I'm not sure what you want them for but the 8430/8433 disks were word channel disks (IIRC) and the manuals for those are on bitsavers:
http://www.bitsavers.org/pdf/univac/disk/8405_8430_8433/
I would expect the command codes and CCW formats to be more-or-less the same between the various word channel disks.
The only Uniservo PRM that I can find is for the U6C which is ancient. The basic command codes are likely the same but this drive likely to be from the 1107 era or there about. If the language seems a bit odd it is because it was translated from aGerman manual using Google Translate and what little I could remember of my high school german.
https://www.dropbox.com/s/xuv9w3p2yroejff/Uniservo%20VI%20C%20Technical%20Description.rtf?dl=0
Enjoy
I also found an manual for an FH880 drum. Not specifically the one you were looking for but ...
chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/http://www.bitsavers.org/pdf/univac/1107/UT-2560_FH-880_Drum_Dec61.pdf
I'm not sure what you want them for but the 8430/8433 disks were word channel disks (IIRC) and the manuals for those are on bitsavers:
http://www.bitsavers.org/pdf/univac/disk/8405_8430_8433/
I would expect the command codes and CCW formats to be more-or-less the same between the various word channel disks.
The only Uniservo PRM that I can find is for the U6C which is ancient. The basic command codes are likely the same but this drive likely to be from the 1107 era or there about. If the language seems a bit odd it is because it was translated from aGerman manual using Google Translate and what little I could remember of my high school german.
https://www.dropbox.com/s/xuv9w3p2yroejff/Uniservo%20VI%20C%20Technical%20Description.rtf?dl=0
On 9/5/2022 8:42 AM, sboy...@gmail.com wrote:German manual using Google Translate and what little I could remember of my high school german.
I'm not sure what you want them for but the 8430/8433 disks were word channel disks (IIRC) and the manuals for those are on bitsavers:
http://www.bitsavers.org/pdf/univac/disk/8405_8430_8433/
I would expect the command codes and CCW formats to be more-or-less the same between the various word channel disks.No. The early disks used an IBM style controller, interfaced to a word channel via a device called a Multi-Systems Adapter (MSA). This device essentially presented a word interface to the channel, but an IBM style channel interface to the peripheral controller. This was used to allow Univac to use peripheral devices available for IBM systems. This
included disks, such as the ones listed above, tape drives, and even a
few printers such as the 770. This is the interface described in the
above referenced manuals.
Later (1977?), Univac developed their own word channel disk controller,
the model 5046 (later the 5056, a compatible but faster controller).
This implemented a totally different (and much more advanced) interface
to the word channel (including things like word addressing and automatic read before write in the controller), but it still supported disks like
the 8430/8433, as well as newer disks like the 8434, 8450, 8470, etc.
Even later, the 5057 controller was the caching controller, and had a different interface yet.
The only Uniservo PRM that I can find is for the U6C which is ancient. The basic command codes are likely the same but this drive likely to be from the 1107 era or there about. If the language seems a bit odd it is because it was translated from a
https://www.dropbox.com/s/xuv9w3p2yroejff/Uniservo%20VI%20C%20Technical%20Description.rtf?dl=0As I mentioned above, Univac (at least later) tape systems were IBM compatible tape systems attached through an MSA. If you look at the
manuals referenced above, but substitute the tape commands (from an IBM
or StorageTech tape system manual) for the disk commands in the MSA
command words, you will have most of it (except for the character translation and formatting stuff, e.g 6 bit Fieldata to/from 8 bit
EBCDIC - used for tape data interchange with IBM systems).
--
- Stephen Fuld
(e-mail address disguised to prevent spam)
On Monday, September 5, 2022 at 9:58:44 AM UTC-7, Stephen Fuld wrote:German manual using Google Translate and what little I could remember of my high school german.
