Modifying Motronic 2.3.2 ECU hardware and software.

[prj]

Yes Rs2 551c boost chip...

The analog input of the tps is put into 0x5d internal RAM and used in many routines (boost control, overboost calculation etc.)

I dont't have much knowledge about the boost chip at the moment. I can tell you from the code which map is read. But i don't know anything about the axes. Just that rpm and tps are used as axes.


Bye

The way it works is, it takes raw ADC and if there is "overboost" it just simply adds a configurable fixed value to the raw ADC value.

So for example your overboost % is 20%.
Let's say throttle is at 60%. With overboost, the total value will read 60%+20% with overboost active.
Overboost activity depends on a number of configurable parameters. IIRC the time counter was rather long at 30 seconds or so by default (it's been a while since I looked at it), so technically the car is always operating in overboost mode. Except one case. When you are flooring it top speed on the Autobahn.
And that's why the overboost mode is there - to reduce the engine heat when on autobahn long WOT runs and make the engine "Vollgasfest".

At WOT, the overboost will push the TPS value over 100%, hence why the last row is also 117% in the maps.
If you want to disable the overboost functionality at WOT you can just make the last two rows read the same in the target boost and N75 base DC maps.

On the RS2 for example, there is no added midrange boost, but after 30s or so of WOT, the boost is lowered at high RPM WOT scenario a little.
Again, this is to prevent overheating of the engine and associated components.

This is all from memory, I have no tuning related material on this PC. Sorry if I made some errors with the exact numbers.

Someday I will write all this up in a pdf... Someday.

[vincentroyc]

Hey PRJ how can i map the inputs of both chip in the assembly? For example where is the boost voltaGe /input is in the boost chip? I guess they would me in the special function register space where there are some constant defined for ram? Can you guide me a little in this?

[prj]

You take the maps that are related to boost cut or similar, find which variable is compared and backtrace the ADC read from that variable.

[vincentroyc]

Ok here are my questions :
1. Are the maps always accessed throught the address of the firt value then the program interpolate to the good address/value from the converted value of the input?
2. What is the process to read an input? Is the input read throught a specific adress than converted with the adc? Or all inputs are read directly throught the adc?

Thanks

[kymi]

Not sure if you're talking about boost or fuel/ign chip. But these answers will do on both chips.
1. Not sure if I understood you correctly. It depends on the map how the result is used. Some of the values are compared to ram values, some values are directly used for something, some values are used with adaptation, etc.
2. It depends on the input or the data. Analog inputs are not the only way feeding the data to chips. Analogs are easy to find, but you just need to find where those are used. You should check interrupt routines, timers, etc. Internal & external ram, what feeds what.

[prj]

I recommend reading the 80C515/535 user manual before continuing to ask questions which are answered therein.

[vincentroyc]

Not sure if you're talking about boost or fuel/ign chip. But these answers will do on both chips.
1. Not sure if I understood you correctly. It depends on the map how the result is used. Some of the values are compared to ram values, some values are directly used for something, some values are used with adaptation, etc.
2. It depends on the input or the data. Analog inputs are not the only way feeding the data to chips. Analogs are easy to find, but you just need to find where those are used. You should check interrupt routines, timers, etc. Internal & external ram, what feeds what.

Hey kymi,

For the first question i was more asking : how does the maps are accessed, i mean are they always refered to with a MOV DPTR, #first_map_value_adress and then based on the values of each axis input it calculate the offset of the wanted value in A and then get the good value with the MOVC A, @A+DPTR?

New question : are the analog input accessed like this :
MOV ADCON1, #channel_number
; and when channel ready
MOV SOME_REGISTER, ADDAT

[kymi]

There's map lookup routines on both chips. This thread has more than enough info how to find those and those are pretty straight forward.

Analog inputs are accessed like that on both chips, with ADCON0.

[vwnut8392]

i know i'll probably hear crickets but i was thinking about ignition map switching again. i already discovered that when grounding pin 5 of the coding connector it will switch over to a different ignition map than normal. what i was wondering is if there is a way to switch to other maps via this wire and resistor or another wire in the coding connector? it just seems there's a who lot of ignition maps and most of them are not really used for anything.

[kymi]

Sorry, I just have to ask. What are you going to do with multiple manually switchable ignition maps? I assume you have found 2 usable maps at the moment, one for your "launch control" and one for normal operation. Isn't it enough? And when switching ignition maps, shouldn't you be changing the fueling also?

[vwnut8392]

Sorry, I just have to ask. What are you going to do with multiple manually switchable ignition maps? I assume you have found 2 usable maps at the moment, one for your "launch control" and one for normal operation. Isn't it enough? And when switching ignition maps, shouldn't you be changing the fueling also?

working on an idea i have and i want to see if it will work thats all. if i can switch to another map im going to use it for a no lift shift in conjunction with the WOT BOX. i will have full ignition cut with the WOT BOX and timing retard if its needed with another map thats flattened at a certain RPM like i did with the launch control. for me this is a bit simpler than trying to understand all the madness going on in the ECU directly.

[kymi]

Ok, if that's the case. Why don't you use that same map for NLS? When launching you have revs limited lower and with NLS you're at normal rev limiter.

PS. It's not madness going on in the ECU.

