Agilent 546xx cannot initialize SFRC2DA-board

[linklocal]

hi!

i purchased a "SFRM72-DU26 720K USB SSD" as a replacement for an original TEAC FD-05HF (1.44Mb) slim line FD.

flashed the most recent FF-FW (v3.25) using a serial-connection with a galvanic isolated USB2SERIAL (and an external 5V power supply) . unfortunately my board has no display - but a single LED showing traffic at PB10 (7seg-DISPLAY-DAT line).

regardless of:

• any jumpers set or not
• factory default settings (both buttons)
• a given FF.CFG
• any modifications to the FF.CFG (pin02 / pin36 / host / interface)

at power on
the above mentioned LED is flashing for 2-3s, while in between the usb-stick-activity-led enables for ~1s; followed by a message on the scope:
system concerns detected: floppy drive failed startup initialization
the same message appears, when removing the floppy at all. no probelms will be detected when connecting the old TEAC-FD.

i tried to update to the DEBUG-version: '''alt/logfile/FF_Gotek_Logfile-v3.25.upd''' to see at least a file of some output. i placed this file in a partitioned USB-stick (partition of 1GB fat32) root-directory. but i don't see, that this FW is been touched in any way - when releasing the two buttons - nothing happens. even observing of the current consumption does not show any changes in activity (~83mA for a single board, ~161mA with an USB stick attached) . while waiting for 2min nothing happened.

as an alternative: how could i flash the debug-version with serial?

TEAK data sheet tells something about PIN02 which offers INDEX .. but this is not select able in configuration. but i don't know if this is the marker for the presence of an FD in the system.

my best regards from linklocal

[keirf]

You can download a serial debug build here: https://github.com/keirf/FlashFloppy/actions/runs/782964740 (you must be logged in to Github to see the Artifact download link)

Debugging issues with no display/logging at all is no fun so that's your first mission. I daresay the Gotek PCB has the header for attaching OLED/LCD if you obtained one. Or you should be able to get serial out which is v helpful.

[linklocal]

found a oled-display in my toy-box and immediately saw the message: FF Update Flash v3.25 .

FAQ tells: https://github.com/keirf/FlashFloppy/wiki/Troubleshooting-FAQ remove jumper JA (short to GND). someone found in the issue #155 -> remove resistor: RA. :)

display switched to:FlashFloppy v3.25 and when placing an empty USB-stick ERROR* 34 which leads to an empty USB space - at this time - ok.

lets see what happened when attaching this HW to the scope ...

[linklocal]

jumpered: JE, S1 and none on JD combined with the following FF.CFG the scope is not complaining a missing floppy drive any more.

FF.CFG

interface = shugart
display-type = oled-128x64
display-order = default
twobutton-action = eject

but no disk could be found from scope, so writing a file will fail.

[linklocal]

meanwhile i created a few disks

# mkfs.msdos -C /media/x/xxxx-xxxx/image_00.img 1440

inside FF this img-file is a legal image of floppy. but i don't know if H/C/S:2/80/18 for a 1.44 is sufficient.

# mount -t msdos -o loop /media/x/xxxx-xxxx/image_00.img /mnt

brings up the possible content of that disk.

but the scope is claiming Disk not initialized ... for my point of view: there is a difference to a real floppy-img (which i don't have)

[keirf]

Is the Agilent known to use PC 1M44 FAT formatted disks?

[linklocal]

yes, it uses normal preformated disks.

after all i have to report a first success.

this settings:

interface = shugart
host = unspecified
motor-delay = 10
twobutton-action = eject
display-type = oled-128x64

combined with a 720kb SD-floppy image
mkfs.msdos -C /media/x/xxxx-xxxx/disk_720.img 720

adding motor-delay was the key which made the communication possible

now i have to find the density or !density signal for enabling 1440kb.

[keirf]

Did you do hardware mod to bring the motor line to the Gotek MCU then? Otherwise motor-delay should have no effect.

[linklocal]

i didn't mod the HW. but i had two observations while peeking and poking around. i prepared a couple of different disk-images in sizes and formats on the same USB-memory - non worked. the other observation was booting the scope with a USB-stick connected failed but worked without.

the scope can only do simple things. like writing an image or reading or writing a configuration. but failed in any case - with different error-messages - when stepping through the possible candidates.

at least i configured track-change = realtime and it started to work on the 720kb-image - but writing was very slow.

finally i changed back and for other possible candidates: write-drain = realtime/instant and even motor-delay = 100 which worked and all other switches set to their defaults. so i guessed motor-delay = 10 is the one. even booting with an attached USB-key is now working - which was the reason for the "hello world"-image.

latest available original instrument-firmware sys5462x.jzp, 700188bytes, Apr 25 2005.
while checking the size i come to the conclusion that the 720kb-limit might be correct.

[keirf]

Perhaps it just provides enough delay when first image is inserted.... I wonder if you need instead something like chgrst=delay-3 and another experiment would be interface=ibmpc

[linklocal]

tried interface = ibmpc which didn't work.
in a second guess i set motor-delay = ignore and added chgrst=delay-3 which worked ... in the same way like the above configured stuff.

thank you for the assistance :)

[keirf]

Ok, that chgrst line is the magic :)

[linklocal]

nevertheless of the closed issue - i would like to add the solution for the 720 / 1440kb problem.

the original TEAK FD-05HF disk drive offers to the host the signal HD OUT on pin9. the data sheet tells if: HD-OUT is LOW for 2DD -> 720kb resp. HD-OUT is HIGH for 2HD -> 1440kb - which i wanted to achieve .

this pin is connected to a pullup-resistor PU14 and could be pulled-down only by a jumper: JE - unluckily not by a signal from the MCU. removing jumper JE and offering a mkfs.msdos -C FD-THREE.img 1440 generated image and everything worked perfectly.

with my best regards