26 Feb
2012
26 Feb
'12
9:04 a.m.
Hi guys,
Does anyone know what the typical range of floppy and ST506/ST412 drive
track-to-track seek rates is?
I'm finding myself having to modify the seek logic in the DiscFerret to
accommodate the Seagate ST-277R RLL drive, which requires that
buffered-seek pulses be spaced between 8 and 200 microseconds apart. Any
more than that and it assumes you're doing a 'slow' seek at 10ms per step.
At the moment the DiscFerret's step rates are set up in 250us intervals,
with an 8-bit divider register. Seek rates can be between 250
microseconds and 64 milliseconds in this configuration.
Feeding the ST277R the 250us step pulses... really screws it up. The
drive deasserts READY and SEEK-COMPLETE and seems to freeze up
completely. Hardly unexpected...
If I change the seek clock to 125us, I get a minimum of 125us and a
maximum of 32ms using the same divider.
Is 32ms likely to be enough for even the slowest drives?
Out of curiosity, does anyone know what the slowest-seeking floppy drive
or MFM/RLL hard drive ever made was, and what its track-to-track seek
rate was?
Thanks,
--
Phil.
classiccmp at philpem.me.uk
http://www.philpem.me.uk/
26 Feb
26 Feb
9:17 a.m.
From: classiccmp at philpem.me.uk
Hi guys,
Does anyone know what the typical range of floppy and ST506/ST412 drive
track-to-track seek rates is?
I'm finding myself having to modify the seek logic in the DiscFerret to
accommodate the Seagate ST-277R RLL drive, which requires that
buffered-seek pulses be spaced between 8 and 200 microseconds apart. Any
more than that and it assumes you're doing a 'slow' seek at 10ms per step.
At the moment the DiscFerret's step rates are set up in 250us intervals,
with an 8-bit divider register. Seek rates can be between 250
microseconds and 64 milliseconds in this configuration.
Feeding the ST277R the 250us step pulses... really screws it up. The
drive deasserts READY and SEEK-COMPLETE and seems to freeze up
completely. Hardly unexpected...
If I change the seek clock to 125us, I get a minimum of 125us and a
maximum of 32ms using the same divider.
Is 32ms likely to be enough for even the slowest drives?
Out of curiosity, does anyone know what the slowest-seeking floppy drive
or MFM/RLL hard drive ever made was, and what its track-to-track seek
rate was?
Thanks,
--
Phil.
classiccmp at philpem.me.uk
http://www.philpem.me.uk/
Most of the drives I've seen with slow seeks will also do auto seek.
This is much faster.
Dwight
10:34 a.m.
On 02/26/2012 12:04 PM, Philip Pemberton wrote:
Hi guys,
Does anyone know what the typical range of floppy and ST506/ST412 drive
track-to-track seek rates is?
I'm finding myself having to modify the seek logic in the DiscFerret to
accommodate the Seagate ST-277R RLL drive, which requires that
buffered-seek pulses be spaced between 8 and 200 microseconds apart. Any
more than that and it assumes you're doing a 'slow' seek at 10ms per step.
At the moment the DiscFerret's step rates are set up in 250us intervals,
with an 8-bit divider register. Seek rates can be between 250
microseconds and 64 milliseconds in this configuration.
Feeding the ST277R the 250us step pulses... really screws it up. The
drive deasserts READY and SEEK-COMPLETE and seems to freeze up
completely. Hardly unexpected...
If I change the seek clock to 125us, I get a minimum of 125us and a
maximum of 32ms using the same divider.
Is 32ms likely to be enough for even the slowest drives?
Out of curiosity, does anyone know what the slowest-seeking floppy drive
or MFM/RLL hard drive ever made was, and what its track-to-track seek
rate was?
Thanks,
slowest seeing floppy was the SA400 at 40ms though you could push them
to 35.
the ST506 was a stepper driven seek and if memory serves I used 6MS.
However the SA4000 (8" HD) was slower.
Allison
10:35 a.m.
On 02/26/2012 12:04 PM, Philip Pemberton wrote:
Hi guys,
Does anyone know what the typical range of floppy and ST506/ST412 drive
track-to-track seek rates is?
I'm finding myself having to modify the seek logic in the DiscFerret to
accommodate the Seagate ST-277R RLL drive, which requires that
buffered-seek pulses be spaced between 8 and 200 microseconds apart. Any
more than that and it assumes you're doing a 'slow' seek at 10ms per step.
At the moment the DiscFerret's step rates are set up in 250us intervals,
with an 8-bit divider register. Seek rates can be between 250
microseconds and 64 milliseconds in this configuration.
