Calling on Motronic experts: IC information

[hardtailer]

I have a known working identical Motronic that I can use as a reference 😀.
That shows the following signals for pin 34. Pin 33 is like the NOK DME.
But not only that, the OK DME does not reset and that is also without an ICV connected to it.






Pin 9 to uC of NOK


ADV11 signals of both:
NOK DME


OK DME



Next, I did a transistor check of T560 and T561, the ON588 npn Darlington transistors in-circuit in both units and got the same results betweenntheb2 units..
Also checked D890 and D891 in-circuit in both units and also those gave the same results each between the 2 units.

Having read that the adv11 signal should be square at ~+3V high, 0V low @~77Hz, I presume that the source for ADV11, is defective but I don't know how to check s704 and s890, let alone s700.

Am I on the right track?

[wazzz]

The picture of Pin 9 to uC of the NOK DME clearly shows that reset pulses are emitted every 36 ms or so. These are produced by a LM139 comparator (1/4 of S890) when no ~80Hz square signal is received from pin 5 of uc S700. This square signal is the same freq. and duty cycle as the ICV signal, but inverted. It has a 0-5V amplitude. So at least you know that this particular quarter of S890 is good, as it does its job by producing the RST pulses. And you don't need to check S704 either, because the lack of square signal is upstream of it.
The lack of square signal from uC pin 5 sounds like the uC is not alive and not waken up by the RST pulses. That doesn't mean that the uC itself is dead, but it sure is lacking basic entries or necessary data to start running.

[ischmitz]

Check clock input for MCU on pin 19 and ALE output to make sure the MCU is at least not completely fried. Next check that it reads and processes code from the EPROM.

[hardtailer]

Originally Posted by wazzz

The picture of Pin 9 to uC of the NOK DME clearly shows that reset pulses are emitted every 36 ms or so. These are produced by a LM139 comparator (1/4 of S890) when no ~80Hz square signal is received from pin 5 of uc S700. This square signal is the same freq. and duty cycle as the ICV signal, but inverted. It has a 0-5V amplitude. So at least you know that this particular quarter of S890 is good, as it does its job by producing the RST pulses. And you don't need to check S704 either, because the lack of square signal is upstream of it.
The lack of square signal from uC pin 5 sounds like the uC is not alive and not waken up by the RST pulses. That doesn't mean that the uC itself is dead, but it sure is lacking basic entries or necessary data to start running.

Merci, wazzz!
Pin 5 does not show the square wave in the NOK DME, whereas in the working DME it does (at f=77Hz).
Based on the frequency being different in the NOK DME, I checked the signals at pin 18 and 19 and compared to working DME.



====


======




The pin comments are correct in all pictures, so I think I'm onto something here! Perhaps one of the capacitors is bad? Or the output by the uC is missing and pointing to a damaged 8051/S700?

[hardtailer]

Originally Posted by ischmitz

Check clock input for MCU on pin 19 and ALE output to make sure the MCU is at least not completely fried. Next check that it reads and processes code from the EPROM.

Also thanks for your reply, Ingo.
See my previous post re pin 19.
ALE/pin 30 see below, as well as some other signals.
Those without a signal I took no picture of.







Do these show expected behaviour?
I will check them against working DME but I'm unsure if I'm able to draw the right conclusion from differences.

Signal on Pin 30 on working DME below:

Pins 31 to 39 all showed data and required 10V/div to be in range of scope.
I seem to recognize the different rows of the maps stored in the EPROM.
Example of pin 38:

[ischmitz]

Hardtailer,

unfortunately your oscilloscope seems to have very limited bandwidth and is sampling to where the signals are aliasing and thus not making a lot of sense. But here are the simple things that need to happen in the following order:

- The MCU needs power (+5V and GND)
- Pin 18 and 19 need the 6MHz clock from the quartz
- Pin 7 RST needs to transition from low to high to bring the MCU out of reset.

With that you need to see the ALE output on pin 30 generating a 1Mhz square wave. If that isn't happening MCU is dead and you need a replacement.

Then the MCU starts reading and processing code from the EPROM and that will generates the 80 Hz square wave ICV signal on pin 5. And that ICV signal serves the watchdog keeping it from resetting the MCU. If the watchdog doesn't see the ICV signal coming from pin 5 it will pull the reset line low for a short time and then release it back to 5V. If that process repeats you are seeing a square wave on pin 7.

