The answer, for the moment, is that Mega-Em doesn't produce sound that would let me call it "working" with the current ROM. It produces random squeaks and noises that aren't too far from the feedback noises you might get from mishandling an amplified microphone. Applications with native InterWave support and IWSBOS appear to work as they should.

I'm not sure what allowed it to appear to work with the single 2 MiB bank, nor why it's failing now. I'll try different instrument sets and more different sizes when I have time, for curiosity's sake. I suspect that, upon detecting an InterWave board, it makes certain assumptions about its features and sets the registers accordingly. If there is a reasonable match, it "works"; if not, you're treated to noise. Discovering what it's doing and how we can change that or work within its limitations may prove to be a challenge.

Could I unsolder this rom to try it on the ARGUS ?

^ nope...

plus one does not destroy a wavetable..dafuq that piece is more expensive than the whole argus^^

Phreeze wrote:

plus one does not destroy a wavetable..dafuq that piece is more expensive than the whole argus^^

Oh... Do I have a rare piece of equipment rigjt there or what?

Well, that's a Diamond Monster Sound 2MB wavetable. These are not rare, but as most wavetables, they are sought-after. You can easily get 50 to 80 bucks for that.

derSammler wrote:

Well, that's a Diamond Monster Sound 2MB wavetable. These are not rare, but as most wavetables, they are sought-after. You can easily get 50 to 80 bucks for that.

More liek $800 on eBay!

Mega-Em seems to behave somewhat strangely, and I'm not quite sure what the source of the problem is yet. Is there someone with an InterWave board who would be willing to post a recording? Naturally, this requires the original ROM to be present on the board.

The test involves the following:

1. Download and install MegaMID.
2. Get rid of the BLASTER environment variable ("SET BLASTER=" at the command prompt).
3. Load Mega-Em by typing "MEGAEM -V100 -M100".
4. Change to the directory in which you installed MegaMID.
5. Get set up to record your GUS output.
6. Begin recording, then use MegaMID to play the included demo file DEMO0011.MID (usually just "MegaMID DEMO0011.MID").
7. Stop recording when playback completes, then upload the file.

It would be more useful if this were done on a somewhat older machine (ideally a 386 or 486), but I'll take any recordings I can get. It would also be useful if you could mention which version of Mega-Em was used (displayed when it loads).

^ I will do it tomorrow with the PnP card we revived with hard1k - I can use my 386DX40 for this , unless someone else beats me to it 😀

I just got my IW78C21M1 from UTSource - confirmed +1, it does have the original contents.
What better way to try it out than to perform your test? 😀

Recording: https://drive.google.com/file/d/1PEhtz_I685NW … iew?usp=sharing
Mega-Em Version: 3.11
Machine is an Am5x86.

Nice song - apart from the very weird-sounding open triangle!

Hope I've done this right!

d0pefish wrote:

Nice song - apart from the very weird-sounding open triangle!

I haven't listened to your recording yet; I'll have to wait until I'm in a quieter setting. Just from your comment, I can guess that you got the same result that I did. That "open triangle" sounds terrible with Mega-Em 3.11. If you unload Mega-Em (MEGAEM -f) and use PLAY.EXE (PLAY DEMO0011.MID), you'll hear it the way it should be played. That's what I was trying to pick out. So even on a card that has never had any other ROM, the result is the same; thanks.

keropi wrote:

^ I will do it tomorrow with the PnP card we revived with hard1k - I can use my 386DX40 for this , unless someone else beats me to it 😀

If you can still spare a couple of minutes, would you mind trying the same thing with an older version of Mega-Em, if you can find one?

What was wrong with that card, if you're willing to share?

That triangle bothered me instantly, and I'd never heard this MIDI before - it's very jarring.
It's almost like it's being faded up towards the end of the note.

I see what you mean, you're absolutely right - it sounds great with PLAY.EXE: https://drive.google.com/file/d/1GuGjvxwDomUf … iew?usp=sharing

Happy to help 😀
(Also thanks for unintentionally putting me onto Mega MIDI - fun program!)

EDIT:
Another recording with Mega-Em 3.03b: https://drive.google.com/file/d/1NQo2K_Qk36Zn … iew?usp=sharing
It sounds very different - but closer to the results of PLAY.EXE than with version 3.11.

I've put all three together here so you can compare: https://drive.google.com/drive/folders/16g3tH … HRk?usp=sharing

The v3.11 recording not only has the broken triangle, but the whole song sounds roughly a semitone down in pitch!

