DME and electronics specialist's thoughts needed
 

I installed a shortened G50 in my Carrera, and used a Patrick Motorsports flywheel/clutch assembly made specifically for short G50's using their custom DME flywheel with the sensor ring, but larger in diameter and thinner to allow for more space in the shortened bellhousing. The reference pin is relocated to the pressure plate for space reasons. The only note I got from Patrick Motorsports was to swap the two sensors (refernce and speed) in their bracket.
Well, to make a long story short, the engine never ran right with the setup. It was acceptable at the beggining but after I accummulated break-in miles and started to rev the car higher, I started to notice the hesitation and missing. I re-adjusted my sensors, as per Jim Patrick's reccomendations, and I got the top end better, but the bottom end was hesitating badly. Playing with the sensors' adjustments made a noticable difference with the engine running, but I could never get it right.
I called Patrick once again, and he told me that the reference sensor must be reading the reference pin AND the three pressure plate friction plate-to-pp housing mounting tabs, thus confusing the DME and engine's timing. He said I would have to remove the pressure plate and mill down the three tabs. So I removed the engine, reinstalled the original flywheel with pressure plate (for starter ring gear) and started the engine. It ran like a charm. Smooth idle and easy fast revving with no hesitation. It is 100%clear that the Patrick motorsports kit I bought was never fully tested, and I am sort of a guinea pig.

Anyway, I noticed that the original DME ring gear has different shaped teeth. The bottom, or valley of the teeth is rounded, (see below)
http://forums.pelicanparts.com/uploa...1064789945.jpg

while the Patrick flywheel teeth are perfectly square (the one nearer the floor is the starter ring gear). I am not sure if this shape difference can cause a bad signal. Any experts know if this differnce is ok? (patrick's flywheel and pp below)
http://forums.pelicanparts.com/uploa...1064790106.jpg

Also, here's a pic of the complete pressure plate with the three mounting tabs visible. I am not sure if i can actually mill enough material between he outside perimetre and the rivets, to actually make their prescence to the ref. sensor unknown. I would hate to have this done and put the engine in with no change in the running of the engine.

What do you guys think???? I'm freakin' out here!!!!

http://forums.pelicanparts.com/uploa...1064790383.jpg

This pic should replace the X above...http://forums.pelicanparts.com/uploa...1064790700.jpg

The reference pin on your new flywheel looks too short and mangled. Also compare the number of teeth for the speed sender from the old to new flywheel.

Mike,
I am pretty sure that the speed signal is magnetic, the sensor head shoots magnetism across the air gap and if a tooth is opposite, the signal is counted and each passing tooth adds to the total for each revolution of the toothed wheel. The tabs look like they counted as one tooth and this would have lowered the count of signals per revolution telling the DME the engine speed was slower than actual.
The air gap between the teeth presents a huge barrier to the magnetic signal so I believe the configuration of the bottom of the gap rounded or flat just doesn't matter. The magnetic reesistance of air is huge compared to steel.
The Bosch DME Motronic book book is a good read.
Ned Monaghan

I don't know too much about the 3.2 DME but here are some thoughts:

- If your original reference pin looks so mangeled maybe your reference sensor was hitting it and got damaged in the process. How does its surface look. If it has a groove milled in it from making contact with the reference mark its time to replace it.

- On a 3.6 the reference signal is generated by one missing tooth on the toothed pressure plate. On a 3.2 there are two individual sensors. The reference sensor should be seeing only one signal per revolution caused by the reference mark. Maybe you really got erratic readings from the mounting tabs coming close?

- The gap between the sensor and the passing theeth is 1.0mm +- 0.1mm per factory manual. I can see where missadjustment causes problems like you are seeing.

- The shape of the teething is not critical. The hall sensor will out out a sine wave anyways.

Good luck,
ingo

Mike: Don't be a noodge, send my your address and I'll send you a scope. :D
Has Patrick offered to make good on the flywheel in a any way? Is there something unusual about the flyhweel "flavor" or is has he been selling this type for a while?

ischmitz: For the 3.2's (and 944's) the gap is .8mm. Is the 1.0mm spec for 3.6 motors?
-Chris

Thanks for the input guys.
Ed, the number of teeth is correct, and you are right, the ref. pin is shorter and is mangled slightly because it hit certain sections inside the bellhousing.
Ned, the speed signal is ok, so the dme knows how fast the engine is turning, but you gring on a good point for the ref sensor. The pressure plate housing is aluminum so the ref. pin signal should stick out, but the ref. sensor is also reading the 3 tabs as they pass.
Ingo, the ref. pin slightly hit the inside of the bellhousing. I replaced my sensors anyway, but the problem got worse. I account this to the new sensors being more sensitive. Thanks for clarifying that the teeth shape is unimportant, that helps.
But now I am beginning to think (after looking at the pressure plate pic above) that there is no possible way to remove enough material as to not get an interuption from the three tabs.
I wish they would take back their set-up and give me my money back!! With the money, I would just add a few hundred bucks more and get myself a Kenedy clutch and SDSEFI stand alone engine management system...

Hey Chris, thanks again for the offer. You have been really helpful, and I appreciate it. I spoke to Patrick twice, the first time he was surprised I had this problem, and the second time he told me what I should do, "and that should correct the problem" But what if it doesnt? I will have to remove the engine yet again, and I don't have too much time. I work 65-70 hours a week as it is.
I don't think he has sold this type of setup before, as most Carrera owners would opt to sell their 84-86 carrera and buy a 87-89 if they wanted a G50 so badly. Short G50's usually go to 930 owners, with CIS or stand alone system.