Problems with File Operations and Disks....

[Andi B.]

Hello,

I have some problems with my disk drives, controller and/or QL - I don't know
As a controller I use a Sandy Board with 512KB and the drives are a "modern" TEAC (1,44MB) and an old TEAC (720KB; Atari ST). The disks I have are used HD disks, formatted to 720KB. And the QL is a German one.

Now the problem is, that I can't load 'some' files.
It's a bit difficult to reproduce the errors because they also occure by accident.

But to get more details I made some tests with disks formatted on flp1_, flp2_ and on the PC drive (Q-emulator):
A written BASIC program like '10 Print "Test"' could be saved on each disk and loaded from each disk.
So in general it works...

But then trouble begins:
It is sometimes not possible to edit the line "EDIT 10" of the loaded test program - the QL freezes when typing this command.

And I have major problems if using downloaded files (e.g. Cavern and the PSION software) which I unzipped and saved on a disk: The Q-emulator can load and run them but the QL makes trouble.
I can 'dir' such disks on the QL and see the directory but I can only load & run the test program I've written. If I load 'Cavern_bas' or 'clone' it freezes the QL.

So what could be the source of this strange behavior?
Does this looks like a problem with the disk equipment (are the disks worn, the drives old, the IC sockets of the controller oxidised) or is it more likely the QL itself...?

Any feedback & help is welcome!

[Jungsi]

Hallo Andi,
have you taped the hole which is on the HD Disks? When I'am using HD-Disks with old computers I tape the hole and then format to 720k......

[Andi B.]

Yes, I taped them...

[RWAP]

And I have major problems if using downloaded files (e.g. Cavern and the PSION software) which I unzipped and saved on a disk: The Q-emulator can load and run them but the QL makes trouble.
I can 'dir' such disks on the QL and see the directory but I can only load & run the test program I've written. If I load 'Cavern_bas' or 'clone' it freezes the QL.

This sounds as though it could be various issues...

a) Unfortunately taping over the hole in HD disks is not always the best solution - many HD disks formatted to DD just cannot be read in other disk drives - mainly due to the stepping rate and head alignment differences as far as I understand
b) Your best bet may be to format the disks on the QL in the DD disk drive (still tape over the hole), or get some DD disks to try.
c) The crashes after you load a program suggests that the slave blocks are being corrupted in some way - this could be a faulty disk controller, or just that the controller is struggling to read the disk reliably and adding extra information into memory.

[Dave]

In my experience working at Sandy back in 84-85, I never saw this type of problem with a SQB. They either worked or they didn't. The only two components likely to fail were the WD1772 or the GAL and both of those were almost exclusively complete failures.

I did see problems with mismatched drives. Some drives had termination resistors on them, and they would need to be removed from the drive positioned mid-cable. If that was wrong the floppy at the end of the cable could be temperamental. Also, memory problems could cause issues like this too by corrupting the slave block area. The transfers would be fine, but using the copied data would cause issues. Problem is, that doesn't match what's happening with your QL either.

It doesn't help that the memory test on the QL is very weak and it is hard to test the memory properly in a running machine.

Do you have the option of testing the card and floppies on a different QL to rule out the machine? To try a different cable to rule out the cable? Can you remove all the RAM from the SQB to force the slave blocks into the QL's on-board memory and see if that changes or resolves the issue?

Finally, with one of the mechanisms being newer, it will have different step and spin-up times than the other drive. These timings cannot be set per drive, only globally. It may be that the default timings only suit a 1980s mitsumi or citizen mechanism (the mechanisms we used exclusively at that time) and that one of your drives is a lot faster. Try changing the step and spin-up times to a range of values and see if it changes the behavior. That said, the symptom is this would normally be failed reads, not hangs. We never ever put mismatched drives on it though.

[Andi B.]

Thank you for the detailed anwer.

Now it looks like this:

The problem could be the QL itself:
Two times QL Cavern started from a disk (copied with Qemulator to a QL-formatted DD disk.
But only two times

Then I tried to load QUILL's clone from an old cartridge and the same (?) problem occurs:
MDV is running for a second, the screen turns blue and that's it...

Maybe I should clean the QL's IC soccets?