Calling on Motronic experts: IC information
 

I am working on a 3.2 Carrera Motronic and found that one integrated circuit (IC) is most likely bad. Its input is low impedance and straps the injector signal to a fixed value close to GND. As a result the injector output stage does not work. Once I disconnect the input pin (4) of the IC the injector signal produced by the CPU looks normal again. Nice 4.3 msec pulses during cranking like one would expect at R418/R419 The IC is used to drive the injector output darlington transistor T402 and the fly-back circuit Y403/T404. My guess is it must be some type of custom op-amp or similar. It has some resistor network and filter components attached to it.

The part number is 0127 and the date code is 8712. It is a DIP-14 housing with logos from Bosch and Siemens on it.

http://forums.pelicanparts.com/uploa...1205695110.jpg

http://forums.pelicanparts.com/uploa...1205695138.jpg

I understand Bosch while never making their own semiconductors used custom-designed chips made by companies such as InterDesign and had other chips custom-labeled by many suppliers. Is there any recourse to find out a replacement part? What do others do to fix a problem like this?

Thanks,
Ingo

Having put some more thoughts into this I came up with more questions:

The 911 3.2 injector configuration is such that all six injectors are wired up in parallel and the injectors are of the peak & hold type with low impedance. Let’s assume the DC resistance of one injector is 2.5 Ohms. This makes the total resistance “seen” by the DME something like 1/R=(1/2.5)*6 or 0.41 Ohm.

Looking at the schematics a little more I expected IC S400 to be a custom peak & hold driver with one negated output for the fly-back control stage. The external timing components and the sense resistor R409 suggest this.

However, with R409 being a 10 Ohm resistor the output stage can not drive more much more than 1 to 1.2 Amperes through the entire injector bank to begin with. At those current levels the resistor alone acts as a perfect current limiter and no peak & hold is required to further reduce the current. What do others think?

Ingo

Ingo,

Is signal at ADV13 from the CPU? Assume so,..but don't have prints. It has to be....


As to the 0127,...would there exist an numbers on the front side of "0127"?

Doyle,

correct, ADV13 is an output from straight from the CPU. It is active low. When operating properly you would see a +5V with short pulses to zero. It is pulled up to +5 Volts on the digital board. During cranking I saw 4.2msec solid pulses (no PWM) with a frequency of about 10Hz making for 600 injection events per minute during cranking.

I am more and more puzzled by the 10 Ohm resistor in line with the injectors. It seems wrong in many ways and yet the thing is there. I measured it and it is 10 Ohm. See my other post here: http://forums.pelicanparts.com/showthread.php?t=398809

Ingo

I believe R10 is 0.1 ohm


Edit: I have asked about that chip many times with no response so you might be SOL.

Rick, I will measure it one more time tonight. You are right, the way it is named in the schematics it should be 0.1 Ohm. At that level I think I might look into the LM1949 as a potential candidate to do the peak&hold current control. It's design is based around a 0.1 Ohm current sense resistor.

OK, my bad - you can all start to call me names. It is indead a 0.1 Ohm resistor. So that makes S400 a peak & hold driver. Now the one-million dollar question is how to find a replacment part. The only thing that I can come up with is the LM1949. Nothing else seems to fit the bill. So my options at this point are to sacrifice a 944 DME or to design a new output stage with the application note coming along with the LM1949. Grrr.

As it so happens we have an eletrical engineer working for us who used to work for Delco when they designed a similar application with a Motorola chip. He was part of the team that spec'd the Motorola chip. Unfortunately, he can not remember any details - much less the chip type they ended up getting...

Any clues out there. I know this information is out there - would some please step up... ;)

Ingo

I take this thread up again, since I had the same problem with a 911 ML3.1 box. Still looking for further information about the 0127, I'm wondering if anybody found news in the meantime.

For the schematic: R409 is labled R10 witch is short for 0R10 or 0.10 Ohm. The injector circuit is not connected through ADV13 with the CPU board. Connector ADV12 is used.

S400 and S100 are the only custom chips in a ML3.1. It would be great for the future of our cars if we could find resources for both.

You may want to ask this question in the Bosch Motronic forum over on the Dorkiphus Porsche board:

Bosch Motronic - The Land of NoVA/DC/MD area Porschephiles

Have you looked through the Bosch Semiconductor website/data sheets ?? ( I can't view the pics in your original post from this PC)

Robert Bosch GmbH - Automotive Semiconductors and Sensors

John

Great link - thanks!