On 9/5/2022 8:42 AM, sboy...@gmail.com wrote:
I'm not sure what you want them for but the 8430/8433 disks were word channel disks (IIRC) and the manuals for those are on bitsavers:No. The early disks used an IBM style controller, interfaced to a word
http://www.bitsavers.org/pdf/univac/disk/8405_8430_8433/
I would expect the command codes and CCW formats to be more-or-less the same between the various word channel disks.
channel via a device called a Multi-Systems Adapter (MSA). This device
essentially presented a word interface to the channel, but an IBM style
channel interface to the peripheral controller. This was used to allow
Univac to use peripheral devices available for IBM systems. This
included disks, such as the ones listed above, tape drives, and even a
few printers such as the 770. This is the interface described in the
above referenced manuals.
Later (1977?), Univac developed their own word channel disk controller,
the model 5046 (later the 5056, a compatible but faster controller).
This implemented a totally different (and much more advanced) interface
to the word channel (including things like word addressing and automatic
read before write in the controller), but it still supported disks like
the 8430/8433, as well as newer disks like the 8434, 8450, 8470, etc.
Even later, the 5057 controller was the caching controller, and had a
different interface yet.
The only Uniservo PRM that I can find is for the U6C which is ancient. The basic command codes are likely the same but this drive likely to be from the 1107 era or there about. If the language seems a bit odd it is because it was translated from a
As I mentioned above, Univac (at least later) tape systems were IBM
https://www.dropbox.com/s/xuv9w3p2yroejff/Uniservo%20VI%20C%20Technical%20Description.rtf?dl=0
compatible tape systems attached through an MSA. If you look at the
manuals referenced above, but substitute the tape commands (from an IBM
or StorageTech tape system manual) for the disk commands in the MSA
command words, you will have most of it (except for the character
translation and formatting stuff, e.g 6 bit Fieldata to/from 8 bit
EBCDIC - used for tape data interchange with IBM systems).
--
- Stephen Fuld
(e-mail address disguised to prevent spam)
Thanks for the reply (and for confirmation about what I recalled about peripheral subsystems).
As for 6-bit to 8-bit translation, I'm not particularly interested in that as it was (IIRC) a non-issue for any of the older 1100s with IOPs thanks to the B-format flag (?) in their CCWs.
On 9/5/2022 10:13 AM, Lewis Cole wrote:
<snip>You're most welcome.
Thanks for the reply (and for confirmation about what I recalled about peripheral subsystems).
As for 6-bit to 8-bit translation, I'm not particularly interested in that as it was (IIRC) a non-issue for any of the older 1100s with IOPs thanks to the B-format flag (?) in their CCWs.Ah you young whippersnappers! [...]
[...] There were successful 1100 series systems
before the advent of IOPs (which IIRC came in with the 1100/80). Specifically, the 1108, 1106, and 1110, and their solid state memory counterparts, the 1100/20, /10 and /40 had only word channels, and no
IOPs (the 1110 IOACU was essentially just a way to repackage channels separate from the CPU).
These systems didn't have CCWs at all, [...]
[...] but used
different instructions (Load Function in Channel, Load Input Channel,
Load Output Channel, and a few related instructions) to instruct the
channel to do I/O. [...]
[...] So, the 6/8 bit character and Fieldata/EBCDIC
conversions were done in the MSA.
--
- Stephen Fuld
(e-mail address disguised to prevent spam)
On Monday, September 5, 2022 at 5:39:40 PM UTC-7, Stephen Fuld wrote:whippersnappers) and I'm content to live with this fiction.
On 9/5/2022 10:13 AM, Lewis Cole wrote:
<snip>You're most welcome.
Thanks for the reply (and for confirmation about what I recalled about peripheral subsystems).
As for 6-bit to 8-bit translation, I'm not particularly interested in that as it was (IIRC) a non-issue for any of the older 1100s with IOPs thanks to the B-format flag (?) in their CCWs.Ah you young whippersnappers! [...]
Guilty as charged.
[...] There were successful 1100 series systems
before the advent of IOPs (which IIRC came in with the 1100/80).
