AndrewMartens

I'm not sure if you've run across my Ten Million Projects for old video/computer sound chips that I've been occasionally mentioning around here. I've got potential ideas for POKEY, AY-3-891x (YM2149), etc. Those chips are generally quite straightforward because they are single-chip audio generators. You stuff some data in the registers, you get some sound output - just like the SID. There's programming to do, but the only major things that go onto a PCB are some shift registers, the sound chip itself, and an output buffer. Essentially one can steal much of the circuitry from the MBHP_SID. The 2A03 is unfortunately a completely different beast. I read up on it a while back when I was going to use the 2C02 (PPU) as the graphics chip in a homebrew Z80 computer. While the previously-mentioned chips are standalone audio generators, in the 2A03 the audio circuitry is just part of a modified 6502 CPU. IIRC the sound registers can only be accessed by the 2A03 itself. As you can imagine, this presents a few problems. A basic sound-generating system based on the 2A03 would need (at the bare minimum) the 2A03 and a parallel EEPROM (8kB would be more than enough, 2kB should be fine), and a 2kB SRAM wouldn't hurt. There's also other support circuitry (power supply, clock / crystal, audio output amp, etc.), but we'll just ignore that since it's trivial for the sake of this discussion. Essentially what you need to build is a small single-board computer with some firmware that would read in note data (from shared RAM, or a serial link, etc.) and pass that on to the audio registers. The hypothetical "MBHP_NES" machine would probably require some shared memory between the 2A03 and the PIC. One could still use a shift register connection from the CORE module, but the "NES" module itself would be rather complex - at least compared to the SID module. Not only would you need at least three major chips (2A03, ROM, RAM), you would either need some kind of bus arbiter to allow access to the RAM (by the CORE module, via the shift registers or whatnot), or a dual-port RAM that can be accessed without any bus contention issues. Such a thing is definitely possible, but due to the integrated nature of the 2A03, using it with MIOS is far more complicated than for most of the other old sound chips.

@jdutcher Got them used at a consignment store ... clearly someone had gotten fed up with the flaky buttons and couldn't be bothered to fix them. I picked them up, read about the problem, checked to make sure that they worked (other than the button issues), etc. One of the LCDs is compromised (a little dim due to an internal leak in the display), but it doesn't appear to be getting worse and I can just ignore the problem. @moebius I'm hopefully getting a digital camera one of these days, I'll try to remember to snap some pics of the MMT-8 panelboard, LCD, etc.. I wouldn't mind finding the 1.11 firmware as well, since mine is also a 1.07 unit.

That reminds me, I need to go and fix up the contacts on my MMT-8 and HR-16... picked up the pair for $20 a few months ago, and the only thing that needs doing is making the contacts reliable again... Thanks for the tip, Smash!

Sweet, thanks moebius! I'm going to let other people do development work until I get life straightened out...

Short answer: yes, with a but. Long answer: TK did some experimentation a long time ago, and I've tossed around some ideas. It can be done, but it depends on whether you want to just do a square wave, or a whole wavetable, etc. There are lots of factors to take into account. I mentioned part of my plan (long-term, not likely to be finished this year) here: http://69.56.171.55/~midibox/forum/index.php?topic=3713.0 Discussed a few of the details with TK in the middle of this thread: http://69.56.171.55/~midibox/forum/index.php?topic=3687.15

I have the following soundcards at home with the YMF262 on them: Mediamagic ISP-16 Creative Sound Blaster Pro 2 (CT1600) Media Vision Pro AudioSpectrum 16 Turtle Beach Tropez Rev 2.0 Also when I had been at the local surplus shop, I found the following cards with YMF262 and the YAC512 on them: ESS AudioDrive (no part number, ASIC was labelled ES688F) Aztech (no p/n, ASIC AZT2316A) Creative Sound Blaster 16 (only marking was "Copyright 1992")

jdutcher has obviously done this many times before :-) My preferred prying tool of choice is either a pair of cosmetic tweezers (they have a nice angle which lets you get under the chip before levering up gently) or a small flat-head screwdriver.

