pdp-11/23 config - test-drb@ccmp.vtda.org

12 Apr 2013

Jerome H. Fine <jhfinedp3k at compsys.to> wrote:
...
  Please be aware that most (maybe now almost all) RD53
 drives have failed with the head stuck problem. (...)
 I would agree to use an RD53 as a scratch drive for
 temporary storage, but not for saving files after the
 drive is powered down. 
Which conveniently also is an application for which their
capacity is reasonable. One could however hope for the
best and use one e.g. as a boot drive, provided I have
the contents backed up elsewhere just in case it decides
to croak.

...
  I have a number of Micropolis ESDI and Hitachi ESDI
 drives which seem to be quite reliable.
 (...) since I doubt that using
 the internal power supply from the BA123 is helpful. 
Hmm, as the enclosure supports no less than four full
height 5,25" devices, shouldn't the PSU be able to
power them as well, or is the 11/83 setup stretching
the limits on its own already?

...
  3x 5MB 14"
RL01,
1x 10MB 14" RL02, 
 VERY reasonable for a small RT-11 system 
And PERIOD at least.

...
  but prone
 to hardware problems (especially filters) in the long run. 
A thorough cleanup and check-out will of course be performed
before power is applied.

...
  (2x 80MB
14" SMD) (I just have a lead on those yet)
  The ESDI drives are more reliable and faster.  ESDI controllers
 were also less expensive and more available. 
Not a problem, I do have a (DILOG IIRC) QBUS SMD controller
(but not a SCSI one yet). And I _want_ to run rackmount drives,
not microcomputer stuff, with a rackmount system.
Not to mention that the CMD Phoenix drives have one removable
platter (16MB) in a cartridge, so unlimited offline storage
(in case I can manage to get more of those landmines).

allison <ajp166 at verizon.net> wrote:
...
  On 04/10/2013 02:25 AM, Arno Kletzander wrote:
  ok, then TU58 emulation is, as I had hoped, going
to float the
 boat for me, at least until either more machines or media to
 be read crop up here.  The novelty weill wear off fast as a drive is only 256K,
and
 its serial line speed and hence sloowwwwwwww... 
Nevertheless it's a start; it's not like I'm running a
production system. I can still upgrade once my patience
has been exhausted.

...
    Thats plenty of space for PDP-11 OSs. 
Unfortunately I have no sound way to (mechanically)
 _mount_ and power those tiny little 5,25" disks in
 the '11 rack, that's why they're staying in the BA123.  I got a rack
shelf and gutted a few PCs for the disk cage. 
Perhaps not pretty, but a pragmatic solution anyway...
I do even have a general purpose rackmount case and PSU
for two full-height 5,25" devices after all. (I had not
considered it as such, as it is currently occupied by
a drive bay mountable SCSI RAID controller and I had
originally not realized that they were separate parts.)
So this is basically an option to use the RD53s with my
RQDX3 on 11/23 as well.
How about the STxxx (unshielded) ribbon cables going from
one rack drawer to the other, won't those be a problem in
terms of length/signal quality)?

...
    Likely the bad one has the stuck head problem
 (...)  I'd not want to try my luck, I'll work on that one
 once I've got an aquarium type glove box set up.  Suit yourself.  I've done
them on the kitchen table.
 I must ahve done at least 8 of them.
 The only failures were not media related (controller
 dies on write and blows the track servo).

 Keep in mind they have a filter and the spin-up is
 to clear the platters and the dust that lands into
 the filter. Good to know they're not that fragile, but every head
crash is one too many in my book. They don't make 'em
any more, after all.

[large drives]
...
  Anything I can lift with one hand is small.
 > 1(2)x RD53 5,25" MFM
 > 1x 300MB 5,25" ESDI, 
...
  Physical large drive: any that want a lift or
 two people to install in the rack.
 > 3x 5MB 14" RL01,
 > 1x 10MB 14" RL02,
 > (2x 80MB 14" SMD) (just a lead on those yet) 
...
  RL02 was far more reliable and more common.
Despite the higher track density in an otherwise
pretty much unchanged design? (I got the (common)
user's guide with the drives.)

What's the story about the "brush cycle" that was
later on removed by an ECO anyway? At least one
of my drives still has the brush arm installed.

...
   ?? BBU??
Unless you mean BBS7 (Io tends to use that).
 Battery backup. Feeding it from an alternate 5V source
 to keep RAM content intact when mains power is off. 
 Oh, It's been a while, I never use it.  No need.  If I
 need the data preserved I have it on disk and the system
 on UPS. 
Nowadays, of course. Back then UPS usually meant a large
noisy expensive MG set with a heavy flywheel accompanied
by an air-started diesel genset in the basement (and you
probably had chosen a MINIcomputer to avoid all that), so
BBU was the next best thing to core to have, when power
was not as reliable as your service needed to be.

