motronic idle surge solution

[landwerlen]

i'm a newbie to the motronic (DME), but this forum, and pelican in general has been an invaluable wealth of information. regardless of experience levels, most everyone seems to have interesting insight. also to note, im usually a taker and not a giver to these forums, so i hope this bit helps someone.

to my problem i encountered: my 1985 carrera began to idle surge during a trip. i looked for obvious discrepancies, and found none. i began searching these forums, and felt fairly educated after a couple of hours or browsing. AFM? maybe. induction leaks? maybe. o2 sensor? doubt it. CHT sensor? doubt it. DME and ICV? definite maybes.

all of the sensors checked ok, with the data i had available. since the problem happened so immediately i kinda ruled out an induction leak, unless it was obvious, and a propane tank yielded no increase in 02 output.

i noticed if i pinched off the bypass tube feeding the ICV the car would smooth right out. clue! also i could suppress the surge by turning the throttle body adjustment all the way in. another clue!

even with the adjustment all the way in idle was in the 1000-1100 RPM range.

i tried to 'center' the ICV by shorting pins B and C. this resulted in no change in RPM. this was my problem. either the ICV or the DME which drives the ICV was malfunctioning.

a previous post from loren to another fellow confirmed this could be a problem. loren's post was helpful, but a bimmer guru had just the info i needed.

using a multimeter to check the operation of the ICV just doesn't work. the ICV uses DC pulses to move the rotary pintle valve from full open to full close. i decided that my 'scope would show me a better picture, even though i had no idea what i would be looking at. after looking at the 'scope it became a no brainer. i had dc pulses on the 'open' side of the valve, but nothing on the 'close' side.

i wish that i would have taken a jpg of the 'open' side while the engine was surging. it was really cool. the dme changes the dwell time (not the voltage) of the pulses based on the desired RPM and the actual RPM. when the desired RPM and the actual RPM are equal, so are the pulse on-off dwell times.

any way a darlington transistor (ON588) was found shorted between the base and collector in the DME. i googled this part and found a replacement and the bimmer guru's following post.

http://www.s14.net/forums/showthread.php?t=6426

i threw in a TIP3055 NPN i had laying around, and viola!! life returns to normal. the BDX53 series of darlington transistors is the correct replacement.

to close, whats is interesting about these ON588's is the fact that there is little to no heat sink set-up on the DME board, so it seems doomed from the start. mine looked great from a visual standpoint, but the board itself showed signs of overheating. im also not naive enough to overlook the possibility that the ICV maybe 'going south' which could be overdriving the DME output. the coils in the ICV still have better than 20 ohms per side, so im hoping it last a while longer.

im posting this without pics, but i'll add them when i figure out how to post an attachment.

[jester911]

Great info for your first post. Thanks.

[landwerlen]

this is the board with the TIP3055 temporarily in place



here is a pretty lousy pic showing the 'pulses' created by the ON588 in the DME.



see ya when the next thing breaks on this car!!!

[KFC911]

These have to be the most technical "first posts" that I've ever seen here...you're even excused for not posting a pic of your car . Thanks for this valuable info and speak up more!

[DonMo]

Now you will really be my hero if you can tell me what is the replacement for the ON946 at T503 and the npn darlington on the heatsink that says 1012 at T504

[nhromyak]

I am curious if you had any other running rich condition when you found this. I had a DME go bad on me with very large surging and running rich (black smoke). It always took about 20 - 40 seconds for it to happen though.

[ZOA NOM]

This should be submitted to Pelican as a Tech Article ASAP!

Outstanding.

[JeremyD]

Excellent contribution to the DME Surging mysteries

[Hugh R]

No friggin idea what you posted but if I have a surge problem, I'll be PMing you. Thanks

[Lorenfb]

"to close, whats is interesting about these ON588's is the fact that there is little to no heat sink set-up on the DME board, so it seems doomed from the start."

Really?

So, Bosch has failed in its worst case design. Right!
Thus, all Porsches (3.2s, 944s) and BMWs (M6s, M4s, M5s) DME ECMs will fail.
Doesn't look good for the owners of these vehicles.

Bottom line: Please, let's not alarm the above owners of unreal problems!

[jester911]

Loren, I knew we could count on you for your usual constructive comments. Please give us another morsel of your wisdom.

[KFC911]

Originally posted by jester911
Loren, I knew we could count on you for your usual constructive comments. Please give us another morsel of your wisdom.