[vwnut8392]

Ok, if that's the case. Why don't you use that same map for NLS? When launching you have revs limited lower and with NLS you're at normal rev limiter.

PS. It's not madness going on in the ECU.

at the current point im not using the stock clutch switch to enable/disable the launch control, im using the set button for cruise control to enable and disable it so i have the option for rolling launch control. the issue with the stock clutch switch is that it works backwards for starters so when the pedal is up the circuit is closed and when the pedal is down the circuit is open so im thinking i have to use a relay to flip the switching. as for the inputs between the ECU and the WOT BOX the ECU uses a ground to switch and the WOT BOX uses 12v to enable/disable. so i have the set button wired to complete the ground circuit to the ECU and a relay that switches on letting 12V to the signal wire on the WOT BOX at the same time.

now to incorporating the NLS into this is where its getting tricky because of how i have the LC wired in. if i just tie the clutch switch in to the circuit i'd have to be holding the set button in to enable either or feature and id lose rolling launch. so ultimately my idea was to add in another whole independent NLS circuit to start and make sure it was going to work right. i wanted to use the on/off switch on the cruise to to enable NLS but it turns out that the set button and the on/off switch are tied together and do not work independently from each other so i ruled that out as an option. over all i wanted to have an on/off switch wired in parallel to the clutch switch to make or break the circuit for daily driving. as for the flipping of the clutch switch i havnt figured that out yet but i know i can invert the signal in the WOT BOX so it would play nice with the clutch switch as is but for the ECU not so much.

all this relay nonsense is why i wanted to just add another independent circuit to start. i've been sick and sitting around over analyzing this whole thing really, i just need to get healthy and take a fresh look at this again from zero really.

The ECU may not be madness to you guys who do programming for a living but for us non programmers its japanese. trust me, i have tried to understand it but i just cant get my brain around it thus why im doing things like this to achieve the same things as you guys would do when patching custom code into the ECU. im just working with what i understand. if there was more real concrete documentation out there im sure i could understand the ECU better but there isnt so that leaves me out in the cold. i do understand that this ECU is an odd one compared to the rest of the motronics developed at its time thus the lack of documentation on it. i hope someday someone comes forward with a self published document atleast on the function of the ECU because it would be great to see more things happening like in the ME7 and up world because i know there's tons of potential in this ECU but it will never be unlocked or flourish with only a few people that have the knowledge of it.

[kymi]

I think you're a bit overcomplicating your needs. If you're unable to build the logic inside of the ecu, why not use some tiny external chip to suit your needs? For example with arduino nano you can easily create map switcher. When certain conditions (resistors, switches, speed, etc) are met it will do the map switching, without huge array of relays. And you can fit that "match box" easily behind the dash.

I have never done a single day of programming for living if you don't count some html back in the 90s. I see this ecu more as a programmed math device. After you know what goes in and what goes out, all you have left is pure math functions and comparing the values. And that's the way you can inject some code. For example really simple launch control: if speed=0, rev limit=x, ign map=y, fuel map=z.

Sorry, this is starting to go a little bit offtopic as we're not speaking modifications underneath the ecu cover.

[vwnut8392]

I think you're a bit overcomplicating your needs. If you're unable to build the logic inside of the ecu, why not use some tiny external chip to suit your needs? For example with arduino nano you can easily create map switcher. When certain conditions (resistors, switches, speed, etc) are met it will do the map switching, without huge array of relays. And you can fit that "match box" easily behind the dash.

I have never done a single day of programming for living if you don't count some html back in the 90s. I see this ecu more as a programmed math device. After you know what goes in and what goes out, all you have left is pure math functions and comparing the values. And that's the way you can inject some code. For example really simple launch control: if speed=0, rev limit=x, ign map=y, fuel map=z.

Sorry, this is starting to go a little bit offtopic as we're not speaking modifications underneath the ecu cover.

i understand what your getting at. as for the ECU i understand what goes in and what goes out as for how and where it goes in and out is one of the problems in front of me. once the signals from sensors hit the main montronic connector thats the end of it for me as there is no schematic for the ECU. when i started tracing the board with high resolution pictures i was scolded and called stupid for doing that. for the math part well its never been my strong suit in general. as for off topic i started this thread a long time ago and i think we're still talking about the ECU just not on a super technical level like the yourself and others have been.

its still on my mind to just go buy another VEMS ECU for my S6 and be done with it. when i started the stock ECU adventure with my swapped 4000 quattro i figured the community would be more friendly and open about stock ECU tuning but i was waaayy off on that. so they lead me to just go buy VEMS for that car because i needed to upgrade injectors and i have no idea where the injector scalar is in the stock ECU let alone what the variable is that the injector scalar is based on. now that im on my last spare MAF on my S6 i fear finding another replacement later so migrating to VEMS for speed density and ease of being able to tune the car freely.

i was told PRJ was working on some sort of daughter board for the stock ECU that would allow a form of open source tuning but im not sure what happened to that project because i did have an interest in having one of those on my S6 vs VEMS. i know the old motronics are just as capable of not more capable of great things like old DSM, honda, etc ECU's but it totally baffles me as to how the old motronic ECU's never evolved like the japanese ECU's did in the tuning department. i have a friend with a 1G talon that's putting down 856whp to all 4 paws on a stock ECU with DSMlink V3. thats pretty impressive to me for a stock ECU thats modified.