Is your crank sensor new or have you tried swapping out for a known good one?

If the wiring is all up to date then I'd ask:

Is the noise present only under compression even with coils unplugged entirely?

You may need to add the relay for ECU power.

Although the starter main wire is isolated, is the trigger from the solenoid still going to the ignition switch? This could be your noise.

The crank sensor is still the original one for this 22 year old engine. BUT ... it is producing pulses AND those pulses seem fairly normal with no compression.

Yes, noise is only when starter is working under load.
Coils and injector circuits are completely unpowered.
Also, the starter is currently on a completely separate battery and the yellow wire is disconnected at solenoid.
In this configuration I am cranking by touching a new wire from the solenoid to the second battery.

I hear you on the ECU relay and I will look at where I can fit one. As I see it, that relay will give clean 12v straight to ECU instead of 12v via the ignition switch circuit, which could be subject to noise.
I'm afraid I can't get excited about this being the solution as I don't believe there is anything noisy happening up there, especially with coils, injectors and starter out of the frame.

I think the next trick might be to try the original Bosch starter.

That or the spark plugs...although the part I cross referenced is resistor, if you have a "racing" version of that plug then you may have no resistor.

True. but in these tests Bill does not have the coils connected and the plugs are not even firing.

True...

Could try relay for the ecu power.

Where are the grounds for the ignition drivers?

All coil grounds are run straight to the engine ground (old distributor mount).
Coils have been unpowered during all latest logs, so there is no chance of interference from them.

http://forums.pelicanparts.com/uploa...1402977653.jpg

I had no trouble with a 60-2 tooth wheel and VR sensor even when I was running 10.5:1. However, it is a Clewett crank wheel and sensor as well as a stock starter. Has anyone else successfully converted a stock 964 VR sensor to MS?

Yes.

spam!

Mustered up a bit more enthusiasm for another crack at this over the weekend, but no progress. :(

Noise seems to be starter related, so I thought I'd try the original Bosch starter instead of the HiTorque. After hours of dicking around, I managed to run a log that is no different than what we had before.
Now I have refitted the HiTorque starter and wired it up to the battery as it should be (ditched the separate battery as it made no difference).

Fitted a relay for the ECU power. I didn't really expect it to make any difference, and it didn't.

Surely there must be others out there using MS3X on a 3.6 engine with the standard crank sensor. Talk to me about how yours is set up.

The areas where I could make changes are becoming limited ...
try switching the crank sensor wires around
try a different sensor (but this one seems to work)
try running a new set of wires from sensor to ECU outside of harness
try switching ignition input capture to rising edge (I keep forgetting to do this when I am logging)
try cranking up R56 and/or R52

Some of these are relatively easy to try, but most seem like clutching at straws to me.

Bill, That "switch the crank sensor wires around" sounds interesting. I have a hall sensor on a 60-2 crank wheel driving MS3, so I've never messed with a VR sensor. But I have seen forum stuff where someone reported a phantom missing tooth that was caused by reversed wiring to the VR sensor. Just a thought -- good luck. I get the sense that you are just a hair away from success. (I went on for months with a problem what eventually was traced to self-induced grounding problem on some VAG coils that caused me to switch to LS-2 coils.)

Brian

http://forums.pelicanparts.com/uploads23/Tooth+log+14a1402830977.jpg

This looks exactly like the pattern in the MS manual where they state that either the wires to the sensor are reversed OR that your capture setting is reversed. I would try reversing your input capture setting first. If that doesn't solve the problem then try reversing your wires to the sensor. You still may have to try either capture setting when you reverse your wires.

Just had a rethink on reversing the sensor wires.
There is a grey wire that is shielded - surely this must be the signal wire.
The other wire is black and runs outside the shield = ground?
Seems pretty obvious which is which, so not much point in pulling harness apart just to try the switch. See diagram in post #72.

Definitely must try capture setting tomorrow night.

Reversing the wires can occur at multiple places, as you probably already know. Yes, the signal should be carried by the shielded wire, but it may have to be switched somewhere down the line in another connection.

Brian

Bill. At this point I would consider get my hands on an oscilloscope. Start at the sensor and start tracking the signal through each wiring junction and then onto the board.

It should not be this difficult to eliminate the noise.

Bill. Is it possible that your VR sensor is seeing the starter ring? This would go back to the question about the position of the crank sensor.

I had the same thought in bed last night! The 964 wiring diagrams and DME troubleshooting section of the manual is below. This is what I used to figure out my wiring of the sensor.
I guess the selection of pins 1 and 2 is what we are talking about. Just need to find an easy way to switch them without undoing a nice sealed harness. :(
Looks like a trip to the auto electrician to pick up M and F connectors for a short jumper.

http://forums.pelicanparts.com/uploa...1403574687.jpg

http://forums.pelicanparts.com/uploa...1403574154.jpg

I am working on this. I found a good one for sale but the seller is a bit dim. He lives about 600km away and is slow to respond but should be getting a price for delivery today.

No. the two sets of teeth are a couple of inches apart.

http://forums.pelicanparts.com/uploa...1403575661.jpg