"Chips suck, case closed" - Lorenfb

Since chips suck, I'm grabbing some popcorn for this one

[rick-l]

Hey cut Loren a little slack. All he said was he thought the transistor was used correctly in the application and a heat sink would yield negligible reliability gains. Since this appears to be a rare failure mode I would tend to agree.

They give thermal resistance from junction to ambient in the data sheet for a reason.

Anyone know what electrical abuse leads to a B-E short in an NPN power transistor?

[KFC911]

Originally posted by Lorenfb
...Bottom line: Please, let's not alarm the above owners of unreal problems!

You mean like the "unnamed alarmist" who has warned everyone about the impending doom of running a Steve Wong chip with no empirical evidence to back up his claims ?

[landwerlen]

well fellas, this is why i have been reluctant to post in the past....i guess i need to be a little more thick skinned. i appreciate the positive responses i have received from most of you. i'm glad the post seemed to help. i guess some have to find a negative in everything. im quite the opposite.

my heat sink opinion i don't think was alarming at all. i was kinda thinking out loud i guess. my expertise is aircraft systems, and ive never seen a TO220 case transistor left to cool with just ambient air (or a PCB in this case), in an aerospace application. i also know that heat kills ic's, whether it be due to excessive current flow (bad ICV) or the rotten luck to be driving in traffic on a 100 degree day in indiana, or even worse, both. which were the exact conditions when the failure existed. again i understand the DME was not designed to meet the same criteria that i might be used to in my vocation, but it was just an observation and i apologize if it alarmed anyone.

it is also my opinion that this failure has happened to more than just a few. it has been documented by more than a few, including loren in this very forum. most guys probably never know the demise of their DME when it fails. the owner throws in a rebuilt DME, and is happy as a clam to be driving again. the DME gets 'rebuilt' with a 59 cent transistor (Mouser), and the cycle begins. this is in no way a knock on those rebuilders, because thats the value of knowledge and the price of a warranty on the existing parts he/she can't control.

here is my bottom line: i don't think this is a truly isolated incident. a person with a 'scope or digital multimeter (with a freq counter function), and very basic soldering skills can verify and fix the same problem i encountered.

im just trying to empower the same folks who empowered me (on this forum). this information is a great resource.

BTW someone above posted something about a rich mixture before the fix. the car ran fine before the failure.

[nhromyak]

I am curious how it ran during the failure, if at all.

PM'd ya. Thanks.

[Otto]

Then let's have all Porsche owners with DMEs open them and add a metal heat sink!

Wait, before this is done, remember:

Power = Integral (V X I ) dt where;

V is the voltage across the semiconductor, and
I is the current being carried by the semiconductor.

Which is basically the RMS Power being dissipated by the semiconductor.

Then using the thermal characteristics (case to ambient) of the
TO220, determine the max chip temperature without a heat sink.
If this results in an acceptable margin of safety, no heat sink is
required.

Maybe the Bosch engineers did the above analysis and determined NO
heat sink was necessary.

[jester911]

Originally posted by landwerlen
well fellas, this is why i have been reluctant to post in the past....i guess i need to be a little more thick skinned. i appreciate the positive responses i have received from most of you. i'm glad the post seemed to help. i guess some have to find a negative in everything. im quite the opposite.

John most of us really appreciate this info. Even though I have posted quite a bit here and a couple other boards I really do know what you mean. There have been many times that I have not posted for the same reason.

At any rate as these cars get older and parts get more expensive and harder to get we will need more info like this.
Not only that if you look at the number of post here that have to do with some kind of idle or misfiring issue that many times don't get solved this may be more pertinent than we know.

[dentist90]

Peer reviewed articles are subject to harsh scrutiny. In this case it sure seems that you know what you're talking about and most who are in the know agree with you. If my DME bakes I hope you don't mind if I ship it to you!! I wouldn't know how to hook up an oscilliscope if I had one!
Thanks for posting.

[KFC911]

Originally posted by landwerlen
well fellas, this is why i have been reluctant to post in the past....i guess i need to be a little more thick skinned. i appreciate the positive responses i have received from most of you. i'm glad the post seemed to help. i guess some have to find a negative in everything. im quite the opposite....

Yep, put on your "thick skin" and keep posting ! It's nice to see intelligent discussions (for most of us) about things that we are not experts on. Loren is very knowledgable, and his arguments about chips (for example) have enabled me to become better informed on a subject I knew little about 5 years ago. Same goes for DMEs...he knows his stuff (as I think you do too). Many times "experts" will disagree, sometimes it doesn't hurt to "over engineer", and most of us are intelligent enough to make our own informed decisions. Thanks again!