Basically, it was a bad jumper setting (volume and hum).
As for the stuck note, that was just init3dxg.exe trying to initialize my Monster Midi with settings meant for a DB60XG .
I will switch to Tiido's excellent looking setymf .

Now excuse me while I go kick myself .

Did you have to use /FORCEPNP option ? I don't think SETYMF would recognize the Mediatrix card due to different PnP ID from other YMF71x cards out there.

Tiido wrote on 2020-04-23, 10:35:

Did you have to use /FORCEPNP option ? I don't think SETYMF would recognize the Mediatrix card due to different PnP ID from other YMF71x cards out there.

Indeed I initially had to use /FORCEPNP . I flashed the EEPROM because I was trying to fix an issue with SETYMF silently overriding my MPU-401 base address choice of 300h (SETYMF kept setting it to 310h) . It was late and I misread what it would do (I thought it would just write settings to the card, not change the PnP ID) . Then I discovered /ALLOWBAD, which fixed my issue .Now the card thinks it is generic . It does not really bother me as it still works with both init3dxg and SETYMF (a nice feature would be EEPROM backup) and I will never sell it . My understanding is that I could reflash it with /SETPNPID if I wanted to, but the syntax needed for the file is not mentioned in the doc . As expected, CTCM complains (I also have an AWE64), but still works for the AWE64 , so I don't mind either .

The 300 not sticking is a program bug that I will investigate. It probably originates from hidden YMF704/721 specific settings that are visible of my own made cards.

/SETPNPID will only change the ANSI ID string but not the actual PnP ID (YMH0021 etc). You make a file named "NEWPNPID.TXT" and any text in it will get written to the EEPROM as the ID string, so your BIOS listing will show that new name and you can briefly see it in Windows as drivers are installed and in SETYMF also when initialisation is underway.
Offering backup of original EEPROM is a nice idea, I will put it in the todo list but I'm reaching code size limits soon so it may not be possible without rewriting the program in assembly or something. I have made a pair of tools to read and write EEPROM of the YMF71x cards, it can be used to backup things and to restore original PnP EEPROM if a dump is found.

CTCM complaining is something I don't plan to deal with for now, it is CTCM not following PnP specification not an actual problem of what I'm doing.

Tiido wrote on 2020-04-23, 17:54:

The 300 not sticking is a program bug that I will investigate. It probably originates from hidden YMF704/721 specific settings t […]
Show full quote

The 300 not sticking is a program bug that I will investigate. It probably originates from hidden YMF704/721 specific settings that are visible of my own made cards.

/SETPNPID will only change the ANSI ID string but not the actual PnP ID (YMH0021 etc). You make a file named "NEWPNPID.TXT" and any text in it will get written to the EEPROM as the ID string, so your BIOS listing will show that new name and you can briefly see it in Windows as drivers are installed and in SETYMF also when initialisation is underway.
Offering backup of original EEPROM is a nice idea, I will put it in the todo list but I'm reaching code size limits soon so it may not be possible without rewriting the program in assembly or something. I have made a pair of tools to read and write EEPROM of the YMF71x cards, it can be used to backup things and to restore original PnP EEPROM if a dump is found.

CTCM complaining is something I don't plan to deal with for now, it is CTCM not following PnP specification not an actual problem of what I'm doing.

Would you would be willing to share those EEPROM reading and writing tools ?

They are found here : http://www.tmeeco.eu/TKAYBSC/SETYMF/
READPNP reads the PnP EEPROM content into EEPROM.BIN file and WRITEPNP writes it to the card.

Tiido wrote on 2020-04-23, 18:11:

They are found here : http://www.tmeeco.eu/TKAYBSC/SETYMF/
READPNP reads the PnP EEPROM content into EEPROM.BIN file and WRITEPNP writes it to the card.

🤣, I somehow missed those .

Thanks again!