The ICV signal can be missing because:

a) MCU is not powered or damaged
b) EPROM is not present or inserted incorrectly
c) Wrong EPROM or corrupted data burned to EPROM
d) The 74LS373 is damaged
e) Another component on the address or data bus causes issues

All this needs to work before you put any flywheel signals into the box

[mysocal911]

Originally Posted by wazzz

The picture of Pin 9 to uC of the NOK DME clearly shows that reset pulses are emitted every 36 ms or so. These are produced by a LM139 comparator (1/4 of S890) when no ~80Hz square signal is received from pin 5 of uc S700. This square signal is the same freq. and duty cycle as the ICV signal, but inverted. It has a 0-5V amplitude. So at least you know that this particular quarter of S890 is good, as it does its job by producing the RST pulses. And you don't need to check S704 either, because the lack of square signal is upstream of it.
The lack of square signal from uC pin 5 sounds like the uC is not alive and not waken up by the RST pulses. That doesn't mean that the uC itself is dead, but it sure is lacking basic entries or necessary data to start running.

Yes, it's pin 9 (reset) & it must be low in the run mode.

[mysocal911]

Originally Posted by ischmitz

Hardtailer,

unfortunately your oscilloscope seems to have very limited bandwidth and is sampling to where the signals are aliasing and thus not making a lot of sense. But here are the simple things that need to happen in the following order:

- The MCU needs power (+5V and GND)
- Pin 18 and 19 need the 6MHz clock from the quartz
- Pin 7 RST needs to transition from low to high to bring the MCU out of reset.

With that you need to see the ALE output on pin 30 generating a 1Mhz square wave. If that isn't happening MCU is dead and you need a replacement.

Then the MCU starts reading and processing code from the EPROM and that will generates the 80 Hz square wave ICV signal on pin 5. And that ICV signal serves the watchdog keeping it from resetting the MCU. If the watchdog doesn't see the ICV signal coming from pin 5 it will pull the reset line low for a short time and then release it back to 5V. If that process repeats you are seeing a square wave on pin 7.

The ICV signal can be missing because:

a) MCU is not powered or damaged
b) EPROM is not present or inserted incorrectly
c) Wrong EPROM or corrupted data burned to EPROM
d) The 74LS373 is damaged
e) Another component on the address or data bus causes issues

All this needs to work before you put any flywheel signals into the box

Wrong MCU, my 8051 datasheet indicates pin 9 & the run mode is a low;

[ischmitz]

Good catch - I am in Taiwan right working with that little company that make the iPhone and Nvidia silicon and not in front of my test bench.

Most modern processors have !RST input (active low) but the 8051 uses RST (active hight) and to reset the processor the line needs to be high for a least two clock cycles and then go low.

[wazzz]

Here is a SPICE simulation of the watchdog circuit which I made just for fun and also to practice LTSPICE. It shows that the RST pulses (blue) begin to be sent after the ICV signal (green) stops being emitted by MCU on pin 5. When normal ICV signal is present, RST stays low.

[mysocal911]

Originally Posted by ischmitz

Good catch - I am in Taiwan right working with that little company that make the iPhone and Nvidia silicon and not in front of my test bench.

Most modern processors have !RST input (active low) but the 8051 uses RST (active hight) and to reset the processor the line needs to be high for a least two clock cycles and then go low.

The OP probably needs the latest 8051 using the 4nm wafer process from TSMC.

[hardtailer]

Thank you, gents.
The scope has indeed way too small bandwidth as it's only 1 MHz...
I will only have access to my 2 (edit: 20) MHz scope next week so till then, work on this pauses.

I must say, I dread the thought of having to replace the 8051 and needing to dive into programming/flashing its replacement...

For sake of completeness, the supply to the microcontroller is a constant 4.76V to pin 40.
The EPROM is a copy of the OEM with only the rev limiter raised, done by myself.

I'll keep you posted and thanks again to all providing input.

[mysocal911]

Originally Posted by hardtailer

Thank you, gents.
The scope has indeed way too small bandwidth as it's only 1 MHz...
I will only have access to my 2MHz scope next week so till then, work on this pauses.

I must say, I dread the thought of having to replace the 8051 and needing to dive into programming/flashing its replacement...

The worst case solution is not that extreme.

Originally Posted by hardtailer

For sake of completeness, the supply to the microcontroller is a constant 4.76V to pin 40.
The EPROM is a copy of the OEM with only the rev limiter raised, done by myself.

I'll keep you posted and thanks again to all providing input.

Another one with access to a "performance" tuning app? Everyone can be a "tuner", just check what's available for a 991,
e.g. dynamic stability mods, paddle shifting changes, and of course DME "performance tuning".

[hardtailer]

Status update from DME in the car:

Engine starts with OK DME and ICV disconnected. Based on this observation and my understanding of wazzz's post with the Spice simulation, my test result shows that even without normal ICV signal present, the reset signal to pin 9 of 8051 still goes to low and the processor starts processing.

Hooked up NOK DME, reconnected ICV and noticed the following:
-Upon ignition on, but not cranking, the fuel pump starts running immediately
This is not correct, obviously. It should do so only after the crank starts turning or key is turned to crank/run starter motor.
-Upon ignition on, but not cranking T504, 1012 in the TO-3 package gets hot within seconds. After 15 seconds it's too hot to keep your finger on it.
-Upon ignition on, but not cranking pin 14 and 15 to the injectors are pulled to ground for the duration of ignition on.
-under above conditions the reset signal is no different from as bench tested earlier.

