'91 ecu troubles

[fixkick]

in almost all cases a shorted COIL primary will burn up the ECU internal, Ignitor driver transistor.
in later years, they used thermal protected IGT transistors to self protect from shorted coils. but not back in 1996. (the red hot and burn up)

is one coil pair still working? your first post made it seem to me both are dead.
and that is key to diagnosis.
after all if both are dead, it implies the ECU did so on purpose. (CKP/CMP dead)

did you try using alien coils on car. never do that. (or attempt so swap left and right coil, putting the shorted coil on a GOOD driver line?,now killing both lines?)
the primary impedance of all coils are different on all cars. huge variances.

the dwell time and current specs are different on all coils and must not be guessed at .. .or boom.

---

i do think now , that somehow, this German ECU is OBD2, as the pin 7 and 10 are wired. for ISO OBD2 signalling, it seems.

[greenfield54]

Both coils are dead but one was still working before. I will post a photo of the relatively clean ecu tomorrow showing the part nos. of the drivers. Do you know of simulators to mimic the signals of the CMP/CKP. Better if Diy. Will log out now as it is almost 10pm down here. Thanks.

By the way. Please feel free to use my photos in your website if they are useful at all. Just indicate my handle in the credits. Too bad I don't have a proper scan tool. That would have made the job easier. This is my first EFI project car. My second is a 1994 Corolla GLi with 4afe engine. This is more dinosaur than the Elantra which has OBD2 and ABS but no airbags.

Again thanks.

[fixkick]

this is the neg lug of the coil (spark coil) (-)neg
this is a good signal.
and on your car there are 2 of these. are both dead?
that is the line below is flat, cranking?
if both are dead, id bet you the ECU did that on purpose... (dead sensors)!!!!!!
if one is good and the other is bad, then the coil is shorted or the ECU is bad on that driver line, OR BOTH. (coil shorted can damage the ECU)
I'd bet the coil can not blow up both lines, as the coils are 2, in a box. and independent.




if both are dead.
I'd bet you that ECU is logging, sensors errors. by you have not scanned it yet, so will never know the truth.(in most cases and for sure replacing parts over and over,, and not getting it to work)

[greenfield54]

Thanks. Why delete the photos? You can use them if you need them. No problem with me.

Yup I too is wondering why they do not use all hall effect instead of inductive sensors. We use hundreds and hundreds of hall effect sensors in the machines I worked on. They are not too critical regarding gap clearance.

Just one question that keeps popping in my mind. Would an engine start with just CMP/CKP connected? Of course all power and grounding should be present.

I really want to bench test everything like what I used to do with all our factory equipment. But sometimes the problem is getting the correct pin outs for each ecu.

I used to simulate a large drying plant that is housed in a 5 floor building with many motors. If you look at the control panel one would think the plant is actually running. I am thinking of using two pulleys linked to each other with a belt just like a crank and cam shaft pulley wherein the cam shaft pulley runs at half of crankshaft speed instead of using timers or signal generators.

[fixkick]

great questions !
only deleted my wrong posts. (all photos are on my web page still or above in the Gallery above, all still there click ,Gallery+other = bingo. your stuff.
REDUX 2:
I really do not know what is in your ECU, for sure im not sure the free schematic you have even matches what you have.
all I know is it does have a CMP and CKP sensor. and you have not told me how many pins are in each.

below is HALL sensor , note it make perfect signals even at 1 RPM... like magic, its Electronics doing that magic inside the CMP here.

http://www.fixkick.com/IGN_timing/video-...Dtest.html


NON HALL: failures.
the ECU uses a device called a zero crossing detector , if it cant see the signal cross zero, the engine will make no spark.
it is that simple, just that.
how can the signal fail, Ill count the ways.... numbered...
1: wrong sensor , or bad sensor., on some cam sensors there is the cam plastic, coupling device, installed wrong, broken or fell off....Hyundai issues.
2: wired backwards.
3: wires bad. (cut, broken, rusted, damaged , crushed, in any way )
4: shields bad, on the coax line pairs.... if bad, the noise cranking will exceed the real signals.
5: bad sensor gap set. per FSM. read that? installed wrong, or calibrated wrong.
6: damage to the TONE wheel. The part the has bumps and triggers the CKP.
7: A scope sees all this.... (if the ECU scans show RPM and CKP dead, it IS)
8: bad ECU, rare for inputs on ECU to fail, but outputs do fail easy.
9 : an engine swap gave you the wrong crank tone ring...
Kinda overwhelming all this , lacking a SCAN tool that tells you, gee, CMP or CKP is dead or intermittent.

each pulse is the tone wheel teeth passing by the sensor on the crank.
this is our USA Beta engine 1996 CKP tone ring, not saying yours is same, i cant do that. we dont have books from your country, here.