I'm going to have to chime in here as well... While we all love boards that are designed by professionals who get paid to do it 100% by-the-book, have labs full of test equipment, etc., that isn't always a reality for DIY projects. If this was a commercial ready-made box that I was paying a bunch of money for (let's say the Sidstation), then I would expect it to have bypass caps / work 100% all the time / etc. However, that's not what I'm going for. What we have is a pretty good design that works well for most people most of the time. Those who want extra features that were omitted (for whatever reason) in the existing design are free to add it in themselves. The rest of us who are happy with something that works well enough, and just want to buy an inexpensive PCB of that circuit... well, we can. If offered the choice of a professionally-designed PCB for $100, and a hobbyist-designed PCB for $10 that has a large community of people using it and helping troubleshoot problems, then I'm going to go for the latter. Why? Partly due to budget, and partly because the SID isn't exactly a crystal-clear chip anyways. Part of the charm is getting some nasty noise along with it! And if I don't like the noise, I can always tweak the design myself. I think that TK has done a great job on all of the MBHP designs. Perhaps he did make some incorrect design decisions along the way, but all hobbyists do. Even a lot of professionals do - it's part of learning. Sometimes one needs to omit features (and possibly reliability) to cut down on cost, which is understandable. I support TK and hope that he continues to work on more cool designs in the future (with or without bypass caps ;) )

I use my Rat Shack 15/30 at home, and it's okay, but for doing anything decent I prefer the WCTPT irons that we have here at work with a nice fine conical tip on them. Good for doing detail work, and I would highly recommend them.

Still working, but as I've mentioned in a few other threads, I barely have time to think about such projects these days - my fiancee has me busy with home decorating / renovating / organizing projects, and there's a wedding coming up in May. My SID is mostly complete - if I can find a weekend to finish it off and put it in a case, then I should be able to get on to breadboarding one of these designs and hooking it up to my spare CORE module. I plan on making this work at some point, but it will probably be quite a while before I get anything going.

Alternately you could buy a few old soundcards that have them onboard, learn to desolder SMT chips on something that isn't the YMF or YAC, and then be very very careful. I picked up a bunch of soundcards before airmailed's bulk order, so that's the route that I plan on taking. It's not for the faint of heart, though - if you don't have complete confidence in your soldering abilities, you definitely won't want to try desolderin ghtme.

TK: parallel transfer is definitely a better idea - I hadn't realized that you had used a parallel interface to the OPL3. A PIC may still be able to handle it - while the frequency registers on the 5503 are 16-bit, the volume registers are only 8-bit (which one would write to with the software-controlled envelopes). Some of the 32 oscillators could be used as LFOs, thus taking that burden off of the PIC. And besides, everyone loves using the PIC! ;-) Hooray! It's a cool project for someone ambitious, that's for sure!

Project is still stalled in the design phase, unfortunately. Lousy home decorating and wedding-related projects are killing all of my free time. The den's almost organized, so I might get back to developing again soon. The hardware end of things should be pretty easy - just loot the design of the MBHP_SID board (power supply, serial interface, output buffer) and modify it appropriately. I made a few notes if anyone is keen on designing it themselves. Unfortunately my ideas tend to be ambitious but I don't have much time to follow through on them.

Hi TK, I did a quick check with P18 and the debugging window, and it was as I figured - the P16PRO40 has all the polarities reversed. So I got out the datasheet for my PIC burner, one for the ProPic2 (which it was a rough clone of), and noted that the key difference between those two was the use of an inverted in the P16PRO40 and just a buffer in the ProPic2. I compared that to the AN589 and Tait programmer which also just have a buffer, and also used a slightly different wiring to control MCLR. Did a few more comparisons, and I determined that I can modify my P16PRO40 to be a Tait-compatible programmer by swapping in a 74LS07 (so that all the controls signals will be the correct polarity) and moving two resistors (to map MCLR properly). That should do it! The '07 just arrived from Digikey today, so with any luck I'll have a chance to modify and test my changes sometime tomorrow evening. [uPDATE]: I took a few minutes at work to desolder the 74LS06, a pair of 4k7 resistors (R10 and R4), and then put in a 14PDIP socket and a resistor across the parallel-port side of the socket. Stuffed a 74LS07 into the socket, connected it up, and behold! the 18F252 I dropped into the ZIF socket was recognized by P18! Once in a while it's nice to have little things like that work out :D

