Vida CEM swapping

[dikidera]

Yes I just read the code. May I ask how long it took you to get a working SBL? The HCS12/M68K arch is much more....nuanced than (SH2)SH7055 ever was. I mean just the assembly made my skin crawl and I've seen my fair share of assembly.

[al1Volvo]

Oh ... at least 2 month working on it by night when my child is sleeping... That's also why I'm sharing it, making win time to people who need it. I understand ! Have you made a SBL for the SH7055 ?

[dikidera]

I have an official Volvo one for my ECM which a fellow user kindly shared and allowed me to customize my ECM to work the way I want it, I spent more than a year analyzing the Denso ECU to find the relevant maps for which no DAMOS existed since it is a naturally aspirated engine. With this I gained a very strong familiarity of SH2 assembly.

But now after I did what I wanted with the ECM, it is now time for the TCM for which I have nothing and will bench test a custom SBL as soon as I have the TCM in my possession. But I have already identified critically important maps which I will modify as soon as I can read/write.

The goal for TCM is to disable slip lockup(maps identified), disable continuous adaptations(more challenging), enable low speed full TCC lockup(maps identified) and maybe increase shift ramp up speed(not identified)

[al1Volvo]

Very nice project! would it be possible to share the SBL for the DENSO ? I did a similar thing on old Fenix 1B (Renault 2L Turbo ECU that I put into a Clio RSi with a mercedes compressor added to the engine) that was based on 68HC12 MCU, I finally was able to rebuild on paper the complete formula for the timing injection depending on battery voltage correction, air temperature/pressure/water temp etc ... I found over 5000 parameters to edit ... That's why I am familiar with HC12 assembler

[dikidera]

I was given a list of at least 100 variables/parameters so my work was accelerated very very much, even then it took over a year to discover the relevant maps.

I dare not fiddle with more parameters as the formulas, the implementation of physical models implemented is beyond me, it takes not only a solid understanding of fundamental engine theory, but also fluid dynamics to comprehend much of it.

For me, the base injector flow rate was a hard coded constant with many things pre-calculated in the constant(or gcc compiler simply replaced the formula with a constant) so I could never determine it's true origin.

It is super complex at times.

Clio eh, I have a friend whose Clio 2 1.6 K4M has idle misfire, those stepper motors do not want to be calibrated....and the ECU does not detect misfires but are clearly felt...

[al1Volvo]

Yes it is a very complicated calculation, even worth now with de-pollution systems. the misefire could be coming from a lot of things ... That's why it is nice to now of the software works when you want to diagnose an issue !

[dikidera]

I managed to accomplish my goal, 140->170hp due to throttle limiters, fuel richness (12.1 afr in upper RPM range), ignition timing advance by 2 or 3 more degrees for petrol and quite a few more degrees for CNG mode, CNG loves more timing but we do not know MBT.

What I could not extract, it wasn't clicking, was how the CAN IDs were calculated. I discovered almost everything, the functions that handle most CAN controller stuff, but never the specific algorithm for the IDs. I regret this very much.

[al1Volvo]

Nice ! Ah yeah, that could be very nice to now that !

[dikidera]

I have received my Temic 29.1 with part code 30735493. But I am away from home so my research/progress will take a few days to start. I will use fenugrec's kernel with modifications to load into the SH7055S(180 nanometer node).

[dikidera]

I am not sure which modification did it, but I managed to overwrite the Expected Checksum in EEPROM. It either happens automatically if the TCM is in the car, and you power cycle via key in POS II and then POS 0 or doing a read of DTCs and then a clear caused it to get overwritten, I cannot verify as I cannot simulate the ignition cycles on the 'bench'. Thus CalculatedChecksum and ExpectedChecksum are now equal.

At first I tried flipping this flag, but did not observe immediately results, so it also might have played a role, but cannot confirm right now

can1 01200005 [8] CF 6E E5 A7 34 AC A7 A5
can1 01200005 [8] CF 6E E5 A6 61 6D 6E F5
can1 01200005 [8] CC 6E 7F BA 31 00 00 00
can1 01200005 [8] CD 6E FA FF AB 90 00 00
can1 01200005 [8] CF 6E E5 A6 61 6D 6E F5
can1 01200005 [8] CF 6E E5 A6 61 6D 6E F5
can1 01200005 [8] CF 6E E5 A6 61 6D 6E F5
can1 01200005 [8] CF 6E E5 A7 34 AC A7 A5
can1 01200005 [8] 8F 6E EE 11 00 20 00 2A
can1 01200005 [8] 4C 00 2F 00 39 00 00 00
can1 01200005 [8] CB 6E EF 11 00 00 00 00
can1 01200005 [8] CB 6E EE 11 00 00 00 00
can1 01200005 [8] CB 6E EE 11 00 00 00 00
can1 01200005 [8] CB 6E EE 11 00 00 00 00
can1 01200005 [8] CF 6E E5 A7 34 AC A7 A5
can1 01200005 [8] CF 6E E5 A6 34 AC A7 A5
can1 01200005 [8] CF 6E E5 A6 34 AC A7 A5
can1 01200005 [8] CF 6E E5 A7 34 AC A7 A5