12 Jul
2004
12 Jul
'04
3:34 p.m.
>>
>> I guess there is more to an LLF (Low Level Format) than I realized.
>> I had thought that a completely blank RK05 media was able to be
>> FORMATed via an RK05 drive. Your answer seems to STRONGLY suggest
>> that the media must first be SECTORed correctly.
>>
>> Otherwise, it should be possible to take and RK05 media and do an
>> LLF on the drive using the system on which the media is to be used -
>> and that it would not be a problem if the media originally had an
>> incorrect number of sectors. Based on your answer, the RK05 media
>> must FIRST have the correct number of sectors.
>>
>> Can you please confirm?
>>
It would be VERY interesting if the drive had the capability to change the
number of mechanical "slits" in the hub on the pack.....
More seriously. I wonder how good the rotational stability is on the drive.
Mny of the reasons for using "hard sectored" media was because a) Drives
"wobbled", and b) Accurate timing electronics were "difficult".
If the drive motor speed is stable (over the course of a rotation),
it*MIGHT* be possible to alter a drive to just use the "home" slit and
synthesize the individual sectors. Never tried this with an RK05 [heck
havent touched one of them in over 25 years!], but I have used this
technique to synchronize otgher items.....
12 Jul
12 Jul
4:44 p.m.
On Mon, Jul 12, 2004 at 08:34:22PM -0400, David V. Corbin wrote:
More seriously. I wonder how good the rotational
stability is on the drive.
Mny of the reasons for using "hard sectored" media was because a) Drives
"wobbled", and b) Accurate timing electronics were "difficult".
I'd say it was because at the time, the predominant methods were either
hard-sectoring (which is easy to key off of), or, for multiple platter
drives, a servo *surface*. When you have only a single platter, filling
the bottom surface with servo data kills your capacity.
Also, keep in mind that the RK11-C, an early Unibus RK05 (RK03?) controller
was made of several dozen "FLIP-CHIP" modules. Any design decisions that
could keep that controller as small as possible would be a big win in terms
of manufacturing cost. The RK8E was several quad-height boards (approx one
square foot each).
The RL01 came out a few years later with embedded among the sectors. This
was a pretty good advance in drive technology. It's handy that the same
packs were used with both the RL11 and RL8A controllers. At the time I
first started getting into OMNIBUS PDP-8s (1984), I chose to go with an
RL01 ($150 shipped) and an RL8A (~$600) rather than an RK8E ($450?) and
adding another RK05 to the one I already had. Even in 1984, finding a
16-sector pack wasn't easy, but RL01Ks were available for nearly nothing
(businesses were switching to RL02s) 5MB per drive (RL01) vs 2.5MB (RK05)
was nice, too.
-ethan
--
Ethan Dicks, A-130-S Current South Pole Weather at 13-Jul-2004 01:30 Z
South Pole Station
PSC 468 Box 400 Temp -98.7 F (-72.5 C) Windchill -134.1 F (-92.3 C)
APO AP 96598 Wind 6.8 kts Grid 103 Barometer 670.5 mb (10985. ft)
Ethan.Dicks(a)amanda.spole.gov http://penguincentral.com/penguincentral.html
13 Jul
13 Jul
4:07 a.m.
>>>> "David" == David V Corbin
<dvcorbin(a)optonline.net> writes:
David> More seriously. I wonder how good the rotational stability is
David> on the drive. Mny of the reasons for using "hard sectored"
David> media was because a) Drives "wobbled", and b) Accurate timing
David> electronics were "difficult".
I think the motor is the kind whose rotation speed is tied to the
power line frequency, so assuming normal stable utility power the
rotational stability will be excellent. Also, the motor power is
fairly low compared to the inertia of the disk, so any short term
variations in motor power have only a small effect on RPMs.
David> If the drive motor speed is stable (over the course of a
David> rotation), it*MIGHT* be possible to alter a drive to just use
David> the "home" slit and synthesize the individual sectors. Never
David> tried this with an RK05 [heck havent touched one of them in
David> over 25 years!], but I have used this technique to synchronize
David> otgher items.....
I suppose that's possible. One way to do it would be to build a PLL
that's locked to the start of track pulse, and derive a 12x or 16x
sector clock from that. Or, analogously, you could use a PLL to
derive a 16 sector clock from the 12 sector clock (multiple by 4/3rd,
not hard with a PLL). That's probably a good solution...
paul
4:43 a.m.
On Tue, 13 Jul 2004, Paul Koning wrote:
David> the "home" slit and synthesize
the individual sectors. Never
David> tried this with an RK05 [heck havent touched one of them in
David> over 25 years!], but I have used this technique to synchronize
David> otgher items.....
I suppose that's possible. One way to do it would be to build a PLL
that's locked to the start of track pulse, and derive a 12x or 16x
sector clock from that. Or, analogously, you could use a PLL to
derive a 16 sector clock from the 12 sector clock (multiple by 4/3rd,
not hard with a PLL). That's probably a good solution...
Hm, and then ignore the "real" sector pulses and feeding the drive
electronics the output of the PLL? Smart, that should work, assuming
one can format the packs given this "sector clock".
Would these then be called hard- or soft-sectored? :)
--f
4:58 a.m.
>>>> "Fred" == Fred N van Kempen
<waltje(a)pdp11.nl> writes:
Fred> On Tue, 13 Jul 2004, Paul Koning wrote:
David> the "home" slit and synthesize the individual sectors. Never
David> tried this with an RK05 [heck havent touched one of them in
David> over 25 years!], but I have used this technique to synchronize
David> otgher items.....
> I suppose that's possible. One way to do it
would be to build a
> PLL that's locked to the start of track pulse, and derive a 12x or
> 16x sector clock from that. Or, analogously, you could use a PLL
> to derive a 16 sector clock from the 12 sector clock (multiple by
> 4/3rd, not hard with a PLL). That's probably a good solution...
Fred> Hm, and then ignore the "real" sector pulses and feeding the
Fred> drive electronics the output of the PLL? Smart, that should
Fred> work, assuming one can format the packs given this "sector
Fred> clock".
Yes, that's what I had in mind -- hook the optical sensor to the PLL,
feed the PLL output to where the sensor used to go. The one obvious
tricky part is to generate the track index pulse, because it isn't
equally spaced along with the sector pulses.
Fred> Would these then be called hard- or soft-sectored? :)
Firm-sectored? :-)
paul
7348
days inactive
7349
days old
4 comments
4 participants
participants (4)
-
dickset＠amanda.spole.gov -
dvcorbin＠optonline.net -
pkoning＠equallogic.com -
waltje＠pdp11.nl