Motronic Ml4.1 Diagnostics.

[davemar 16]

I've just fitted a Motronic ML4.1 setup to my Mi16 engine after it being a Jetronic setup, so I'm trying to get it all up and working. So far it doesn't want to fire up, so I decided to try using a diagnostic switch and light (using pins 4 & 17 on the ECU).

 

According to Mr Haynes, there's two modes of diagnosis that can be done. One requires you to press your switch (i.e connect 4 & 17 to ground) after the ignition is on for 3 seconds to get a code (sequence of flashes) for sensors. The other mode is to press the switch while you turn you ignition to test the actuators.

 

I decided the try and test the sensors by pressing the switch after turning the ignition on, but it behaves like the second mode which buzzes the fuel injectors and gives my a 81 code (fuel pump relay). Is there any way I get it to diagnose the sensors without it actually running the engine (which it is failing to do)?

 

Also if the fuel pump relay is a problem (the relay is brand new), how do I know what sort of thing it is counting as faulty? How does it check it, the resistance of the relay coil?

Edited February 20, 2012 by davemar

[welshpug 1,657]

got a crank sensor and suitable flywheel fitted?

 

no experience of the diagnostic side of these systems, but it would be easy enough to test if the relay's are working as they should, i.e do you have power at coil and injectors, fuel pump etc.

[davemar 16]

Yes, I've got all the right bits and bobs for the Motronic setup: flywheel, crank sensor, AFM, distributor, ISCV, coil. I ought to just go round and check all the voltages for each sensor/device/relay/etc before getting too bogged down with diagnostic codes.

 

I'm curious to whether the ECU 'sees' something it doesn't like with that relay and just does a "computer says no" and refuses to let the engine run at all.

[welshpug 1,657]

I doubt it, its just a relay.

[davemar 16]

I've just had a bit more of a poke around measuring voltages on the ECU pins. Everything measured up well expect for pin 1 which is supposed to be the signal to the coil amp. According to CAPS this should be nbv (+12v) when the ignition is on but I measured 0v. This may suggest the ignition amplifier isn't connected correctly? The amplifier was working fine on the old loom, so unless its fried itself somehow, it is a little weird.

 

Just looking the wiring diagram it shows ECU pin 1 going into pins 2&3 of the amplifier. I'm guessing the input to the amplifier is going to be a high-ish impedance input, so measuring the DC voltage there when its not got a signal on it isn't going to be too helpful. It would be nice to know what the circuit diagram of the amplifier is, and whether it has a capacitor coupled input. Anyone know?

 

I also observed the waveform from the crank senor, which you can see in the image. This the signal measured at the ECU pins (so I know the wiring is OK) with the engine being cranked. The signal looks OK, and I've caught the gap in the teeth there too. My only concern with this is whether the signal is too small, it's about 3.5-4.0v peak-to-peak for the non-gap part of the waveform, though I would expect this to be enough.

Edited February 21, 2012 by davemar

[davemar 16]

Just checked the injector signal while cranking and it is showing pulses, so clearly the crank sensor signal is getting used. I disconnected the ignition amp from pin 1 and checked the signal on it (in case the amp was pulling it to ground) and it was still sat at 0v when cranking. Maybe the ECU's pin 1 output has died?

[ade 4wd 11]

Check that pins 5,16 and 19 on the ECU are earthed. If so, then it looks like your ECU is faulty. No output on pin 1.

[matyasg 0]

 

Just looking the wiring diagram it shows ECU pin 1 going into pins 2&3 of the amplifier.

I don't know, I have a different wiring diagram - see attached (10 - is the ign.amp.)

 

(

[ade 4wd 11]

Wiring diagram above is correct.

Ignition amp pinout is,

1 - output to coil

2 - earth

4 - switched 12v

5+6 - input from ECU

[pug_ham 244 3 Cars]

Check that the ignition amp has an earth.

 

If the ignition amp plug has a small wire with ring terminal coming out of it then make sure this is connected to a suitable earth point, I worked on one Mi16 with a no spark problem quite a while ago & this was the cause.

 

g

[davemar 16]

Wiring diagram above is correct. Ignition amp pinout is, 1 - output to coil 2 - earth 4 - switched 12v 5+6 - input from ECU

 

My ignition amp is wired correctly, I just read the pin numbers in the wrong direction, hence 2&3 instead of 5&6. I've ye to check the amp's earth.

 

On the ECU pins 5 & 16 are earthed. Pin 19 has no connector, and I think that's because it is usually an output pin for other uses.

[pug_ham 244 3 Cars]

Pin 19 on the ML4.1 loom should be an earth, are you sure there isn't a wire into the ecu plug?

 

g

[davemar 16]

The loom didn't have pin 19 connected, but I managed to get a connector and connect it to ground. The good news now is that I now get a spark and the engine sort of runs. The bad news it doesn't do it very well. Here's the symptoms:

It runs for a few seconds and then gives up.

The rev counter doesn't want to get much above 2300rpm even though I could rev it well beyond that. The needle was a bit bouncy too.

When it did idle, it was at about 1800rpm.

 

So there's something still fundamentally wrong, any ideas for the initial checks? Its a little difficult to test as it runs for such a short time before cutting out, so I haven't got much chance of probing things while its running.

[davemar 16]

Had another attempt and the engine ran continuously now, and eventually settled to an idle around 1000rpm once warmed up a little. I didn't have a chance to drive the car as the loom was hanging all over the place, so couldn't put any load of the engine to see if it was really running properly. It certainly revved OK.

