LisaList2

Quote from: AlexTheCat123 on August 03, 2026, 02:31:27 PMThe updated core is now available on GitHub, so go ahead and flash that to your board if you want to get LOS 1.0 running and unlock full Twiggy support!

Will that update also give Sony 400/800k support under the other ROMs?

Thanks for all of the great work!

Very cool thank you!

As promised when I released the LisaFPGA boards, I've gotten to work on ESFloppy after my break from the project, and I have some really good progress to report.

Sony 400K/800K emulation is fully-working at this point, with reads, writes, and formats tested in every Lisa OS/environment that I could get my hands on. And it appears to be faster than the Floppy Emu too, which has noticably-longer seek times.

Twiggy emulation is also pretty much done. As far as I know, this is the first emulator capable of emulating a Twiggy, so I'm pretty excited about this! Reads, writes, and formats are fully tested and working on Twiggy, and while it can only emulate the lower Twiggy drive right now, the full framework is in place to add upper drive capability so that both can be emulated simultaneously. This should honestly be quite easy to add; it'll likely be done by today or tomorrow.

Twiggy emulation was significantly harder than I expected for a variety of reasons. Even though Twiggy uses the same sector encoding as Sony, just with a different track count and different numbers of sectors per track, it's got some seriously-annoying quirks:

- This one's probably pretty obvious to anyone who's worked with Twiggies before: Given the odd positioning of the heads, when one head is on track T of the disk, the other one is on track 45-T, so you have to read in two separate tracks from the disk image for each seek. And to make matters worse, in the vast majority of cases, each side will have a different number of sectors too...
- Twiggy has a set of "speed sync" bytes that have to be sent out before sector 0 on each track that the controller uses to fine-tune the spindle speed before a format op. This isn't present on Sony at all.
- Twiggy has a weird "track -1" that's used as a timing track during a format. Before the format begins, the controller seeks to track -1 and writes a pattern of 14 sectors, then 15, and then 21, measuring and adjusting the disk speed each time. This requires making a "dynamic track" that allows storage of different numbers of sectors instead of hard-coding it to the expected number of sectors per track like you can for all the others.
- During a format, the Twiggy controller expects each track to take a very specific amount of time between consecutive starts of sector 0. If the track is too long or too short, the format will fail. On an emulator, you generally want to send data out as quickly as possible with minimal padding in between sectors (which works fine on Sony), but on Twiggy you're forced to use very specific (and different) amounts of padding on each track to ensure that the track time is exactly what the controller expects.
- And that just scratches the surface!

Here's a video showing the Twiggy emulation in action; I didn't bother making a video about Sony emulation since that's boring by comparison: https://www.youtube.com/watch?v=8sN7Cv-MONM

The big thing left to do (after I emulate the upper Twiggy) is to make the UI that runs on the OLED. This won't really be hard (the hard part was all of the optimization necessary to get the actual disk routines to run fast enough), but it will certainly be tedious, and will probably take a little while.

During my efforts to get Twiggy emulation working on ESFloppy, I discovered that the hang in LOS 1.0 is NOT because Twiggy drives aren't connected and was actually because of a bug in the LisaFPGA floppy controller. The 6504's stack page was set to the wrong place in the FDC's memory and was clobbering actual FDC data as more and more stuff got pushed onto the stack, causing the FDC to get stuck in weird infinite loops, randomly reset itself, and sometimes fail to respond to the 68K. Moving the stack page to where it was supposed to be fixed the problem, and now LOS 1.0 will boot regardless of whether or not Twiggies are connected! Make sure you have the 40 ROMs selected though; LOS 1.0 will still crash with the A8 ROMs and that's by design.

Oddly enough, this bug didn't expose itself with the A8 ROMs for whatever reason; only with the 40 ROMs. I guess 40 either pushes more stuff onto the stack or stores more sensitive info near the invalid stack location. Who knows?

The updated core is now available on GitHub, so go ahead and flash that to your board if you want to get LOS 1.0 running and unlock full Twiggy support!

I have a profile image that I'd like to keep using, but lisa terminal is not licensed for this lisa.

Is it possible to just reinstall from floppy?

Quote from: TorZidan on July 30, 2026, 05:10:15 PMThe floppy disk interface does have "clamp" and "unclamp" commands
...
I wonder if these commands are sent as is to the twiggy floppy drive, or are they converted by the controller

Clamp and Unclamp are among the small set of command codes sent by the 68K to the Floppy Disk Controller (the 6504 on the I/O Board).

The FDC performs such commands by manipulating the phase wires of the Twiggy carriage's stepper motor and "head select", watching for responses on the "sense" line.

As described above, Clamp and Unclamp are mechanically performed by literally moving the stepper via sequences on the phase wires. Other commands such as checking for Eject button press, write protection, and disk-in-place are done in the same way, but don't move the carriage.

The Unclamp command is also used as the Eject command for the 3.5" drive, although in the 3.5" case, it does not require moving the carriage.

The floppy disk interface does have "clamp" and "unclamp" commands (they are used only on Lisa 1):
https://github.com/arcanebyte/lisaem/blob/master/src/lisa/io_board/floppy.c#L233C1-L234C1 .
LisaEm now supports twiggies. You can add "fprintf" log lines to the code, recompile and run it, use I/O rom "40" in your Preferences, boot LOS 1.0 from a Profile image or boot from a bootable twiggy disk image, and observe what calls are being made from LOS to the floppy disk controller.
I wonder if these commands are sent as is to the twiggy floppy drive, or are they converted by the controller to e.g. "seek track 0".

Compilation image DOES work once 2/10 jumper is installed on the FPGA! 

ried's build works great! Thanks!

Still not getting the LOS Compilation Base.image to work even with the sm fcc180 00ff 0055 00aa PRAM reset.

Going to try the jumper next...

Quote from: AlexTheCat123 on July 29, 2026, 11:54:21 PM

Quote from: bmwcyclist on July 29, 2026, 08:11:58 PMI have used a copy that works on my 2/10 "

That's probably why! Either use an LOS image that was made on a 2/5, or install a jumper wire between GPIO0 and 3V3 on the board to trick it into thinking that it's a 2/10.

Thanks! I love this thing, I am learning quite a bit I never knew about LISA and the LOS!