I've previously looked at the possiblity of using the 5503 DOC-chip in a custom synthesizer. Supplies are somewhat limited, though they can be obtained here-and-there for about $15-50. I'm sure that one could create a 5503 synth module interfaced in a similar manner to the SID (a serial shift register that feeds a parallel output for DATA and ADDRESS lines, with one or two additional chip select lines). It would require a closer look at the datasheet and programming interface to determine how easy/complex this would be, but should definitely be possible. You really shouldn't need any reverse engineering to use the chip - just what's provided in the datasheet should be fine. As a starting point, check out the MBHP_SID schematic to see how TK interfaced the SID, and then the MOS6581 datasheet to look at the control registers. Then compare with the 5503. IIRC the 5503 has a 16-bit address bus multiplexed with the 8-bit databus, which may need to be buffered from the input coming from the MBHP_CORE module. I really like that you can load up your own wavetable into memory - that's a pretty cool feature, kinda reminds me of the wave-synth I have on my todo list. The fact that you have to interface up to 128kB of RAM is a bit of a pain, but not impossible - it just makes the design more complex. The 5503 is on my "cool old digital synth chips to play around with one day" list, but that probably means that I won't actually have time to even think about making a design with it for several years. That reminds me, I should probably buy one or two so that I actually _can_ build something with it when I get around to it... Regarding the $500 US for a beat-up IIgs, that's way off the mark. Check out http://www.vintagefunworld.com/items.asp?ID=1 - I bought a SID, two Atari POKEY ICs, an AY-3-8914 and a YM2149F IC off the guy and he was a pleasure to deal with. He has a bare IIGS for $30, or $40 with a keyboard, mouse, and power cable. Buying an old Ensoniq would be instantly more usable, but you could probably get into programming the raw chip more easily with the Apple. One datasheet for the 5503 is at http://www.buchty.net/ensoniq/files/ics1261.pdf - the chip also goes by the name ICS1261, along with 5503, DOC-chip, Q-Chip, etc.

Looks good! Once one of these designs is finished and stable, I look forward to building something - then my secondary PC will finally have MIDI capabilities!

TK, glad to hear you've been having success with it. I've been using IC-Prog in the past with my programmer (the "Kit 96" programmer, aka "P16PRO40", etc.). Sadly I don't think that it's AN589 / Tait compatible and won't work with P18. My German is a bit rusty, so I'm going to have to try to determine whether or not P18 can be made to work with it. Either that, or I'll wait until you finish your new PIC programmer, and I'll just buy one of those boards from Smash :) That's cool that Microchip has a USB bootloader available for the 18Fx550 PICs... I'm looking forward to building a nice USB-based IO setup to mount in a 5 1/4" bay, and whip up some simple Labview code to make a poorman's multi-channel oscilloscope! :) No huge loss that MIOS applications won't work with the USB PICs... it's still nice to just have the MBHP_CORE board available as a base platform for adding extra hardware and whatnot to a system.

It's odd... just the other day I had been looking at this quite independently when I noticed that these new PICs were available, and ordered samples of the 18F4550 and 2550. IC-Prog doesn't appear to support it yet, but it's only a matter of time...

Congrats on hitting the v1.0, TK! Those new samples really go to show the amazing power of the OPL3 - who knew that chip could actually make some cool sounds when programmed properly. I hope that in a month or three SmashTV will have an MBHP_OPL3 kit ready. By that time I might actually be finished with wedding plans (ARGH!) and unpacking (STILL!) from the move, and I'll finally have time to finish off my SID and build one. :-)

Does this also take into account the different transfer rates of the connections (ie, does the Core have an auto-detect function)? Otherwise you're going to run into trouble with MIDI being 31250 bps and serial being... well, whatever your PC's serial port is set to. Or is this just designed as a switch for the physical connection, and you otherwise have a device that does 31250 bps over an RS-232 link?

What an awesome solution to a shortage of user-contributed patches. This is pure TK genius!

@DrBunsen: I haven't looked into the DSPics yet, but I'll check it out. Mostly I was thinking what I could do with parts on hand that I know my cheap PIC programmer can handle :-) @ illogik: one day, some day, I'll have an osc sample available. I got it to pseudocode a while back, but since then I've been distracted by home improvement projects, wedding plans (those last two go together), and Sid Meier's Pirates! when I have any extra time. Part of the home improvement projects includes organizing the den (well under way), which means I'll end up with my studio and electronics equipment setup again.

I got a 130 kB/sec download from Vancouver, BC (west coast of North America), and the forum seemed quite responsive.

Answers to some of your questions, since I can't answer all of them... :-) 4. There's enough variance between units (not to mention your filter capacitors) that you should be able to hear it. At the very least, it will give some character to the stereo sound. 5. POTX and POTY were intended for analog/joystick inputs on the chips. The inputs from POTX and POTY are mapped into the SID's registers. Since the SID is only written to by the CORE module (with a serial shift register), the pins aren't used at all and have no effect on the sound. AFAIK they should be N/C or just tied off to ground, but one should confirm that by checking the schematic.

I'm not quite sure what you mean by "dynamic" here... are you looking for a 2D+pressure sensitive touchpad, like the one in the Moog Voyager? Back in university, I was experimenting with such a beast, made by a company called Tactex. In fact, I was using it to control the cutoff frequency and resonance of the SID (it's the SIDgroove project referenced in TK's SID links). IIRC the Voyager uses a touchpad made by them, but I'm not 100% sure. They have both an OEM product and a commercial touchpad available - see http://www.tactex.com/STC.htm Their products were always quite expensive, though, so I'm not sure if this is exactly what you're looking for.