 

However, the rev counter is still not working properly, not getting above 2300rpm or so and being a little erratic. Is the 205 rev counter compatible with the Motronic ML4.1 setup? The ignition amplifier is the same, so I assume it would be.

[davemar 16]

I tried driving her very short distance and it wasn't good. A lot of misfiring and feeling rough. It seems the ignition side of things isn't working properly. So the erratic rev counter and limiting seems to be either the ECU's signal (from pin 1) or the amplifier. I had a quick look at the signal coming from pin 1 and it seems to be a clean square wave of about 3.5v, with the +ve phase of about 5ms and the 0v phase varying with revs. Nothing look obviously wrong, but it difficult to really tell.

 

Any suggestions welcome, as I'm getting very stumped here!

[matyasg 0]

I would try a different (known good) ignition amp.

 

The other thing I would try to add a thick cable between the engine and the chassis to be sure that there is a good earth.

[pug_ham 244 3 Cars]

Did you check for an earth for / from the ignition module?

 

Last ML4.1 loom I worked on had a seperate ring terminal out of the ignition amp plug that connected directly to earth & without it the car ran badly with the rev counter giving erratic indication.

 

Is the engine / gearbox earth lead tight on the stud?

 

g

[davemar 16]

Just doing some maths here. The +ve pulse from ECU pin 1 (signal for ignition amplifier) is 6ms, which is the dwell time. From what I can see it fires twice per engine revolution, so rpm = 30/period. Where period = 6ms of +ve and x ms of 0v.

So at 1000rpm, period = 30ms, so x = 30-6 = 24ms.

So at 7000rpm, period = 4.2ms, so x = 4.2-6 = -1.8ms!!!

 

So with a dwell time of 6ms, anything about 5000rpm isn't going to work. Am I overlooking something very fundamental here, or does this 6ms dwell time drop above 3000-4000rpm or so?

[davemar 16]

I added a new earth lead from the ignition amp to the main gearbox earth and that made no difference. I also tried another ECU and that was exactly the same.

 

I also rigged up a little circuit that generates a square wave to simulate the ECU's output. I could vary the frequency and keep a constant time (4ms) on the positive phase like the ECU does. Used a spark plug on the coil output, so I could test it without needing the engine moving. The spark plug fired away quite happily, and the rev counter moved smoothly through the rev range. The spark makes a noise and that sounded clean too. So it suggests the ignition circuit is fine, and its the ECU that is sending a dodgy signal, or that there's something interfering with it all when the engine is running.

 

The AFM and coolant sensor voltages seem fine, so I'm rather stuck now!

[matyasg 0]

I added a new earth lead from the ignition amp to the main gearbox earth and that made no difference. I also tried another ECU and that was exactly the same.

 

I also rigged up a little circuit that generates a square wave to simulate the ECU's output. I could vary the frequency and keep a constant time (4ms) on the positive phase like the ECU does. Used a spark plug on the coil output, so I could test it without needing the engine moving. The spark plug fired away quite happily, and the rev counter moved smoothly through the rev range. The spark makes a noise and that sounded clean too. So it suggests the ignition circuit is fine, and its the ECU that is sending a dodgy signal, or that there's something interfering with it all when the engine is running.

 

The AFM and coolant sensor voltages seem fine, so I'm rather stuck now!

 

Have you tried the engine with the alternator disconnected? Disconnect all the wires from it, you can remove the belt as well.

I saw several times when the noise came from a faulty alternator.

 

Sorry, it's just a guess again.

 

By the way.

Are you sure that the signal from the ECU is always 4ms long? It should be varied I think, because of the different coil dwell time. AFAIK these ignition modules are not the smart ones, so the dwell values are coming from the ECU.

[davemar 16]

Haven't tried the alternator yet, but will bear it in mind.

 

I would expect the ECU to reduce the pulse length as revs rise, otherwise they will overlap. I chose 4ms for my little circuit as I wanted to keep it simple and have a period that will allow full revs. I measured the pulse at 6ms from the ECU at idle, and maybe it isn't reducing as revs increase. I don't think this is the ECU's fault as I've tried a different ECU and the fault is the same.

 

I'm wondering whether the ECU is getting confused over a sensor, even though I have checked through them all?

[davemar 16]

I've checked the output on pin 1 of the ECU more throughly by recording it on my laptop via the soundcard (with some intermediate electronics to get the signal levels down), and doing the same with the signal from the CAS sensor. The signal on pin 1 is a clean square wave which matches the frequency of the CAS sensor perfectly, so it looks like it is sending out a good signal.

 

So I'm now thinking it is how the ignition amplifier receives the signal. Does anyone know if pin 1 on the ECU is open-collector? If so, maybe the ignition amp isn't providing a pull-up resistor and not getting a good signal. My little signal generator which worked well with the ignition amplifier was an emitter follower, so maybe gave it an easy source.

 

Maybe some ignition amps aren't made the same as others. It would be nice to know what the circuit diagram inside the ignition amplifier actually is; I have googled for it, but without any joy.

[davemar 16]

Solved it! It was the coil in the end. I replaced it with a new one and now it revs smoothly and the rev counter responds accordingly. The reason I didn't try this earlier was that the old coil worked fine when it was running on the Jetronic setup (same ignition amplifier too), so I assumed it was a good coil. Clearly there's a subtle difference in the way it gets driven in the Motronic setup that exposed a flaw.

 

Right, now I can strip out the old Jetronic loom and get the Motronic one properly fitted!