2.8" QD Disk Emulation

[Jeff]

Ok i have just fixed something. Try this firmware (v3.4.7.6) :


EDIT : Removed !


If you still have the "verify error" message after a write, update to the debug version :

EDIT : Removed !

Copy the DBG_DUMP.BIN to the stick, write to some qd files, and post back the DBG_DUMP.BIN here.
(Note : The debug infos are saved when you change/eject the currently loaded image)

[etaoin]

I still have the verify errors.

I've used the debug version and written to five images

EDIT : Removed !

[Jeff]

I still have the verify errors.

I've used the debug version and written to five images

Thanks : So apparently we have 6 write gate assertions (3 blocks * 2 retries i suppose) and 4 Motor-on assertions.
And most notably : No reset event at all (unless the reset line isn't wired correctly ?)

@AMSynths

I really need to understand how is this short ready is possible on the real QD without the use of the reset line :



Ideally : Three screens on the real QD: One with /RY + /RS and another with /RY + /MO, and finally /RY + WG, with the short ready event fitting the oscilloscope screen.
(Think to increase the oscilloscope sample rate if possible).

[etaoin]

Thanks : So apparently we have 6 write gate assertions (3 blocks * 2 retries i suppose) and 4 Motor-on assertions.
And most notably : No reset event at all (unless the reset line isn't wired correctly ?)

For one disk I did a disk change by mistake while the display still read 'verify error' and it immediately retried the save, maybe it didn't stop the motor there.

I double checked QD pin 8 to pin 20 so that should be ok (unless I missed a change in the wiring somewhere in this thread).

I do have a working real drive as well, but unfortunately only an analog scope so I can't really help you there.

[etaoin]

By the way, I also have a Sharp MZ-800 with a QD drive, which I will try later. The QD interface will probably be identical to the MZ-1500 as the main differences between the two are in graphics and sound.

[Jeff]

Thanks : So apparently we have 6 write gate assertions (3 blocks * 2 retries i suppose) and 4 Motor-on assertions.
And most notably : No reset event at all (unless the reset line isn't wired correctly ?)

For one disk I did a disk change by mistake while the display still read 'verify error' and it immediately retried the save, maybe it didn't stop the motor there.

I double checked QD pin 8 to pin 20 so that should be ok (unless I missed a change in the wiring somewhere in this thread).

I do have a working real drive as well, but unfortunately only an analog scope so I can't really help you there.

This ready behaviour is a mystery if the reset signal isn't asserted... This doesn't match to any of the schematics i have here.

Anyway the probable cause if the the "verify error" message is that the ready signal is asserted for a too long time and the machine start to verify before the track start.

Even with an analog scope it is may be interesting to measure the ready signal assertion time and period on the real QD drive ? Today i am still unsure about the normal Roland QD drive rotation time and ready signal assertion time. I have got many conflicting informations about this so far.

[etaoin]

Even with an analog scope it is may be interesting to measure the ready signal assertion time and period on the real QD drive ?

I'll see what I can do. Biggest issue is lack of memory so any one-time signals will be impossible to measure.

I do have a cheap logic probe though, I'll have a look at what that produces...

[Jeff]

I do have a cheap logic probe though, I'll have a look at what that produces...

yes this may be a better solution !

[etaoin]

This is a write trace.
It actually did three passes, where the third fell just off the 20s window. Not sure why three, it looks like it's only doing a write in the second pass.

Note, this is an A/B sample, so it will ask you to change the disk between saves. That might have something to do with the additional read cycle. I'll do one that is single-sided.

Let me know if you want to see something specific.

[etaoin]

Addition: the first read pass is of course the overwrite check where it checks if there's already something on the disk.

[etaoin]

Write cycle after overwrite check (I waited for the drive motor to stop before confirming overwrite), so this is a write and verify cycle.

So you're right on the Reset line. It doesn't seem to do anything at all.

[Jeff]

ohh thanks a lot ! So this confirm that you have also this short ready state here :

viewtopic.php?f=2&t=773&start=375#p21408

And we can also see the 3 blocks write .

Can you add the others signals on the others channels ? (MO, RD data, WR data...)

Is it on the real QD drive ?

Can you export these data to a file ? i want to check some precise timing.

[etaoin]

Is it on the real QD drive ?

Yes, this is on the real drive.

I was just wiring it up for the full 8 channels.

[etaoin]

Can you export these data to a file ? i want to check some precise timing.

What format do you want?

[etaoin]

full probe