I just completed a set of tests of David's MFM emulator on my Pro 380.
Summary: everything works right. Very impressive device. My compliments to
David for an amazing piece of engineering.
Details:
1. I built and tested it (rev C board) per the instructions and all that worked
nicely. A few minor points in the instructions, quickly clarified by David and
already updated on the web page. I used an old 2 MB BeagleBone Black (the kind
that was shipped some years ago with the abandoned Angstrom distribution), that
fits just fine.
2. I read the three drives I have, one RD51, two RD52 (Quantum). All worked
fine. They were identified by the analysis tool as "Elektronika_85" which
makes sense since that's a Pro clone.
3. On the RD52, the last cylinder cannot be read. The reason is that the DEC
standard formatting tool does not format the last cylinder except on the RD50.
I'm not sure why; the comments say it is "reserved for the FCT" but I don't
know what that is. In any case, ignore those errors; the OS does not use that
cylinder so nothing bad happens.
4. In spite of what the XHomer documentation says, Pro disks have 16 sectors
per track, not 17. It may well be that the drive is physically capable of
holding 17 sectors per track if you have an RQDXn controller, but the Pro
format is definitely 16 sectors. And 4:1 interleaved to account for the
performance issues of programmed I/O rather than DMA.
5. I set up auto-start of emulation mode using one of the files created by the
disk reader. That works fine, the OS boots, identifies the drive type
correctly, and runs happily.
6. I also tested creating a new empty RD52 image (i.e., 512 cylinders, 8
heads), and running a P/OS 3.0 install to that emulated disk. That works also;
as Chris Zach suggested, the installer includes a low level formatter tool and
invokes that automatically when it detects it is needed.
7. Poweroff data saving works nicely. I can watch the BBB keep running after I
turn Pro power off, then after 5-10 seconds the BBB power light also goes off.
8. I looked at the extracted data files created by step 2: they are good block
level data images of the disks.
A note on testing and drive copying: at first I tried to do this using a spare
Pro power supply. That does not work because that supply requires a
substantial minimum current. Even with a real drive plugged in alongside the
emulator, the current draw is too small and the supply shuts down almost
immediately. Instead, I temporarily connected an old small PC-type power
supply I had lying around (rated at 50 watts according to the label); that was
plenty for the emulator and also good enough to power it along with the drive
to be read.
paul
