4 Oct
2004
4 Oct
'04
2:11 a.m.
Does anyone know how an RM05 drive assembly fastens to the massbus
control cab beneath? Doesn't seem likely that the drive assembly just
rests on top - but there's no obvious way of securing the drive/cab
together. I assume something's supposed to thread through the bracket at
the top-front of the control cab and attach to the drive, but what/how?
And then what secures the drive toward the rear?
We almost got out Decsystem 570 completely rebuilt yesterday (it was
totally dismantled and parts have been spread everywhere for the last
few years) - just got the battery backup units and comms brackets to
bolt on, and a couple of minor bits of trim to find and it's done.
I'd stick some pictures up somewhere but I've just realised I have no
idea how to get them off the camera with this new version of linux -
d'oh! :-)
cheers,
Jules
4 Oct
4 Oct
3:45 a.m.
On Oct 4, 2004, at 8:11 AM, Jules Richardson wrote:
Does anyone know how an RM05 drive assembly fastens to the massbus
control cab beneath? Doesn't seem likely that the drive assembly just
rests on top - but there's no obvious way of securing the drive/cab
together. I assume something's supposed to thread through the bracket
at
the top-front of the control cab and attach to the drive, but what/how?
And then what secures the drive toward the rear?
Jules,
The drive assembly and massbus ctlr cabinet are side-by-side. If you
have a
second drive, the controller cabinet is sandwiched in between; there's
space
in the controller cabinet for 2 controllers. Bitsavers has RM05
documentation:
http://bitsavers.org/pdf/dec/disc/RM05
See the picture on the front cover of RM05_Handbook.pdf
Mike (Who scraped his knuckles on DEC stuff too often 15 or so years
ago...)
4:14 a.m.
On Mon, 2004-10-04 at 09:45 -0600, Mike Cesari wrote:
On Oct 4, 2004, at 8:11 AM, Jules Richardson wrote:
Interesting. Maybe we've still got part of our Decsystem 5 to locate,
then :-) (it'll be in the rooms of unsorted stuff *somewhere*, it's
just a case of finding it)
Does anyone know how an RM05 drive assembly fastens to the massbus
control cab beneath? Doesn't seem likely that the drive assembly just
rests on top - but there's no obvious way of securing the drive/cab
together. I assume something's supposed to thread through the bracket
at
the top-front of the control cab and attach to the drive, but what/how?
And then what secures the drive toward the rear?
Jules,
The drive assembly and massbus ctlr cabinet are side-by-side. If you
have a
second drive, the controller cabinet is sandwiched in between; there's
space
in the controller cabinet for 2 controllers. Bitsavers has RM05
documentation: See the picture on the front cover of
RM05_Handbook.pdf
OK, so what we have are three drives looking like the outer pair of
cabinets in that image, although one is marked as dead (I believe it
just needs a new blower motor though, so no big deal). What we haven't
found (yet) is the middle of the three cabinets in that image.
However, all of our cabinets which look like that outer pair are
seperated into two bits - the drive mechanism that sits on top, and the
assembly beneath (with the hinged front door and containing the massbus
sockets - what I assumed was a controller/massbus interface for the
drive above).
There's no clear way in which the drive sits on top of this lower
section; no obvious mounting holes on the underside of the drive unit.
Whilst there are a lot of holes on the top of the lower cabinet and this
bracket at the top-front, there's no obvious way of clamping the drive
to it. I assume there must be some metalwork that we're yet to locate -
the drive presumably vibrates quite a bit and not having it secured to
the lower cab is asking for trouble!
Mike (Who scraped his knuckles on DEC stuff too often
15 or so years
ago...)
Heh, I managed to get blood all over the rack after taking a chunk out
of my finger on it... plus after moving the RM05s, rack bits, CPU, PSUs,
memory crate, RK05 and the comms crate on Saturday and then lifting it
all into place and assembling yesterday my joints are feeling it a
little today :-)
cheers,
Jules
6:22 a.m.
On Mon, 2004-10-04 at 16:14 +0000, Jules Richardson wrote:
On Mon, 2004-10-04 at 09:45 -0600, Mike Cesari wrote:
> On Oct 4, 2004, at 8:11 AM, Jules Richardson wrote:
>
> >
> > Does anyone know how an RM05 drive assembly fastens to the massbus
> > control cab beneath?