Feeding the ST277R the 250us step pulses... really screws it up. The
drive deasserts READY and SEEK-COMPLETE and seems to freeze up
completely. Hardly unexpected...
If I change the seek clock to 125us, I get a minimum of 125us and a
maximum of 32ms using the same divider.
Is 32ms likely to be enough for even the slowest drives?
Out of curiosity, does anyone know what the slowest-seeking floppy drive
or MFM/RLL hard drive ever made was, and what its track-to-track seek
rate was?
Thanks,
slowest seeking floppy was the SA400 at 40ms though you could push them
to 35.
the ST506 was a stepper driven seek and if memory serves I used 6MS.
However the SA4000 (8" HD) was slower.
Allison
11:38 a.m.
On 26 Feb 2012 at 17:04, Philip Pemberton wrote:
Out of curiosity, does anyone know what the
slowest-seeking floppy
drive or MFM/RLL hard drive ever made was, and what its track-to-track
seek rate was?
I don't know if it was the absolute slowest, but the early single-
sided Micropolis drives (1015/16) were pretty darned slow--about 30
msec. track to track, if my memory is correct. You would actually
wait a bit to get from track 77 to track 0. I suspect this was due
to the 4-step per track scheme that Micropolis used.
I just dredged up a copy of the driver that I did back in 1978 or so
for a system. The WD1781 r1r0 bits were "11" for the positioning
commands, so it seems as if 30 msec. is correct.
Last-gasp Micropolis floppies actually used a buffered-seek
positioning scheme--but it wasn't enough to save the line.
Even the Memorex 650 drives were only 10 msec. track-to-track.
There may be some super-low-cost drives used in typewriters or game
consoles that were slower, but I'm not aware of any.
That being said, the Micropolis drives were very consistent in their
seeks and were nearly indestructible. And they were expensive.
--Chuck
12:59 p.m.
On Sun, 26 Feb 2012, Chuck Guzis wrote:
I don't know if it was the absolute slowest, but
the early single-
sided Micropolis drives (1015/16) were pretty darned slow--about 30
msec. track to track, if my memory is correct.
Certainly quite slow, but not as slow as 40ms Shugart SA400
You would actually
wait a bit to get from track 77 to track 0. I suspect this was due
to the 4-step per track scheme that Micropolis used.
Yes, but it ended up positioned right for the track.
2:49 p.m.
On 26 Feb 2012 at 12:59, Fred Cisin wrote:
On Sun, 26 Feb 2012, Chuck Guzis wrote:
I'd forgotten about the stupid 35-cylinder SA400 and the cam-follower
scheme. I hated it. It seems odd that anything could be slower than
a 4-step per cylinder positioner, but there you go.
What's strange is that at the time, most, if anot all 8" drives
positioned faster, and some even used a voice-coil setup.
And you're right--after things had worn in a little bit, the
positioner was a little less than accurate. The usual procedure in
case of error was to recalibrate (return to track 0), seek to
cylinder, retry 3 times; recalibrate and then to seek one cylinder
past the target and then step one back, then repeat 3 times; finally,
recalibrate then seek to one cylinder just before the targe cylinder,
then pause for a rev., then step one cylinder in and retry.
All of this to battle the backlash in the positioner mechanism.
--Chuck
I don't know if it was the absolute slowest,
but the early single-
sided Micropolis drives (1015/16) were pretty darned slow--about 30
msec. track to track, if my memory is correct.
Certainly quite slow, but not as slow as 40ms Shugart SA400 
27 Feb
27 Feb
1:07 p.m.
What's strange is that at the time, most, if anot
all 8" drives
positioned faster, and some even used a voice-coil setup.
The Persci? Lovely drives!. Were there any other 8" voice coil floppy
drivces made?
I only ever came across one 5.25" drive with a sort-of voice coil
mechanism (not quite the same mechancially, but very much the same
principle) and that was the Epson 1/3rd height ones used i ntheTF20,
QX10, etc. Again really nice drives.
I have never seen a 3.5" or 3" floppy drive with a voice coil positioner.
Were there any?
-tony
4:16 p.m.
On 27 Feb 2012 at 21:07, Tony Duell wrote:
The Persci? Lovely drives!. Were there any other
8" voice coil floppy
drivces made?
Not that I'm aware of.
I only ever came across one 5.25" drive with a
sort-of voice coil
mechanism (not quite the same mechancially, but very much the same
principle) and that was the Epson 1/3rd height ones used i ntheTF20,
QX10, etc. Again really nice drives.