CKP ring. note the MISSING tooth?

[fixkick]

did you check for spark using new spark wires at the coil and test spark lugs, x2 at each coil set?
you do know that an open spark wire, kills both plugs right, in wasted spark fashion.
even a bad spark plug can kill the matching other plug.
say plug 1, is open can fire (its too old) and is not shorted, its open.
that kills cylinder 4 too.
see? this is because they are wired In series and any break in the line, kills 2 plugs not one.
if there is truely no spark we scan the pcm to see why.
if the PCM can see the error we check the , 2 coil pins for the COIL CHARGE TIMES, THEN MAKE SURE THE PULSES LOOK NORMAL.
If only one side is bad. then the coil may be bad. (hope) if the coil primary shorts this can wreck the ECU driver on that one line, not both.
if both lines (coil neg) then the ECU is cutting spark on purpose (see why? ) that is the ECU normal logic.
the ECU has a set of complex rules for making spark,, worst case is below!
1: ckp good, signal (if ckp is too weak it will never start)
2: ckp notch present (missing tooth seen by ecu) the ECU cuts spark if the gap can not be detected !
3: cmp good. signal
4: CMP timing not grossly off time, like when the cam belt slips..... if way off the ECU is protecting the engine for damage, by not allowing 2 cylinders to fire , fighting each other.
on brand new cars the ECU has a code for #4, your probably does not. back in 1996. or the CMP coupling device, is bad, lose. broken, missing or damaged in some why.
or cracked cam shaft, such tha CKP is off time.
or any of those parts related to CAM timing or CMP on the end, are off time for any reasons at all...


here is your tone wheel (well it ours, 1996 beta1 engine ) not saying yours is same , i can't do that. lacking your FSM book.
ours is 58/2 6degree tone wheel.
usa ECU , so yours can be different. but if not , now you know what the scope view will be. 58/2 pulses

---

home made mechanical sensors?
you will be hard pressed to make a set of pulley do this,
but can with a disk punched full of 58 holes. (if in fact it is 58, on KDM cars)
and and optical sensor,

[greenfield54]

The sensors I used are brand new with the same part nos. as the originals. Now I understand how very important shielding and good connections are important particularly with CKP. Now this is what I will do.
1.Pull plugs of all injectors.
2.Remove spark plugs to lessen load on starter and battery.
3.Monitor signals from CMP/CKP.
4.Monitor trigger signals to S102976 ICs.
Question. What pins of the ic should I monitor? I cannot find data sheets or at least pin identification of the ic.

   

   

   

At the moment I cannot find the ICs here so I asked for quotation from Alibaba.com. Problem is the minimum order is 10 pcs. I am thinking that I can use the others on points type distributors which are still in use down here. I will drive the IC with the points and the IC fires the coil. Do you think this is possible? Hope you can provide a simple diagram.

Thanks again. Though my knowledgeon these things are not of the same high quality as yours, I am learning a lot.

[fixkick]

the TI S103238 is a custom Ti processor , special ordered by Siemens.
S102976 ignition driver chips.
some serious potting removal there, hard work.


many of these ECU's use non standard chips.... or ASIC. and other custom ordered chips. to meet the makers needs.
i see a real intel flash 28F512 chip there ,

That 976 is a ST German chip.. alot of that alibaba stuff is clone junk, (some work but not for long or not at all, you need to buy St.com real part if bad)

my guess, is you traced the 2 ign. coil wires. to the 976s? is that your questions. related just to the drivers.?

S102976 listed as and ignition driver chip.
that S10 number seems to be a private Siemens contract number.

the clue to that, is two fold, non STI numbers there, non commercial standard, numbering used. Non JEDEC numbers . Non Mil spec, numbers..
and clue 2 , no Google hits. and that be hard to do if the number was commercial grade, and standard. (impossible and very big clue that)

here is the real company, that made those chips (st.com )


http://www.st.com/web/en/catalog/sense_power/FM1965

http://www.st.com/web/en/search/partNumberKeyword
type coil driver

you can be sure they dont make them now. they moved on to thermal protected devices now. BU xxxxx parts

in many cases we contacted the makers, and asked what part this is , and was told its secret between, Siemens and STI. brick wall.
this is why many major ECU rebuilder hoard ECU's and then take one, ECU and use it as a parts source. (cannibalized)

id contact STI and ask them for the data sheet, maybe you can get lucky.

i bet the S code is a contract code. for Siemens. just my bet.

the chips might be just stress tested to last located near an engine. "HOT stress or burn-in tested" , just like MIL parts are done, and with special numbers there too.
good luck finding them....

id call them STI, would.


that potting is to keep parts cool near the engine, removed, they overheat.
it's not to hide a secret, as some things are...

