NefMoto

Hi,

Im planning on to take an Me7.5 ECU and desolder the processor and replace it with a shield to connect complete diffrent controller that i have. Is it some how possible to debug which pin on the ECU connector that goes to the corresponding I/O on the processor...  For example i want to know which output from the processor controls the injector #1 for example..  The idea is to use the ECU as mainboard as I/O..  Is this possible or is it to hard to do sort this out?
 ;D

You may just have to hand trace each PCB track from the output pin on the harness connector, back to the MCU. However, bear in mind the MCU will not drive the pin directly; it will go through some output driver IC (transistor, thyristor etc.) Another pitfall is that some PCBs are multi-layer, so whilst you may trace the top and bottom tracks, it's much more difficult to trace the ones that are sandwiched in between.

Good luck.

To give you an approximate idea of what you're looking at: http://nefariousmotorsports.com/forum/index.php?topic=67.0title=

If you can't figure this out based on the data sheet of the existing cpu i don't understand what you are asking for.

In general, though, the io pins are the least of your worries. Swapping a non pin compatible cpu on an existing board is a fools errand.  It would be easier to write your own code from scratch for the proper ecu than do that level of trace modification. And even if it was possible, the mess of wires you'd end up with would be completely unreliable.

How do you intend to wire the rest of the pins that are not the same? Do you even have the correct clocks and voltages available?

Yes this was my first concern..  I had an idea to connect an LED to 12v and connect it to an injector pin. Then with a piece of wire connected to ground go by hand and ground the different pins on the processor to se if I get a hit. This is my understanding of how the processor it self works.  Would this be a method of sort of "mapping the pins" ?



Nyet i already have a board that has its own circuit to drive an CPU with its own code and is tunable via online interface. Now I want to try and se what this can do.

Thanks nubcake that was something very intresssting..  I also have to learn how map other ECU´s aswell..   This could be awsome if it works.. :)

Injectors will not be driven off the CPU directly - they draw too much current.

How do you plan on grounding out that single injector signal?  IE: how do you stop other sensors/injectors also providing a ground source?

My plan is to desolder the whole Processor... I have alot of ECU´s i can try and error..   My test plan would be to connect a lamp/LED to 12V and connect it to an INJ pin on the connector then eiter send 3.3v via the pad or ground... Im not sure if the processor grounds or sends 3.3v  PnP or NpN?

The controller would be an arduino based Speeduino.. This is how the speeduino I/O looks like.
Heres how its tuned aswell. https://www.youtube.com/watch?v=Q475JN-t63E.

I have tested this and made a DIY mainboard.. but it would be sick to use existing ECU board and harness.. Mainly to use with older engines as a very cheap alternative to standalone ECU,.

Pin #   Function
1   Injector 1 - Pin 1/2
2   Injector 2 - Pin 1/2
3   Injector 3 - Pin 1/2
4   Injector 3 - Pin 2/2
5   Injector 4 - Pin 1/2
6   Injector 4 - Pin 2/2
7   Ignition 1
8   Ignition 4
9   Ground
10   Ground
11   MAP Sensor (0v-5v)
12   Ground
13   5v
14   Fuel Pump Relay (Gnd)
15   Thermo Fan Relay (Gnd)
16   Low current output 3
17   Low current output 4
18   Low current output 5
19   Coolant (CLT)
20   Inlet Air Temp (IAT)
21   O2 Sensor
22   TPS input
23   Ground
24   Cam Input / VR2+
25   Crank Input / VR1+
26   VR2- (Not used for hall sensor)
27   VR1- (Not used for hall sensor)
28   5v
29   Idle Stepper 2B
30   Idle Stepper 2A
31   Idle Stepper 1A
32   Idle Stepper 1B
33   Ignition 3
34   Ignition 2
35   Boost
36   Idle 2 (For use with 3 wire idle valves)
37   PWM Idle
38   VVT
39   Injector 2 - Pin 2/2
40   Injector 1 - Pin 2/2

Don't you think it Will be easier to cut the pcb off And solder speeduinio/megasquirt pcb to ecu connector?  If you want to use different processor on me7 pcb you have to define everything on the stock pcb And make it work with new processor.  Way too much work for nothing. For 150/250 euro you buy ms2 based pcb Inc ign And inj mosfets etc.


