Thread: Button interface to Bordmonitor on BMW E39 and E46

Art wrote:
You wrote there that one has to find out which kind of knobs are used. You don't have to, I can tell you that This are just simple optical rotary encoders giving relative rotation as phase differences on their outputs. On the "reverse engeneered" schematics (which I posted before) of the button pcbs, I put the schematics of how I have connected those rotating knobs. For whatever reasons Alpine (bordmonitor is built by Alpine) decided to use two different knobs, one with 5 pins, the other with 6 pins. However, the functionality/technical specification of them seems to be the same. The one with 6 pins need an Vcc input (5V) and gives phase moved peaks on the J2.2/3 pins. The other (with 5 pins) don't need an extra Vcc. So there one just need to pull-up (to Vcc) both pins J1.1/3 and you get the output. On the E39-forum.de somebody found out the manufacturer of compatible rotating knobs for the 4:3 monitor. I think for
the 16:9 it could be the same. The original knobs where a special development for Alpina. The compatible one is that one: Grayhill 61C22-01-04-02

So for that part we know already the solutioon
cheers,
art

Excellent! Found a data cheet here.

But I doubt that it is optical. I got pure 0 ohm and --- ohm without power supply. Seems mechanical to me.

I think the sw would really benefit from some logic interfacing. Also with debouncing maybe.

If you detect these sequencies and generate a clockwise and counterclock pulse if you see what I mean, it would really offload cpu and require much less real time performance.
clockwise
A 000-1-111-0-00
B 000-0-111-1-00

counterclockwise
A 00-0-111-1-000
B 00-1-111-0-000

So,
1) debounce with nand (here)
2) generate clock from any change in A and B to pulse a latch
3) detect red marked conditions above in two differnt signals. One-Step-Up and One-Step-Down. You can then just have a single positive edge interrupt on these inputs.

Hardware should do what it do best, and sw what it does best. Gates are cheap. :-)

I will post Eagle file with schematics for above logic, and mechanical dimensions for both E46 boards later tonight.

Hi Art

I have been looking at the E39 button pcb today. I am making an Eagle project that contains both E46 and E39 in one for reference. Common electronics, and two set of connectors. One set for E46 and one set for E39. This way it will be very clear how to make our pcb to fit both.

There is one concern that I have, and that is regarding the diodes on the button pcb. All inputs on E39 has diodes to LED GND, and some inputs on E46 has the same. Any push button terminal cannot have higher voltage than 0.7V (above LED-)

So with no lower voltages in the system but 0V that is the voltage on LED's, we have to detect 0V when button is closed and 0.7V when open. Very special IO will have to be used to read that. Please tell me if I am wrong.

I cannot see any way around this but to remove the diodes, do you?

I have looked ATMEGA32A briefly, and has found that it has 8 channel 10 bit ADC. Is this the correct spec? http://www.atmel.com/dyn/resources/p...ts/doc8155.pdf

To save IO I think using the AD inputs is a good ide. Ex. Use 0k, 2k, 6k, 14k, 30k and 62k in series with row 1-6 on left button board. Together with 2x1k serieresistors you will then be able to poll that chain and get unique combinations so that two buttons pressed simultaniously will not be detected as a third button.

If AD is set correctly only 6 bits will be significant, and 4 bits can make up for noise and tolerances. A "round robin" of button chains will have to be implemented in code and hardware. A current generator will have to be used to get a uniform voltage stairway. (A transistor with emitter resistor will do fine)

I have made a layout and will order boards and components this week

pcb received! edit in first post

Pcb built. Now its time to debug code. Edit in first post.

excellent work. how's this going?

Thanks Sama
Well, I do not like strapping pcb's so I am building rev 2 at the moment :-)
I went Art's way with the buttons using a PCF8574 instead of reistor ladder as the buttons resistans varies a lot.
A lot of code tested, so I will sone be sone with my combined E46&E39 board.
Art has finished his board for E39!

I've just actually been in touch with Art to figure out how on earth he "cut" the screen. As it turns out, he cut the footprint by cutting the frame, but not the glass itself.

I'm going to be doing the same with mine (just bought an E46 vert) so I can get a transflective high res in there, however I'll be keeping the buttons as they are as I want to use the existing OEM setup. I'm hoping to achieve that by using an RGsB converter for the OEM stuff and getting that on the higher res monitor, then switching to use a PC. Worklog currently here

So just to understand 100% your project, the reason you're replacing the buttons with a custom board, is so that you can fit a bigger screen in there?

Will the buttons broadcast on the ibus directly or will they be CarPC inputs, which then you're free to do what you want with, including ibus broadcasts?