E38 Documentation attempt - Starting from the DTC functions.

[Xnke]

Gave Claude a new BIN it had never seen, told it to apply all the tricks and sneaks I taught it to use with 4088, and gave it the BIN and XDF for 12619078. Took about an hour to bring it up to the same level as the 4088, so not too bad!

[kur4o]

If you want to boost, the better approach will be to switch to e67, Almost same as e38, but more input/outputs, and have native boost support + it is used on v6 engines. Code wise there will be no big differences. Feeding AI with some e67 can give you an idea how close these are.

[Xnke]

You are correct, and I will eventually do that. The reason I didn't start there is that I didn't have one in the beginning, and they are not cheap at all, due to everyone wanting them for swaps in VPW based vehicles.

(One of the reasons I need one too-the 05 Colorado is a VPW comms vehicle!)

Made good progress on the 9078 auto BIN today, found a discrepancy that I can't quite decipher yet with TPU-A channel 11 and the "EV FAN SPEED" tags. I'll get into it more after dinner. Built a test harness with an ESP32 to feed real engine data in and a second ESP32 to handle canbus data feeds. I have a sack of salvaged ESP12 and ESP32 modules i removed from E-Scrap during covid

[Xnke]

Roughly 23% through the 2008 "9078" codebase now, using 11 different BIN files for the same OS. I started building a new tool specifically to manage all the fueling and ignition tables, although I don't think I'll be including flashing the OS, I am going to try to get data logging and data tracing working.

Right now, the baby tool opens, has the user select BIN and XDF, and doesn't crash if you ask it do anything weird. The XDF tables open up with an information frame along the bottom of the main window, and a table frame above that, and scalars and labels just point you directly to the hex address and give you the raw bytes to look at.

[Xnke]

Man, there is so much conflicting information about the OS in these PCMs.

9078 has a full speed-density system (like 4-5 tables) that has no calls from or to the torque management system, that is just turned off. If I could locate all the other tables that would interfere with it, I could just turn this system on, route all the fueling calculations through existing functions, and get down to 5 tables: Main VE, Accel Enrich, Temp Compensation, Load Compensation, and a Derate function. The original MAP calibration data doesn't seem to match up with the calibration coefficients published for any of the GM map sensors, and it doesn't match up with the HPTuners published info for the MAP sensors either. It's really close to the 2Bar MAP sensor though, just the stock "9078" OS, 2008 Sierra 2500HD 6.0L/4L80e calibration.

Do any of you want to have a look at the BIN tuner that I'm building, give me some input on where I've labeled things badly or allowed Claude to make assumptions that are wrong?

(Claude is still dead set that this PCM does not use a big polynomial fueling calc, and the evidence is pretty strong, but then again...everyone else seems to say that it does.)

[kur4o]

Do you know the addresses of these tables, I can take a look at the utility you are working on. MIght help making it better.

[MPC001]

9078 has a full speed-density system (like 4-5 tables) that has no calls from or to the torque management system, that is just turned off.

Do any of you want to have a look at the BIN tuner that I'm building, give me some input on where I've labeled things badly or allowed Claude to make assumptions that are wrong?

(Claude is still dead set that this PCM does not use a big polynomial fueling calc, and the evidence is pretty strong, but then again...everyone else seems to say that it does.)

There are some 17 x 17 "throttle VE" tables in most/all E38's that are full of zeroes. No idea what gets them to function in lieu of the VVE setup. Populating them has no effect on engine operation that I can recall. Have seen it suggested they could be used for alpha-N operation.

Maybe that is what Mr Claude is referring to. Never seen a 9078 without VVE being setup. Especially a VVT engine like the 08 Sierra.

Happy to help out & make some time to look at your bin tuner .

[MPC001]

There are some 17 x 17 "throttle VE" tables in most/all E38's that are full of zeroes.

4 "Throttle VE" tables that I can ID in 12619078:

1e974c
1e9bd0
1ea054
1ea296

Axis:
RPM & TPS.
Seems to be more than 4.

[Xnke]

I'm still learning how to get the Bin Tuner packed up and available to share, haven't had a lot of time to play with it this week. Teaching kids how to use machine tools without machining off their own fingers can be rough at times.

Another oddball thing that Mr. Claude cannot get through their head - Apparently the MAF input is also tied up with "EV Fan Speed" somehow? Everytime I go through the MAF input chain to detail out more of it, the AI always tries to pull in that EV fan speed direction and I have to push back against it.

I'm at work right now and don't have the exact table positions to hand, but I'll put some time in tonight to share more information.

[Xnke]

Ok. Keep in mind, this is NOT TO BE USED ON A LIVE BIN.

You are welcome to play with it but absolutely do not expect it to do anything correctly. Be pleasantly surprised that it functions at all.

There is a 9078 BIN and the current XDF packaged with the Zip. The program will grab these bins as the default, but you can open any BIN and XDF, no guarantees that it won't scramble whatever file you open.

There are some tools to play with injector sizing, engine displacement, and number of cylinders. It should unzip and run on any windows system, and everything is packaged together, no other files are needed.