I have never seen a 3.5" or 3" floppy drive
with a voice coil
positioner. Were there any?
The Imation Superdisk (LS-120/240) and Caleb 3.5" drives use voice-
coil type positioning, but maybe that's not what you meant by "floppy
drive". Still, both accept 3.5" DS2D and DSHD media.
I don't know about Sony HFD.
--Chuck
26 Feb
26 Feb
11:44 a.m.
Hi guys,
Does anyone know what the typical range of floppy and ST506/ST412 drive
track-to-track seek rates is?
I'm finding myself having to modify the seek logic in the DiscFerret to
accommodate the Seagate ST-277R RLL drive, which requires that
buffered-seek pulses be spaced between 8 and 200 microseconds apart. Any
more than that and it assumes you're doing a 'slow' seek at 10ms per step.
A number of hard drives are like that. The worst one I've seen is the
SA4000 (Shugart 14" winchester) where there's an aceptable speed range
for 'buffered seeks' and a slower range for track-at-a-timne seeks. But
if you get between them, the drive will mis-step and end up on the wrong
cylinder. With no warning.
At the moment the DiscFerret's step rates are set up in 250us intervals,
with an 8-bit divider register. Seek rates can be between 250
microseconds and 64 milliseconds in this configuration.
Feeding the ST277R the 250us step pulses... really screws it up. The
drive deasserts READY and SEEK-COMPLETE and seems to freeze up
completely. Hardly unexpected...
If I change the seek clock to 125us, I get a minimum of 125us and a
maximum of 32ms using the same divider.
Is 32ms likely to be enough for even the slowest drives?
The slowest I have seen were some of the full-height 5.25" drives which
had a 30ms seek time, So 32ms would eb OK for those.
Is the problem in providng a greater reage of times the length of the
divider or the numbero f bits needed to sleect the time? If the latter, I
mighty be tempted to have a doubling series onece toy get to, say, 10ms
(that is, have 10, 20, 40, 80 ms only). I can't think that would cause
any problems.
Most single-chipl FSCs only allowed for a few different step rants.
-tony
2:33 p.m.
On 26/02/12 19:44, Tony Duell wrote:
A number of hard drives are like that. The worst one
I've seen is the
SA4000 (Shugart 14" winchester) where there's an aceptable speed range
for 'buffered seeks' and a slower range for track-at-a-timne seeks. But
if you get between them, the drive will mis-step and end up on the wrong
cylinder. With no warning.
Oh, this one has a great failure mode.
If you seek at 300us/step, the drive sits there as dead as a doornail
until you power-cycle it... INDEX pulses, but READY and SEEK_COMPLETE
are stuck inactive...
I suspect it'll probably reset if you do a deselect-and-reselect, though
I haven't tried that yet. What I'd really like is a copy of the ST277R
OEM manual (aka Product Manual)... I bet I'm doing something that it
doesn't like.
The 'seek wraparound' feature is a nice touch. Seek past the maximum
cylinder, and it seeks back to track 0. Bleh. The CDC Wren and Alps 20MB
(rebadged as Amstrad) just flag a seek error and wait for a
deselect/reselect cycle.
Is the problem in providng a greater reage of times
the length of the
divider or the numbero f bits needed to sleect the time? If the latter, I
mighty be tempted to have a doubling series onece toy get to, say, 10ms
(that is, have 10, 20, 40, 80 ms only). I can't think that would cause
any problems.
The main issue is getting enough resolution to specify the 'buffered
seek' step rates while still being able to specify the
tens-of-milliseconds values needed for slower drives.
--
Phil.
classiccmp at philpem.me.uk
http://www.philpem.me.uk/
12:03 p.m.
Out of curiosity, does anyone know what the
slowest-seeking floppy drive
or MFM/RLL hard drive ever made was, and what its track-to-track seek
rate was?
For 5.25" floppy drives:
The Shugart SA400 had a 40ms track to track time.
http://www.swtpc.com/mholley/SA400/SA400_Datasheet.pdf
Anything much faster than that would have "popped the needle out of the
groove" of the SPIRAL (NOT HELICAL) flat disk positioner (designed like a
record (NOT "record" as in a file, but record as in ancient sound
recording)
It would be tempting to not bother supporting that, since the 35 tracks
of the SA400 are a subset of the 40 tracks of any "360K" "standard"
drive.
BUT, the SA400 mechanism (SA390 without the logic board) was what Apple
used for the Apple ][ drive. Eventually replaced by other drives.
The need to recover data from disks that also have bad alignment might
force the need to use the origianl drive.
