Cloudschatze wrote:

I wouldn't say it's impossible; it's just not as simple as leveraging EMM386's port-trapping capabilities in a manner similar to SoftMPU, as I'd suggested earlier. It does seem like TEMU's redirection could be used as a model...

OK, i took a look into TEMU's documentation included with its source code. It states that it won't work on any V86 mode software, which makes sense so yeah... No EMM386 and no JEMM or QEMM as well. But a XMS driver like HIMEM can still be used.
I don't know jack shit of assembly so i'm really useless to try and understand what TEMU actually does.

Now theoretically, what we need is a circuit with a SN76489 sound generator that sits natively at 1E0h and then have a TSR to redirect the C0h range calls so that the SN76489 can play whatever is called. We could essentially even give the capability to respond to whatever port calls the PS/1 audio card makes and play via the SN76489. I also don't know much of electronics but i'm positive that it should be a very simple circuit. And get someone well versed in x86 assembly that can give us a hand.

Perhaps instead of making this an ISA card we could do a serial or parallel port dongle instead?

Then the output question. The final output would probably sound a little bit different via an audio output jack through speakers or headphones when compared to an actual Tandy speaker. Tandy 1000's have an "ear" jack output right? How different does the output sound through the ear jack when compared with the actual speaker?

It seems the SN76489 chip uses only one IO address, so basically it only interferes with the second DMA controller at address C0h.

Consider that C0h is the port for DMA channel 4 memory address, and DMA channel 4 is never used as it is the cascade channel (8-bit channels 0-3 of the first chip goes through the cascade channel), could it actually work if the audio chip just sits at C0h?

The SN76489 is write-only and if the C0h port writes are also seen globally on ISA bus, what could go wrong?

If that can't be used, maybe games can be patched. I know I am not good at trapping IO accesses at CPU level (a la SoftMPU), but patching sound drivers should be a piece of cake. In that case, the sound chip could sit even on LPT port, but it would require three IO accesses instead of one.

How about putting the sound chip to IDE cable, if one has extra IDE port? In older computers, IDE bus is basically nothing more than a buffered 16-bit ISA bus with only 3 address bits and other address bits pre-decoded as two chip selects.

carlostex, Jepael
Wait, wait wait! I think it is already new thread starting here. Everything about Tandy sound card need to go to new thread.

I have some great pictures for you.

My SID on new PCB.

Connecting audio output to another card (AWE64).

Preparing parts...

Parts for assembling one card (parts not soldered, just inserted into the board):

Fully assembled cards (9 pcs).

Bottom left: my own card, enhanced "debug" version with DIP sockets.

Kits:

Left: not completed kits (6 pcs.). Not enough parts.
Center: almost complete kits (3 pcs.). No jumper headers left in my parts bin. Oops. 😲
Right: full kits (7 pcs).

But I'm still low on some parts.
I have only ONE SID (my own). And I have only 5 mounting brackets, didn't managed to get some more yet.

excellent, love the pics 😁

Ooo Weller time .. 😁

AlphaC wrote:

Ooo Weller time .. 😁

Actually, "goot" time. 🤣

And some more, multi-card photos:

For fully-assembled cards I also need to know, what SID will be installed in it. Because with 8580 there must be a 78L09 reg and filter caps 6.8nF, but for 6581 there must be short on +12 and 1nF caps.

^^ All those pics look like electric candy!

I'm interested in what you and bristlehog mentioned earlier, So please can you setup one card for 6581 and one for 8580. That would be great, cheers.

I need my card set to 6581 btw

Cloudschatze wrote:

I wouldn't say it's impossible; it's just not as simple as leveraging EMM386's port-trapping capabilities in a manner similar to SoftMPU, as I'd suggested earlier. It does seem like TEMU's redirection could be used as a model...

From what I can tell, TEMU basically does what SoftMPU would have to do without EMM386's port-trapping capabilities. Given that the TEMU source is available, one could certainly work from that.

Today I've sent two parcels for Russian buyers and got some information about cost of shipping abroad. Seems like it does not depend much on buyer's country. And total shipping price will be around $7.

so when do we send the monies? 😁

so when do we send the monies? 😁

He contacted me yesterday and I sent him payment!

