3 Feb
2003
3 Feb
'03
12:57 p.m.
Does anyone remember what was required to convert an IBM Selectric
Typewriter to a computer printer? I seem to recall a kit of some sort
was available, but my memory is
*really* hazy about that.
3 Feb
3 Feb
1:15 p.m.
I converted a 2741 (20 yrs ago) since it already had all the
solenoids intact. I think I used an ancient Byte
article that gave instructions and some sample 8080 code -
mostly lookup tables indicating which solenoids needed
energizing for each char. It took, in all something like
12-17 coil drivers - including "shift". The 2641 did caps by
first shifting, then printing - needed a delay subroutine.
Perhaps it's listed in a Byte index or Best of Byte books.
I don't think it would be feasible to do on a non-2741
unit, as you would have to find and connect to all those
linkages - unless you connect a solenoid to ~each~ key ;)
Get yourself an ASR-33 - much more fun and better noises!
(I once had an ASR than moved itself off a table with each
carriage return- until it went "thump" one too many times"
and fell 2 feet to the concrete floor - still kept printering
and worked flawlessy afterwards and continues to do so AFAIK...
Heinz
----- Original Message -----
From: "Marvin Johnston" <marvin(a)rain.org>
To: "ClassicCmp" <cctalk(a)classiccmp.org>
Sent: Monday, February 03, 2003 9:54 PM
Subject: Selectric Typewriter conversions
Does anyone remember what was required to convert an
IBM Selectric
Typewriter to a computer printer? I seem to recall a kit of some sort
was available, but my memory is
*really* hazy about that.
2:22 a.m.
New subject: Selectric Typewriter conversions - looking for ASR-33
Sadly, true. I gave mine to a friend with a 1 transistor
rs232 to 20mA loop interface built into the DIN plug,
to connect directly to the TRS COCO. A single BASIC
poke for the bitbashed software UART made the 110bd work.
When Radio Shaft came out with OS-9 for the COCO, I
immeditate bought it and hooked up an opto to make the
ASR-33 into a terminal. With OS-9 on the the ~.9MHz COCO,
it was a matter of whoever hit enter first got the floppy
drive!(multi-user!)
I was the first with a multi-user operating system on the block ;)
It functioned as great printer but slow for many years....
I've never been able to find another, more less in that pristine shape
- ie clean oil, and not a speck of dirt!
Anybody have a stash of workting ASR-33s out there? and know of a
cheap shipping company ;) It would make a lovely PDP11/10 console ;)
Heinz
"Kent Borg" <kentborg(a)borg.org> wrote:
On Mon, Feb 03, 2003 at 10:11:49PM -0500, Heinz Wolter
wrote:
Get yourself an ASR-33
Easier said than done.
-kb, the Kent who still hasn't found one. 
3 Feb
3 Feb
1:16 p.m.
On Mon, 3 Feb 2003, Marvin Johnston wrote:
Does anyone remember what was required to convert an
IBM Selectric
Typewriter to a computer printer? I seem to recall a kit of some sort
Modified Selectrics were used, ISTR, as consoles on several models of
mid-60's BIg Blue Iron. They spoke, of course, EBCDIC. There were kits in
the early BYTE mags to make these units transcieve ASCII.
Making a 'manual' Selectric into an ASR or even RO machine is
"mechanically non-trivial" as they might say.
IMHO, of course, knowing a little bit about how the Selectric tosses
it's element around. Metallic VooDoo I heard one tech refer to it.
Cheerz
John
4 Feb
4 Feb
1:19 a.m.
IMHO, of course, knowing a little bit about how the
Selectric tosses it's element around. Metallic VooDoo
I heard one tech refer to it.
Agree completely. I took one apart years ago and deep
inside were two compound lever assemblies, which could
sound complicated, but what they did was find a kind of
breathtaking simplicity. The lever mechanism worked exactly
like the way a Mobile hangs and balances several inputs to
one output. (picture below).
I'd like to congratulate the inventor of that one.
(begin fixed width)
+
|
+---+---+
| |
+ +--+---+
| |
+ +
(end fixed width)
John A.
2:42 a.m.
On Tue, 4 Feb 2003, John Allain wrote:
Agree completely. I took one apart years ago and deep
inside were two compound lever assemblies, which could
sound complicated, but what they did was find a kind of
This arrangement is called a whiffletree, and has been in use since
Antiquity - often to evenly distribute the tractive effort of teams of
draft animals. Also, pipe organs used them as a binary-to-analog
convertor, turning the small motions of several wind pouches into a
smooth-ish mechanical motion, eg. for swell shades, etc.
The Selectric was elegant as hell, but VooDoo nonetheless.
Cheers
John
11:28 a.m.
