If you haven't assembled the PCB you're probably fine. Desoldering the RC network resistors on the coil outputs was kind of a pain. I lifted several of the pads but luckily none that required traces to be repaired. If I had a soldering iron that wasn't made by fischer-price I might have been able to do better.

The coil indicator LEDs are probably not a dangerous amount of load for the AVR if you're using the same coils I am. More details on the coils later.

I'm posting my math for everyone to double check, please don't hesitate to do so and post any corrections.

Voltage drop over the 1k ohm current limiting resistor driving the red LED indicating the spare output (I configured it for a tach driver) measured about 3.2 average - peak of 3.6 at turn-on. Sticking with average that's 3.2 / 1000 = 0.0032 A or 3.2 ma. If my math is correct there's no worries here. With that said I'd still eliminate / omit the coil indicators - they were really only intended for experimental / diagnostic purposes. If anyone thinks they'll prove useful I can split their grounding network on a re-designed pcb so they can be configured with a removable or soldered jumper.

Moving on to the coil igniter circuits, I replaced the 470 ohm resistors with 100s. Previous testing showed 0.425ma draw. Getting a clean sync on the signal was difficult but I was able to get a faint image about every other engine revolution. Across the 100 ohm resistor I was seeing 60 millivolts. This waveform had a surge at the beginning then settled down to a flat voltage level - the peak was probably 80mv but I'm going to use 60mv. So 0.06 / 100 = 0.0006 A or 0.6 ma. This is well within the capabilities of the AVR even with coil indicator LEDs and still leaving headroom for other loads.

I am somewhat baffled by this current figure. The reason I decided to research this topic further is because I was afraid the coil igniters weren't getting driven into saturation and spark energy was suffering as a result. Unfortunately this critical piece of data doesn't appear readily available on the interwebs. What would really be useful here are some standardized specification data. Search as I might, I've found nothing similar to a datasheet on any of the GM CNP ignition packs. All I have to work with are anecdotal reports.

My primary source of data was the oft-mentioned megaquirt 2 sequencer manual [link]. In all references to the igniter circuit draw, "a few dozen milliamps" is given. Also salient is the mention of 1000 ohms in series (internal to the sequencer) with the output circuits being used to drive the igniters.

Another tidbit I found in the beginning and then re-discovered a few days ago is [this] post on allaboutcircuits.com's discussion forum by a college student struggling with a very similar project for his 99 Infiniti. Seems like he had problems with connected and / or induced interference. But the recurring theme seemed to be 20ma at TTL levels on the igniter pin.

So, barring someone coming forward with veritable treasure trove of technical data on coil packs, I suppose what's needed is independent testing not only for igniter current but ignition energy relative to dwell time.

Based solely on appearance, I purchased 8 coils off eBay for a very reasonable price that I believed to be these LS2 "hot" car coils.



GM PN: 12573190
AC Delco: D514A

But I found just a few days ago they are in fact these.



GM PN: 12658183
AC Delco: 12674754

Based on application data it seems like the coils I have might be a considerably newer design than the 4ms "hot" coils. But the current draw on the igniter input indicates they're anything but similar to the 12573190 coils if the anecdotal data mentioned previously is accurate (20 to several dozen milliamps). This is a huge difference indicating the coils I have utilize a shielded FET igniter instead of a more standard IGBT.

I've found a source for some of the D514A coils on A website for about $20 apiece so I'm seriously contemplating buying a set. Also logical would be testing the "legacy" LS1 coils with the exposed laminated core.

For those of you planning to install the test setups I'd recommend you pull the brakes if you've already chosen a coil pack, as I'd really like to test the controller driving at least one of each of the variants, especially these beasts.



GM PN: 190005218
AC Delco: D585

So barring any volunteers willing to loan me hardware I guess I'm going to build a test platform and start collecting parts.

I was able to drive it some today and it was better, but not absolutely impressive as back in May. I'll try to post more info relevant to drive data in the "humidity" thread [link], but it may be several days.