Won't you need some sort of a microcontroller to select instrument patches and change perameters and such?

I've thought about that, and although that would make it completely stand-alone, it is a LOT of work. Controlling it from PC is much easier and works just as well. I'll only be using it with the PC anyway.

Here are some pictures of the power supply. It's probably a bit overkill.

Other side: http://img841.imageshack.us/img841/4346/g6bz.jpg

It looks like a mess but that's just what you get when using perfboard.

So you're essentially making it into a synth module then. I see. 😀

That's what I meant, yes 😀

I'm currently attempting to separate the OPL3 signal from the YMF719... again. Now you might wonder what that has to do with this project. Well, I want to use the XR385 as effects processor for the OPL3 only, without the PCM sound.

Now I've done this before, but that was a really quick and dirty setup, and it's a miracle I ever got a decent noise level out of it. I got some better op-amps now, but I can't seem to get much more than about 60dB S-N ratio. Of course, it's all on breadboard now with long wires, and I only have to do as much as wave my hand over it to increase the noise level.

I did manage to measure the sampling rate of the YMF719's OPL3. I'm 99% sure it is exactly 49516Hz, 200Hz lower than the YMF262.

interesting project!
what are your thoughts on housing it?

keropi wrote:

interesting project!
what are your thoughts on housing it?

I was thinking of making a steel case with a wood front panel, with the right tools that's not very hard to do. Unfortunately, I only have very basic tools so I'll have to see how good I can make it look.
If it doesn't end up looking the way I want it, I'll just buy one of those generic cast aluminium boxes used for electronics projects. I'll be aiming to make it about the same size as the X-fi elite pro breakout box, so it'll fit nicely on top of that.

And I need some text on it... But I don't know how. I have an embossed label printer but that's pretty ugly. Anyone has any ideas on that?

I just tried powering the OPL3 circuit with my own power supply (the one I posted earlier). Noise levels went down from about 55-60dB to 70-75dB (and up to 50 when I switch on my unshielded CRT 🤣). Looks like I need a new PSU, and a perhaps mainboard recap.

Here's a pretty "clean" sine wave from the YMF719s OPL3:

As you can see, it's exactly 16 samples long, so calculating the sample rate from that was pretty easy 😀

Here it's playing my Tubelectric remix: http://imageshack.com/a/img22/4865/hgkg.jpg

Also I discovered you can actually make it clip using only 5 channels: http://imageshack.com/a/img40/8318/xq3e.jpg

Some of the noise I'm seeing: http://imageshack.com/a/img41/8940/esbh.jpg (edit: sorry, forgot to resize that one...)
Don't know what this is.. The spikes are spaced about 8ms apart (125Hz).

I'm now using a differential amp design which gets the noise down to -75dB, both with my own power supply and with the PC PSU. I guess that's good enough. Maybe a recap can bring it down even further.

I'll also add a PC speaker input, so I can mix that with the OPL3 signal and send both through the XR385 effects. I just tried that and it sounds pretty neat. Here's a quick example: https://app.box.com/s/zkvgzz0924gg4922dqjh

wow, speaker sounds great!!!
don't forget to post LOTS of pics 😀

Here's what I got for the OPL3 & PC speaker line-out stuff:

Using a differential amp circuit here might seem like an unusual idea at first. But I found that it does help to bring the noise level down. I don't really like it though, as now the noise rejection depends upon how well the resistors are matched together. Maybe someone has any better ideas for that?