[greenfield54]

My question is related only to the drivers and how to test them and what signals they need and also the pin identifications (B+,Gnd,Out,etc)

Yup it took me more than a week to remove most potting compound. The ecu on this elantra is located on the upper left hand of the firewall beside the TCM inside the cabin. Good thing that at 62 years old I can still contort my body to reach the mounting bolts. The corolla I am working on has the ecu on the center console under the AC box. And no potting or encapsulating compound. Very easy to access but also easy to get wet. We encounter very severe flooding here during the rainy season. Water and electronics equals kaboom!! LOL.

I have another question. What type of trigger signal does the ecu inject into the driver ic? Maybe I can use discreet components like transistors to drive the coils. I will house them in a box outside the ecu like old time efi and use only the trigger signals from the ecu. Can you point out the pins on the ic where the signals come in? Is it safe to assume that there are other coil drivers that can be substituted? Maybe the 976 has a substitute.

I registered at the ST Micro website and submitted an online request regarding the correct catalog part no. and or substitute just in case this is out of production.
Maybe free samples? (fingers crossed) LOL!!

By the way. The second link (http://www.st.com/web/en/search/partNumberKeyword) does not work.

[fixkick]

cab temps in cars can hit 200F easy parked in the sun,, windows up. this proves deadly to many parts in ECU.. (all our ECUs here die from that heat, every one)
and water, sure, that too, is bad. (we lost ECU here for simple window (windshield) cracks. the potting also protects from vibration damage (3 way protection)

signals. (inputted?) (you can't use a test lamp to measure TTL signals. you do know that ,right"?) or what ever it is.. 74HC173? (guessing at '1')? 8bit latch?
https://en.wikipedia.org/wiki/Transistor...stor_logic
we don't know any of this yet, we don't t even know what processor that is. (it too is a contract part, see the number? see that "S" prefix?)
does that other chip the 173 driver to the main driver, called a pre-driver? connect up?


http://www.st.com/web/en/catalog/sense_p...pe=keyword

they make processors in different voltage classes now. lower (VDD), is faster and less heat.
but if the uP is 5vdc then the signals will be that, 0 to 5v, or TTL highs and lows. TTL is about (ViL) 0-0.8 and when high 2v to 4 v (Vih).
see page 18 here for all families
http://www.slideshare.net/aMiERuLz/logic-families-1

that driver probably is a near discrete component (modern type) the newer parts have current limiters inside, thermal overload protection, and that coil of yours has huge back EFI spicks (read damaging) to deal with, so will have a back EMP clamp.

in the olden days, all those parts were on the outside (not seen on yours), so ST.com made a new part that does all that magic in one part.
i put the link to the typical ST.com part used for ECU.'s
Other drivers( well which one? there are many VSV and solenoids, and ISC motors used)
the other drivers, are usually in a special form or chip or SIP , and many today are obsolete.. do no assume anything, get the data sheet and read it.
only that works.
and what if the ECU cut spark on purpose? (didn't like the timing relation ship, cam to crank,) what if? be sad to learn that after a new ECU installed.
happens like 90% on the forums... that, (the poster not understanding the ECU logic rules on spark, some ECu are very strict here)


here is one part (coil driver, common)
http://www.st.com/web/en/catalog/sense_p...pe=keyword

HVc is the coil pin. note the complex paths.!
the modern coil is a nasty beast.
my scope views showed that, it can blow up most simple discrete transistors fast. with that huge spike and currents.
VB526SP-E (search for that part , as 1 example)
this part (not saying it's yours) is classic part here. in many millions of cars.

the input has a pull down, so the device does not go active, if the input is disconnected and burn up the cars coil.
page3 in the data sheet shows the input requirement and the schmitt trigger thresholds. 0.4v.
Vil 0.3 to Vih 4v. for this part . using 5v. this device trips with input 1.9v or less.
but as stated above, the signal will be fast.
page 13 in the data sheet PDF shows the uP to driver connections.
this part is a 125watt part. (limited by junction temps inside)

your driver is not a simple transistor. unless there is 2 in one package. but did you trace all 6 wires there? do 2 pins go to the main connector
some ECU can to 4 and 6 cylinders, with one ECU, (a prime goal today) and more. (eg, one ecu fits many engine, and has flash uploads)
but again, we dont have the data sheet, above is me pretending we did.... as an example

if the driver is say a 6 pin dual channel twin transistor. see those 3 pin pairs. what if? and then it has external clamp diodes and its bad, the new part will fail pronto.
i see SIP packs right next to the driver chips, are they related? are they custom clamps for the driver? who knows?, lacking a full trace out....
the other connections can be in the center layers of the card,,, making all this hard to discover. lacking xray machine.