I think he will poke everyone individually as he completes the replicas.

Beegle wrote:

I think he will poke everyone individually as he completes the replicas.

Exactly.
I'm going through the list, one by one, in chronological order.
First five non-Russian parcels were sent today. 😎

PM received and answered! 😁

Just to inform everybody.

First, right now I have no mounting brackets left. And still no SIDs. Trying to find some, but still in progress.
One of the goals of re-designed board was to be compatible with pretty common mounting bracket that can be found on many cheap cards. Like this one, or this, or even this.
So right now it's the way to go, I still can't find source where I could get boards with those brackets fast, cheap and in bulk. Instead I have to drive all over Moscow (it's about 35 km in diameter) through a heavy traffic. It takes a lot of time and I don't have that much free time now.

Second, I don't have enough parts for some boards, but that is not a big problem: parts are already ordered, I'm waiting. Just need to calm down and wait.

Third. Reminding prices and some other stuff.
"Fully assembled card" = $45 (USD)
"PCB+parts kit" = $35 (USD)
"Bare PCB" = $25 (USD)
Shipping = $7 (USD)
- SID is not included in prices above (assuming "w/o SID" option).
- "Fully assembled card" gets tested: playing *.MID with PX player and *.SID with alfa-version software. Also joystick port gets tested under Win98.
- Kit can not be tested (you know why 😀 ), but I check ICs with TL866A programmer. It performs some basic functional test with different logic ICs. Not so thorough, but dead ICs will be sorted out.
- I don't know speed of Russian post from my country to others yet. Opposite direction (when I'm buying on eBay, or AliExpress, or whatever) is not lightning quick: typical delivery time is about 3...4 weeks. Sometimes even 2-3 month, but it can be 1 week on another occasion. Who knows.
- I send parcels with international tracking. Don't know how good it works yet, we will see.

Are you completely sure about PCI soundcard bracket compatibility with the replica?

bristlehog wrote:

Are you completely sure about PCI soundcard bracket compatibility with the replica?

I think too that was a mistake to put the PCI picture.
Screws on the bracket will be located on the top side (while ISA needs them at the bottom)
And the orientation of the db15 port is inverted too.

It can still work though, if the db15 opening in the bracket is wide enough, and you hold the bracket in place with the db15 port hex screws.
But still, less than ideal.

bristlehog wrote:

Are you completely sure about PCI soundcard bracket compatibility with the replica?

Beegle wrote:

I think too that was a mistake to put the PCI picture.
Screws on the bracket will be located on the top side (while ISA needs them at the bottom)

Yes, probably PCI cards are not a good variant. But it can be suitable if there will not be any additional mounting screws on a bracket (so PCB orientation will not be an issue) and hole for DB-15 will be O-shaped instead of D-shaped.

One more notice about assembled boards: by default SID clock is set to 0.89 Mhz (standard for original SSI-2001). If you need 1.02 MHz (C64 NTSC) you can re-configure jumpers on your board, help is on the board itself. Jumper position is dependent on installed clock divider: К555ИЕ6 or 74HC192. Help on the board covers it all.

And some interesting info. If anyone will be interested in manufacturers of ex-USSR ICs, here is link with some data.

Beegle wrote:

I think too that was a mistake to put the PCI picture. Screws on the bracket will be located on the top side (while ISA needs th […]
Show full quote

I think too that was a mistake to put the PCI picture.
Screws on the bracket will be located on the top side (while ISA needs them at the bottom)
And the orientation of the db15 port is inverted too.

It can still work though, if the db15 opening in the bracket is wide enough, and you hold the bracket in place with the db15 port hex screws.
But still, less than ideal.

Oh you had to tease with the AdLib Gold, huh? 😈

Fagear wrote:

For fully-assembled cards I also need to know, what SID will be installed in it. Because with 8580 there must be a 78L09 reg and filter caps 6.8nF, but for 6581 there must be short on +12 and 1nF caps.

Would it be possible to design this in a way that you can easily switch between SID types using a jumper or something?
I can imagine that people may want to swap between a 6581 and a 8580 from time to time.