I made a couple of variants on my sound card circuit and they all exibited the same perculiar not-quite working characteristic. Upon further investigation I renamed R10 through R17 to R8 throuh R15 and suddenly everything works. Even on the card that's hugely violating the timing specs. Eh!!? Go figure... I'd have thought that the register numbers should match the datasheet. Am I just being thick here or what.

EDIT: OK, I am being thick. It does mention it in the datasheet but I still don't know why the hole in the register names between R7 and R10. Maybe it's a legacy thing being as it was designed originally for a now obscure processor.

Octal designators?!?!?!

Mike B.

Now that's a good point. I can count the number of times I've needed octal in my entire life on one hand if I ignore the time spent just messing around.

As I understand it octal was much more popular in days gone by. Especially considering base and wordlength weren't exactly standard.

Of course! The original datasheet for the AY-3-891x series numbered the registers in octal, probably just to confuse everyone! Newer ones use decimal numbering for the registers. It has been so long since I ran into this problem that I had forgotten all about it!


Octal was especially popular with systems such as the IBM 700 & 7000 series with an 18-bit wordlength. 18 bits fits exactly into 6 octal digits but uses only 4.5 hex digits.

_________________
Shift to the left,
Shift to the right,
Mask in, Mask Out,
BYTE! BYTE! BYTE!

I was told last week that octal also had the advantage that the all-numbers representation on [five-hole] paper tape doesn't need continual shifts between letters and numbers. Triva: "The original CompuServe user IDs consisted of seven octal digits in the form 7xxxx,xx - a legacy of PDP-10." "Yes, the origin of CompuServe user IDs were TOPS-10* PPNs"

If you're still trying to get this working, I have just found a possible solution. In the August 1981 issue of Compute! there's an article and schematic on how to interface an 8/910 to 6502/6800 systems.

_________________
All my stuff

Good find: here's a link https://archive.org/stream/1981-08-comp ... 1/mode/2up

Thx BigEd for that link, I was going to scan the article if DigitalDunc was still interested, wasn't sure if it was available online. Now I can fill the holes in my collection.

_________________
All my stuff

I got it working alright. I ended up with three versions. One which operates outside the AY's spec (I think I need to go pray to the gods of magic smoke on that one), one which uses a VIA (even though I didn't want to tie up 16 locations) and one that uses a CPLD (Overkill). Honestly, the one with the VIA allows you to break other timing specs etc but scales with processor clock easiest. I hadn't thought about it before but when WDC VIA is reset I never needed to use all the registers for this application and so I could just not map them. I might be wrong though.

Interrupts are giving me a headache at the moment and I was trying to use the systems other VIA to generate a steady tick for use in such things as music, but it looks like I have an issue there which is going to suck a good solid and more important **FREE** weekend to crack. I'm up by half four this weekend bailing my company out again so no play 'till next WE.

Oh and Archiving is a very valuable thing. What we don't protect properly will die with the last of us.

Well, I got back from work and even though I felt cream crackered I couldn't resist a little tinker. Found out why my ISR didn't work (missing INX in the BRK detection). Now is it worth writing an ISR in cc65 or sticking to assembly?

Sorry to bring up an old topic.

I'm a HUGE fan of the SID, AY-3-8910 (and variants), SN7648 and just about every other PSG from the 80's/90's.

I've built several AY/SN circuits using real PSG's and Arduino's. I have dozens of these things including many YM series, etc.

I've also built a couple Mockingboard clones on a breadboard and got it working in my Apple IIe.

Anyway, I'd like to mention another alternative. Something I plan doing myself with my 65C02 computer.

And that is to use the Parallax Propeller for audio. The Propeller has *VERY* accurate SID, AY-3-8910 and SN76489 emulation using only ONE COG each. A "COG" in the Propeller is a dedicated core. There are 8 in the Propeller and each one runs independently of the others.

The Propeller comes in 40 pin DIP and only requires a small EEPROM and external crystal to run it (along with a serial programmer that can be bought for $14 if you don't already have one).

I can talk for hours on the Propeller.

So, what I plan on doing is having the 65C02 talk to a Propeller and *hopefully* access it like the vintage parts. In fact, you could devote another COG to translate your OWN addressing scheme! You OWN timing, etc.

You could have, for example, 2 COGS devoted to AY's (6 channels in stereo), 2 COGS for SID (another 6 channels!), etc.

That's my plan, anyway.

_________________
Cat; the other white meat.

Good plan.. I have a number of DIP Propeller chips in a drawer somewhere (in addition to various boards and miniboards). Very easy to interface, very low threshold for doing something useful with them.
SID emulation on propeller: https://www.youtube.com/watch?v=EJcbxrdErkY