NefMoto

So I'm converting old audi I5 turbo with hitachi ECU to use typical VAG coilpacks (kind with integrated transistor).

Can someone give me some insight on dwell on ME7 for this, so I can apply to old hitachi ECU?

I saw some map, base dwell - seems to be ~3ms most slots. But there must be another map for battery correction, right? and at what battery voltage is the base map correct?


The way the hitachi ECU works is pretty simple, but I think it can work. It has a very simple UB compensation map, * RPM = dwell in ms. (so for a given batt voltage dwell is constant (in ms), unless the time between cylinders is too short to meet that target, then it is as long as possible). So simply adjusting the factor there to get amount of time ME7 uses is no problem, I just need to know the battery correction map values in typical ME7.

However, below 400 rpm, it uses fixed 'cranking' ANGLE dwell instead of time. (27 degrees).
so
200rpm = 22ms
300rpm = 15ms, etc, until it hits 400 and fixed time takes over.

Which is a lot of dwell. That will only happen during cranking. I wonder if it is too hard on the coils and I should fix it (this will be quite a bit more work, unless just chopping it to a smaller fixed angle is sufficient).

thanks for any advice.

During cranking there is a problem with precision, also battery voltage is very low, so you need to have "stupid" numbers, actual numbers will not be as high, but:
1. During cranking start with multiplying the map by 0.3 or so.
2. For dwell, unless you have a massive turbo, aim for 2.5ms at 14V and do some interpolation, so you have roughly 2.0ms at 16V and 3.5-4ms at 10v.

M2.3.2 I5 only uses angle, you can look at my settings for TFSI coils here: http://m232.org (http://m232.org) - they are converted to ms in the xdf.

Ok, thanks for info. I think that is why hitachi goes to fixed angle.

I guess I can characterize coils with oscilloscope to see what time it takes to saturate at cranking voltage.. Not sure what cranking voltage is, though. Probably pretty damn low with battery in the back seat, I'd think.

Ill check through 2.3.2 stuff, if for nothing else curiosity. Hitachi uses the same 3 trigger setup, but uses 6800 instead of 8051. (and no fuel, of course).


Hitachi setup is really primitive... There's a timer with latches, it counts flywheel teeth.
> Wait X flywheel teeth after last spark (dwell holdoff) - set the latch, start dwell.
> wait Y flywheel teeth after 60BTDC pin - reset latch, fire coil.

the X being derived from RPM and batt voltage, and Y from load/rpm maps + temp