'91 ecu troubles

[greenfield54]

Sorry mine is actually a 96 with no oxygen sensor installed. Reason is that during that time all gas sold here still use lead. This unit came fully assembled from the factory. The heavy thermal insulation and heating system are still there. These I removed after gutting the car in 2013 when we bought it. It is OBD2 but I found the check engine light disabled as indicated in the schematic. A blind which is confirmed original from the factory was used to cover the light with no bulb installed. When it was still running fine we used a scan tool for my nephews 2012 Hyundai but registered no fault codes even if I intensionally pulled the coolant temp. sensor. The Haynes manual is fine but is not what most would expect from a service manual being for U.S. units with all the bells and whistles. And U.S. units don't use Siemens ecu so I was in limbo when finally a friend emailed me a schematic for a MAP based Siemens ecu. I suspect that fault logging software was omitted during the programming. I have experience programming industrial Programmable Logic Controllers (PLC) and we can omit fault monitoring/logging if the system does not really need it. The first time I went under the hood it started (roughly) but died almost instantly. There was an intermittent weak spark from the plugs so I proceeded to do a bench test of the dual coils which produced no spark. It drew very heavy current so I concluded that the primary windings shorted out. A new original coil made no difference. I monitored the signal that triggered the coils with a light bulb. At first it flashed dimly but now no more signal. The ecu still cycles the fuel pump after switching on the ignition and flashes the check engine light so this means that the microprocessor chip is still alive. The injectors also cycles when starting but the engine fails to start due to no spark. A friend replaced two electrolytics which appeared bloated but left the 100 mfd alone including the tantalums so I wonder, can this be the problem? The ecu was embedded in one inch thick potting compound which took me more that a week to peel off. I hope this helps give a clearer picture of what I'm against. We are already contemplating to part out this car. Thanks. The engine is a Beta 1 G4GR of Mitsubishi design.

Regarding the timing belt and cam, everything is in order. I changed both the cam and crank position sensors with original ones (very expensive). I checked for 5vdc and continuity of the wiring harness. Checked all grounding (sensor,logic,B-). All signals to the ecu (B+,Ign,Trans in park signal,etc.) are present. I have not pulled the valve cover though to see if the cam is actually turning. I only removed the belt covers to check.

[greenfield54]

This car was imported direct from Korea by a legit (cannot remember the company name) car dealer when Hyundai was still starting out down here. The diagnostic port is a 16 pin D-connector. If you have an email ad, I can email you the schematic that I used. This site allows only 500KB. I am currently working on a 1994 Corolla with a 4afe engine. This car has a diagnostic port in the engine bay and use flash codes being obd1.

The G4GR has two coils (wasted spark system). The CMP/CKP sensors and ignition coils are brand new. The coils have three primary wires. One wire from each coil connected together (B+) and the other two are separately connected to the ECU ignition coil signal outputs. And 4 HV wires. It is not COP(Coil On Plug) which we also have down here on newer cars. As I said I have not removed the valve cover to actually see the cams turning but I did remove the timing belt cover to check the timing. I will remove the valve cover but I doubt the cams not turning is the problem. If it is, this will be the first time. I am quite sure about the health of the engine.

   

   

   

[greenfield54]

The CKP and CMP sensors are fixed and have no way of adjusting them. I took them off in 2012 when the car was dismantled and put them back in after 6 months with one bolt with no problems. The car ran fine with many, many miles before this happened. I have no compression tester so I will need to remove the valve cover and actually see if the cams are turning. This would be easier for me to fix than an ecu gone bad.

OK let's say this is actually OBD1, then why is the 16 pin connector under the dash? Very difficult to put jumpers in right?

Regarding the ignition coil. There are 2 in one package. One of each primary wire are connected to 12 volts and the each of the other two are connected to the ecu. No firewall mounted drivers. I like a system with all the drivers outside instead of inside the ecu. Looks cluttered but easier to troubleshoot. Agree?

Regarding the ecu don't worry. No SMPs were damaged. That is why it took me more than week to remove most of the potting compound. The pcb is clean with no signs of acid damage or burned tracks. As I said only two small electrolytics were changed (Nos. 1 and 4). I would change nos. 2 and 3.

   

[fixkick]

the top pins 28 and 29 are the 2 ignitor pins.
one transistor is blown. end of life, that ECU.

[greenfield54]

You are amazing man!! You really know your stuff. Thanks. Maybe a lot of people will learn from this thread despite the fact that I literally hijacked the OP's thread and for that I humbly apologize.

OK. Back to business. Regarding the missing sensors on the schematic, I deliberately erased them because they are not present on my engine but I will send you the original anyway. I agree. This ecu is weird. It must have been custom made just to get their cars down here. Would you believe that this car has no charcoal canister and vents fumes from the tank straight into the atmosphere? Believe me. I am telling the truth since I was the one who dismantled every nut and bolt of this car. Tomorrow I will go to where the car is parked and check the camshafts. Will also take some more photos.

Again thank you. I learned a lot and I hope the others too.

