> On Jan 17, 2023, at 3:45 PM, P Gebhardt via cctalk <cctalk@classiccmp.org>
> wrote:
>
>
>
>
> Am Montag, 16. Januar 2023 um 17:57:46 MEZ hat Chuck Guzis via cctalk
> <cctalk@classiccmp.org> Folgendes geschrieben:
>
> Hi Chuck,
>
>> The 844 drives date from the early 70s. I worked for CDC on a military
>> project where these were brought in to replace the 821s that were bid
>> (yes, I know there's no information on those--they're essentially a
>> high-capacity unit build on an 808 chassis and unreliable as hell). A
>> typical installation might have used over 100 of the units on a 4-CPU
>> Cyber cluster. They worked well, unless one got a bad pack, which would
>> clobber the heads on a drive; using the drive on a new pack would result
>> in creating another head-clobbering pack. I recall an overnight report
>> issued by an operator where he succeeded in trashing several packs and
>> multiple drives in his attempt to get something to work. It was a
>> blow-by-blow report somewhat akin to the Gerard Hoffnung bricklayer story.
>
> Wow, thanks for sharing this story!
> Did the 844 drives have the same hydraulic-actuator approach like the MMD 841
> drives?
No, 844 drives use linear voice coils (linear motors), which seems to have
become the norm in the early 1970s. RK05 is another example. The term "voice
coil" is used because they look like oversized versions of the coil that drives
the speaker cone in loudspeakers. These all are some sort of servo device; in
many of them the servo signal is on the media, but not always. For example,
the RK05 uses a pattern of lines on a glass substrate attached to the actuator;
the "fault" light on the drive indicates a failure of the light bulb that
illuminates the servo pattern.
A common feature of linear motor actuators is an emergency retract circuit,
powered by rather large capacitors, to pull the heads off the disk if there is
a power failure.
paul