But I'm afraid that our boxes and chips are a little bit outdated for the actual Bosch development. Anyway - very interesting info on that site... :)

I have searched high and low and this DIP14 is not and was never available anywhere. It is a custom design and Bosch never sold it.

Modern injector drivers are realized with power MOSFET or IGBT stages rather than Darlington BJT. These have much higher conductance and thus less power dissipation issues. Often these stage are integrated into a programmable low-side switch.

Looks like both chips are a kind of copy protection for the motronic boxes.

The S100 is a simple wave->square converter for the position mark and the teeth. They added a simple flip-flop to keep the position mark signal up until the cpu can react on it.

Looks like we have to spend a little more time developing replacement circuits for both... ;)

IC info
 

Guys, it may not be a pin for pin match, but I am sure that it can be replaced with the LM1949 peak and hold injector driver chip.

LM1949 - Injector Drive Controller

The LM1949 is an 8 pin DIP so not sure what the 14 pins are for unless it incorporates some additional circuitry.

The MSEFI injector tester by B&G uses the same chip to drive the injectors under test.

Read about that here. www.bgsoflex.com/FI_tester_doc.pdf

I think it works more like this preliminary datasheet part from 1988.

L584 Datasheet pdf - MULTIFUNCTION INJECTION INTERFACE - SGS Thomson Microelectronics

The driver in the DME might be a little over designed.

http://forums.pelicanparts.com/3868545-post194.html

I'm not sure this part experiences a high mortality rate.

I agree with Rick - the L584 is very similar to the 0127 used in the Motronic. Unfortunately it appears to be discontinued.

At the time IGBJ and PowerMOSFET devices with their extremely low RDsON weren't readily available and that is why Bosch had to go to such extreme in their design. It was to protect the Darlington stage that itself became a necessary design element as the only high-gain switch-like device available at the time.

Keep in mind the 3.2 Motronic drives 6! low-impedance injectors wired in parallel in batch mode. That isn't trivial even by todays standards.

Ingo

Great info - thanks for sharing...

It looks like the 0127 was used in all ML1.x boxes too. These are cheap to buy at ebay.

Another point is that this IC seems only to fail in 911 3.2 boxes. Never saw any Porsche 944 or BMW M3 ML 3.1 box with this problem...

The IC really only fails when the tachometer gets hooked up wrongly (mostly in 3.2conversions).

The fuel consumption output of the DME comes straight out of the IC. It is the injector pulse width signal. And the tachometer input of a 3.2 tach has an internal pull-up resistor to drive the open-collector tachometer output of the DME. If the fuel consumption line gets hooked up to the wrong tachometer input with the pull-up resistor it drives the IC to full open and kills it in the process.

Ingo

Hello, hoping to revive this thread to resolve a fueling issue on my 88' 924s. The ground side of the injector circuit (pins 14 & 15 @ DME connector) is getting about 10.5v when the DME is supplied power. The harness is good, and this problem still persists when I power the DME on a workbench. I tested the rbdt65a transistor and it seems to be functioning. Pins 14 & 15 seem to be giving the same exact voltage coming from pin 10 of the 0127 IC. Is the IC supposed to be supplying voltage here? Pin 10 on the IC also branches off to a couple capacitors, so I'm thinking either one of those have failed, or its the 0127 IC. I can find no info on this damned chip, so if anyone knows what the function of pin 10 is that would help immensely, thank you.

I use to do some repairs on BMW computers. it got to where I just did not have the time to keep it up.
I use to rape other DME's to fix others.

I still have some, not sure if any parts are the same.

First off the injector output of the DME is a current sink. As such measuring the voltage of these pins without a load connected has little meaning.

In other words you need to connect a load (the injectors, or a resistor with the other side hooked up to +12V). When the DME wants to turn the injector on it will turn on the transistor T402 through the driver IC. The IC is a custom Bosch peak&hold driver. It measures the current through the load and switches to a hold mode after a set time has elapsed. This is to drive the low-impedance injectors wired in parallel on the 924. Pin 10 in the current-sense input and tells the IC how much current is going through the injectors. This information is used to regulate the current in the hold phase of the injector cycle.

