I'm a little confused. DSS is auto-paced at 7 KHz irregardless of the host data rate. Wouldn't you need the strobe signal to know when a new data byte is written?

Your confusion is justified. According to the programmer's guide the Sound Source uses pin 17 (C3/SELECT) as its strobe signal, presumably to avoid confusing a printer on the pass-through port.

And this is the power supply and amplifier part:

However, I do not know some of the capacitor values nor the exact type of PNP transistors and only the one and two digit component designators are the same as on the original board.
Anything with a question mark is guesswork.

Doesn't look like it would be too critical anyway. Q3 and Q4 are in the audio path, if I'm following the design correctly. It looks like Q1 and Q2 are a power switch for the LM386. C101 is a bulk cap for the V+ rail. C103 is a decoupling cap for the audio. 10uF should be OK for either, though either one could be fudged by a factor of 10 and would probably work fine. You'd lose bass response if you went too small, but it's not exactly a hi-fi DAC....

I've been interested in picking one of these up to fiddle with, but like a lot of retro stuff, they're going for far more than they're worth on Ebay. Best bet is definitely to RE it, and spend twice that much on manually cloning it. 😈

The most critical part is the totally unknown value of C4 and C6.
Especially C6, which looks like the output low-pass.

C6 is part of zobel network and usually is a 0.1µF film cap.

Tiido wrote:

C6 is part of zobel network and usually is a 0.1µF film cap.

Well, now that you say it: The data sheet does show a capacitor connected to the output with a 10Ω resistor to GND, but all the example circuits use 0.05µF.
So I guess I'll use 0.05µF, as well.

But what is C4? Some kind of filtering for the volume potentiometer, maybe? Wait! That must be part of an input high-pass.

There is already a working clone 😀

shock_ already made a dss clone, now also with stereo support and switchable covox/dss/Stereo mode

and it does not require a MCU 😀

https://dosreloaded.de/forum/index.php/Attach … ovox-krams-jpg/

Do you plan to add covox mode for Stereo ?

matze79 wrote:

There is already a working clone 😀

shock_ already made a dss clone, now also with stereo support and switchable covox/dss/Stereo mode

I'm aware:

Benedikt wrote:

Recently, I realized that there is no real replica of the Disney Sound Source, yet.
There is the thread "Reversing the Disney Sound Source", but the compatible hardware suggested therein looks quite different.

It's a lovely project, indeed, but quite far away from the real thing. Replacement dongles and amplifiers would allow people to mix and match as needed.

matze79 wrote:

and it does not require a MCU 😀

I don't really see how replacing a custom chip with an MCU would be a disadvantage in this case, because the MCU that can technically be programmed via parallel port sits in a parallel port dongle, already.
You therefore wouldn't even strictly need a programming adapter.

matze79 wrote:

Do you plan to add covox mode for Stereo ?

Stereo would not be possible with this approach, but I might be able to include dumb LPT DAC compatibility by programming the ATtiny to just mirror the input while the FIFO is empty.

And this is a schematic of the power supply and amplifier part:

Thanks for your help, keropi!

There have not been any significant modifications since the previous posts, but in case anyone wants to build e.g. the amplifier unit, the design files are now online.