Correct, optimal dwell time is not variable, meaning the amount of time it takes to charge the coil for the given wiring configuration does not change. However, some very smart EFI controllers will vary dwell time and shorten the dwell for light load engine conditions where you may not need as much spark energy. But I think this practice is debatable, I believe it's best to simply fully charge the coil if time permits. One other dwell time change we have not touched on is battery voltage compensation, if battery voltage drops like during startup cranking dwell time goes up. The dwell maps in the EFI has 2 axis, one of which always is battery voltage. The inverse also holds true if voltage goes up optimal dwell time must go down.

Most standard coils charge fully in about 3.5 to 6ms, the 84-89 coil is in the 4ms range, and they vary a bit from car to car. I've seen some at 3.8 while others as high as 4.3 most likely the age of the coil and the circuitry driving them also aging. Plus the voltage at the coils '+' post may not be exactly what the voltage within the DME is seeing. You always have some percent error.

I keep coming up with ~3ms plug firing at 7200 RPM, how do I reconcile this with 6ms to charge the coil?

At 7200 RPMs you have 2.78ms between fire events. Most coils typically need more than 3.5ms to FULLY charge. So at 7200RPMs the coil will never have enough time to fully charge. That's the issue with distributor based ignition systems.

Here's the math:
7200 / 60seconds = 120 revs per second
we have 1000milliseconds in 1 second
so
1000ms/120rev = 8.33ms per rev
and then we have 3 spark events per rev
so
8.33ms / 3 = 2.78ms per spark event

Basically after 6000RPMs the ignition system can no longer fully charge, the plug still fires but much weaker spark and the engine looses efficiency and starts loosing torque after 5800RPMs so cyl pressures decrease and the plug can still do the job since you don't need full spark energy if cyl pressure decreases.

Sal,
What is the current scale on your probe? Is that 5 Amps/div?

That probe is crude inductive device that clamps around the wire. It works out to about 1v per amp but it's inverted, the bottom of the slope is high current the top flat area is 0amp. I don't use this method much any more these days, that was a screen shot from some years ago. Now a days I simply watch voltage between ground and the coil's '-' post, this is the voltage drop across the transistor and the two resistors that are in series with that transistor. As current builds voltage increases across those components, much simpler way to get an idea of what current is doing. I never really cared all that much about the current, I mainly watch for the shape of the charge curve and the knee of the curve where I can see coil is saturated. I suspect these coils draw 5-6 amps at full charge.

OK

I thought I could see the current flatten out and thought that could be due to the 8 amp current limiter inside the DME.

I still can't believe they put all those discrete parts in for that over all the years and never used it. That would compensate for a bunch of variables if you bounced it off of that for a hundred micro seconds every spark

Based on the scope image, the dwell time, i.e. the time the transistor is turned-on,
is about 6 ms. The current trace at 4 ms reaches a constant value and most likely
is where the current limit was set in the ECM. The stock coil inductance on the 911
3.2 is 3.6 milli-henries. Based on that, the peak current can be determined:

I (coil) = T (dwell time before current limit) X V (battery voltage - voltage drops) /
L (inductance)

I = 4 ms X 10 volts / 3.6 mh = 11.1 amps

Once the transistor reaches its current limit the voltage across the coil goes to zero
and the transistor voltage across it becomes the battery voltage (minus the voltage
drops and the current limit resistor). So the power of the transistor becomes approximately
10 volts X 11 amps or 110 watts. That's versus the Vsat times the current before the
current limit is reached or about 22 watts (2.0 volts X 11 amps). So increasing the
dwell time to 6 ms as indicated in the scope trace provided no additional stored coil
energy but caused the transistor to dissipate more power.

The stored coil energy at 11 amps is about 110 milli-joules (11 X 11 X 3.6 x 10^-3 / 2).
The minimum energy for a spark is considered to be about 30 - 40 milli-joules. Assuming
that for minimum energy, the minimum coil current can be determined, which is
about 4.7 amps. Then the minimum dwell time becomes about 2 ms.

110 watts, 2ms.... 0.22 watts average power dissipation.

Are you certain this was with the factory coil...?

On the hand this may have been a design "hedge" in case an owner unknowingly used a non-factory coil.

Why would the factory not use a coil that could fully charge in 2ms..?

11 Amps???

Unless I really screwed up the analysis that long tailed pair in the DME forms a diff amp that limits the current to 8 Amps.

You guys both told me that was never used.

The power is actually calculated based on the duty cycle of the peak power.
So at 1000 RPM (spark every 20 ms) the duty cycle for power becomes 10%
and the transistor now dissipates about 11 watts (110 X 10%).

THX. :o

I also neglected coil primary resistance.

That peak current calculation is based on a very accurate measurement of the coil's
voltage, i.e. the actual voltage to the coil minus the voltage on the transistor's collector.
The best determination of the peak current is to actually measure the current sense resistor's
voltage divided by its resistance. Better yet, use a very accurate current probe!

Correct. Which also reduces the power of the transistor, which (coil resistance) was
assumed to be zero, i.e. not the actual value.

What????
neither one of you guys has a $2000 current probe?:) Security wouldn't like it if I drove my car down the hallway to the lab at work.

Anyone know German?
 

What does Abgleichuiderstand mean in technical (EE) talk? It says equalize resistor but it sure looks like a typical Select In Test component.

There are two of these (R526 and R508) that set the coil current limit to the coil. Again it seems like a lot of trouble to go to for something that is not used.

It's a trim resistor to account for tolerances in the other resistors. Abgleich can be translated as adjustment or trimming.

Your assumption! The current probe signal posted up-thread indicates that the coil
current plateaued, which is the result of the current limit feature and/or the coil resistance.

If the schematic I got from the FRWILK sight is anywhere near accurate I can pretty much guarantee all that circuitry in the upper right hand corner of the analog board limits the current to 8 Amps. What value do you come up with for the current limit?

It's time for you to accurately measure it using a calibrated current probe while
making sure that it's not the coil's resistance that limits the current. The current
limit feature was designed into that coil driver circuitry so that if the primary of the
coil ever shorted, the driver transistor would not be damaged. Most likely with the
original Bosch coil and its resistance, the primary resistance limits the current before
the current limit circuitry has an effect.

Anyone want to review their nodal analysis skills?
reference is presented to the positive input of the amplifier and the current sample to the negative.
http://forums.pelicanparts.com/uploa...1450635329.jpg
I do this kind of stuff with microsoft equation because I make mistakes copying and with signs.