Early IIGS bodges
Already a lot of people are using AppleSqueezers now, and I've received a lot of positive feedback, which is great! Thank you so much for all the kind words. However, four users ran into an interesting problem: when they use the AppleSqueezer, it boots irregularly, and whenever they use anything sound-related, they often (almost always) get a blue screen with "Unclaimed Sound Interrupt 08FF". Clearly, not how it's supposed to work! :) While issues like this aren't fun for those affected by it (as well as making me feel bad!), they do present a bit of a mystery which can be interesting to work on. I will show you part of that here!
First, let me start by saying that this issue never happens to the big majority of users, and personally I never had it happen to me on either my ROM03 or my ROM01 machine. Actually, I wish I did, because that would make it much easier to fix! That said, after studying photos received from affected users, I realized that they have something in common: they are all early logic boards with revision 820-0167-B. This number can be seen underneath the CPU, when you remove it, and on the edge of the board. My ROM01 board has revision 820-0167-C, and the problem doesn't happen there.
Something else these boards have in common is situated on the back side of them, where Apple installed a bodge, as can be seen here:
If you zoom in, you can see Apple had to cut a trace on the PCB (marked by the red arrow), and bodge in an RC delay network. Clearly, they had some unforeseen timing issues with this revision, and bodged it up to fix it. Interestingly, when you look at the schematic for the ROM01 IIGS, and compare it with the one for the ROM03 schematic, they indeed added this bodge to the schematic. Let me show it to you here in the ROM03 schematic below. The ROM01 schematic doesn't have this resistor/capacitor there, and connects that R/W line straight up.
Another point is that this bodge is related to the sound (and so is the problem these users are having!) because it's connected to the Sound GLU chip. This R/W line is responsible for telling the sound hardware to either write or read from its registers and memory. Normally, you wouldn't have to delay this line, but perhaps they were dealing with noise issues and it was sometimes inadvertently writing instead of reading, causing issues.
Another interesting thing is that I found several posts from IIGS users in the 90s dealing with this exact same problem (obviously not AppleSqueezer related), many of them referring to the ZipGS. So apparently this issue especially presents itself when combined with accelerators :) For those interested, you can read about them here:
Now, I don't know how common these bodged motherboards are, nor do I know if this sound interrupt issue happens on all of them, or only a subset. Also, considering that the later board revisions (including ROM03) included this fix into their motherboards, ideally they would behave exactly the same way. But possibly they fixed some additional noise issues on those, who knows. It's possible that these bodged boards only barely worked, and as soon as you push them slightly too much in a certain way, they fail. By the way, it's also entirely possible that this bodge has nothing to do with it, and it's something entirely different on this board revision that's causing this.
Now, it's my job to fix it :) And since it's so rare, I can't just go on eBay and buy a random ROM01 machine, hoping the issue would happen on there, so I could experiment with it and fix it that way. So, what I'm left with for the time being is educated guesses, and the help of affected users who can then try out new core versions that may or may not make a difference. I created just such a new core version today, and will send it to some of these users to see if they have any luck with it. If nothing else, it will rule out possible causes, and hopefully bring us all closer to the real fix!
And if you happen to have this board revision, but it works fine for you, I'd also love to hear from you!
For those affected, thank you very much for your patience and help! And if you'd, understandably, rather not wait, then I'm always happy to send you a full refund!
To be continued...