The Micropolis, with HELICAL LEAD SCREW was a little faster, and MUCH more
reliable.
Later drives, such as the MPI B51 (and eventually Tandon TM100) with split
band positioner were much faster (5ms track to track) and more reliable in
their stepping. Tandon TM100 became the defacto standard for 5.25"
Qume-Trak 142 (5.25" half-height) broke the tradition of quick stepping.
When used on the IBM "Portable" (luggable) and PC-Jr, When used at the
default step times of PC-DOS 2.00, it would get excessive seek failures.
It was actually one of the reasons why PC-DOS 2.00 had to be replaced with
2.10 (with slower step times).
The first hit on Google claims 12ms, but at least the original early ones
definitely could NOT manage better than about 25ms.
--
Grumpy Ol' Fred cisin at xenosoft.com
27 Feb
27 Feb
12:24 p.m.
For 5.25" floppy drives:
The Shugart SA400 had a 40ms track to track time.
http://www.swtpc.com/mholley/SA400/SA400_Datasheet.pdf
That was a _really_ horrible design. Whoever came up with it needs to be
shown the clue-by-four....
Anything much faster than that would have "popped
the needle out of the
groove" of the SPIRAL (NOT HELICAL) flat disk positioner (designed like a
record (NOT "record" as in a file, but record as in ancient sound
recording)
Actually an _ancient_ sound recording might well be a helical track
(think 'wax cylinder' or similar).
-tony
26 Feb
26 Feb
12:54 p.m.
Philip Pemberton wrote:
Does anyone know what the typical range of floppy and
ST506/ST412 drive
track-to-track seek rates is?
When I worked at Apparat around 1981, the
company's main products were
for the TRS-80. Shugart tried to get us to sell a cost-reduced floppy
drive. I don't recall their model number, but it had two major
differences from the normal SA400. It was built on a bent sheet metal
frame instead of an aluminum casting, and instead of a stepper motor for
head positioning, they used a solenoid mechanism similar to an 8-track
player. The direction line on the interface did nothing, and each step
pulse moved the positioner inward by one track, until it reached track
34. The next step pulse would cause the positioner to return to track 0.
If I recall correctly, the normal track-to-track time was specified at
around 400 ms, and the "restore" time at around 1000 ms.
You couldn't use the normal seek and restore commands of a 1771/179x
controller with this, so you had to rely on the step in command. Normal
FDC drivers in TRSDOS, NEWDOS, etc. wouldn't work, so we would have had
to supply a modified OS with it.
Our conclusion was that even though the drive was slightly less
expensive than a normal drive, there would be essentially *zero*
customer demand for it.
I don't think I've heard of a "real" floppy drive with more than 15 ms
track-to-track seek time, nor an ST-506 interface drive with more than 5
ms. There is also a settling time required after a seek, which can be
up to 15 ms, but that isn't required for each step.
1:56 p.m.
> Does anyone know what the typical range of floppy
and ST506/ST412 drive
> track-to-track seek rates is?
On Sun, 26 Feb 2012, Eric Smith wrote:
When I worked at Apparat around 1981, the
company's main products were
for the TRS-80.
One of the things that I remember about Apparat (slightly before that) was
that NewDOS (was it only in NewDos80, or all the way back to APRDOS?) gave
the user the capability to set the step time, to be able to take advantage
of the 5ms step time of the Tandon, etc. drives, rather than have to
always step everything at 40ms to be able to handle the original SA400
(the original drive shipped with TRS80).
Shugart tried to get us to sell a cost-reduced floppy
drive.
"cost-reduced" is a helluva euphemism for POS!
Our conclusion was that even though the drive was
slightly less
expensive than a normal drive, there would be essentially *zero*
customer demand for it.
We did have one administrator pushing HARD to have us switch the entire
TRS80 lab over to using Exatron "Stringy Floppy"! Sometimes people will
put up with amazing things to save a few pennies!
I don't think I've heard of a "real"
floppy drive with more than 15 ms
track-to-track seek time,
SA400 (at 40ms) was still considered to be a
"real" drive in those days,
but not for much longer.
[Did I give you my Micropolis 35 track with one of my model 1s?]
3:18 p.m.
On 02/26/2012 01:56 PM, Fred Cisin wrote:
One of the things that I remember about Apparat
(slightly before that)
was that NewDOS (was it only in NewDos80, or all the way back to
APRDOS?) gave the user the capability to set the step time, to be able
to take advantage of the 5ms step time of the Tandon, etc. drives,
rather than have to always step everything at 40ms to be able to
handle the original SA400 (the original drive shipped with TRS80).
