Emulation assumes currently maximum track seek length to be 80 tracks, this is incorrect and over 80 track seeks only plays 80 tracks of seek sound sample.

That's corroborated by the FDC datasheets - any seek beyond 80 tracks needs a second seek.

Also, you added a seek complete interrupt but that's wrong - seeks do NOT interrupts, that's why the SENSE INTERRUPT STATUS command exists.

Edit: I may be wrong about the 80 tracks (one byte would indicate up to 255 tracks, though the PC87306 supports an extended mode with up to 4095 tracks), though that's for seek, recalibrate is limited though, to 77 or 80, depending on the controller.

PCJr required longer seek-times. All systems working correctly now.

Fixed the second FDD motor kept running, when only first drive was accessed.

PCJr still has odd timing issues. Usually it works fine, but sometimes at failes to load disk. I need to tweak the timings more for it.

The motor always runs when the relevant motor ON bit in the DOR register is enabled, at least from what I understand.

This is from the Winbond W83877TF datasheet:

And as I said, Microsoft DMFWRITE.EXE actually relies on that - it turns on both drives' motors at the same time and then switches between them.

I'm not sure if you understand the problem here correctly. Each drive's motor can run separately and individually and can be commanded to do so, as stated by the documentation provided. The issue here is that, when I access drive B: and it starts the motor, then I access the drive A: before the drive B: motor is commanded off, the A: motor starts. At this point both drive's motors are running. If I then keep accessing ONLY the drive A: like for 1 minute, the drive A: motor keeps spinning for that minute, but so keeps drive B: motor. That is incorrect. This is what is fixed now. The problem could be observed in BIOS POST test routine, which caused incorrect spinning behavior in vanillla 86box implementation when using two drives. Also this could be observed in DOS, by accessing drive B: then back to drive a: and then just pressing enter on drive a: prompt would keep drive A: motor running, but it did also keep B: drive motor running.

Without audio cues, you wouldn't notice this at all, unless you'd look in the FDD/FDC logs very closely. It did not affect the technical behavior of the drive but was apparently wrong when sound was added.

Of course, could be that I'm missing the point here.

I just looed at the code changes and... the new DOR behavior seems to be basically the same as before, except for the indexes, and the motor not being disabled if it was already disabled before.

I think the issue was with drive swapping. We were using wrong index at some situations, which would leave some drive motor bit as not cleared.

This or that, based on my tests it functions as expected and the drive motor shutdowns work correctly when there is no accessing the drive for a second or two - this is controlled by the BIOS, the drive index was incorrectly used / determined.

I just looed at the code changes and... the new DOR behavior seems to be basically the same as before, except for the indexes, and the motor not being disabled if it was already disabled before.

My bad. Don't know what happened before, but both drives still seem to spin during boot, even when B-drive is not accessed during that time.

I suspect that this behavior varies from BIOS to BIOS - some switch drive motors on and off when probing different drives, some have them all spinning at the same time.

@Domppari So, for my notes, what's left to be done before this is ready for merger?

I'll add TeacFD 5.25" drive sound samples. Make UI changes to FDD settings, so that you can select what sounds do you want for each drive separately. The selection would include three drives: Generic Mitsumi 3.5" 1.44MB drive, Panasonic 5.25" 1.2MB drive and Teac 5.25" 1.2MB drive. Some small timing tweaking still do be done and additional system testing.

I'll tell you when I think it's ready for merge from my side.

Teac 5.25" drive samples added and implemented drive specific audio selection to FDD settings. Changed the audio sample path to look into roms/samples/fdd instead of '/samples' and root folder.

So for me, the PR is ready to merge at your will. But the samples should be moved to the roms repository and removed from 86box source - if that's ok. Should be do it now, or?

Better now than later imo

CodeQL found a problem here: #6166 (comment) .

I tried this but I can't hear the sounds - how do I enable them?

I tried this but I can't hear the sounds - how do I enable them?

Copy samples to roms/samples/fdd folder and set audio for drive in the fdd settings.