Specifically, the 1108, 1106, and 1110, and their solid state memory
counterparts, the 1100/20, /10 and /40 had only word channels, and no
IOPs (the 1110 IOACU was essentially just a way to repackage channels
separate from the CPU).
Yes, the 1108 started it all (thankfully for all the paychecks that I was to collect for years afterwards).
But I showed up around the time the 1100/60 was getting ready to go out the door and my "home" system at the time was an 1100/80A.
I could have used a lonely 1108 that sat in the corner of the production floor as my own "personal computer" if I wanted to, but I never did.
Nevertheless, by that time, no matter how the actually appeared as hardware, us young members of the software priesthood could treat them the same as IOUs since that's where they appeared in the MCT (that's Master Configuration Table you even younger
These systems didn't have CCWs at all, [...]
Well, IIRC they had ACWs and I *THINK* there was at least a flag for data chaining so things were at least beginning to evolve towards the light.
[...] but used
different instructions (Load Function in Channel, Load Input Channel,
Load Output Channel, and a few related instructions) to instruct the
channel to do I/O. [...]
There was some sort of nursery rhyme I was supposed to learn long ago ... "Ticks and Tocks, Licks and Locks ... ".
I'm afraid I didn't bother to memorize it though as I liked to just say "SIOF!" and later "SEND!".
[...] So, the 6/8 bit character and Fieldata/EBCDIC
conversions were done in the MSA.
Well, being a whippersnapper, I didn't bother to care about tape labels until the 1100/90 when the existence of ICLP made it possible to have labelled boot tapes.
Thanks for your recollections of The Good Old Days.
On 9/5/2022 7:31 PM, Lewis Cole wrote:< snip >
Yes, there were ACWs - "Access Control Words <snip>
BTW, if you are interested in all the gory details, the 1108 Processor
and Storage Reference manual is on Bitsavers.
There was some sort of nursery rhyme I was supposed to learn long ago ... "Ticks and Tocks, Licks and Locks ... ".
I don't remember that. If you remember more, please do tell. Load
Input Channel was indeed called "lick" as a "shortcut" as its
instruction name was LIC. Similarly for LOC. But TIC (transfer in
channel) was, in fact I believe still is, the channel instruction
"Transfer in Channel", but perhaps it just meant the ticking of the
clock. :-)
On Wednesday, September 7, 2022 at 10:50:56 PM UTC-7, Stephen Fuld wrote:
On 9/5/2022 7:31 PM, Lewis Cole wrote:< snip >
Yes, there were ACWs - "Access Control Words <snip>
BTW, if you are interested in all the gory details, the 1108 Processor
and Storage Reference manual is on Bitsavers.
Thanks for the pointer.
I'll keep it in mind although ISTM that the 1108
IIRC, if I really wanted to know how word ESI worked, I had to look at the earlier 1110 (or was it 1108?) manuals to find out anything ... the 1100/80 manual was virtually useless.
Since the removal of the IOP information from the M-Series Processor and Storage manual, the obfuscation seems complete.
But getting back to my OP, I'm more interested in peripheral subsystems.
I can understand the disappearance of anything having to do with word channel tapes as they didn't seem to last too long, but drums and disks lasted much longer.
Of course, you could have pointed out that I messed up the model numbers of the drum subsystems and didn't mention a controller number (all of which is true), but even using the correct model numbers doesn't get me anywhere either.
Despite their usage on 1100/2200 Series systems, the most information about them (from a programmer point of view) comes from 494 documentation.what can/should be done.
I'd like to think that all I'd have to do convert that information (in particular the EF and EI formats) is to pad/strip 6-bits to the high order end of a 30-bit word to come up with the 36-bit equivalent, but I can't find anything that says that is
But disk subsystems are a different matter.
They existed and were used for quite some time, but I can't find anything about them.
Even the 1100/50 word disk stuff (DSACs?) which were an integrated word channel and 5046/5056 CU just doesn't seem to exist as far as documentation goes.