What is the issue you are having? Are the injectors always on and dumping fuel or are they not working at all? If none of the above I'd look elsewhere.

Ingo

I would guess that Pin 10 is one of the inputs to a differential amplifier that measures the injector current across R409 along with R404, C405, R405 and C407

+1 on not measuring the floating input

This is my guess at how it works. They went to a lot of trouble to rapidly close the injector discharging through the 47 Volt zenner. The design was probably dropped as being over engineered and too costly

Thanks for the quick responses. They are dumping fuel, the car won't start. It has been sitting for about 13 years, and fueling is my last issue. If back the adjustment screw on the airflow meter all the way lean, the car putters and tries to start.

OK, so I'm not entirely sure what I did, but it's running again. I tried to clean up the boards on the DME after probing around, as there were a couple rusty spots, and slapped it back together. I'm not complaining though lol, thanks again for the quick responses.

In the UK, you can pick up Motronic units from cheaper GM cars for very little money. E.g. A 'Vauxhall CalIbra' has a near identical ECU. There was a guy over here who had one running on a 3.2 Carrera. it ran terribly with the wrong maps but you can bet the chips are the same.

Really?

Yep, have a read:

ECU woes, the continuing tale of the poorly 3.2 - Ignition - Impact Bumpers

From what I understand, the s400 0127 IC is exclusive to Porsche for control of fueling, hence why it would run poorly to use the DME from another manufacturer. We have a ton of old BMW 3 series floating around parts yards here that use a similar DME

I guess, if you had an EPROM flasher you could switch it over

I bet the injectors were stuck from sitting so long.

Hello Pelicanites!
I'm pulling up this thread as it describes to a large extent what i'm observing and I therefore would like to tap into your collective expertise regarding the Bosch ML3.1 of my BMW M3 E30 to repair its Motronic, which I damaged due to a short circuit while measuring with an oscilloscope, whilst the engine was generally running reliably apart from lean mixture at partial load, which had a mechanical cause in the form of leaks in the intake tract as I found out afterwards...

Anyway, since the short circuit, the ignition spark is absent; the fuel injectors inject, but possibly continuously, as the engine is flooded after unusually few revolutions.
Engine runs fine with another motronic when I swap to that immediately after, it just requires full throttle to clear the unburnt fuel.

In an initial repair attempt, I replaced T505 with a BC548C, as I thought I noticed visible damage. However, that didn't help and the replaced tested fine out of circuit.

In the meantime, I found the circuit diagrams on here, checked the diodes and transistors (using the diode check on the DMM), and all showed a forward voltage drop of 0.5 to 0.6V with no short circuit.
However, that is not to say I have overlooked something...

I also replaced the 12V to 5V voltage regulator and restored the trace to the 12V supply with a wire, as both were broken. Unfortunately, that hasn't resolved the issue either.
I don't see any further damage and I've re-soldered the usual suspects, meaning the semiconductors in TO-220 and TO-3 packages prior to this mishap and it has been reliable up until my shorting it.

With the DME removed from the car and 12V applied to pins 18 and 35, and ground to 17 and 19, I measure at pin:
1: +0.6V
9: +4.9V
14, 15: +0.6V rising over several minutes to +10.4V
20: +1.3V rising to +3.5V
all relative to ground.
Only pin 9 of these shows correct voltage.

How can I track down the defective part? I have a single-channel signal generator to emulate a speed signal (the S14 engine has 116 teeth on the starter gear, different from the 911) but unfortunately, I currently have no way to emulate the signal from the reference mark sensor.
Is it true that after the first or only recognition of the reference mark, the Motronic could fundamentally do without this signal, as it calculates the ignition timing based on the pulses in the speed signal?

Thank you in advance for your support!

P.S. As a first contribution in this first post of mine w.r.t. the Opel/vauxhall motronic 0261200376 mentioned above in post #28 is NOT an alternative to the ML3.1 as it has 55 pins and therefore is not a ML3.1 but most probably a M1.3 using a 60-2 crank wheel.
I reckon the motronic referenced to in post 28 had only the lid of that Opel/Vauxhall...

Can you post a picture of the unit in question and explain what you shorted?

