Opel Corsa D ECU Repair

[digik]

Hi, i'm trying to repair the ECU of my 2008 Corsa D. The problem of the car is basically a very rough idle, like it only runs on 3 or 2 cylinders. This ecu is well know of suffering from water ingress due to the poor location right under the windscreen scuttle panel, BUT in my case since it was previously sealed (because i've already replaced it years ago) it was absolutely dry inside with absolutely no oxidation. So, after 5 different mechanics checked injectors, fuel pump, egr valve, etc. they all told me that the problem is in the ECU.

So i opened and started some very basic checks as i don' t have fancy test gears, just a multimeter. The injectors are driven by 8 STB60NF06 mosfet, which are all fine. Then i've measured the gate voltages and found that in two of them are 0.50V and when the car shakes the voltage drop to 0.30 - 0.40V and then returns to 0.50, but is very unstable. The gate voltages of the other mosfet is 0.60 and are much stable, although a little jumpy too. Now those voltages are obviously the low side output of a MC33883DW H-bridge driver and there are two of them: one for the first 4 mosfet and another for the other 4. So in conclusion the low sides of one MC33883DW outputs an unstable 0.50V, while the low sides output of the other MC33883DW has a much stable 0.60V. Is it enough to call one MC33883DW bad? Or should I check something else? This IC is also very rare to find, if not impossible so can you suggest me an alternative?

[digik]

...another ecu from trash with same problems, maybe even worse. No, thanks I don't want to spend another € 600,00 for nothing. plus the working principle of those h bridge driver should be well known to some forum members and yes, i don't want aliexpress shit on my car, so i need some compatible part, but before that i need to know if the driver could be bad based on my measurements

[magic]

What a DMM shows you is of course the average voltage, and in reality the gates are driven with square wave of a few volts peak but fairly low duty cycle, hence the average is low. The voltage that you see is likely to vary depending on engine RPM and throttle.

There is definitely something suspicious about your observation, although with this kind of control systems you can never know if the signal is getting lost somewhere along the way or perhaps the controller itself generates a weird signal for some weird reason. I'm not familiar with ECUs enough to tell what's the more likely scenario here.

You could similarly measure average voltage at the inputs of the drivers - if the inputs are constant, but the outputs change, and the engine struggles when the outputs get weird, that would point towards a problem with the driver. Either the IC or maybe some support components around it, like capacitors.

[digik]

thanks magik, i have also suspected bad caps, but those are super small smd ceramics with no values on it, and of course without schematic it's impossible to determine their values. also is very unlikely that those caps are critical. regarding the inputs, i've measured them also and there is a small difference between the goods and the bads of around 6mV, i think not enough to cause issues, but correct me if i'm wrong. the inputs goes straight inside a custom microcontroller (by magneti marelli), so if this is shot i can throw the ecu in the bin. what do you think, should i try one of those driver from ali? i'm quite shure that they are all fake and honestly i'm not confident to drive with such an important component ready to fail. is there a compatible drive that i can try?

[magic]

To be honest, I have only ever taken apart one (old and primitive) ECU and the injectors were all wired together to 12V on one end and individually switched to ground by a single transistor per injector. So I'm not sure why there are 8 FETs and H-bridge drivers on yours. Can you post a schematic of this injector drive circuit with measured voltages written on it?

6mV may or may not be a big difference, depending on whether it's 600mV vs 606mV or 6mV vs 12mV. Also, are these voltages very precisely equal on "good" cylinders or is this magnitude of variation normal between them?

If a custom MCU is braindead then yeah, into the bin the whole thing goes. But the MCU could be generating unusual control signals due to wrong data from sensors, or faults in some sensor input circuitry, or even correct data from sensors indicating legitimate problems caused by something else (maybe even those gate drivers, or FETs). Are there any sort of sensors which are per-cylinder? Perhaps some crankshaft angle sensor?

Does the ECU support computer diagnostics? Was this tried? Any errors reported?

[digik]

yes, it supports diagnosis and it throws all sorts of errors about injectors, which have replaced, tested and coded into the ecu. before injector replacement the car stalled completely on time and didn' t turn on. no sensors on injectors afaik

[fzabkar]

Datasheet:

https://www.nxp.com/docs/en/data-sheet/MC33883.pdf

[Gyro]

There are companies, in the UK at least, whose entire business is repairing automotive ECUs (of all kinds, abs etc.), I would be surprised if there wasn't at least one in Italy. Garages (except franchised dealers, obviously!) use these all the time as an alternative to the customer paying potentially thousands for a brand new one from the car manufacturer. They have the  large amounts of equipment necessary to simulate an engine, all of it's sensors, actuators, injectors etc. together with the appropriate rework facilities.

I'm sorry to say that from reading the thread, I don't think you have the equipment level or experience to repair one of these. I'm just being realistic. These repair companies are unlikely to accept an ECU that has been tampered with, you could probably simply say that you opened it to check for water ingress. This path will at least result in you keeping your own ECU with any relevant codings, safely re-sealed, and the repair will include a warranty, not something you will get from a scrap yard or aliexpress.