Agree completely. I took one apart years ago and
deep
inside were two compound lever assemblies, which could
sound complicated, but what they did was find a kind of
breathtaking simplicity. The lever mechanism worked exactly
like the way a Mobile hangs and balances several inputs to
one output. (picture below).
I'd like to congratulate the inventor of that one.
I have a reprint of an old (1912-ish) book on Telegraphy, and it uses a
similar linkage for a binart to 1-of-n decoder to position a typewheel
for a printing telegraph machine. So the idea is pretty old.
Mildly off-topic, but of course such a linkage works with analogue
'signals' as well. I have in front of me a camera which uses a similar
linkage with the appropriate lengths to combine the settings of the
shutter speed dial/film speed control, lens aperture ring, and maximum
aperture control so as to correctly position the reference circle for the
exposure meter. I said 'mildly off topic' because this is, IMHO, a very
simple analogue computer and it's well over 10 years old.
-tony
11:37 a.m.
Tony Duell wrote:
I have a reprint of an old (1912-ish) book on
Telegraphy, and it uses a
similar linkage for a binart to 1-of-n decoder to position a typewheel
for a printing telegraph machine. So the idea is pretty old.
I've seen a 3-bit-binary to one-of-eight decoder built out of
huge brass and bakelite relays. It's in the London Transport Museum,
and was used to drive the train indicators on the platforms of
the London Underground (subway). The train indicator tells the
passengers where the next train is going to by lighting up an
arrow next to the name of a station. Only three wires need be run
along the tunnels from the signalbox, giving seven possible
destinations and an "off" state.
--
John Honniball
coredump(a)gifford.co.uk
12:08 p.m.
I've seen a 3-bit-binary to one-of-eight decoder
built out of
huge brass and bakelite relays. It's in the London Transport Museum,
and was used to drive the train indicators on the platforms of
There wasa similar thing used to decode the 4-bit code as to what train
was expected next, used in signal boxes, etc on the London Underground.
That was certainly (large) mechanical relays.
The 4-bit train descriptor codes were stored in an electromechanical FIFO,
consisting of a drum with a number of rows of 4 pegs around it. Each peg
could be in one of 2 states (towards the spindle or shifted away from the
spindle. There was a fixed solenoid mechanism to set the pegs, the whole
drum then moved round one position (equivalent to incremementing the
write pointer in a software FIFO). There was a separately revolving
contact assembly to sense the position of each row of pegs and feed it to
the decoder. This was stepped on as each train came through -- equivalent
to incrememnting the read pointer.
The whole thing is _exactly_ like the classic software FIFO we've all
implelmented many times. I don't know when it was built, but around 1920?
It's described in detail in an old book I have called 'Modern Electrical
Engineering' which alas has no dates on it.
-tony
5 Feb
5 Feb
2:07 a.m.
There was a lot of major work on the Underground, ca 1923-1926. It might
well date from that period.
On Wed, 5 Feb 2003, Tony Duell wrote:
M. K. Peirce
Rhode Island Computer Museum, Inc.
Shady Lea, Rhode Island
"Casta est quam nemo rogavit."
- Ovid
I've seen
a 3-bit-binary to one-of-eight decoder built out of
huge brass and bakelite relays. It's in the London Transport Museum,
and was used to drive the train indicators on the platforms of
There wasa similar thing used to decode the 4-bit code as to what train
was expected next, used in signal boxes, etc on the London Underground.
That was certainly (large) mechanical relays.
The 4-bit train descriptor codes were stored in an electromechanical FIFO,
consisting of a drum with a number of rows of 4 pegs around it. Each peg
could be in one of 2 states (towards the spindle or shifted away from the
spindle. There was a fixed solenoid mechanism to set the pegs, the whole
drum then moved round one position (equivalent to incremementing the
write pointer in a software FIFO). There was a separately revolving
contact assembly to sense the position of each row of pegs and feed it to
the decoder. This was stepped on as each train came through -- equivalent
to incrememnting the read pointer.
The whole thing is _exactly_ like the classic software FIFO we've all
implelmented many times. I don't know when it was built, but around 1920?
It's described in detail in an old book I have called 'Modern Electrical
Engineering' which alas has no dates on it.
-tony
3 Feb
3 Feb
5:06 p.m.
New subject: Interest in IBM Token Ring NICs?
Was there someone on the list recently looking for token ring stuff?
This past weekend I picked up three IBM Token Ring NICs, part
number 83X9144A. These are full-length 8-bit ISA cards.
Contact me off-list if anyone is interested.
Cheers,
Dan
www.decodesystems.com/wanted.html
6:39 p.m.
Does anyone remember what was required to convert an
IBM Selectric
Typewriter to a computer printer? I seem to recall a kit of some sort
was available, but my memory is
*really* hazy about that.
