1987 911 3.2 no start no spark

[mysocal911]

Originally Posted by rick-l

Dwell angle in a points car set the charge in the coil and varied with rpm.

The DME sets the current to 8 amps for every spark so dwell time has nothing to do with the energy in the spark.

Of course the dwell time affects the energy stored in the coil.
You need to solve the differential equation and understand how each variable
affects the stored energy:

Voltage (coil) = L (inductance) X d I (change of current) / d T (change of time)

[rick-l]

You need to take an microcircuits class and figure out what T502, T505, R528 and R518 do in the DME.

[scarceller]

Rick, the DME has coil charge time (Dwell Time) I know this 100% for sure because I know were the Dwell map is and how it works. I actually tune the Dwell when using non stock coils I also slightly tuned it for the stock coil at lower RPMs since the stock table is a bit shy of fully charging the coil with the stock Map.

But you are also correct that the darlington transistor has over current protection as well but that's used as a fail safe in case the coil shorts out.

I scope the coil and alter the Dwell map to watch charge times and saturation. I can easily stop charging the coil far before 8AMPs if I wish, the DME has both things built in, a Dwell Map and over current protection but the Dwell map comes first in calculating when to start charging.

[rick-l]

At idle is it hitting the current limit? This is the only place I looked at it.

[mysocal911]

Originally Posted by scarceller

Rick, the DME has coil charge time (Dwell Time) I know this 100% for sure because I know were the Dwell map is and how it works. I actually tune the Dwell when using non stock coils I also slightly tuned it for the stock coil at lower RPMs since the stock table is a bit shy of fully charging the coil with the stock Map.

But you are also correct that the darlington transistor has over current protection as well but that's used as a fail safe in case the coil shorts out.

I scope the coil and alter the Dwell map to watch charge times and saturation. I can easily stop charging the coil far before 8AMPs if I wish, the DME has both things built in, a Dwell Map and over current protection but the Dwell map comes first in calculating when to start charging.

That's correct!

[mysocal911]

Originally Posted by rick-l

At idle is it hitting the current limit? This is the only place I looked at it.

No, or the coil driver would be over-heated, i.e. Power ~ = 8 amps X 13 volts - 8 amps X 8 amps X coil resistance.

[rick-l]

Originally Posted by mysocal911

No, or the coil driver would be over-heated, i.e. Power ~ = 8 amps X 13 volts - 8 amps X 8 amps X coil resistance.

For 0 di/dt the power in the darlington would be 12*8 or 96 Watts which it won't do very long. I was under the impression from everything I've read and seen they contro the dwell to just hit the current limit.

[scarceller]

Originally Posted by rick-l

At idle is it hitting the current limit? This is the only place I looked at it.

No, the dwell map is designed to never allow it to hit current limit. As a matter a fact at idle and lower RPMs they charge the coil to about 80% but at higher revs they charge to 90-95% maybe they felt they did not need a full charge at the lighter loads and cyl pressures?

I actually use a scope and increase dwell time till I saturate the coil then drop back to about 90-95% charge.

[scarceller]

Originally Posted by rick-l

For 0 di/dt the power in the darlington would be 12*8 or 96 Watts which it won't do very long. I was under the impression from everything I've read and seen they contro the dwell to just hit the current limit.

Rick, a lot of what's written often is incorrect. I disassembled the entire Program code for the 89 chip and I know what the software does and how it works. I made no assumptions here as I found the Dwell map and the portion of code that calculates charge times and I know how it does it. I also often tune dwell times wit the scope and alter the map on the fly with my chip emulator. I have shorten dwell times on purpose during idle till I cause mis-fires just because I was interested in knowing that I could do that to be sure I understood the code and the Dwell Map.

I think the engineers added the over current sensing as a durability feature for the DME just in case the coil shorted out or the drive line was somehow shorted to 12vdc.

[scarceller]

Dave,

First, my sincere apology. You are correct that your method for testing coil pulse works. I tried it and it works exactly as you say then I thought about it and I see why it works.

I just tend to use LED test light for everything since the LED works great even on Speed and Ref sensors.

Once again sorry.

Originally Posted by mysocal911

1. Who said to use a LED test light?
2. When using an incandescent test light that test works perfectly.
3. What do mean try it? I've done that for years, i.e. grounding the test light. You need to try it!

"Reason is that when a coil first start to charge it has 12vdc across it as it charges the vdc starts to reduce to 0vdc once fully charged.
So when the DME commands the coil to start charging it grounds the coils '-' side and the coil has 12vdc across it and voltage starts to
decrease from 12 to 0 as it charges."

That statement is not correct. The coil does not saturate (zero volts across it), as the coil driver is switched off before saturation.
Besides once the coil driver turns off the coil voltage rises to about 350 volts on the primary which is more than enough voltage
to flash an incandescent test light.

A coil does not function like a resistor, although it does have a very low resistance, i.e. the voltage does not vary as the current
changes, only when it reaches saturation. Although it's series resistance does have a small effect on the voltage of the coil,
the rate of change of the coil current determines the coil voltage. Time to put an o-scope on the coil signal. The DME ECM
calculates the charge time (dwell time) of the coil driver.

Voltage (coil) = L (inductance) X d I (change of current) / d T (change of time)

[rick-l]

Originally Posted by scarceller

I disassembled the entire Program code for the 89 chip and I know what the software does and how it works. I made no assumptions here as I found the Dwell map and the portion of code that calculates charge times and I know how it does it.

Can't argue with that. It must be a feature the Bosch engineers put in and found impractical.

They sure went to a lot of work to do it however. That circuit is a work of art for the 80's. Look at how they limit saturation, the select in test resistor to set the current. And if you ever use it because the timing gets lost it burns itself up.

[porsche.racer]

Update:

Definitively the no-start was caused by the DME.

Anyone have a good used DME (1987 911) for sale.

[porsche.racer]

Thanks to all, for the help!!

[Jesse16]

If you're sure the DME is the culprit, why wouldn't you just send it to one of the folks including Pelican who could repair it ? There are several as I remember who can test and re-solder parts or joints on DME's, cruise modules, etc.. Then you know what you're getting.

[ischmitz]

Agreed - I suggest a bench test and inspection would be on order next. Let me know if you need help.

Ingo

[porsche.racer]

That's the plan! Sending it out to Sal for repair and upgrades.

[JJ 911SC]

Originally Posted by porsche.racer

Update:

Definitively the no-start was caused by the DME.

Anyone have a good used DME (1987 911) for sale.

Originally Posted by porsche.racer

That's the plan! Sending it out to Sal for repair and upgrades.

This sound pretty definitive... Maybe the DME is BER!!!