There's a fully assembled card now ! Although testing is another matter, I don't actually have a test machine here yet and really no place for one either. This is another logistical challenge to overcome 🤣

Also, fancy bracket, courtesy of Rabanik on these forums :

Not a problem mate, I have plenty of machines ready and I would gladly test it for you! 😁

Congratulations on your progress. I am now following this with great anticipation.

Boohyaka wrote on 2023-10-13, 10:29:

Not a problem mate, I have plenty of machines ready and I would gladly test it for you! 😁

Hahaha, I do need oscilloscoping and possible logic-analysering and definite CPLD JTAGing which is a bit of a bigger logistical challenge. This thing will not work as is yet 🤣

I began writing the new SETYMF, this time it is a full assembly program that is much more compact, much faster and much more flexible than the old QB45 one, which has grown big enough that I am starting to hit its 64KB code limit which won't let me really add any new stuff to it. There are some teething pains but the most important problems are now overcome ~
One thing I do miss is the really easy string management of QB45 🤣

Once there's a bit more of the software done I'll see about getting a temporary test setup going, the workdesk is still clean enough that I can set up my old test machine on it and at least be able to scope stuff to debug things as the need comes. I am sure it is going to, especially when I will try to do the dual OPL2 simulation and in particular getting level matching done. OPL3 inside YMF71x has essentially fixed level which I will have to try to match on the one in YMF704/721 which level I can control, hopefully without needing any funny resistor values but that is something I will worry about when I actually have some level of dual OPL2 compatibility implemented. It will definitely need prior setup to achieve the compatibility, it is definitely not going to be an automatic plug and play matter.

I did simulations on the gateware and discovered a few bugs, including a fatal one (inverted enable signal polarities in one place 🤣) and after fixing them I finally dared to put the card into a computer for a test.

At first I got no sound, all the tests didn't error or anything but there was no sound... turns out I put the sound cables to the wrong sound card (I had to work blindly 🤣), after fixing that, things started to make sound and it was good ~

Not everything is tested as the gateware is not finished and neither is the software but the card is definitely functional ~

I'm now going to a norwegian language course on 4 days of the week and my time has dramatically reduced but I'm trying to get something done every day, eventually this stuff will be finished.

Thanks for the update and glad to see this project coming along! 😀

I am also looking forward to this, already found a donor board to steal the YMF721 from 😁

New software is slowly shaping up, and there's been a little bit of future proofing done. When the software encounters a model it is not familiar with, it'll show this message :

New hardware is not expected to be compatible with old, and it is also the case with 04 and 05. Old SETYMF will not be able to initialize 04 past what is in YMF71x itself, but unlike the new SETYMF, it actually has no version check for any possible newer cards. Nothing bad will happen but it will certainly not be able to support any extra or changed features and if EEPROM is being updated, it will "corrupt" it, making new software not be able to function right as identification etc. will be off.
This is also improved in the new software, it is not going to allow to updates to EEPROM unknown versions. There's also EEPROM size check and updates won't be allowed in cards with 256 byte EEPROMs either, which is something that caused problems to someone few pages back in this thread.

Next up is actual initalization of the chips involved, once that is had the painful process of making an UI will start once more.

I'm really interested in this card but I'd definitely would like it with YMF721. Unfortunately I've never had a OPL4 sound card so no donor board available for me. But I'm sure it is a great card even without one.

I definitely cannot provide a card with one, they're (very) expensive (nearly)unobtanium parts more or less. You'll have to find a donor in the meantime, YMF704 can sometimes be found on other YMF71x cards but YMF721 is a much harder task. There are some Intel motherboards with it and a friend saw one on medical equipment of all things but still $$$...

Some more updates on the software end of things :

It might look like just text output, and it is, but DOS is not really making it easy to do fancy text and especially in color without huge performance loss. Today and yesterday has been spent on writing a fancy text output that is synced up with DOS/BIOS output, but does all the fancy stuff directly, including insertion of numeric values directly and via address, strings and fills, and also some navigation stuff.
Single text output string can do all sort of fancy stuff, which in the long run makes my life dramatically easier. I can place variables directly in that string for example and get things like numbers update automagically without any special effort on crafting text output calls or making some strings magic.
Some keener eye will see that things did not quite go to plan near top of the image lolololol

I'm quite pleased with the results of today. Tomorrow I'll be updating all the existing text bits to the new stuff but now I will need to do preparations for another day of Norwegian language course tomorrow morning.

it needs an /abuse switch, you get verbal abusive messages as initialization progresses:
- found YMFxxxx you d0rk
- resources are port $$$ irq $$$ you weird0

🤣 🤣 🤣

keropi wrote on 2023-10-23, 18:42:

it needs an /abuse switch, you get verbal abusive messages as initialization progresses: - found YMFxxxx you d0rk - resources a […]
Show full quote

it needs an /abuse switch, you get verbal abusive messages as initialization progresses:
- found YMFxxxx you d0rk
- resources are port $$$ irq $$$ you weird0

🤣 🤣 🤣

That could actually be quite funny.

Maybe one day, after all else is done 🤣

The summary screen is now done, it took some effort to get to this point as a bunch of under the hood stuff had to be made and every once in a while I'll add some new bits to it to make it even more useful 🤣

Next up are functions that talk all the settings to the hardware and after that all the UI for adjusting the settings. Plenty of work ahead...

Sound card itself works nicely but good chunk of it remains untested for the time being.

Looking good!

Exciting Tiido! I hope that the software and hardware teams are collaborating well 😁

It is just one person so there's little reason for any drama lololololol
But I am sick and seeming to get worse, some days GF had a fever and all the other funstuff, and now I have that too. I may have to miss the language course tomorrow...

Today I was implementing things for mixer locking and dual OPL2 simulation... aand then at the end I press compile button and I'm greeted with the error that design won't fit...
I have some things to try but I may not be able to bring the extra stuff into the mix unfortunately 😒

EDIT: After some struggling, I have managed to make mixer locking and dual OPL2 simulation both fit into the design !

This CPLD is quite filled up, all the pins are used, 75% of the logic elements are used and close to all the interconnect resources are used. The latter are what gave me all teh headache, I had to shuffle things around and in the end also cut down a few things to a smaller size and then after lot of trying the compiler actually managed to get a fit. Simulations seem to indicate things should work so the coming steps are to flash the CPLD and add support to the new spec in SETYMF, write some test code to verify things work and then I can get to a point where I can call the hardware finished ~
Hopefully I won't have trouble getting matched levels between YMF71x internal OPL3 and one in YMF704/721. They both go into the same input on YMF71x and in worst case I can adjust gain on YMF721 DAC to get a match, hopefully without needing some strange value resistor 🤣

Mixer locking prevents games from issuing a reset to the mixer (and with it make all levels go to shit) and change to master volume when enabled, YMF71x needs it to be at "1" or volume range on other things is not full. SBDAC, Line, CD, FM+WB+MIDI levels can be locked separately when locking is enabled. This should cover all the bases ~

Tiido wrote on 2023-10-27, 15:32:

EDIT: After some struggling, I have managed to make mixer locking and dual OPL2 simulation both fit into the design !

The mixer locking thing will be great, currently I have bat files to re-initiate my YMF719 based card after any game that messes around with the levels. This works fine but it's a bit messy.

This project looks promising, I would love to get my hands on one!