There were two ways to do that.
The more common method was to put a thing with one solenoid per key
over the Selectric's keyboard. This had the advantage that no
modification to the Selectric was required.
The more clever (IMHO) mod was to attach seven solenoids to the
Selectric between the keyboard encoding mechanism and the print
mechanism, to directly control the tilt and rotate. And two more
for the carriage return and paper advance. This has the advantage
that you can still use it as a typewriter without having to attach
and remove the solenoid box, and it's potentially less expensive
since it takes nowhere near as many solenoids. I seem to recall
that Don Lancaster wrote a do-it-yourself article about this
approach.
Either way, though, you don't end up with a particularly great printer.
The Selectric typewriters weren't meant to be pounded on continuously
at anywhere near their full speed, since that wasn't how typewriters
were used (even by *very* good typists).
IBM made "I/O Selectrics," which had heavy-duty mechanisms intended
for continuous output. Model numbers that come to mind are the
1052 and 753, though my memory regarding these seems more than a
little fuzzy.
There were also Selectric-based terminals like the 2741, which were
AFAIK not quite as heavy-duty as the 1052, but still better than a
typewriter.
Trivia: although the Selectric typewriter wasn't introduced until
1962, the first production units were delivered to customers in 1961,
as part of the operator console for the IBM 7030 Data Processing
System (Stretch). Since it wasn't an announced product, all the
publicity photos of Strech had a person, sign, or other item
strategically placed to obstruct the view of the typewriter.
4 Feb
4 Feb
3:54 a.m.
The more clever (IMHO) mod was to attach seven
solenoids to the
Selectric between the keyboard encoding mechanism and the print
mechanism, to directly control the tilt and rotate. And two more
for the carriage return and paper advance. This has the advantage
that you can still use it as a typewriter without having to attach
and remove the solenoid box, and it's potentially less expensive
since it takes nowhere near as many solenoids. I seem to recall
that Don Lancaster wrote a do-it-yourself article about this
approach.
Lancaster's "TV Typewriter" book includes a description.
Either way, though, you don't end up with a
particularly great printer.
The Selectric typewriters weren't meant to be pounded on continuously
at anywhere near their full speed, since that wasn't how typewriters
were used (even by *very* good typists).
IBM made "I/O Selectrics," which had heavy-duty mechanisms intended
for continuous output. Model numbers that come to mind are the
1052 and 753, though my memory regarding these seems more than a
little fuzzy.
This is disappointing. I'd been on the lookout for a Selectric for the
purpose of performing this hack (Powerbook G4 USB->Parallel->Selectric),
but chances of finding one of these models looks pretty slim.
Maybe I'll dig out my Xerox MemoryWriter and try to interface that...
Tom
Applefritter
www.applefritter.com
4:23 a.m.
> Does anyone remember what was required to convert
an IBM Selectric
> Typewriter to a computer printer? I seem to recall a kit of some sort
> was available, but my memory is
> *really* hazy about that.
On Tue, 4 Feb 2003, Eric Smith wrote:
There were two ways to do that.
The more common method was to put a thing with one solenoid per key
over the Selectric's keyboard. This had the advantage that no
modification to the Selectric was required.
The two most popular ones of those were the Rochester Dynatyper (which had
an interface board with two edge connectors - one for plugging into Apple
][ bus, and other for TRS-80 model 1 expansion connector), and the KGS-80,
(which had a Centronics compatible parallel port input).
The more clever (IMHO) mod was to attach seven
solenoids to the
Selectric between the keyboard encoding mechanism and the print
mechanism, to directly control the tilt and rotate. And two more
for the carriage return and paper advance. This has the advantage
that you can still use it as a typewriter without having to attach
and remove the solenoid box, and it's potentially less expensive
since it takes nowhere near as many solenoids. I seem to recall
that Don Lancaster wrote a do-it-yourself article about this
approach.
I remember there being an outfit in Walnut Creek that sold a kit for
mounting under the keyboard of a selectric.
Either way, though, you don't end up with a
particularly great printer.
The Selectric typewriters weren't meant to be pounded on continuously
at anywhere near their full speed, since that wasn't how typewriters
were used (even by *very* good typists).
Some (most?) selectrics were rated at a maximum of 14.8 characters per
second. That is oft cited for the reason for existence of the 150 baud
rate on many modems.
I knew one professional typist who could AVERAGE 150 wpm on an 8 hour
day! She could "wear out" a selectric during the "warranty" period.
IBM made "I/O Selectrics," which had
heavy-duty mechanisms intended
for continuous output. Model numbers that come to mind are the
1052 and 753, though my memory regarding these seems more than a
little fuzzy.
