Spark Latency

[DynoDan]

Hi all,

I've got a little question in reference to spark latency.

Background:
American V8 with MSD distributor which has hall effect outputs for crankshaft position [1 per cylinder, 4 per crankshaft revolution] and a phase tooth.
Running sequential fuel on an S6 with all ignition events sent to one output which is being sent to a ISL9V3040P3 ignitor external of the Syvecs which fires the MSD Master Blaster 2 coil. Spark is good and strong with tapering dwell vs RPM to allow for sufficient charge+flame at high speed.

It has previously run on a D585 coil, but the coil struggled to keep pace running 8 cylinders at 6000rpm+.
Ideally I'd like to run more trigger teeth for crankshaft position but I'm working with what's available and that isn't going to change - I'm triggering on the falling edge without issue, idling at 800rpm there is no spark scatter, likewise at higher rpms.

All is well currently, the car starts and idles well and performs as expected OTHER than ignition timing.

If you manually lock the spark to 20 degrees and wind the engine up to 3000rpm you can see the ignition timing diminish in linear form as revs rise. I can't remember the precise figures but there's not going to be much timing left at 5000rpm. I realise some of this is going to be dwell related...

On other hardware systems I've worked with latency values to compensate for this, but I wonder 2 things:

What is the inherent issue that causes this?
Is there a way to compensate for this within the Syvecs SW?

[stevieturbo]

Throw the MSD in the bin and fit one coil per cylinder.

If you have a crank sensor and phase sensor, you have everything in place to throw the MSD and dizzy away ( at least in terms of spark distribution ). Buy another 7x 585's and fit them.

Job done.

In the meantime, the "issue" may be related to which edge of the tooth you are triggering of. Being hall effect I might prefer the rising edge. Or if a dialback light it could perhaps be related to that if you are flashing at anything other than zero dialed in ?
Or it could be related to the external amp you are using.

IMO keep it simple and get rid of all the crap.

CNP fired in wasted spark pairs via 4 ecu outputs.

[DynoDan]

I'm 100% with you, but this car has to retain the setup it has for various reasons that are beyond this conversation OR professional advice.
Trust me, I'd rather run it on a proper trigger wheel with D585s etc.

It's not timing light related - it's reading right.

I'm pretty sure it's due to the decreasing dwell with rpm that's causing the timing to drop off but looking for someone to confirm my thoughts and/or offer a solution. The dwell is decreased against RPM based on some simple maths,

At 7000rpm with the coil firing 8 cylinders [4 per crankshaft rotation] allows 2.143ms for complete charge and discharge of the coil per cylinder.

Allowing for a 0.25ms spark time at 7000rpm [relatively standard points derived time for flame between electrodes] means that there's 1.89ms to charge the coil before the next firing event. If you aim for a more reasonable flame time of 0.5ms then the charge time for the coil needs to be 1.64ms...

If you set a usual voltage dwell table then with a 3ms dwell you hit the limitation of the ignition system at 4.5k..

I'll say it again very clearly though - I fully understand that this is a not ideal setup. It isn't my design/desires but the setup isn't going to change.

[stevieturbo]

I can see no reason why decreasing dwell would cause the timing to change, and you seem to be describing quite a dramatic change ?

But it would also be simple to test it on the rising edge crank trigger to see if anything changes. This edge should not be affected by the rpm or size of the tooth, whereas the falling edge could be ?

[DynoDan]

Just to conclude this for my application.

I spent some time today going through and checking the triggers - with a scope at high rpm the falling edge had somewhat of a knee in it after 1.5v with a noticeable slower response that the rising edge at the same RPMs. This is where my issue lay, at idle and sub 1600rpm the knee wasn't noticeable so the calibration was set to utilise the falling edge with trigger points both sub 1v.

I've adjusted the calibration to utilise the rising edge and now locking the timing the engine can maintain 800-6500rpm without variance in timing.

[stevieturbo]

Good job.

[RICE RACING]

Just to conclude this for my application.

I spent some time today going through and checking the triggers - with a scope at high rpm the falling edge had somewhat of a knee in it after 1.5v with a noticeable slower response that the rising edge at the same RPMs. This is where my issue lay, at idle and sub 1600rpm the knee wasn't noticeable so the calibration was set to utilise the falling edge with trigger points both sub 1v.

I've adjusted the calibration to utilise the rising edge and now locking the timing the engine can maintain 800-6500rpm without variance in timing.

x 2, very good work