Right, I'm confuzzling myself here.
So, the upper part of the RM05 just contains the pack and head assembly,
whilst the cabinet beneath contains just PSU, blower unit, and interface
cables? There's no read/write logic in that lower cabinet (or the drive
above) at all?
Then there's a seperate cabinet (which we need to find) which actually
has the read/write logic for the drives and in turn hooks to the back of
the 570?
What's confusing me is that there are four connectors (dual port,
in/out) inside the cabinet that sits beneath the drive assembly. Those
look to be exactly the same as an output connector on the back of the
DECsystem. That's what was making me think that the back of the
DECSystem plugs straight into one of those connectors on the lower drive
cabinets (I assumed the out on one cab then chains to the input on the
next one etc.), and that there was no 'missing cabinet' and that all the
read/write logic and massbus interface was in the cabinets below each
drive. Seems odd they'd use the same connector for two different
applications (machine to control cabinet, then control cabinet to each
drive)
Yark!
Jules
ps. every Linux pdf viewer I can find takes an age to flip between pages
when looking at docs the size of the ones on bitsavers - to the point
that it makes casual reading completely impossible :-(
7:24 a.m.
On Oct 4, 2004, at 12:22 PM, Jules Richardson wrote:
On Mon, 2004-10-04 at 16:14 +0000, Jules Richardson
wrote:
There should be a card cage for r/w/servo/etc. electronics. Cables
(memory
fuzzy here on details) go to the heads and positioning servo. 2 ribbon
cables
(certain here!) go to the massbus interface. (A *modified by DEC* SMD
interface.)
On Mon, 2004-10-04 at 09:45 -0600, Mike Cesari
wrote:
> On Oct 4, 2004, at 8:11 AM, Jules Richardson wrote:
>
>>
>> Does anyone know how an RM05 drive assembly fastens to the massbus
>> control cab beneath?
Right, I'm confuzzling myself here.
So, the upper part of the RM05 just contains the pack and head
assembly,
whilst the cabinet beneath contains just PSU, blower unit, and
interface
cables? There's no read/write logic in that lower cabinet (or the drive
above) at all?
Then there's a seperate cabinet (which we need to find) which actually
has the read/write logic for the drives and in turn hooks to the back
of
the 570?
Fat round grey (standard massbus cable) between the RH70 and the RM05
massbus
interface.
What's confusing me is that there are four connectors (dual port,
in/out) inside the cabinet that sits beneath the drive assembly. Those
look to be exactly the same as an output connector on the back of the
DECsystem. That's what was making me think that the back of the
DECSystem plugs straight into one of those connectors on the lower
drive
cabinets (I assumed the out on one cab then chains to the input on the
next one etc.), and that there was no 'missing cabinet' and that all
the
read/write logic and massbus interface was in the cabinets below each
drive. Seems odd they'd use the same connector for two different
applications (machine to control cabinet, then control cabinet to each
drive)
Yark!
Jules
Aha!
Sounds like you have an RM03, not an RM05. The RM03 has 6 heads and the
RM05
has 20. In the RM03, everything fits into one cabinet. Not the case for
the
RM05. This is because space in the RM05 is taken up by the deeper disk
pack
enclosure and the larger motor.
Isn't this fun?
Mike
7:58 a.m.
On Mon, 2004-10-04 at 13:24 -0600, Mike Cesari wrote:
Aha!
Sounds like you have an RM03, not an RM05. The RM03 has 6 heads and the
RM05
has 20. In the RM03, everything fits into one cabinet. Not the case for
the
RM05. This is because space in the RM05 is taken up by the deeper disk
pack
enclosure and the larger motor.
Aha - I think you might be on to something there. the chap who actually
knows about this system hasn't touched it since it was dismantled ten
years ago, so his memory's naturally a little hazy, and I'm not a DEC
person at all!
Isn't this fun?
Actually it's a pretty good challenge - the system was *totally*
dismantled and spread over several rooms, so we've done well to get it
all reassembled and (nearly) all the bits found. Currently we're missing
two door hinges and a little bit of plastic trim from the rack front
that attaches with velcro-like pads, but we seem to have located
everything else.
The CPU seemed to fire up OK after we'd tested the power supplies, but
we need the front panel key before we can check operation properly
(luckily that's at someone's house and not lost in storage).
