ZX81 audio.

[1024MAK]

Yes, the normal practice is that it’s the low to high transition on /WR that clocks the data into the flip-flop/latch. But using a ‘138 as I described does the same thing.

The primary reason I prefer using a ‘138, is that means it’s easier to add additional ‘377 chips for more outputs with no extra components. Farnell (U.K.) have thousands of 74LS138 in stock at £0.19 (excluding VAT / tax) at the moment, so they are not expensive. 1N4148 signal diodes (through hole type) cost between £0.05 to £0.09 (excluding VAT / tax). So there is not that much saving by using the Sinclair type minimal circuit.

No need to slow any of the control signals. Wired AND or wired OR circuits with pull down resistors can sometimes be tricky if using 74 or 74LS chips though, as their inputs tend to float high when unconnected and are harder to pull down to a logic low. Hence using a 74HCT377 may be better (and the resistor can be raised to 10k to reduce the loading on the control signals).

But it’s up to you which you prefer.

Mark

[marste]

@MAK
Sorry, I didn't see your answer on the second page until I submitted this reply!
Regarding circuit with nowdays prices with 0.10$ you can buy a microcontroller more powerful than the full ZX81 with high quality audio interface already onboard! (e.g. https://www.cnx-software.com/2022/10/22 ... -packages/), but I would try to be aligned with the time of the machine . Below I expanded also the not gate with diode logic. A part from the resistor network (that cannot be simplified much) the number of components is really minimal even without the 138

@Moggy
Yes, the quality was not great but the idea is the one of the tracker. The stereo will depends just on the data you'll send to the two channels. You could even play nonsensically two completely different music at the same time one for each channel

-------------

I've done another schema, maybe a bit more clear and probably working:

Components needed (for the stereo the IORQ will be "ORed" with A6 for the CLKEN of the second channel):
- 1 74HCT377 (2 for stereo)
- 1 transistor NPN
- 4 diodes (6 for stereo)
- 3 resistors (4 for stereo) plus the ones needed for the resistor network (should be 16 for channel with the R/2R schema, or 8 for channel with the exponential resistors value)
And surely connector and board

[Moggy]

Yes Stefano I see now.

This is a video of the better known tracker for the spectrum I think.

https://www.youtube.com/watch?v=3MNHP-hBMLI

I think the reason for the lesser sound quality with these trackers is because no DAC is used. Your design using two DACs is able to produce much superior sound quality, this I know from my own experience with Adams juke box.

Whilst i cannot help with the technical side of your project I can however offer some input with the actual audio/samples should you require it. Audio has been part of my life's work for many years and is one of the few things I have knowledge of and especially the last few months working on the audio side of Adam's project. I have real instruments (guitars,drums bass ,keyboards etc ) that could be sampled if you want real instrument sounds.

I hope this is as successful for you as the 1k chess.

[marste]

Yes, this is what I was thinking about! Very nice music, and as usual Yerzmyey is an artist!

4 bits doesn't seem bad, one 377 may suffice even for the two stereo channels in case!

The tracker can even be ported if there are the sources or there is the author around.

In the meantime I had a new idea for the simplification:

3 diodes, one resistor and one simple 74xx IC, this time I think we are at the bare minimum (more can be just to remove the IC I think).

From the datasheet schematics seems that data lines are staying in place enough to allow clock transition to write inside flip flops.

For one week I'll be away, but coming back home I might try my poor soldering skills

[Moggy]

Hopefully the guru's can help with the technical stuff ( my self the audio ), as for Yerz he got married and disappeared from the scene sadly.

[1024MAK]

In the meantime I had a new idea for the simplification:

Oh, I think it’s possible to get even simpler if you don’t mind doing it the “dirty” way…

A5 to /E (pin 1) on the 74xx377
/IORQ to CP (clock pulse)(pin 11) on the 74xx377

No other chips, diodes, or resistors (for the control lines) needed.

Just ensure that the software does not try to input (read) from any Z80 port address with bit 5 low.

Although keep in mind that I have not tested this (yet)…

Mark

[Freya]

https://www.farnell.com/datasheets/91703.pdf

https://www.ebay.com/sch/i.html?_from=R ... 8&_sacat=0

I think these are really interesting... Maybe make someone nd offer to see if you can get a better price!

No harm in trying.

Freya

[Freya]

The texas instruments chip is just a cheap op amp I think, to make sure you can hear it loud and clear.

Freya

[marste]

Oh, I think it’s possible to get even simpler if you don’t mind doing it the “dirty” way…

A5 to /E (pin 1) on the 74xx377
/IORQ to CP (clock pulse)(pin 11) on the 74xx377

No other chips, diodes, or resistors (for the control lines) needed.

Just ensure that the software does not try to input (read) from any Z80 port address with bit 5 low.

Yesss!! Not dirty at all (at least not much more than already in use...)!!
The bare IC, Clive would have appreciated for sure!!

[Freya]

This one is hiding but is only $3 with free shipping and they are accepting offers apparently!

https://www.ebay.com/itm/321744316013?m ... media=COPY

Freya