Amiga Floppy Disk layout

[Giants]

Hi !

In france they are a old edition book from Micro Application : LA BIBLE DE L'AMIGA
They're exist 2 Tome (Tome1 and so, Tome2 that is a the same version with correction of tome1)

in Tome 1 I have this picture :



I think there is an error.
The first tracks (so 0), start on Side0 (which is upper, which is the FACE1 on the picture) and, on Amigados have 11 sectors/side

If we use the sector information, the eleventh (yes because we start from 00->10 = 11 blocs) sector/bloc is the first sector on other side


So for me, it's a mistake in this picture...

Blocs 11-21 must be on Face2 on this picture
and
Blocs 00-10 must be on Face1 on this picture.

If anybody confirm this ?

[phx]

No. Block 0 (the boot block) starts on the lower side of the disk. Block 11 (first sector of track 1) is on the upper side.

[Giants]

Are you sure about this ?

Because when we start a doscopy with X-copy, it start first whith 'UPPER SIDE',
So x-copy don't start in block 0?
Strange isn't ?

With MKIII, a RT 0 2 50000 , go in your way, Start reading Lower side and after Upper.

Why X-copy start from upper side ?
And burstnibler start from lower side... -_-'

And, i think they are not really 'standard' name for Head (0, 1 sometimes even 2).

Best is to talk about 'upper' and 'lower'

[phx]

I'm pretty sure about it, because I wrote a trackloader which sets CIA-B port B bit 2 (DSKSIDE) to select the lower head, whenever I access an even track. And it reads the data I expect it to read.

According to the Hardware Reference Manual DSKSIDE=0 means upper head and DSKSIDE=1 means lower head.

[Giants]

#Block 0 (the boot block) starts on the lower side of the disk.
#...DSKSIDE=0 means upper head and DSKSIDE=1 means lower head.

So Start begin from DSKSIDE=1 !?

[phx]

Originally Posted by Giants

So Start begin from DSKSIDE=1 !?

Yes. Which is the lower side. So your graphic is correct.

[Giants]

OK, so This =>

[ross]

Warning.

I arrived quite randomly on this page (I was looking for a utility for MFM conversion ..) and I noticed that former picture
can be misleading for those that approach the Amiga floppy system.

In the Hardware Reference Manual, as already specified by phx, it's clearly indicated that DSKSIDE*=zero implies the upper head.
But this, like many other CIA's signals, is negated.

So actually to select lower side you need to write 1 to DSKSIDE*.
Sectors count up from cylinder 0 side 0 sector 0 (outer disk position, lower side).
Alternatively Amiga uses tracks (where track 0 is [cyl.side]0.0, 1 is 0.1, 2 is 1.0, 3 is 1.1, 4 is 2.0 and so on).

The correct version:

Blocks 11-21 (track 1)   [Upper head] Side 1 | DSKSIDE*=0
                         --------------------------------    (cylinder 0)
Blocks 00-10 (track 0)   [Lower head] Side 0 | DSKSIDE*=1

[Giants]

Exact !
Thks

[nocash]

Originally Posted by ross

In the Hardware Reference Manual, as already specified by phx, it's clearly indicated that DSKSIDE*=zero implies the upper head.
But this, like many other CIA's signals, is negated.

To me, that's still totally unclear...
The manual doesn't say if "zero" means "0 volts" or "side 0".
The asterix in DSKSIDE* would imply active low, so 0 volts would mean side 1.
But, in my copy of the Amiga Hardware Reference Manual, it's called DSKSIDE, not DSKSIDE*.
And anyways, I would be very surprised if Commodore did actually really know whether they had stored block 0 on the upper or lower side : )
(unless they had a good reason to care about sides, like needing to load the bootblock from single-sided disks).

The only way to be sure would be to remove or disconnect the lower drive head, or somehow block/erase the magnet field on the lower side. And then test which blocks are still accessible.

---

In the UNWARP software, the sides are called TOP and BOT, with the bootblock in block 0 being on TOP side (but there's a 50:50 chance that they got that wrong, and TOP is actually bottom).

The trackdisk device functions are also confusing, ETD_FORMAT does use Offset=(track*2+side)*1600h. But ETD_RAWWRITE seems to use Offset=(track*2)+(1-side). Either one seems to have "side" inverted in software, and it might be further inverted in hardware.

---

Also in the UNWARP software: That program defaults to allow to write to Track FFFFh (Track -1) up to Track 0052h (Track 82). I know that one may access hidden tracks in range 80..82.
But Track -1, is that also possible? Can one issue an extra step, after triggering the track 0 sensor, and before hitting a hard boundary?

EDIT: Warp is using Track FFFFh as dummy track number in Text chunks (added via +TEXTFILE in warp commandline), I guess it does never use Track FFFFh in actual data track chunks.