[greenfield54]

Interesting!! I'll try what you suggested. In the meantime I went this morning to where the car broke down and is parked. Quite far from where I live. I pulled the valve cover and inspected cams and timing. All are as it was when I set it up in 2012. It ran perfectly from then to 2015 after many, many miles on the odometer. Nice car to drive so I am sad it is now sick. Attached are photos I took this morning.

   

   

   

   

   

---

Some more photos.

   

   

   

   

   

[fixkick]

id say the harness and ECU do not match that schematic at all,
looks like the CMP is fixed. good. less variables there. very good. that.
chain linked cams, better.
then belt that must be timed ok. not early or late timed.
now that 16pin connector is shows its not "PWM, or VPW, or can-bus" (2004 is that) but can be ISO9141/14230 (comms) pin 7 only version.
lets look at the populated pins, on car this old (got air bags< got ABS? ?)? 2,3,6,9?,10,11,13. sure looks OBD2 to me.

observations on this DLC.
i see the pins are numbered correctly , just an artifact if the connector maker.
why all that grease on that connector.? pin 9 is impossible to see pin there.
ISO9141/14230
this is pin 7 only scans.


so if it is OBD2, and the ECU is OBD2.
then you connect any modern (old scan tools skip pins and can fail on old cars)
the newer scanners support all pins, every one, listed in oBD2 pages.
seen here
https://www.scantool.net/images/diagrams...pinout.gif

but this publisher skips a critical pins.

4 and 5 must be grounded. or the scan tool will never work or comm.
pin 16 is fused 12vdc, if not, most scan tools go DEAD.

check for 12v on pin 16 yet key on, that be first.
of its oBD2 compliant it will work.
how a USA car got to your country in 1996,ive not a clue, maybe it just USA harness and ECU, only , added.....at some time.
it sure does not match that KDM body tag. at all,

but he connector to me looks OBD2 setup, for ISO

ISO9141/14230

is there a data code on that ECU?

wow a GM map sensor on 1996 Korean car? hummmmm odd. seems it be Korean or Mitsubushi or Denso to me.


got a scan tool? obd2.
check for 12vdc on pin 16, it must be there key on.
plug in tool
key on, magically it will connect just key on, and can even watch you move the throttle and report throttle angle.
crank engine see RPM leave 0 and it, 300 cranking, if not the CKP or CMP are dead.
crank for 5 seconds, got P03xx errors. BINGO.

that is all you do on OBD2 cars, just that.
but the scan tool needs to be wired to pin 7 and support ISO9141/14230 OBD2 signalling rules. (all newer scan tools do, old 1996 many didnt)

what you must do is make the ECU happy
working blind with no matching scan tool makes that near impossible...... on any car with EFI.
see?
even a bad CKP coax can kill spark. just that........ bad wiring.... and must be correctly wired with shields just like new... (no ground loops by accident)
bad grounds to ECU.
power pins dead.

the $10 ELM327 based, scan tools work here. pin 7/15 wired. $10 is ebay clone price.
ever seen the ELM?
https://www.scantool.net/elmscan5-compact.html

pin 4,5 wired to ground
pin 16 ,battery voltage, 12v
pin 7 and 15 from tool to connector good, no grease on pins. ever.
key on, and it communicates.

if we had a real schematic matching the actual ECU in car, this be far more easy, i think yours is wrong. that drawing.

[greenfield54]

Some more photos and recap.

   

Below is the series of events as it happened:
1.Car started normally but hesitated for a moment.
2.At a traffic light, engine ran roughly as if only 2 cylinders firing. Car was towed back to garage (not mine).
3.A mechanic (not me) replaced all plugs. Car started but still rough.
4.Mechanic above pulled each HV wire and found only 2 cylinders firing.
5.When I checked, engine started the died immediately.
6.I monitored signal from ecu to ignition coil input using small pilot 12v pilot lamp.
&.There was dim flashes then nothing anymore.
7.Bench tested ignition coil and found no spark on both and drawing very heavy current. I concluded the coils on LV side shorted out. This coil has no transistors because it is not COP.
8.Replaced coil. Still no spark.

What I did next:
1.Checked for 12v at pins 24,32 and 52. All ok.
2.Checked for 5v on all sensors. All ok.
3.Checked power,logic and sensor ground. All ok.
4.Checked pin 9 switch drive position. OK.
5.Checked continuity of harness wires to sensors from ecu plug to sensors. All ok.

The trouble as you can see occurred in stages and deteriorated in a no start condition.

Finally I want to thank you for your patience and I promise to go thru the procedures you indicated. Will report back in about a week since the car is quite far from where I live.

[fixkick]

the shorted coil burned up the ECU , ignitor transistors.

this and injector drivers are the 2 most common failures of ALL ECU.
in 1996 they didnt have overload detecting transistors like today.
end story, bad ECU.

---

the oBD 2 scan tool should show P03xx codes, for lost of spark.

[greenfield54]

the shorted coil burned up the ECU , ignitor transistors.

this and injector drivers are the 2 most common failures of ALL ECU.
in 1996 they didnt have overload detecting transistors like today.
end story, bad ECU.

---

the oBD 2 scan tool should show P03xx codes, for lost of spark.

That is what I have been suspecting all along. Correct me if I am wrong but slow blow fuses could have helped in protecting the transistors.