Deeper and Deeper in... then OUT of the Weeds
It's been a busy Summer so I can't say I've spent the entire time frustrating myself with my non-starting project car and my inability to sort it out. But it feels like a massively disproportionate share. Frustration and defeat are very de-motivating, as is a 98 degree garage, so even when I did have time it wasn't always easy to get myself out to the garage and get to work. I can almost convince myself that I worked at the right pace because if I'd crammed this amount of annoyance into a shorter time frame my head would have exploded two or three times by now.
After a lot of fruitless investigation, and fixing and replacing things that weren't broken, I finally tried spark test mode with ALL of the spark plugs out in the air. This revealed that on the Ignition1 channel, ALL FOUR plugs were firing - and firing very weakly at that. Ignition 2 was firing cylinders 2&3, as it ought to have been. So what's up with Ignition1?
The oscilloscope was hooked up to the ECU's ignition out leads, and showed just what was expected when testing the ignition channels individually - a spark impulse to one coil pack, quiet on the other side. So the ECU seemed to be doing its job.
The first generation Miatas had both coil packs driven by a single igniter which made for some strange wiring contortions. A few minutes with the multimeter showed there was no conduction between the wiring for Coil A and Coil B, so what's making A and B fire together? Bad igniter was an obvious possibility, so I tried a new one - same mysterious behavior. Now I had a real puzzle on my hand, and one that baffled the Miata.net forum, the micro/megasquirt forum, and the ECU tech support people.
I consulted with Uncle Brother Ray the Engineer and Crack Shot Andy (who is also an engineer, and both are car hobbyists). They applied their tremendous brain power and extensive vocabulary of made-up terms like 'induction coupling' to the problem which yielded the only plausible theory: it must be electrical noise from one channel triggering the other channel.
So I looked in to filtering capacitors and shielded wire and then belatedly noticed (right side picture above) that I had neatly bundled the low energy ECU wires for both channels right next to the high energy coil pack wires and, for good measure, added a sleeve to make sure this ideal noise coupling arrangement stayed in place. Distance is better than any kind of filtering or shielding (inverse square law) so I un-bundled these wires and arranged them as far apart from each other as possible... and still got the same maddening inexplicable "four weak sparks all at once" behavior.
Then I randomly tripped over the fact that the 1999-2000 Miatas had coil packs with their own built-in igniters. This would simplify the wiring dramatically by eliminating the detours to the igniter and then back to the coil packs.
Modifying the 1990-1997 coil pack bracket to accept the 99-00 coil packs was a small project, but nicely tangible, with no invisible forces or electrons that don't really move moving at the speed of light or need for imaginary numbers like √-1 to calculate what's going on. This made the wiring so much tidier and more satisfying. I would have done it this way from the beginning if I'd known this was a thing.
Better still, it worked. Spark test above. If you look closely you can see two plugs firing at the start of the video, then the other two when I switch channels, and NEVER all four at once. Not really clear from the video, but these sparks are visibly stronger/brighter than any of the sparks from the previous arrangement and dramatically so compared to the weak 'all four at once' gremlin sparks.
Even better still yet, the engine now starts. I was able to get it to start immediately after this change, and with a slight timing adjustment, it was capable of idling with the throttle closed and kept running until I turned it off.
So halleluljah and praise be!
This has been a very prolonged frustrating and discouraging ordeal. The engine still needs a lot of tuning, but "Why the &$(&*&@@!!! won't this *$(&@ing thing start?" and "needs a little tweaking" are totally different worlds.
There is a lot of lifter noise (the clickety-clack sound). These are brand new lifters, installed almost two years ago, and only now being run for the first time. So they'll probably quiet down with use. First generation Miatas are notorious for lifter noise so this may never go away completely.
Mentally editing out the clickety-clack, the sounds qualities are most satisfying. The exhaust end is a very quiet low-end rumble. The intake side is an exuberant whoosh when the taps are opened - a little loud now, but should be just right when it's under a hood. Not bad, considering the intake is my own redneck contraption design and the exhaust is partly first gen, partly second gen, and a few parts I made myself.
There is good oil pressure. It seemed to get up to temp very quickly and the radiator fans never came on (albeit, it was only running for about five minutes total). So next time I start the motor I'll watch the temp gauge. And turn on the data logger, for that matter. My dashboard gauges are of uncertain interpretation since I made my own graphics for them and installed the pointers somewhat randomly - but when the laptop is hooked up to the ECU I get actual live data so this won't matter in the long run.
Comments
Post a Comment