There were also Selectric-based terminals like the 2741, which were
AFAIK not quite as heavy-duty as the 1052, but still better than a
typewriter.
And there were special dedicated heavy duty ones for cranking out dead
tree spam, that would permit storing a document, and pausing for a typist
to fill in blanks. "Dear ___, of all of the families living on ____
street, ..." That included the MTST ("Magnetic Tape Selectric
Typewriter") and later the MCSC ("Magnetic Card Selectric Typewriter").
I remember one fellow writing an early microcomputer mail merge program
that he wanted to call "the FULL ST", but nobody got the pun with MTST.
Sellam probably still has the service manual from my MTST.
Trivia: although the Selectric typewriter wasn't
introduced until
1962, the first production units were delivered to customers in 1961,
as part of the operator console for the IBM 7030 Data Processing
System (Stretch). Since it wasn't an announced product, all the
publicity photos of Strech had a person, sign, or other item
strategically placed to obstruct the view of the typewriter.
Trivia2: There were numerous type balls for the selectric, including APL!
4:56 a.m.
On Tue, 4 Feb 2003, John Lawson wrote:
On Tue, 4 Feb 2003, Fred Cisin (XenoSoft) wrote:
Sorry
It has truly been a long time, and that part of my memory hasn't been
refreshed in several decades. But surely not everybody went along with
IBM's mandated nomenclature!
Trivia2: There were numerous type balls for the
selectric, including APL!
^^^^^^^^^^
TSK! TSK! TSK!!! ELEMENTS!!!
I mean, honestly - we are *supposed* to be experts here, right???
:}
11:28 a.m.
Trivia2:
There were numerous type balls for the selectric, including APL!
^^^^^^^^^^
TSK! TSK! TSK!!! ELEMENTS!!!
I mean, honestly - we are *supposed* to be experts here, right??? 
7:33 a.m.
on 4/2/03 3:36 AM, Eric Smith at eric(a)brouhaha.com wrote:
IBM made "I/O Selectrics," which had
heavy-duty mechanisms intended
for continuous output. Model numbers that come to mind are the
1052 and 753, though my memory regarding these seems more than a
little fuzzy.
For what it's worth, my memory (no less fuzzy than Eric's) remembers that as
a model 735.
But even with the more robust "I/O Selectrics" there was a hidden gotcha:
when you sent data to the mechanism you had to time it right. The Selectric
needed varying amounts of time to tilt and swivel the type element,
depending on the letter chosen. To cope with that, lots of the early
computer interface projects just stuck a software delay loop in the driver
code to wait plenty of time for each letter to finish printing before it
sent the next character code to the Selectric. This apparently caused a lot
of mechanical wear-and-tear as the clutch relentlessly cycled in and out,
and even the heavy duty "I/O" mechanism would eventually crap out from all
the lurching. The right way to do it was to make sure your driver polled
the state of some microswitches that IBM had thoughtfully buried in the
clockwork to provide feedback on the progress of the print cycle. That way
you could start sending a new character at a magic moment just _before_ the
previous print cycle fully finished, which provided the fastest possible
printing and extended the life of the clutch.
Like John said, VooDoo.
Regards,
Arlen Michaels
3 Feb
3 Feb
6:46 p.m.
On Mon, 3 Feb 2003, Marvin Johnston wrote:
Does anyone remember what was required to convert an
IBM Selectric
Typewriter to a computer printer? I seem to recall a kit of some sort
was available, but my memory is *really* hazy about that.
I have in storage a modified Selectric console/terminal and interface box
that works with a TRS80 model 1. I got both from the original owner years
ago along with the model 1. The interface box was sold assembled, and
wasn't a kit. It also seemed to be very well made.
-Toth
4 Feb
4 Feb
7:51 a.m.
Don Lancaster's "TV Typewriter Cookbook" had all the
details. I sold my copy last September unfortunately.
Lawrence
On 3 Feb 2003, , Marvin Johnston wrote:
Does anyone remember what was required to convert an
IBM
Selectric Typewriter to a computer printer? I seem to recall
a kit of some sort was available, but my memory is *really*
hazy about that.
lgwalker@ mts.net
7991
days inactive
7993
days old
20 comments
15 participants
participants (15)
-
allain＠panix.com -
ard＠p850ug1.demon.co.uk -
arlen＠acm.org -
at258＠osfn.org -
cisin＠xenosoft.com -
coredump＠gifford.co.uk -
dan＠ekoan.com -
eric＠brouhaha.com -
h.wolter＠sympatico.ca -
jpl15＠panix.com -
kentborg＠borg.org -
lgwalker＠mts.net -
marvin＠rain.org -
owad＠applefritter.com -
tothwolf＠concentric.net