Unfortunately I cannot remember exactly were/what I shorted as it was about 2 years ago but it was on the PCB to which the 35 pin socket is attached.

http://forums.pelicanparts.com/uploa...1750546128.jpg
http://forums.pelicanparts.com/uploa...1750546030.jpg
http://forums.pelicanparts.com/uploa...1750546078.jpg
http://forums.pelicanparts.com/uploa...1750546100.jpg

The pads and tracks for the dc/dc convertor Unfortunately didn't survive the replacement and were substituted by the grey wires.
New spring clamp rivets will be fitted or I'll replace therm with screws and nuts.

The yellow arrow marks the spot where the +12V track burnt away.
http://forums.pelicanparts.com/uploa...1750547316.jpg

Ah - that is very close to the 3.2 Porsche DME. Same Bosch 0127 peak&hold driver scheme for the injectors, likely with a different current sense resistor. Same flywheel decoder and ICV driver and same ignition driver. Just different inputs into the ADC because of the water-cooling and a different chip.

Does the DME show brain activity when you turn the ignition to RUN in the car. The ICV needs to vibrate. If that’s a yes the next step is to find out if the flywheel sensors put the DME into run mode. Check for fuel pulses when cranking or are the fuel injectors wide-open?

For bench testing purposes you need both the speed and reference signals to simulate the flywheel turning and then use an oscilloscope to confirm the 8051 MCU puts out spark and fuel pulses. After that it’s troubleshooting the analog portion to see if the fuel and spark signals are property generated to the outputs.

Thank you for your quick response, Ingo!
I was hoping that amongst others you would chime in 😀.

I am not where the car is but I can emulate RUN on the bench easily, however only without connecting an ICV, unfortunately.
Attached the scope views of the outputs to pin 33 and 34 and for reference, to ground.

http://forums.pelicanparts.com/uploa...1750550394.jpg
http://forums.pelicanparts.com/uploa...1750550413.jpg
http://forums.pelicanparts.com/uploa...1750550268.jpg

I expect no buzzing ICV when less than 10mV is put out to it.

The ICV outputs are current sinks and measuring a voltage has little meaning unless you had a pullup resistor to Vbat hooked up to both outputs.

However, the fact that you see a signal with a frequency around 100 Hz might indicate that the MCU is generating a signal train. If that is the case it'll confirm the MCU is alive and running code. The ICV signal also serves the watchdog and makes sure the MCU doesn't get reset. Check the reset pin to confirm it is steady.

Next, once the flywheel decoder sees the flywheel signals it will transition into run mode where the MCU will issue pulses to the ignition and fuel output stages. That is the next step to confirm

This BMW DME ECM is basically identical to the Porsche 911 3.2, with exception of the EPROM (now a junk "performance" chip).
The way to begin troubleshooting, is to monitor the data & address I/Os of the microprocessor (uP).
You need to check pin 9 (reset) of the uP for ground, and all the data/address lines of the 8K EPROM.
The 8 bit latch circuit (S702) is critical for ADC & EPROM data to the uP.

It's simple to generate the speed & reference signal by using a sine wave generator & a pulse generator to initially to develop TDC.
If the S105 chip, RPM/TDC discriminator, is damaged, there are none available unless you find another DME ECM.
This also the case for 0127 (injector IC). All the other ICs, e.g. same ADC as the 911's, are available.

Don't waste efforts on the analog board until the uP & support ICs are running. You obvious need to have a good +5V supply
to the digital board. The analog board is simple to troubleshoot!

Thanks for chiming in too, Loren.

I have solid 5V power supply to the digital board.
I used the 1 channel function generator as input to s100 first to pin 4 and 7 and then to 2 and 3. Outputs from 8 and 10 respectively looked different from the input signals and from that I guess that s100 is not damaged.
Am I right in thinking so?

Pin 9 on the uC is not steady to ground but keeps showing a square wave going high. I take it that means it keeps resetting due to not seeing both rpm and reference signal simultaneously.

I will only able to continue work on this after I get a pulse generator, which will be next week at the earliest and will then continue checking the digital board as well as putting it in the car to check icv buzzing as well as check if the injectors are pulsing during start.

I'll update as soon as I've got new results.

Thanks again for your input guys!
Bernie

If pin 9 of the MCU keeps getting reset signals the watchdog circuit isn’t served by the ICV signal. That has nothing to do with flywheel pulses and needs to get fixed before attempting to put the MCU into run mode to generate fuel and spark signals.

Thanks for the pointer. I'll remember to check that first, once with the car.