Whether the drives are still operational after ten years of non-use is
another matter (all sound deadening / filters have naturally all
decayed). The system was operated by Shell and provided a multi-user
document system to about 40 users, complete with a primitive form of
hyperlinking - and apparently was totally network-aware and could hook
up with other such systems across a WAN to share information. Would be
interesting to get it booting (we've got all the disk packs, including
engineer tests, and those have all been in sealed storage so should be
OK still)
I'll have to hunt out a 'scrap' disk pack to try initially in case
anything lets go in a big way (worst case we've got sealed original-
stock packs so could sacrifice one of those)
cheers for the pointer - think you're spot-on with the RM03 theory!
Jules
5 Oct
5 Oct
8:26 p.m.
Jules Richardson wrote:
ps. every Linux pdf viewer I can find takes an age to flip between pages
when looking at docs the size of the ones on bitsavers - to the point
that it makes casual reading completely impossible :-(
I was trying to read them at work, in Acrobat on Windows 2000. It was
far too slow to attempt. Printing them out is proving to be a problem,
because some of the page orientations are wrong. The landscape-format
circuit diagrams only print as half a page, because the bounding box is
the wrong way round.
Gordon.
6 Oct
6 Oct
1:27 a.m.
>>>> "Gordon" == Gordon JC Pearce
<gordon(a)gjcp.net> writes:
Gordon> I was trying to read them at work, in Acrobat on Windows
Gordon> 2000. It was far too slow to attempt. Printing them out is
Gordon> proving to be a problem, because some of the page
Gordon> orientations are wrong. The landscape-format circuit
Gordon> diagrams only print as half a page, because the bounding box
Gordon> is the wrong way round.
Turn on the "adjust paper orientation to match page" flag in the print
options.
paul
1:07 a.m.
>>>> "Jules" == Jules Richardson
<julesrichardsonuk(a)yahoo.co.uk> writes:
Jules> On Mon, 2004-10-04 at 16:14 +0000, Jules Richardson wrote:
> On Mon, 2004-10-04 at 09:45 -0600, Mike Cesari
wrote: > On Oct 4,
> 2004, at 8:11 AM, Jules Richardson wrote:
> >
> > >
> > > Does anyone know how an RM05 drive assembly fastens to the
> massbus > > control cab beneath?
Jules> Right, I'm confuzzling myself here.
Jules> So, the upper part of the RM05 just contains the pack and head
Jules> assembly, whilst the cabinet beneath contains just PSU, blower
Jules> unit, and interface cables? There's no read/write logic in
Jules> that lower cabinet (or the drive above) at all?
Jules> Then there's a seperate cabinet (which we need to find) which
Jules> actually has the read/write logic for the drives and in turn
Jules> hooks to the back of the 570?
Jules> What's confusing me is that there are four connectors (dual
Jules> port, in/out) inside the cabinet that sits beneath the drive
Jules> assembly. Those look to be exactly the same as an output
Jules> connector on the back of the DECsystem. That's what was making
Jules> me think that the back of the DECSystem plugs straight into
Jules> one of those connectors on the lower drive cabinets (I assumed
Jules> the out on one cab then chains to the input on the next one
Jules> etc.), and that there was no 'missing cabinet' and that all
Jules> the read/write logic and massbus interface was in the cabinets
Jules> below each drive. Seems odd they'd use the same connector for
Jules> two different applications (machine to control cabinet, then
Jules> control cabinet to each drive)
Sounds like Massbus connectors...
But the docs on bitsavers.org (pdf/dec/disc/RM05) clearly show a
separate "adapter" cabinet. My memory isn't good enough to comment
further.;
Jules> ps. every Linux pdf viewer I can find takes an age to flip
Jules> between pages when looking at docs the size of the ones on
Jules> bitsavers - to the point that it makes casual reading
Jules> completely impossible :-(
I noticed that too. Get Adobe's Acrobat Reader for Linux. That one
does the job right. Ghostview is pretty decent, too, but acroread is
clearly the best. One issue I fought with for a while: acroread was
confused about wanting a UTF font and not finding it. (Why not? I
have lots of fonts installed...) Anyway, defining "LANG=C" (for
acroread, at least, not necessarily globally) fixes that.
paul
1:46 a.m.
On Wed, 2004-10-06 at 09:07 -0400, Paul Koning wrote:
Sounds like Massbus connectors...