IIRC, NEWDOS/80 allowed the step rate to be changed on a per-drive basis
using the PDRIVE command. In fact, you could change an amazing number of
parameters for the drive or the filesystem format using that command.
I don't think NEWDOS had such a command, though there might have been
ZAPs (patches) to change the step rate globally.
3:37 p.m.
> [Did I give you my Micropolis 35 track with one
of my model 1s?]
On Sun, 26 Feb 2012, Eric Smith wrote:
I don't think so. There was a hole in the monitor
where a floppy drive
had previously been fitted.
I had at one point had a TM100 in that one.
My Micropolis was in a stand-alone horizontal single drive (blue) case/ps
made by Micropolis.
I never got around to trying the Micropolis OS (that had come with it) on
the TRS80 :-(
27 Feb
27 Feb
12:41 p.m.
Philip Pemberton wrote:
Lets see.. Suhgart cam up with that spiral disk positioned that Fred
mentioned
They came up wit hte SA4000 hard disk which has a lovely bit of misdesign
in the logic meaning it will end up o nthe wrong track if you feed it
stepping pulses that are too slow for buffered seeks but too fast for
individual seeks
And then they come up with this....
ARGH!
Does anyone know what the typical range of floppy
and ST506/ST412 drive
track-to-track seek rates is?
When I worked at Apparat around 1981, the
company's main products were
for the TRS-80. Shugart tried to get us to sell a cost-reduced floppy
drive. I don't recall their model number, but it had two major differences from the normal SA400. It was built on a
bent sheet metal
frame instead of an aluminum casting, and instead of a stepper motor for
head positioning, they used a solenoid mechanism similar to an 8-track
player. The direction line on the interface did nothing, and each step
pulse moved the positioner inward by one track, until it reached track
34. The next step pulse would cause the positioner to return to track 0.
That;'s horrible!. DOes anybody have one? I would actually quite like
one, not to use for storign data but to demonstate as a way not to design
things...
If I recall correctly, the normal track-to-track time
was specified at
around 400 ms, and the "restore" time at around 1000 ms.
You couldn't use the normal seek and restore commands of a 1771/179x
controller with this, so you had to rely on the step in command. Normal
FDC drivers in TRSDOS, NEWDOS, etc. wouldn't work, so we would have had
to supply a modified OS with it.
Our conclusion was that even though the drive was slightly less
expensive than a normal drive, there would be essentially *zero*
customer demand for it.
Given tha all copy-protected software would break, any self-booting disk
would break, disk untilites would break, any OS you didn;t supoort would
break, etc, I can understand that!
-tony
3:08 p.m.
On 02/27/2012 03:41 PM, Tony Duell wrote:
SA400(35track) and SA400L (40 track version) and the SA450 (two sided 40
track)
they were slow and ugly in every way. But they were the only
reasonable game
in town in 1977 and for a while after. Step rate was 40ms or if you had
a controller
that could 35ms was as fast as it could go. You had to wait an
additional 40ms for
the head to settle after stepping.
The Tandon TM100 was a major improvement over shugart.
Same for MPI.
The bulk of the 5.25 inch drives never got faster than 6ms/step. The 3.5"
most of them were 6 or even 3mS.
I still run enough non-PC systems with 8, 5.25 and even 3.5" drives to
appreciate
what just a faster step rate can buy in response time if the OS knows that.
Another slow poke hard disk was the SA1000, it's used a stepper.
By time Tandy got into the act with a disk the market had already
started to
turn on shugart for the problematic SA400. But they used it anyway for
cost
and the first year with the EI and disks (never mind DRAM timing issues)
was
seriously painful as the data jitter on some SA400s made disks unreadable
using the 1771 internal data separator. Using the internal 1771 data sep
was stupid to the extreme but the SA400 really was bad.
I was there for all that. It wasn't pretty.
Allison
Philip
Pemberton wrote:
Lets see..
Suhgart cam up with that spiral disk positioned that Fred
mentioned
Does anyone know what the typical range of floppy
and ST506/ST412 drive
track-to-track seek rates is?
When I worked at Apparat around 1981, the
company's main products were
for the TRS-80. Shugart tried to get us to sell a cost-reduced floppy
drive. I don't recall their model number, but it had two major
4462
days inactive
4464
days old
19 comments
7 participants
participants (7)
-
ajp166＠verizon.net -
ard＠p850ug1.demon.co.uk -
cclist＠sydex.com -
cisin＠xenosoft.com -
classiccmp＠philpem.me.uk -
dkelvey＠hotmail.com -
eric＠brouhaha.com