine is running - just disconnect before starting/ restarting after engine warm and see what happens.

[tryskriel]

I made a change to the program and it became an error so i had to reload the code and problem have gone after that.
No Resync. Don't know why it came fine but the good is that is working right now.

The harley manifold is so little plus with this two big pistons, the MAP is going crazy, i did smooth the value reducing it to 20% but i will have to tune it in alpha-n blend SD to make it run stable AFR at idle. I have to use the SD feature because it's turbo.

I will keep you updated with a vid when working fine

Thanks anyway for your help
Fred

[tryskriel]

I did tuned the bike on the dyno with a car (just one wheel on the dyno) to start de bike.
The bike is very powerful but all the install seem useless if we cannot start it by his own.

RPM AVERAGING LAG FACTOR
I did change the rpm averaging lag factor to 90%, it did little effect.

PREDICTOR ALGORITHM
Last interval. don`t know how to set-up the alpha beta filter

TACH SIGNAL MASKING
Time Mask 0msec
Perctentage mask 0%

NEXT-PULSE TOLERANCE
Cranking 100%

Do someone can help me on this one to know which other parameter would change my cranking problem. When cold, sometime it will start right away but other time the engine seem to not be able to sync itself, i think it would need to turn around two times min, because the cranking setting are set to last interval.

I did try to connect my crank sensor away from the bike with shield wires, i did install the crank sensor closer to the trigger wheel but right now i`m a little lost.

I remove the plugs and engine see a steady RPM, we did install some decompressor on the heads but it seem to not be enough

Thanks
Fred

[grippo]

With 32 teeth on your wheel that should give you plenty of timing accuracy. I would like to see a datalog of one cold start that doesn't start right away. Or you can look and see if it is getting and losing synch, or if synch is solid and trigger +/- stays at 0. As far as what to change, put rpm lag back to 50% or less, this can't help, and I would reduce the crank tolerance to 75%. I doubt changing to alpha-beta-gamma will help, but you could try trigger rise, but again, you have enough teeth, so I don't think that will help. Do not try trigger return - that will not work for you. It may be that the VR signal during starting is not strong enough because it cranks too slowly. This all assumes it is not a fuel problem. You don't need priming pulse or that sort of thing, do you ? Does it start better if it is already hot ? Using delay teeth of 11 means it won't fire a spark until 11 teeth after synch at the end of the missing gap. If you could move the sensor so you only have to delay a few teeth, that would help. In fact, just reducing delay from 11 to 9 and setting trigger offset to -6 instead of -28 might help a little, but if you could move the sensor 8 or 9 teeth, so delay teeth was 2 or 3, it would save you a 1/4 turn.

[tryskriel]

Hi
Thanks for the reply
I did make those change (i did set the delay teeth to 9 and set a trigger angle of -6)
I did set those other percentage setting like you suggest me

The bike did start one time on 7 try but we did broke the starter another time (4 th) after this try
I didn`t record this start but i did record the starting we did broke the starter so give it a look if you see something special about it

To change the teeth on the trigger wheel, i would have to complete dissasembly of the engine because it`s built around the crank

Thanks
Fred

[24c]

...
The bike did start one time on 7 try but we did broke the starter another time (4 th) after this try
I didn`t record this start but i did record the starting we did broke the starter so give it a look if you see something special about it

Broken starter? You don't have too much ignition advance do you? I only mention this because my engine was like this at first, and it turned out my engine was not at the right position, that is TDC in my MicroSquirt was not TDC of the engine.

[tryskriel]

My engine is timed with a timing light but i was giving the same advance than cars but harley do start at 0 degree and i was starting about 6-10 degree
I did retard it to -3 and it seems a lot better, Microsquirt have difficult to sync the trigger wheel so it start one time for 6 try.

Here's some datalog of starting run
Maybe someone will see something

Thanks
Fred

[tryskriel]

Another datalog