GM5 MIDI out message problem (SOLVED)

Testing/Troubleshooting
1

Hi, long time not posted here, but I`m still using Midiboxed keyboard all the time, occasionally SID or FM.

Unfortunately I can’t find any similar problem to mine at forum so decided to ask.

Yesterday I tried to use first time soldered many time ago GM5 modules (I have three of them, all has the same problem).

Problem: MIDI output is sending otjher messages than I want.

I send messages from GM5 MIDI OUT and receive it on

  • other interface (novation Remote ZERO SL)

  • directly at GM5 (via MIDI loop from MIDI OUT to MIDI IN).

In both case I receive at MIDI IN garbage messages, not what I sent (it was simply note on/off or CCs).

For checking MIDI IN ports at GM5 and second interface I send to them messages from my MIDI keyboard, and both interfaces works good.

But when I send the same simple message (not sysex) via MIDI out in GM5, I receive garbage on both inputs (GM5 and novation interface).

Example: instead of note on 90 30 64 , I received two Song select messages: F3 76 and F3 02. Other example: sent B0 45 40, received 3 messages: F3 52, FD, F3 01. If it will help I can putt more results.

Firstly I think it is an error on input side GM5, so I changed 6N138… and tried others tricks, but finally I recognized that GM5 receive everything properly, even long sysex. So half of GM5 is working good. Problem is that GM5 is outputting other messages that I order in MIDI OX or MIOS Studio.

Interesting problem, tried to find source, but without any succes.

Have anybody idea where to look to find solution? Will be grat to use GM5 as interface.

Strange is that I have the same bad results on all three GM5 modules MIDI output, even on other computer. All computers with XP, GM5 driver installed. Without driver problem is also.

Please help. Thanks.

2

Hm… nobody reported such an issue before.

Wrong MIDI values normally indicate a baudrate mismatch, e.g. because the wrong crystal is used (GM5 requires 16 MHz).

But this would also lead to wrongly received MIDI events, and the correct events would be received on a direct MIDI OUT->IN loopback.

So, the only explanation I have (yet) is, that you are using the wrong resistor value for R38 and/or R50. Both should be 220 Ohm.

A higher value would lead to unstable transfers. In your examples more 1s are received than 0s, which confirms my hypothesis (1 is the idle state)

If this isn’t the reason, please try following experiment: put at jumper at J6:I1 and J6:O1 to short the In/Out connection (don’t worry, this won’t lead to damages) - is the loopback working now?

Best Regards, Thorsten.

3

Thank you Thorsten for trying help.

About Crystal: I cannot read value on it because I see only X8301JT (symbol), but you said that it must be good one since I can receive from real MIDI world.

About resistors: Probably you mean R55 and R38. Both are 221 ohm (measured before solder, and also directly on the board after soldering). But question: should it be solder both from top and bottom side? I make soldering resistor mostly from bottom side. Can it introduce this problem?

If this isn’t the reason, please try following experiment: put at jumper at J6:I1 and J6:O1 to short the In/Out connection (don’t worry, this won’t lead to damages) - is the loopback working now?

I removed MIDI loop cable and installed jumper like you said. With this configuration I see on input monitor software exactly what I sent (good result)!

I don’t know what is next step now but I`m happy that at least GM5 chip is working good.

p.s. adding solder from top side to both mentioned resistors not improved situation. But I see that most of my part are not soldered from top side (besides GM5 chip of course). Is this wrong?

QBAS

4

Hi Thorsten,

I would like to thank you very much, like usual. Of course you helped a lot giving me an inspiration where to look for source of problem. It was stupid mistake, sorry that I waisted your time. From other side it was nice to hear you speaking directly to me :rolleyes:

I found that my midi cables was connected in wrong way: ground was connected to wire on the side. :whistle:

Stupid basic mistake eat my all Sunday.

Now my GM5 is working correctly.

I wish all the best for you Thorsten and for other people who are taking care for this forum and site. :sorcerer:

QBAS

5

That’ss indeed an unexpected error mode! I’m glad that you got this fixed :slight_smile:

Best Regards, Thorsten.