Your summary of the history of the 5046/5056/5057 is about as good as anything I've run across about them.
On Friday, September 16, 2022 at 10:08:43 AM UTC-7, Stephen Fuld wrote:
On 9/8/2022 12:19 PM, Lewis Cole wrote:
On Wednesday, September 7, 2022 at 10:50:56 PM UTC-7, Stephen Fuld wrote: >>>> On 9/5/2022 7:31 PM, Lewis Cole wrote:
BTW, if you are interested in all the gory details, the 1108 Processor
and Storage Reference manual is on Bitsavers.
Thanks for the pointer.
I'll keep it in mind although ISTM that the 1108
1100/80 ?
was the start of the Univac/Sperry/Unisys to keep outsiders in the dark
when it comes to how I/O was done.
But getting back to my OP, I'm more interested in peripheral subsystems.
I can understand the disappearance of anything having to do with word
channel tapes as they didn't seem to last too long, but drums and disks
lasted much longer.
Yes. Drums were on the way out by the time of byte channel
availability.
Of course, you could have pointed out that I messed up the model
numbers of the drum subsystems and didn't mention a controller
number (all of which is true), but even using the correct model
numbers doesn't get me anywhere either.
I forget the model number of the drum controllers (IIRC there were three
of them, one for the FH drums, one for Fastrand drums, and one for the
8460 (actual physical disks emulating Fastrands). From memory, each had
a one word EF that gave the function, the unit number and the starting address. The length wasn't explicitly sent to the controller. The
transfer stopped when the host stopped sending/taking words for a
specified (pretty short) amount of time. This happened when the count
in the ACW was exhausted. The status word was also one word.
But disk subsystems are a different matter.
They existed and were used for quite some time, but I can't find anything about them.
Even the 1100/50 word disk stuff (DSACs?) which were an integrated word channel
and 5046/5056 CU just doesn't seem to exist as far as documentation goes.
I know nothing about the 1100/50.
Again, from memory (as I can't find my copy of the 5046 manual). Most External Function words (EFs) were two words long. Exceptions were
things like Sense and Read IPL which were one word long.
The first word contained the unit number (4 bits), the function (e.g. read/write/preseek) 6 bits (?), and (for the read and write) the
transfer length in the low order 16(?) bits. The preseek was the same
except instead of the length, the low order two bits gave a value about
how much time to allow between the seek completion and the expected time
of the disk's rotation to reach the start of the requested record.
On Friday, September 16, 2022 at 10:08:43 AM UTC-7, Stephen Fuld wrote:
On 9/8/2022 12:19 PM, Lewis Cole wrote:
On Wednesday, September 7, 2022 at 10:50:56 PM UTC-7, Stephen Fuld wrote: >>>>> On 9/5/2022 7:31 PM, Lewis Cole wrote:
< snip >
BTW, if you are interested in all the gory details, the 1108 Processor >>>> and Storage Reference manual is on Bitsavers.
Thanks for the pointer.
I'll keep it in mind although ISTM that the 1108
1100/80 ?
was the start of the Univac/Sperry/Unisys to keep outsiders in the dark
when it comes to how I/O was done.
No, I meant 1108.
While the Company seemed to try to document everything in sight in the Good Old Days (in particular in the Processor and Storage manual) in the beginning,
I think that changed ... I don't recall them mentioning anything about the designator bit that allowed the 1108 to address beyond 262K.
The implications for I/O seem obvious to me.
If you want to say that this was covered in something other than the R1 manual (UP-4052) that I happen to have and that I should RTFM, I'll accept your observation defer to your expertise.
< snip >
But getting back to my OP, I'm more interested in peripheral subsystems.
As luck would have, someone recently sent me a copy of the PRM for the 5056/8470 (UP-9122).
I won't mention who sent me the manual so they don't get inundated with requests for old manuals, but thanks a lot, Masked Man.
Thanks to everyone out there who helped.