[ischmitz]

You’re potentially causing more damage putting the bad DME into the car. If the main fuel injector and ignition coil output transistors are driven into saturation or short you will dump fuel into the engine and burn the current-sense resistors on the analog board.

If it was me I’d bench-power the DME and sort out the watchdog before anything else. For that you only need power to the DME.

[wazzz]

Originally Posted by hardtailer

Status update from DME in the car:

Engine starts with OK DME and ICV disconnected. Based on this observation and my understanding of wazzz's post with the Spice simulation, my test result shows that even without normal ICV signal present, the reset signal to pin 9 of 8051 still goes to low and the processor starts processing.

Hooked up NOK DME, reconnected ICV and noticed the following:
-Upon ignition on, but not cranking, the fuel pump starts running immediately
This is not correct, obviously. It should do so only after the crank starts turning or key is turned to crank/run starter motor.
-Upon ignition on, but not cranking T504, 1012 in the TO-3 package gets hot within seconds. After 15 seconds it's too hot to keep your finger on it.
-Upon ignition on, but not cranking pin 14 and 15 to the injectors are pulled to ground for the duration of ignition on.
-under above conditions the reset signal is no different from as bench tested earlier.

ICV signal is still present. Even if you disconnect the ICV unit from the ECU, the ICV driving signal is still output by the MCU on pin 5 and correctly seen by the watchdog. Hence the RST signal is kept low on MCU pin 9 and that allows it to run normally. This is for the OK ECU.

Now for the NOK ECU, as stated by Ingo, you should really first address the absence of ICV driving signal on MCU pin 5, which is the indication that the 8051 MCU is not running. That's why the watchdog circuit (LM139 comparator) keeps sending it RST pulses to wake it up, to no avail.
You don't need to connect the ICV to the ECU for these checks, as the ICV driving signal monitored by the watchdog is upstream of it

[hardtailer]

Thanks again for your replies.

I won't connect it to the car again until it's fixed.

Using my 20 MHz scope I measured the following on the pins of 8051, S700 on the NOK DME, powered on the bench:
Pin 5 4.85 V 33 Hz
Pin 18 0.8 V 33Hz
Pin 19 0.9 V 0 Hz
Pin 30 2.36- 2.40V, square wave at 33Hz
Pin 40 4.8 V
Pin 9 kept resetting at 33 Hz.

I left the working DME with the car, so no data using the 20MHz scope on that.
I take it that the 8051 is damaged and must be replaced?

(Edit: correct bandwidth of scope is 20MHz, added Pin 30 signal)

[ischmitz]

Are you sure you're doing this right? You need +5V power on pin 40 relative to pin 20 (GND). Then confirm you have a strong 5V p-p 6 MHz clock signal on pin 19 and a weaker version of the same signal on pin 18.

Maybe check the good DME to confirm your scope is set up properly to detect the 6 Mhz signal.

[mysocal911]

Originally Posted by hardtailer

Thanks again for your replies.

I won't connect it to the car again until it's fixed.

Using my 40 MHz scope I measured the following on the pins of 8051, S700 on the NOK DME, powered on the bench:
Pin 5 4.85 V 33 Hz
Pin 18 0.8 V 33Hz
Pin 19 0.9 V 0 Hz
Pin 40 4.8 V
Pin 9 kept resetting.

I left the working DME with the car, so no data using the 40MHz scope on that.

I take it that the 8051 is damaged and must be replaced?

Or one of the other ICs connected to it.

[hardtailer]

Originally Posted by ischmitz

Are you sure you're doing this right? You need +5V power on pin 40 relative to pin 20 (GND). Then confirm you have a strong 5V p-p 6 MHz clock signal on pin 19 and a weaker version of the same signal on pin 18.

Maybe check the good DME to confirm your scope is set up properly to detect the 6 Mhz signal.

NOK DME (copy from above) and OK DME bench test values written thereafter
Pin 5 4.85 V 33 Hz OK: 9.2V 50Hz
Pin 18 0.8 V 33Hz OK: 2,4Vpp (0.3-2.7V) 3.3 - 3.8 MHz
Pin 19 0.9 V OK: 1.2V- 1.3 V 16.7Hz
Pin 40 4.8 V OK: 4.95 V DC
Pin 9 kept resetting. OK: 0.435V DC
Adv11 OK 7,3V 50Hz

Scope is a Hameg HM412



Sanity check, 1kHz signal from signal generator

From that I calculate 6.75 waves in 10 div × 1ms/div= 6.75waves in 10ms --> 1000ms/10ms x 6.75=675Hz....
I calculated the frequencies of the DME signals the same way. What am I missing?