ah so it was a case of multiple mpus 🤣
nice job Bancho! Enjoy and thanks for your support!

Are you still looking for someone to implement this into a modern microcontroller?
I am an avid FOSS supporter and the author of several widely used FOSS projects. I would be willing to spend some time doing this for free, under the condition that whatever is made is released as open-source hardware & software.

I need some help keropi.

I have my PCMIDI card installed on a Socket 3 PCI motherboard, and I'm trying to troubleshoot some issues with my mt32-pi. It seems that the constant clock signal from the card (metronome) is messing the connection up. MQSETUP.EXE from the Midi Quest software pack should be used to stop this as I understand, but it asks for being called from AUTOEXEC.BAT and says will do what SETUP.EXE tells it to. When I run SETUP.EXE it complains about the card not being available on IRQ2 even though it is (and it actually runs fine). MSD shows IRQ2 being used by Intel 8259A so I'm guessing the PCI bus won't just let it go. Should I switch to IRQ3 or something? How can I stop the metronome / constant clock signal?

The signal you mention is not the issue here , I have so far two different mt32-pi devices and they work fine with PCMIDI
(not to mention that mt32-pi creator has a PCMIDI since many years and now also has an Orpheus) .

Almost certain your issue is IRQ/System-related... if you put the mpu+mt32-pi in some older system do they still have issues? I bet they will not.

keropi wrote on 2021-05-05, 14:52:

The signal you mention is not the issue here , I have so far two different mt32-pi devices and they work fine with PCMIDI
(not to mention that mt32-pi creator has a PCMIDI since many years and now also has an Orpheus) .

Almost certain your issue is IRQ/System-related... if you put the mpu+mt32-pi in some older system do they still have issues? I bet they will not.

Same PC with same PCMIDI using same USB interface can send SysEx commands to a Windows PC no problem
Another PC with same PCMIDI using same USB interface can send SysEx commands to same MT32-Pi no problem

When this particular PC, USB interface and MT32-Pi come together, it just doesn't work. It seems to be an issue/bug with the MT32-Pi.exe software though, I think I can get SysEx messages to the Pi with DUMP.EXE now. Thanks keropi 😀

P.S. Should I expect any incompatibilities due to running the card on IRQ3?

weird stuff , I admit I only connected mt32-pis directly to pcmidi and stuff just work...
you will only face issues with IRQ3 if some game is hard-coded to using 2 - I started putting some patches for this on pcmidi site in the files section , so far all there is are kykypyza's patches but I will keep adding stuff that is getting known to me

keropi wrote on 2021-05-05, 19:01:

weird stuff , I admit I only connected mt32-pis directly to pcmidi and stuff just work...
you will only face issues with IRQ3 if some game is hard-coded to using 2 - I started putting some patches for this on pcmidi site in the files section , so far all there is are kykypyza's patches but I will keep adding stuff that is getting known to me

The issue was a software bug in the program and got resolved - it seems to have something to do with mt32-pi-control not behaving the way Roland's SysEx handshake method expects or something: https://github.com/dwhinham/mt32-pi/discussio … ncomment-700991

It's a long thread if you care to read up on it, but now I can finally send SysEx commands to my mt32-pi with the PCMIDI card. If anyone has this issue, read the above git discussion 😀

I've opened an issue over on mt32-pi-control's repo with a summary of this problem: https://github.com/gmcn42/mt32-pi-control/issues/2

If there are any experienced MPU programmers reading who could check over the code that hits the MPU in mt32-pi-control and maybe see what might be wrong, it'd be very much appreciated.
I think it may be a sequencing problem with the reading/waiting of the MPU status bits as it's timing-sensitive, but I can't pinpoint it.

Keropy, you may soon have no more sound card to clone 😀

for sure, mt-32s are now out of cloning range 😀

I did a test with a pcmidi and scc-1 on my 386DX and I could replicate what was discussed above (since I did not understand it initially and in case anyone else is also wondering, mt32-pi.exe is a DOS program that controls mt32-pi from a reset to switching modes)
Using the new version d0pefish posted that added a tiny delay fixed the issue completely.

MT32-Pi is getting better and better by the day , I honestly cannot tell the difference from a real mt-32 when using a "normal retro setup" (meaning 70e speakers and a random 386), the only difference I can tell is that the mt32-pi sounds a LOT cleaner. Maybe using some lower Hz for output one can almost replicate what a mt-32 sounds like 😀
amazing work again d0pefish!!!!

keropi wrote on 2021-05-06, 11:58:

I did a test with a pcmidi and scc-1 on my 386DX and I could replicate what was discussed above (since I did not understand it i […]
Show full quote

I did a test with a pcmidi and scc-1 on my 386DX and I could replicate what was discussed above (since I did not understand it initially mt32-pi.exe is a DOS program that controls mt32-pi from a reset to switching modes)
Using the new version d0pefish posted that added a tiny delay fixed the issue completely.

MT32-Pi is getting better and better by the day , I honestly cannot tell the difference from a real mt-32 when using a "normal retro setup" (meaning 70e speakers and a random 386), the only difference I can tell is that the mt32-pi sounds a LOT cleaner. Maybe using some lower Hz for output one can almost replicate what a mt-32 sounds like 😀
amazing work again d0pefish!!!!

Glad to see this bug fixed, and yes mt32-pi is awesome.

keropi wrote on 2021-05-06, 11:58:

amazing work again d0pefish!!!!

Thankyou for the kind words!

mt32-pi stands on the shoulders of giants, however - Munt and FluidSynth have been available for a long time now and the people behind those projects deserve utmost respect.
All I've done is tried to build a dedicated "appliance-like" platform for these synthesizers and allow them to be used in new and interesting ways.

Hopefully we have begun to achieve that. 😀

mt32-pi-control now has a 1.0.0a MS-DOS build that is compatible with PCMIDI and Orpheus cards, rejoice. Tested working by @dopefish and me 😀

We got a report that BIOFORGE had troubles with MIDI playback when using a PCMIDI MPU (original or our clones) .
Thanks to ripsaw8080 there are now 2 patches for versions f1.01 and f1.02s
(version b1.14 does not need any patch as it works as expected)
I am attaching the patchers to this post , will also link them to the PCMIDI homepage in the files section.

Thanks again to ripsaw8080 for fixing this! 😀

PCMIDIs are now back in stock and build on-demand
this time the color is classic green 😀

I recently installed the nice X2GS on my PCMIDI-Card and it works beautifully. One tiny issue is that I have also a MT32Pi connected to the MIDI-Port of the card.
With some games a couple of Instruments sound wrong from the MT32. Is there a way to disable the Wavetable Module on the PCMIDI? I suspect the issue comes from that

there is no way to "disable" the wavetable header
but there is also no way for the ports to interfere like that: the WT header is a receiving-only connection so it cannot send anything back to the system and mess things
I suspect any wrong instruments are a software error - perhaps too fast system?

Hmm I imagine maybe the rom file on my MT32Pi-SDCard is corrupt due to turning it off while reading or something. I have to check at home.

The system is not too fast. It's a Compaq Deskpro 486/33M EISA-System with DX2 Overdrive.

just for closure, the issue was indeed speed related... very weird speed issue though the kind that only DOS provides 🤣 🤣 🤣