On 9/19/2022 1:35 PM, Lewis Cole wrote:
On Friday, September 16, 2022 at 10:08:43 AM UTC-7, Stephen Fuld wrote: >>> On 9/8/2022 12:19 PM, Lewis Cole wrote:
On Wednesday, September 7, 2022 at 10:50:56 PM UTC-7, Stephen Fuld wrote:
On 9/5/2022 7:31 PM, Lewis Cole wrote:
< snip >
BTW, if you are interested in all the gory details, the 1108 Processor >>>> and Storage Reference manual is on Bitsavers.
Thanks for the pointer.
I'll keep it in mind although ISTM that the 1108
1100/80 ?
was the start of the Univac/Sperry/Unisys to keep outsiders in the dark >>> when it comes to how I/O was done.
No, I meant 1108.OK. I never looked at any manuals for systems earlier than the 1108.
While the Company seemed to try to document everything in sight in the Good Old Days (in particular in the Processor and Storage manual) in the beginning,
I think that changed ... I don't recall them mentioning anything about the designator bit that allowed the 1108 to address beyond 262K.Correct, but that is because it didn't exist! The 1108 could only
address of to 262K. There was a third party company that offered to
upgrade an 1108 to 524K by making use of the bit documented as not used
in the 1108 P&S manual, but wasn't Univac,
Later, with the introduction of the 1100/10 and /20, which were
basically the 1106 and 1108 CPUs with semiconductor memory replacing the earlier machines core memory, addressing up to 524K was supported, and documented in the relevant manuals which are on line at bitsavers. They
used the previously unused bit, and redefined some other bits to expand
the base register fields.
At the same time as that announcement, they allowed 524K support for
1106s but not 1108s. I believe that the faster cycle time of the 1108
didn't allow time to run with the larger memories. In fact, in order
to make timing with 524K on the 1100/20, they had to slow the cycle time
to 875ns from the 1108s 750ns.
The implications for I/O seem obvious to me.No, you are right. The 1108 P&S manual didn't discuss that, because it
If you want to say that this was covered in something other than the R1 manual (UP-4052) that I happen to have and that I should RTFM, I'll accept your observation defer to your expertise.
didn't exist!
< snip >
Good. I believe the 5056 was compatible with the 5046, but it wasAs luck would have, someone recently sent me a copy of the PRM for the 5056/8470 (UP-9122).But getting back to my OP, I'm more interested in peripheral subsystems. big snip
faster internally in order to keep up with the faster 8370 disks.
I won't mention who sent me the manual so they don't get inundated with requests for old manuals, but thanks a lot, Masked Man.You might consider asking bitsavers if they would scan it, so others can access it as well.
snip
Thanks to everyone out there who helped.As I said before, you are most welcome! It gave me a pleasant stroll
down memory lane! :-)
--
- Stephen Fuld
(e-mail address disguised to prevent spam)
As for passing along the 5056/8470 manual to Bitsavers, I thought about that, but I get the impression from what I've read that they might not want it ... which I hope is not the case.
On Monday, September 19, 2022 at 11:25:44 PM UTC-7, Stephen Fuld wrote:
As for passing along the 5056/8470 manual to Bitsavers, I thought about that, but I get the impression from what I've read that they might not want it ... which I hope is not the case.
On 9/26/22 2:06 AM, Lewis Cole wrote:
On Monday, September 19, 2022 at 11:25:44 PM UTC-7, Stephen Fuld wrote:
As for passing along the 5056/8470 manual to Bitsavers, I thought about that, but I get the impression from what I've read that they might not want it ... which I hope is not the case.trying again. I'm interested. I still have a bunch of scans in the backlog as well
| Sysop: | Keyop |
|---|---|
| Location: | Huddersfield, West Yorkshire, UK |
| Users: | 546 |
| Nodes: | 16 (2 / 14) |
| Uptime: | 45:13:57 |
| Calls: | 10,392 |
| Files: | 14,066 |
| Messages: | 6,417,260 |