Verzonden vanaf mijn iPhone met Tapatalk

http://www.k-data.org/kdfi-pnp-vw-agu-18t.html

Or buy this And call it a day

Well that would be to easy :)   


Atleast i made a try :)

Are you referring to a ground loop?
I think OP wants to replace the processor alone. So he can still use the injector driver(bosch ic 30344) There is a spec sheet of this chip posted on this forum, it communcates using its own serial interface, if I remember correctly.

@OP- why change the processor? If you're writing your own code on a different micro, why not use the c167 and write the code from scratch?

I don't understand why you take this so seriously, OP could not hack his way out of a wet paper bag, much less write any code.
Also, for ME7, at least the ST versions, all the ECU schematics have leaked, so they can be used as a reference, making this thread a waste of time.

Yes some sort of ground loop.
Use the processor was also an idea but I don't have that deeper knowledge to write a complete new code..  I was thinking to use the already tested and well known Speeduino software as controller. It uses an ARM processor with the code already tested and can be used with TunerStudio. But not sure if it will work.

Yesterday i desoldered the CPU and made a test to see if I could managed to set an INJ out. But didn't manage to get any results. I think the issue was that i didn't have all the needed 12V and GND connected to the connector. This is needed to drive the output drivers (I think). Need a better wiring diagram.

I don't see the point in this at all.  If you're going down that road; just build the Speeduino as it was intended.  Build a custom harness?

Someone has already said that the injector drivers require a serial interface.  You can't just smash 12v into it and expect it to work.

ME7.5.10

Injector drivers serial? Sounds really complicated. Is this only newer ecus?
I have one speeduino done and tested but just wanted to se what i could with a simple arduino and an ecu board. Think i have around 20 ecus lying around here. Looks like they will end up in the trashbin.

Looking at a few driver chips that pop up on Google, they're all SPI... Not saying that's what Bosch used, but most likely. They're not hard to control.

You really need to learn how to program if you want to do anything correctly.

If you are going to make hardware changes, some digital design education is also a must.

LoL read the thread.. It was a test to see if it was possible. I have some programming knowledge but not very deep in embedded systems. Im educated as automation engineer but we didnt learn deep into embedded to bad.. :)

SPI bus im familiar with.. Arduino supports it.. But to much headacke to solve :)

Ironically, control via SPI bus is probably easier than some crazy PWM or current source driver...

You're an automation engineer but didn't learn about embedded systems?

Isn't automation going to be entirely embedded systems?

Well in some cases yes but mostly we work with Siemens/Mitsubishi PLC´s and ABB Robots. It more High Level programming. In school we learned only basics in embedded and after 10years its almost forgotten. I could have chosen the to take the course that was more towards embedded but decided to go for Automation instead. We work with advanced regulation and process control but we never need the knowledge to know whats inside the PLC system. The PLC systems are mostly programmed in Ladder or SFC logic. Some SCADA systems requires C# knowledge. So what we have is knowledge of control systems but not the embedded circuits that do the magic. And the worts part is that we have very few standards to follow.. One factory uses their framework and another their framework. But in the embedded C++ code world it seems like the programmers are more shaped and follows one standard on how to struct things.

Why I like to test thing is to learn and have fun. A friend once said to me that in other parts of the world they have a different approach to take on complicated stuff.. "you leave that to a PRO". It seems like in the U.S you dont try and error you send your car or ECU to the certified Pro and he does the job.. But in my culture in Sweden we just try alot of stuff and se what happens without anyone calling you stupid or wasting time... They just call you crazy and appreciate that people learn and try stuff as long as they do it for their own purpose. I think ideas can come from most crazy projects and people as well. Im not doing this to get a medal or get Paid just for fun :)

Now it seems like the Me7.x.x system is a quite complex system to take on... I have some older systems that might be easier.. For example i have and mercedes M104 engine with harness and an VDO ecu.. Might that be simpler to try with just for fun.