[heat issues]
...
   A hard
"CPU dead" damage without any warning?  (...)
 Mine ran usually for months at a time, if the building
 AC hadn't failed the cpu baord would still be in there
 (it is actually and repaired returned to my spares box). 
So out of interest, what component(s) actually got fried
in the incident?

[old-time networking]
...
  PDP-11s were networked before NICs using DDCMP and
sync
 lines. (...) Of course Unix had UUCP via serial lines
 and modems.
 After NICS Ether pipe.  RT11 does little with it and
 its mostly useless for that RSTS and RSX in the later
 versions were full fledged nodes (DECnet Phase III). 
So my idea of it was pretty much correct, by and large.

...
   (2x 80MB
14" SMD) (I just have a lead on those yet)

  The ESDI drives are more reliable and faster.  ESDI controllers
 were also less expensive and more available.
 [TK50 tape subsystem]
...
  I just can't say enough nice things about TK50.
 (set user/mode:sarcasm=off)

 But they were the common transfer media and big for its time.
 that and I have a small bin full of Compact Tape I media with
 stuff on it.  Tk70 was the one to have. jkunz and I had a bunch of (negative) fun
with CompacTapes last
VCFe, when we tried reading some on a VAXstation IIRC.
We had a bunch in which the binder had degraded; they started
shedding oxide and/or sticking to the heads, sometimes in a way
that even broke the tape.

...
  There is also TLZ04 tape. DAT, usually regarded
as a "Write once/read never" medium if
I'm not mistaken?

David Riley <fraveydank at gmail.com> wrote:
...
  If you manage to find a QBUS SCSI card (...) you can
use a Zip drive
 for removable storage (...)  It's not exactly "classic",
 but they are plentiful and cheap and fairly reliable if you stick
 with the 100 MB versions.  They're also BIG (in storage size) compared
 to an RL02, and they're much better about being dropped. :-) 
Three problems:
1. no free QBUS SCSI cards (as opposed to MFM, ESDI, RL0x and SMD),
2. bad personal track record with ZIP drives (click of death),
3. going more for a period system than for ease of use.

...
  You could always run your -11 from the BA123. 
It's a bit of
 overkill, but you can fit all the disks you want in there and
 pretty much never worry about running out of slots for boards.

 Of course, if you're actively using the MVII, don't do that. 
(shakes head) Totally wrong approach, I'm basically building
the 11/23 "just" to populate an empty BA-11S shell that has
that designation on it...

The VSII is a very nice system all by itself which I got in
a period configuration (CPU, 16MB RAM, VCB02 board set, RQDX3)
and rounded out accordingly (8-port SLU, DELQA), and as it
fits all together so nicely and works (even has the correct
"VAXstation II/GPX" case sign too), I'm not changing that.

(Neither of those systems is going to see much use anytime
soon, much to my own regret...)

[system monitoring board]
...
  That's a fun idea.  When I'm running my 11/23
in an open backplane,
 I make sure I have a decently powerful fan blowing through it and
 over the top.  Those boards tend to dissipate between 10 and 20
 watts each, which is nothing to sneeze at, and the CPUs can get
 quite locally hot. 
So I probably need two fan airflow sensors (I've seen that done with
components that look like glass case small signal diodes - these work
by dissipating a small amount of power and measuring the temperature
increase from that, so they should detect when either the air flow
fails or the ambient air is too hot to begin with, and a few stick-on
probes (NTC?) for neuralgic components. Adjustable levels for
forewarning and protective cutout; monitoring the DC rails for
overvoltage and excessive ripple could be added for completeness.
Just needs a breaker contact in line with the front panel switch.

...
  The QBUS LAN cards definitely have more of a
"meant for VAX" feel
 to their documentation, IMO, but they do work fine in an -11.

 If you can, I'd recommend getting a DELQA instead of a DEQNA.
 The DELQA is backwards-compatible with the DEQNA and is a lot
 more reliable (mostly due to lower power dissipation).  I'm
 given to understand it takes a lot of load off the CPU as well.
 The DELQA can use the DEQNA cable kit, so no need to worry
 about that. 
Not standard for an '11, so not really of great interest for a
start.

Allison <ajp166 at verizon.net> wrote:
...
  The DEQNA has the mop boot for PDP11. seems that is a
PDP11 thing.
 The DELQA deleted that.

 The CPU load this was added around V5.* VMS as they discovered
 a DEQNA bug that could corrupt packets in a uVAX system so the
 packet has a secondary CRC16 added and does the reverse on
 the other side to check the data and thats all in software (real pita). 
So basically DEQNA (if anything) for an '11, DELQA for a (?)VAX.

...
  The CMD SCSI card is the way to go. 
Nah...SCSI is for Microcomputers, ain't it? ;^)

...
  Get another...  ;) 
If anything, I'd go for the radiator style (BA23) microPDP now,
that's the package I don't have yet. Damn, you made me want
one...

Arno