carPC steering wheel controls

[vwboy308]

Hey man keep up the good work.

 

[dbduke]

Awesome work man. I'm hoping you can learn us on how to splice this into the factory harness for those of us without sat radio.

 

[wireb]

That bus goes quite a few places. Sat tuner, Premium amp, Blue tooth module, radio, and the data bus on board diagnostic interface. Once I get it to work remind me and I will put together a list with pin numbers and wire colors where you can find it.

 

[wireb]

Update: I think I am back online in perl. Looks like Win32::API is not supported in ANY x64 version of perl yet..... Swapped back to ActivePerl-5.10.1.1006-MSWin32-x86-291086 and its included by default. Yay... I guess.

More in a few hours. Going to test out all the SPI code now

 

[wireb]

Ok had a bad MCP2515 (don't ask) replaced that and now the SPI bus is fully functional.

Also found out the 12V power to the sat module is always on (wanted switched power. Will have to find another location to tap off of to charge my Ipod)

took a SWAG at the clock config and turned the card on and got received message buffer full bits set but also had MERRF and ERRIF set (error bits). They could just be normal (ie one time things when I turned the car on) or a issues with my bit time settings.

Tomorrow I will work on the code to read the buffers and service errors. Hope to have some dumps from the car tomorrow afternoon.

 

[dbduke]

I don't have any idea what any of that means but its awesome.

 

[wireb]

Got some dumps off my car grab them here:
http://www.wire2wire.org/process.zip
open the .csv files in excel or any other spreadsheet program you like.

Getting quite a few errors some I expect (overruns due to the serial link to the pc slow) but others I have a question into microchip on what exactly sets that flag.

Either way long as the messages I did get are valid there should be enough here to figure it out. Once we know which message has the button data in it I can set the filters on the chip to just pick up those (and greatly cut down the # of messages I am getting)

Each file has 20 seconds of bus traffic in it. All were taken with the ignition on but the car not running (unless noted otherwise)

audmute.csv = MFD set to the audio page and me smashing the mute button
audphone.csv = MFD set to the audio page and me smashing the phone button
caroff.csv = car off and key out of the ignition (ie nothing going on)
caron.csv = ignition on but car not running and nothing else happening
menu.csv = smashing the menu button to flip though the MFD pages
mute.csv = MFD on the MFD page and me smashing the mute button
ok.csv = MFD on the MFD page and me smashing the ok button
phone.csv = MFD on the MFD page and me smashing the phone button
radoff.csv = MFD on the MFD page and the radio turned off
radoffphone.csv = MFD on the MFD page and the radio turned off while smashing the phone button
sat.csv = MFD set to audio and radio set to sat and me smashing random buttons
updn.csv = MFD set to MFD and me smashing the up and down menu buttons

Need some help picking though the data and finding any patterns (ie address / data only called in files where I press button xxx). If you find a good candidate let me know and I can update my filters to look just for those messages and test out the theory on my car.

 

[wireb]

Ok got a response from microchip and trimmed the "data" down to just the bytes received and re-uploaded the zip file today (nov2). If you already grabbed the files go get a new copy it will make more sense.

 

[GTI_Fahrenheit_859]

This is looking sweet. I am in.

 

[wireb]

Take a look at the dumps and see if you see any correlation between what is in them and what I was doing at the time

Right now work is being a lot of work and have not had much free time to dig thought the data.....

 

[staulkor]

Bump! I am very interested when you get this working. I just ordered all my car pc bits.

 

[Darkside]

what exactly is that data and why did you format it as HEX and not binary? Also, what exactly is shown in the txt?


I did the following for the audmute.csv file: Sorting through the data buffer field gives me the following possible messages

4c
4c105e
4c105e63
4c105e63dfef
4c105e63dfef92
4c105e63dfef927c
4c1061
4c1061bc
4c1061bcca77
4c1061bcca77de
4c1061bcca77de67
4cfddc
4cfddcdc
4cfddcdcffef
4cfddcdcffefcf
4cfddcdcffefcf73

seeing a pattern there? Imo, this leaves 3 possible "full" messages:

4c105e63dfef927c
4c1061bcca77de67
4cfddcdcffefcf73

Now, searching through the csv files gives me:

MacBook-Pro:process Alex$ find . -iname '*.csv' | xargs grep ',4cfddcdcffefcf73,' -sl
./audmute.csv
./audphone.csv
./menu.csv
./sat.csv

MacBook-Pro:process Alex$ find . -iname '*.csv' | xargs grep ',4c1061bcca77de67,' -sl
./audmute.csv
./audphone.csv
./caron.csv
./menu.csv
./mute.csv
./ok.csv
./phone.csv
./radoff.csv
./radoffphone.csv
./sat.csv
./updn.csv

MacBook-Pro:process Alex$ find . -iname '*.csv' | xargs grep ',4c105e63dfef927c,' -sl
./audmute.csv
./audphone.csv
./menu.csv
./sat.csv

So. We can probably discard the 4c1061bcca77de67 message since it seems to occur everywhere.

The other two messages (4c105e63dfef927c and 4cfddcdcffefcf73) are interesting, however. They only appear in a few files:

• audmute.csv - Well, obviously
• audphone.csv
• menu.csv
• sat.csv - Obvious here too since you smashed random buttons

Now, I'm not sure whether this helps, and why the data is in the audphone and menu csv files but if you tell me more about what that data is, exactly, I'm sure I can help figure out what the buttons do.

 

[wireb]

what exactly is that data and why did you format it as HEX and not binary? Also, what exactly is shown in the txt?

Binary files are a PitA to process with a text editor or spreadsheet. Since I am asking for help I put it in the simplest format I can.

The .txt files are the raw dumps coming off the Can bus. Each valid message is separated by a line return.

The .csv files are the massaged data from the .txt files. The chip is not smart enough to not return buffers it did not use so you get some bogus data that needed to be stripped out and translation of known codes and data fields. I think most of the bus errors are due to me trying to capture everything and the hardware cannot keep up. Once I put some message masks in they should go away.


I will try and get that loaded up in my message mask filter and get some new dumps and see if there is any rhyme or reason to the messages.

 

[Darkside]

Binary files are a PitA to process with a text editor or spreadsheet. Since I am asking for help I put it in the simplest format I can.

Can't argue with that :biggrin:

The .txt files are the raw dumps coming off the Can bus. Each valid message is separated by a line return. The .csv files are the massaged data from the .txt files.

right! I see what you did there. Do you know what the flags mean (EF EFRTRR)?

The chip is not smart enough to not return buffers it did not use so you get some bogus data that needed to be stripped out and translation of known codes and data fields. I think most of the bus errors are due to me trying to capture everything and the hardware cannot keep up. Once I put some message masks in they should go away.

I will try and get that loaded up in my message mask filter and get some new dumps and see if there is any rhyme or reason to the messages.

Fair enough. Also, I suspect it won't be enough to look for matching full messages. We can probably expect to get a different message when pressing and holding two buttons at the same time.

So when you capture your data the next time, could you try to make sure you only press exactly one button at a time, and, if possible, in evenly spaced intervals. As in, press the button every 2 seconds exactly 10 times. If we knew exactly how many button presses were in one file, that would probably be helpful.

 

[Darkside]

I just thought of something. Depending on where you attach your PIC, you might not get all button messages. I'm pretty sure the MFD and Up/Down buttons are only routed through the CE CAN, and not through the Multimedia-Mini-CAN.

Any idea when you can get the new set of messages? I'd love to have a look.