Modifying Motronic 2.3.2 ECU hardware and software.

[kymi]

There should be no problem understanding the 3d interpolation if you know 2d interpolation. When you look at the "read_3d_map_rpm" function, "read_map_value" interpolates every read and then those two map values will be interpolated. Things are stored in temporary ram_4x locations. Interpolate function returns the same accumulator it always does. No knowledge, no motronic, just pure code and math.

[vincentroyc]

There should be no problem understanding the 3d interpolation if you know 2d interpolation. When you look at the "read_3d_map_rpm" function, "read_map_value" interpolates every read and then those two map values will be interpolated. Things are stored in temporary ram_4x locations. Interpolate function returns the same accumulator it always does. No knowledge, no motronic, just pure code and math.

Ok so the main fucntion the reads 3D maps by passing in the map lookup table address to find the right map alsways return the final value via the A register? So after every lcall of this function A should be the final value interpolated right?

Also, i think that RAM_67 may be Raw ECT value and RAM_68 the IAT?

How are the knock sensor read. ViA ADC? Or is it a interrupt generated multiple time depending on knock intensivity?

Thanks

[1gcrazy]

Just wanted to revive this. I was following along when VWNut was doing this a while ago. Looking to catch up but see there hasn't been much activity. Are there any updates?
I'm a bit more advanced in my ECU knowledge than before so hoping to understand a bit more this time around.

[kymi]

Should there be any updates? If you need updates on this thread just ask if you have something on your mind.

[curamrdan]

I will try to ask. Can anyone help me with conversion dwell table to msec and correct shape of it?

So far i have this , raw values with correct axis. It doesnt make me a sense if higher value means lower dwell , then its not correct to have smaller dwell at lower voltage.
If higher values mean higher dwell, then doesnt feel correct to have higher dwell with higher rpm. These are my toughts if i compare some dwell tables of other newer or older motronics.

Thanks.

[blackS2]

I will try to ask. Can anyone help me with conversion dwell table to msec and correct shape of it?

So far i have this , raw values with correct axis. It doesnt make me a sense if higher value means lower dwell , then its not correct to have smaller dwell at lower voltage.
If higher values mean higher dwell, then doesnt feel correct to have higher dwell with higher rpm. These are my toughts if i compare some dwell tables of other newer or older motronics.

Thanks.

This map is more complex than ignition or fuel maps. Its not simply converted by a constant. I could post the entire formula but it spent to much time on my own to solve this. It uses also rpm to convert the values, so you get different dwell times for the map value 157 at different rpm for example.

[kymi]

Dwell table values are based on the engine/crank angle. Higher rpm needs more degrees to achieve the same dwell time.

[curamrdan]

Intresting. Thanks guys.

[blackS2]

Dwell table values are based on the engine/crank angle. Higher rpm needs more degrees to achieve the same dwell time.

[curamrdan]

Can somebody show correct shaped request boost + wgdc boost maps ? Doesnt matter which software.
I dont know WGDC value conversion, just comparing boost and wgdc maps side to side -> more dc = more boost.
The AA softare make sense, but RS2 no. C software have reguested boost(16x10 table) last 5 rows same boost(just first left - highest? rpm collumn is different) but wgdc has same only last 2 rows, and last collum with dropped boost is not dropped in wgdc. B software has same wgdc table as C, but request is different.
Maybe axis are not same?

[filip.krticka]

Can somebody show correct shaped request boost + wgdc boost maps ? Doesnt matter which software.
I dont know WGDC value conversion, just comparing boost and wgdc maps side to side -> more dc = more boost.
The AA softare make sense, but RS2 no. C software have reguested boost(16x10 table) last 5 rows same boost(just first left - highest? rpm collumn is different) but wgdc has same only last 2 rows, and last collum with dropped boost is not dropped in wgdc. B software has same wgdc table as C, but request is different.
Maybe axis are not same?

With Vems, MAxxECU you can download free software and maps are widely available.

[curamrdan]

With Vems, MAxxECU you can download free software and maps are widely available.

Im asking concrete on factory ecu to confirm my findings or find what i assuming wrong.
No basics how boost tables may look like in general.

[curamrdan]

Made two pictures, to better understand. 551A you can see WGDC coresponding to Request. Not as RS C software.
Tables are in raw dec values(we can roughly estimate request for 250/300kpa sensor, but its not the case now).

Axis are the same for both tables? And are the same in both softwares?

[kymi]

It seems to me that you haven't done much or any disassembly? If you know how those maps are read, you have at least the raw values of the axis. There are now too much assumptions and even the easiest things missing.

I checked a few binaries and they all have identical axis maps in the binary.

[curamrdan]

Thanks. I have lack of knowledge to dissasembly by myself alone. So im decided to shoot in the dark sometimes, realtime testing by emulator. But i dont have latch board only external ziff socket, so i would be happy for any more info or will end to use external/mechanical boost solution.

Do not care exact axis numbers, now.
Are the values in wgdc table real duty cycle numbers? Is so, why it not correspond to request...if axis are the same?