Finding and buying old parts, can be very hard, (dispite the cloner sellers claims)
most parts are out of production, those parts were made 20 years ago, and the technology moved on,,,, today we expect the devices not to blow up
when a simple spark coil shorts. back in 1996 , they did go boom on most cars , back then.
but not all, did that. Maybe Siemens is more advanced, then... I dont know. until the data sheets arrive.


lets now talk Dwell: (rule one, infinite dwell burns up all coils)
the coil dwell time , is about 15mS max, this is long to reach saturation of the coil (- lug) (the coil maker dictates this, and if exceeded it burns up)
15 Milli seconds. 15/1000th of a Second.
on modern cars, it varies, from cranking to max RPM, to keep heat down and to make dwell longer at high rpm. not over 15mS,
so the coil pin is at 12vdc when the coil is off, a test light glows then.
then the ECU charges the coil for say 10mS, you cant see that with out a scope.
if you crank it the light may, flicker,? this dont prove it's working, not at all. or with normal eyes, you might see nothing..
math:
your engine turns 250rpm cranked. divide by 60 to get seconds. (wasted spark rules)
so is 4.1 RPS revs per sec. call it 4. and coil 1/4 fires together, 2 times, per 4 sec. (a wasted spark deal) 1,3,4,2 order but fires 1/4, the 2/3, then 1/4 again, in one turn of crankshaft.

250 ms per turn (=4PRS). during that 250MS period the coil is charged 2 times, at 10mS each turn. (typical)
that works out to 12v for 230ms and 2 negative pulses for 10mS each. FAST that is, 10/1000th second.... fast..
most humans can not see light flash that fast. (why movies are called flicks) Movies are best seen at 20 frames a sec. or faster for comfort.
some can , some cant see the flick. one person see a movie the other sees torture. (1920 Charley Chapman era, seen today at excess speeds.)

the scope cures this. we can see 3 things, the full pulse, any noise there, and the fact that the pulse is NOT WEAK.
The WEAK thing, only scopes can see. never some test light.
the scope will allow me to see a leaking transistor, say causes the voltage not to reach full battery volts. 0v, 10v, 0v 10v. like that. and would be bad.
it's an analog world, to a tech. and is not digital , until digitized.
or the low signal does not reach 0v. that too is bad driver.


substitutes, why are you doing that, and the part is still unknown
you cant do that, yet.
what if the part has read back pin.?
if you dont get that right, the ECU will shut down the coil.

Good luck finding out what contract parts are used.... that be super good to know,
most companies, have a non disclosure agreements, with contracts, and ST.com is bound by that .... a signed contract.
them might even say, we are not going to read ours, just for you.
in most cases, they tell me that, it's a contract and that is all we are saying,,,,
that is why I send my hard nut cases to CARDone.
they dod all this hard work for me.....
but yes, after full rev. engineering i can replace a dead obsolete, VSV solenoid driver, in a SIP pack, i cut the pins off the dead transistor
and add my one, discrete trans. but its just a SIP pack with 6 transistors, NPN and i know the beta and other specs. too. so can sub out easy
but as i say, you dont even know what you have. yet.



Cardone may say, sorry, can't do Siemens anything....

---

2 wire CKP sensors only, (is yours 2 wire?)
https://www.youtube.com/watch?v=2tHkV3v6XIo
see the first seconds here, of the VR CKP? first 3 seconds of video, repeat this... over and over. see?
see how weak it is. then it starts, and voltage just to +2 volts and -2v swings... (faster you go the more volts it does)
so.... that is how we check CKP
there is NO other ways.
sorry. and is 1st check. when spark is lost.

here is same deal, but note its only 0.2v up and down. weak as a cat.
see? great example of weak CKP, yours is new, and im sure strong. but is the timing right, does it match the cam? !
https://www.youtube.com/watch?v=7Ofa8Gcn7S8

see coil as amps. , see each charge up.
here
https://www.youtube.com/watch?v=F8Pr81mH...c7fHhuPJIV