I just tried extracting the wave data for the triangles straight out of the ROM image. I won't post that here, for obvious reasons, but it sounds like they aren't honouring all of the parameters in the ROM when using Mega-Em. Basically, they are trying to take some short-cuts, and the results aren't great.

I think my results were better with the 2 MiB bank I tried recently. I'll try re-programming my guinea-pig ROM module with that instrument set again, and see if that really is the case. If it is, the challenge will be to figure out why that is so.

Another disappointing discovery is that Mega-Em only seems to support 8-bit instrument banks (or has some other limitation that makes it seem that way).

For the picture-loving sorts, here is a glary picture of my Frankenstein 4 MiB ROM module while a chip-erase was in progress:

It basically creates its own version of the BKSELx# signals based on ROMCS#, OE# and RA21. It works seamlessly, and doesn't seem to have had any problems with corruption so far.

I am beginning to wonder if the problems with slower flash ROMs aren't exclusive to newer InterWave chips, like the AKC versions that shock__ is using on the ARGUS. Not that this is a problem; we just avoid the cheaper 90ns+ ICs.

640K!enough wrote:

keropi wrote:

^ I will do it tomorrow with the PnP card we revived with hard1k - I can use my 386DX40 for this , unless someone else beats me to it 😀

If you can still spare a couple of minutes, would you mind trying the same thing with an older version of Mega-Em, if you can find one?

What was wrong with that card, if you're willing to share?

I will search for an older version and see what I come up with

The PnP repair wasn't anything spectacular: new capacitors, a VERY good cleaning, reflow of the F32 at U10 because it looked weird (still does!) and re-flash of the eeprom. After all these the card was revived

Considering some of you had slight problems finding backplates for the card, I've found out that the backplates off ESS1868F cards work just fine (albeit the labels being wrong). Found it kinda weird how the leftmost jack would line up, while the other 2 wouldn't - the space between all jacks are exactly the same.
https://i.imgur.com/Zg8JY0d.jpg

On another note, I was recently approached by someone regarding implementing SPDIF output - lemme copy the E-Mail I got:

I would like to submit a suggestion on the design of the PCB of the ARGUS sound card The Am78c201kc and Am78c201akc have an exte […]
Show full quote

I would like to submit a suggestion on the design of the PCB of the ARGUS sound card
The Am78c201kc and Am78c201akc have an external serial port.
For output signal the pins are :
pin 54 : ESPSYNC - External Serial Port Sync signal
pin 60 : ESPCLK - External Serial Port Clock output signal
pin 61 : ESPDOUT - External Serial Port Data Out signal
(and for input : pin 41 : ESPDIN - External Serial Port Data In signal)
The external port signal should be able to be converted into a S/PDIF digital audio signal.
However, the ICMPTI register should be switched to enable the External Serial transfert mode.
In the design of your PCB, it’s possible to access pin 54 by R14 and pin 61 by R15,
The pin 60 is not accessible. (except to try soldering a wire directly to the Am78c201akc chip).
Can you extend the track corresponding to pin 60 to a surface so that you can weld a wire and get the signal ESPCLK ?
This modification will make it possible to use ARGUS sound card to design an ARGUS digital audio output add-on module.
Ideally, it would have been cleaner to add a connector on PCB with the signals ESPSYNC, ESPCLK, ESPDOUT + GND. But at the stage of progress of the project, it is too desired.
Best regards,

Any input on that? I wouldn't mind adding a header or even a full implementation - but have to admit that I've never worked with SPDIF before, so some guidance might be required.

have you also considered giving the caps a little bit more space ? around C718

shock__ wrote:

Any input on that? I wouldn't mind adding a header or even a full implementation - but have to admit that I've never worked with SPDIF before, so some guidance might be required.

I have hardly ever used a S/PDIF-enabled device, much less designed anything involving one, but I question the feasibility of this endeavour. Looking through the datasheet and programmer's guide, it quickly becomes clear that it would be far more involved than setting a register once and connecting those signals to an appropriate encoder. The only way it seems even remotely feasible still involves going through the on-chip analogue stage before getting to the CODEC module. Otherwise, you seemingly won't have access to the full card's output. Furthermore, the serial output is unavailable when certain other FIFO modes are used, if I'm reading correctly, and would seemingly also require driver-level changes for Windows to manage that register setting. My vote would be to ask this person to provide some form of proof-of-concept implementation, or at least more concrete circuit and software design guidelines, before going ahead with unnecessary layout and routing work. I'm not claiming that it definitely can't be done, but it doesn't look like it would work as intended.

I also have 0 experience with spdif, never used it