Help needed in getting MD working!

[heckstall]

Hello all!
I've build MD with 22 inputs Atmega644/20MHz, latest firmware (20100126) and MIDI interface to PC.
Unfortunately I have a lot of problems I cannot solve and I mention them here hoping someone can help me (I have 30 years experience as an electronics designer so that should not be the problem).
1. I Cannot program the PIC18F2550 using COM port. I measured a (clamped) voltage on the TxD-pin when connected to the PC of -0.6V. Correct? (Used two PC's). For the time being skipped using USB-interface.
2. Could not use MIDI interface (External MIDI to USB convertor: M-Audio midisports 1x1). I found out using an oscilloscope that the signals coming out of the SN138N (I used several devices) were distorted (no sharp edges). This one has been solved in adding a 10k resistor between pnt.7 and ground (I found also in the internet midi-interfaces with this resistor added!). Probably this has (also) something to do with the brand I used (cannot figure out the logo). Now I was able to load the firmware with MIDI-OX (MCT did not work).
3. When MD is switched on almost all 22 VU-meters are give full screen response with no pads connected. ???
4. When using MCT loading MD the dynamicly addressing of the muxes (4051) stops!
5. When trying to lower the "HighLevel" with the MD-keyboard value changes but suddenly jumps back to higher values.
6. When trying to set "All Gains Low" to "Yes", again with the MD-keyboard, the "firmware screen" is displayed with no change to "All gains low".
Thanks for helping me out!!

[Firelord]

I Cannot program the PIC18F2550 using COM port

This problem is very common. See this thread. Provide more information.

SN138N

I've failed to find this device. The part number should be 6n138.

3. When MD is switched on almost all 22 VU-meters are give full screen response with no pads connected. ???

I can think of several reasons for this. 1) Wiring problems; 2) Shorts on the boards; 3) Dead/defective 4051 MUXes; 4) Wrong F/W; 5) Some other H/W problem.

5. When trying to lower the "HighLevel" with the MD-keyboard value changes but suddenly jumps back to higher values.

Since there is evidently some problem(s) with the signals absorbed by ATMega's ADC's this is normal, that is if the HighLevel is in auto mode -- it dynamically adapts to the incoming signal.

[dmitri]

3. When MD is switched on almost all 22 VU-meters are give full screen response with no pads connected. ???

I can think of several reasons for this. 1) Wiring problems; 2) Shorts on the boards; 3) Dead/defective 4051 MUXes; 4) Wrong F/W; 5) Some other H/W problem.

Most likely the schematic used is for 32 or more inputs but firmware is for 22 inputs.
Also possible resistors on the ADC inputs of Atmega are missing or well above 4k7.

[heckstall]

I've failed to find this device. The part number should be 6n138.

Sorry, you are right I meant the 6N138.