DIY LTCC or similar system for LT1s

[spfautsch]

I had a chance to do some preliminary testing on the D580 and D581 coils over the weekend, and the results were very promising. I wasn't able to perform any measurements other than observing the timing and quality of spark produced, but nothing thus far leads me to believe any of the common GM coils have significantly different igniter current requirements.

Someone knows a good, DYI method of measuring dwell time and the current that the coil draws from the pcm.
I think that the resistors are kind of current limiting device protecting the chip. So how much is needed to ignite the coil and can that be measured reliably without a scope or some fancy equipment.

I managed to confirm the resistors value on the 1mb pcm with blue color. It is again 430 ohms.

Those resistor values are only relevant for the igniter circuit - think of it as a low-level trigger. These draw an incredibly small amount of current at TTL levels (0 to +5v). On the other hand, the current being referenced in what you quoted is primary current (the 12v input) on the coil and this is the amplified signal that results from the igniter IGBT / MOSFET and will register a significantly larger power level. This is what puts power into (aka "charges") the coil and is fed by the circuit I was mentioning might need a larger fuse due to more than 2 coils being charged at one time.

I have a methodology mostly fleshed out for testing things, and will try to document my process very thoroughly. And believe it or not I actually have some formal training in electronics engineering, unlike all my other "skills" which are completely self-taught.

I have an adjustable power supply on it's way (from China) and the v1.1 circuit boards should be here tomorrow, so with any luck I should be able to do some fairly thorough testing this coming weekend and through next week.

The method you quoted for measuring charge time until what they refer to as "current limiting" is mostly accurate, but somewhat flawed. Most of these coils have built-in dwell limiting logic that takes into account all the variables (system voltage, temp, etc.) and starts to clamp charge current after reaching some pre-determined maximum charge state (measured in Joules). Ideally, we want to find the amount of charge time (dwell time) that provides the maximum discharge energy well before incurring the coil's dwell limiting. So to jump off on a tangent, with coils that have built-in dwell limiting it would be incredibly hard to actually cause a coil to fail due to over-dwelling. It's not clear whether the D580 coils have this function, but I intend to test and hopefully determine scientifically for these as well as all the others.

I wish I had the time and patience to try to convey everything I've learned about how inductive discharge systems work over the past year, but it's been a long day filled with idiots complaining about their inability to make phone calls, so I'm going to decompress now with a double (ok, quadruple or sextuple) bourbon. But believe me when I say I'll do my best to clear up all the mystery after I've been able to give all five of the coils I have a thorough shake-down with instruments and what-not.

The baseline of measuring is 12 volts. At 12volts the voltage correction must be zero. Below 12 volts the correction is positive and above 12volts the correction is negative.

We're making the tables from scratch, so it doesn't really matter what we consider "nominal" system voltage - 12, 14.4, 13.7, etc. it's all a linear equation that arrives at the same result. As evidenced by the GM dwell vs temperature tables, using 0.864 (the 2014 corvette bin) as a default dwell multiplier is evidently more meaningful than 1.0. So clearly I'm stupid in this regard. But I have two twenty-something children, so I've become fully accustomed to being "stupid" since each of them turned fourteen.

You can be suprised but the end of injection target calculations in ls1 pcms is borrowed from $ee code, so it is continuous code upgrade in the years. If you missed the beginning, it will be very hard to understand the later code.

That doesn't surprise me at all. I've worked with computer-illiterate types for the past 25 years, and learned first-hand how "we've always done it this way" becomes scientifically acceptable conjecture in such scenarios.