But the docs on bitsavers.org (pdf/dec/disc/RM05) clearly show a
separate "adapter" cabinet. My memory isn't good enough to comment
further.;
aha, well it was my fault (well, I'd been fed wrong information as I was
told they were RM05's, plus I'm a DEC novice :-)
They're actually RM03 units, not RM05 - the RM03's don't have the
seperate cab and have the relevant guts inside the cabinets beneath the
drives. Thanks to Mike C. for realising that :-)
So we're not missing a cab after all. That means there's only a little
bit of trim and two door hinges to find and we've got a complete system
again...
Still don't know how the darn drive units mount though. I'll be able to
have a proper flick through some of the pdf files when Adobe's viewer
finishes downloading though.
Jules> ps. every Linux pdf viewer I can find takes
an age to flip
Jules> between pages when looking at docs the size of the ones on
Jules> bitsavers - to the point that it makes casual reading
Jules> completely impossible :-(
I noticed that too. Get Adobe's Acrobat Reader for Linux.
Will do. 8MB download for a PDF viewer though - ouch! Odd that they have
one version for Linux, yet list different downloads for every version of
Windows; I'm surprised that there isn't a 'one version fits all' version
for Windows as there apparently is for Linux.
cheers,
Jules
2:49 a.m.
New subject: Acrobat grrr (was: Re: DEC RM05 mountings)
On Wed, 2004-10-06 at 09:07 -0400, Paul Koning wrote:
Get Adobe's Acrobat Reader for Linux.
Word of warning - the current version (5.0.9) installer is somewhat
broken. When run, it chmods /tmp to 755, changes ownership to an id in
the ten thousand range, and changes group to users. Consequently other
things on the machine get upset rather quickly when they can't write
to /tmp.
So that's nice of it.
I only noticed after I'd tried to get the reader to work post-install;
at first I thought it was the reader doing it, but turns out it's their
installation script. How that one slipped through their testing I'm not
sure!
That one does the job right.
Yes, it runs very quick. Shame it demolishes the rest of the system at
the install stage though :-)
cheers,
Jules
3:28 a.m.
New subject: Acrobat grrr (was: Re: DEC RM05 mountings)
In message <1097074189.18515.110.camel(a)weka.localdomain>
Jules Richardson <julesrichardsonuk(a)yahoo.co.uk> wrote:
Word of warning - the current version (5.0.9)
installer is somewhat
broken. When run, it chmods /tmp to 755, changes ownership to an id in
the ten thousand range, and changes group to users. Consequently other
things on the machine get upset rather quickly when they can't write
to /tmp.
So that's nice of it.
Methinks someone needs to perform the Clue Induction Ritual on Adobe's Linux
AcroRead developer team... Make sure to do it properly - with the Book of
Clue and a Clue-by-four.
As far as brain damaged code goes, I've just spent a good chunk of the
afternoon trying to figure out why my Panasonic Toughbook was ignoring the
Hibernation partition. Seems the BIOS only checks the partition pointed to
by the second entry in the partition table. If it ain't a hibernation
partition, the BIOS assumes there isn't one. Took a good hour (and about half
a dozen repartition/reformat cycles) to figure that one out.
I also learned that DOS FDISK is totally brain-damaged. I ended up digging
out a copy of Ranish Partition Manager and repartitioning the drive manually.
Bleagh.
I only noticed after I'd tried to get the reader
to work post-install;
at first I thought it was the reader doing it, but turns out it's their
installation script. How that one slipped through their testing I'm not
sure!
"Hang on a minute - the installer's clobbering the /tmp directory!"
"Aww, leave it. We're already late for the company picnic, just release it as
it is"
:P
Later.
--
Phil. | Acorn Risc PC600 Mk3, SA202, 64MB, 6GB,
philpem(a)philpem.me.uk | ViewFinder, 10BaseT Ethernet, 2-slice,
http://www.philpem.me.uk/ | 48xCD, ARCINv6c IDE, SCSI
... If you really want to know, you won't ask me.
7334
days inactive
7336
days old
11 comments
5 participants
participants (5)
-
gordon＠gjcp.net -
julesrichardsonuk＠yahoo.co.uk -
mcesari＠comcast.net -
philpem＠dsl.pipex.com -
pkoning＠equallogic.com