I tried doing a few things and didn't see any audio issues in windows 98 with any disc swaps recently. I think some recent code changes made it too fast for Windows 98 to see the disc change though, but either way-- can you try this build out when you can? Note you'll need to be logged into github to perform the download. This is a "dev" version.

https://github.com/danifunker/usbode-circle/a … uns/20016432101

Should be available in about 10 minutes. I'm heading to bed now. Thanks again for your help with troubleshooting and testing this. I actually had no idea that re-loading the same disc a few time could fix the issue, so that helped me in trying to get this thing working.

danifunker wrote on 2025-12-08, 04:15:

https://github.com/danifunker/usbode-circle/a … uns/20016432101

No changes, clicks in DOS when changing images, as before.

I forgot to say that if clicks appear, they will also sound on the external DAC when playing test sound from the web interface. Even in winxp and win98. Re-loading images helps in this case too.
Tested the dev build on USB 1.1 and 2.0

@danifunker and/or any moderators:

I assume "ODE" in the topic title is a typo and should be "IDE", right? Can one of you maybe correct that?

Thanks. 😊

USB_Optical_Drive_Emulator. USB interface only, not IDE.

DangerMouse wrote on 2025-12-08, 09:55:

USB_Optical_Drive_Emulator. USB interface only, not IDE.

Ah, okay. Never mind. I assumed thit was something with a Pi Pico interfacing with the IDE bus.

Thanks for clarifying. 🙈

digger wrote on 2025-12-08, 10:53:

DangerMouse wrote on 2025-12-08, 09:55:

USB_Optical_Drive_Emulator. USB interface only, not IDE.

Ah, okay. Never mind. I assumed thit was something with a Pi Pico interfacing with the IDE bus.

Thanks for clarifying. 🙈

Incidentally, the PicoIDE project also exists, and the developers promise it's open source. However, this USB device has its advantages and works surprisingly well on retro computers, despite its interface. At least I no longer have any desire to have an IDE CD emulator for redbook in DOS.

I still want to get to the bottom of this issue.

1) Which usbaspi are you using and version?
2) how are you loading it? Devload or config.sys?
3) how have you been converting multi-bin cue files?
4) Which audio hat are you using?
5) are you on 32 or 64bit of usbode?

Initially I was able to recreate the issue (randomly) but I haven’t been able since making many of these changes.

We had another person in the discord with similar issues a while ago, they ended up going with a different dac and a pi4

1. usbaspi v2.27 now but I tried multiple versions.
2. When testing, I try both methods.
3. I don't use multibin, only one bin and one cue files. Converting in CHD via chdman.
4. Chepest one on PCM5102A, in attach.
5. I tried both.

Pi4 is too big and I'm already all set to build a zero 2w in a case with a volume control.
I also don’t see any problem with changing the disk image several times if clicks appear during the first mounting, because if they don’t appear, they won’t appear in the future.

Are you using a display as well? If not try to disable the display in config.txt (set displayhat=none). This is great information, thanks a lot.

In general I recommend staying with 64-bit.

I'm also going to look at some other coding solutions this week. I'll reply here once I think I figured something out.

displayhat=none was set in web interface, just tried it in config file, nothing changes.

I've been working on this for the past few days non-stop and believe I resolved the issue. Also, if you encounter the issue with bad playback, just re-load the disc again. I've changed the audio initialization logic again, and we are also gracefully clearing out the DMA buffers between every disc swap.

Are you able to test this new release out?

https://github.com/danifunker/usbode-circle/r … ses/tag/v2.16.0

If you think it's valuable, I can look at creating a flag to do an test on the sound system right after we load up the disc (before you execute a game). I also proposed a special patch to the Circle developers, and I will see what they have to say about it, so this might not be an issue much longer.

Everything works perfectly, the clicks during playback are gone, thank you! Your workflow is amazing, I've seen you pushes build after build to github.

There's no need to add the verification flag yet. If any glitches arise, I'll report them.

The only thing missing to be completely happy is a soft-off button on the unused GPIO.

Thanks for the feedback! I think GPIO 23 works with all of the different HATs and configurations we have. Can you verify? I believe that might be PIN 16. But also, when are you rebooting this? If usbode hangs then I don't think the reboot button will work.

What is needed is not a reset, but a soft power off. I managed to burn out one of the microsds while testing: dozens of disconnections from the usb port without turning off the device via the web interface. Now, just in case, I completely turn off logging so that nothing is written to the flash drive.

GPIO23 looks convenient, but ideally it would be possible to set an arbitrary pin through the config.

Maybe I'm overcomplicating things and it's enough to use the standard pads on zero 2 to turn off without a software power off?