If you can't find any obviously cracked solder joints or components by visual inspection alone under a magnifier, I would send the ECU to one of these companies. If they find the board is covered in meter probe marks, they probably won't touch it.

[minifloat]

My wife had a Corsa D. Same symptoms as described. The ignition coil was the culprit. During the 160000km / 100000mi it went through the factory-original and two replacement ign coils. As you already said, water ingress and thus corrosion has been the problem.

br, mf

[fzabkar]

Are you sure you are not looking at the throttle actuators rather than the injectors?

https://www.eevblog.com/forum/repair/need-replacement-for-an-obsolete-mosfetcomponent/msg4415851/#msg4415851

[nali]

So, after 5 different mechanics checked injectors, fuel pump, egr valve, etc. they all told me that the problem is in the ECU.

That often translates to "We can't think of anything else"

If the engine is running OK except idling I doubt it's your injector drivers, more likely a sensor reading somewhere (which granted still could be the ECU). Also I heard that Vauxhall/Opels were notorious for EGR valves sticking which resulted in rough idling... you could try blanking it off completely as a test.

[magic]

yes, it supports diagnosis and it throws all sorts of errors about injectors, which have replaced, tested and coded into the ecu. before injector replacement the car stalled completely on time and didn' t turn on. no sensors on injectors afaik

Understanding reported errors should be the first step. I don't know how human readable, standardized or well-documented ECU diagnostics are, but try to get maximum possible information out of them. Hopefully more is possible than just "lots of errors of all sorts".

Example: in the ECU I repaired, I was told by the mechanic that it complained something about "wrong injector voltage". This was fairly quickly traced to a defective power switching relay with excessive contact resistance.

[digik]

not only at idling: it goes also rough at high speed, like if it runs at 3 cylinders.

[fzabkar]

@digik, could we see a photo of the PCB?

[amyk]

Do you think it's definitely a fueling and not ignition or even other mechanical issue?

You'll need an oscilloscope if you want to see the waveforms.

If you suspect fueling and this isn't a timed EFI, you can try swapping the signals between injectors to see if the dead cylinder moves.

[tatel]

Really cheap oscilloscope that would probably be enough to check all waveforms in your car (min sensitiviti 20 mV/div): https://aliexpress.com/i/1005004972406356.html

(Probably you should get the model with probe and battery)

[magic]

One quick check with a DMM that could still be made is to measure the LR voltage of the suspect chip, since you say that it's the low side channels which behave weird.

Gyro also has a valid point - the more you mess with it, the more chance of screwing something by accident. There may very well exist people repairing such things for a living in your area, they will have all the test equipment and experience, they will hopefully repair it quickly and without unnecessary collateral damage.

[darkspr1te]

This model has the all in one coil pack , i would deff swap that as a test, known failure point. for the SA model they actually swap it out for individual ones and a second loom to fit the original harness. 


darkspr1te

[digik]

One quick check with a DMM that could still be made is to measure the LR voltage of the suspect chip, since you say that it's the low side channels which behave weird.

Gyro also has a valid point - the more you mess with it, the more chance of screwing something by accident. There may very well exist people repairing such things for a living in your area, they will have all the test equipment and experience, they will hopefully repair it quickly and without unnecessary collateral damage.

Why the LR voltage? for what is used? the datasheet just says that is the output of a linear regulator

[magic]

Carefully study the block diagram on page 3 of the datasheet

[mikerj]

Are you sure you are not looking at the throttle actuators rather than the injectors?

https://www.eevblog.com/forum/repair/need-replacement-for-an-obsolete-mosfetcomponent/msg4415851/#msg4415851

I agree, I have never seen injectors driven by an H bridge, they typically have a single low side driver.  The H bridge that the OP is looking at is far more likely to be for the throttle actuator.

[digik]

Carefully study the block diagram on page 3 of the datasheet

Yeah, sorry 

Are you sure you are not looking at the throttle actuators rather than the injectors?

https://www.eevblog.com/forum/repair/need-replacement-for-an-obsolete-mosfetcomponent/msg4415851/#msg4415851

I agree, I have never seen injectors driven by an H bridge, they typically have a single low side driver.  The H bridge that the OP is looking at is far more likely to be for the throttle actuator.

Then why the car is shacking at idle when the gas pedal is not pressed?
My throttle body sould be like this:



so an entire H bridge only for this sh*t?

[digik]

This is the back of the ECU with 3 MC33883. Note that this is NOT my actual ECU, but an identical one, just to give you an idea. This one has water damage, mine NOT!



The front side:

[Ranayna]

My Corsa D also got the ignition coils replaced. Apparently this is a known issue that the car gets very rough on idle.
I trust my mechanic with that, i do not remember how much that cost, but the repair was definitely not "ECU Replacement" expensive.

[digik]

My Corsa D also got the ignition coils replaced. Apparently this is a known issue that the car gets very rough on idle.
I trust my mechanic with that, i do not remember how much that cost, but the repair was definitely not "ECU Replacement" expensive.

Apparently on this car EVERYTHING can cause a rough idle.
I've already spent € 500,00 and the car is worse than ever. No mechanic has solved the problem and I won't give anyone a cent again.
The car is really a piece of garbage engineering, a mix of FIAT parts put together by some drunk engineers.