Spark cutting out: Please help

[PeterCip]

I didn't know where this should go, so i'm retrying over here..

MSNS-E029j - V3 board, MS-1, 2.25MT

Ok,
I am having a problem with what i believe to be the output circuit for my ignition circuit. I have been having problems lately with the spark cutting out randomly. The car literally stops sparking for a second or so at a time. WHat i have observed is:
1. Nothing out of the ordinary from the logs
2. MT thinks its still giving spark (reads whatever the map value of tming is)
3. AF ratio goes whacked (reads open air)
4. Serious afterfires when spark returns
5. Measured with timing light - no spark from coil
6. Measured with timing light - no activity from the ground lead from coil (pin 36?37?)
7. IRQ trigger light remains the same, Spark Output LED stays flickering.

What i can tell you, is that I ran the Ignition driver with spark non inverted for a little while when i first started tinkering with timing..

All these things considered, I am thinking that the output circuit is somewhat screwed up, and the VB921?(ignition driver) is not grounding the circuit. Also:
1. It rarely happens when car hasn't been running long..
2. It frequently happens if i have been keeping the engine in higher revs.

Could my Ignition driver be over heating? What could be causing the circuit to not ground? I'm also attributing small misses and pops while the car is running to electrical hiccups.. since my timing light pulses when they happen.
Thanks,
Pete

This is a 60-2 wheel decoder setup, with a board that was purchased from DIYautotune.

[Keithg]

I never saw this with 029J, but do know thaty James fixed a number of things in the final release 029q.

KeithG

[jsmcortina]

As you observe the spark output LED continuing to flash but get no signal at the coil it seems safe to assume the VB921 is faulty. I would replace it and try again. Run with the Megasquirt case off and feel if the VB921 is getting warm/hot. If it is then reduce the dwell.

James

[PeterCip]

As you observe the spark output LED continuing to flash but get no signal at the coil it seems safe to assume the VB921 is faulty. I would replace it and try again. Run with the Megasquirt case off and feel if the VB921 is getting warm/hot. If it is then reduce the dwell.

James

Thanks James,
I plan on getting a replacement in a couple days and will re-install the VB921 then. In the mean time, I'm gonna drive the car tomorrow and check the driver to see if it gets hot.
I'll check back in with a log, and more information.
-Pete

[PeterCip]

I replaced my VB921 last night, and its running cool to the touch. In the half an hour i drove around, i didn't notice any missing sparks. ;)

I also lowered my dwell to 2.6ms running, and 4.0 cranking. Is there a way to tell when the dwell is too low? car is an inline 6, reving to 7000; single coil.

Dwell is confusing..

[Bernard Fife]

Is there a way to tell when the dwell is too low?

PeterCip,

Yes. If the spark misses at mid rpm and high load, then dwell likely needs to be a bit higher.

Dwell is just the amount of time current passes through the coil. This current charges the coil. But it only charges the coil as long as the current is increasing. The trick is that the current eventually levels out, and the coil is no longer charging, but instead the current is just heating up the coil. So you want the dwell to match the point where the current is no longer increasing.

If you set the dwell too short, you will get a weaker spark (and possibly a mid-range mis-fire, as mentioned above). If you set dwell too long, you will overheat and possibly damage the coil. So you can see you want it as short as you can get it, as long as the engine still runs well under all conditions. In that respect, it is like setting the PWM% for low-impedance injectors.

What I do is lower the dwell until I first notice a mis-fire, then increase it by about 0.2 to 0.3 milliseconds.

Lance.

[PeterCip]

Lance,
I was figuring that was the easiest way of determining dwell setting.. The only problem with waiting till i feel a miss is that it might be difficult to determine whether it was the timing valules, or the dwell.

Does 2.6 seem low for an inline 6?
Thanks,
Pete

[Bernard Fife]

PeterCip,

Yeah, the problem with this, and most other methods of tuning, is that it assumes everything else is right.

So in the real world, you have to keep going back to things to check them. So after you've dialled in the VE table some, then go back and see if the dwell is working well (and vice versa!).

2.6 milliseconds dwell seems in the normal range. For example, a good value for a HEI external coil is 2.5 milliseconds, and for the HEI 'in-cap' coil it should be about 3.5 milliseconds. The dwell depends mostly on the coil design, not the engine.

As an aside, the shorter the correct dwell, and the fewer the number of cylinders, the higher the rpm you can run while achieving full dwell (and consequently full spark intensity). So for the HEI systems, the external coil is a better piece for rpms above 3500 or so on a V8 (since it maintains the full dwell period). The same coil would maintain full dwell up to 7000 rpm on a 4 cylinder engine.

Lance.

[PeterCip]

PeterCip,

Yeah, the problem with this, and most other methods of tuning, is that it assumes everything else is right.

So in the real world, you have to keep going back to things to check them. So after you've dialled in the VE table some, then go back and see if the dwell is working well (and vice versa!).

2.6 milliseconds dwell seems in the normal range. For example, a good value for a HEI external coil is 2.5 milliseconds, and for the HEI 'in-cap' coil it should be about 3.5 milliseconds. The dwell depends mostly on the coil design, not the engine.

As an aside, the shorter the correct dwell, and the fewer the number of cylinders, the higher the rpm you can run while achieving full dwell (and consequently full spark intensity). So for the HEI systems, the external coil is a better piece for rpms above 3500 or so on a V8 (since it maintains the full dwell period). The same coil would maintain full dwell up to 7000 rpm on a 4 cylinder engine.

Lance.

Is there any quantitative way to measure the coil? IE, take its primary, secondary resistance.. input them into a Calculation in order to find its optimal dwell?

The problem with these coils.. they have no ratings on them what-so-ever. All I know is that its bosch #0 221 118 335

Thanks,
Pete

[Bernard Fife]

Is there any quantitative way to measure the coil?

PeterCip,

There sure is. It requires measuring the resistance and inductance of the coil's primary circuit (which means you need an LC or LCR meter, unfortunately). There's a fair bit of info on this here:

http://www.msefi.com/viewtopic.php?t=7281

Lance.