Honeywell terminal ASCII keyboard

MMcM

24 Dec 2020, 23:15

A Honeywell keyboard from 1982 with SD-series switches.
Honeywell-BKBD-case.jpg
Honeywell-BKBD-case.jpg (432.9 KiB) Viewed 3385 times
Similar or identical to just the keyboard in this topic from earlier this year.

The label on the back says BKBDUSWA-002 PR.
Honeywell-BKBD-back.jpg
Honeywell-BKBD-back.jpg (332.16 KiB) Viewed 3385 times
I am not certain how to parse that. KBD might be keyboard and US the key layout.

The Micro Switch keyboard subassembly is 91SD30-3.

The auction listing didn't say anything more specific about the terminal, either. It did say that everything was working, which perhaps meant that they had access to the rest. Since the keyboard cable has a DB-25 and they said it was RS-232, I was a little bit worried that someone might try to connect it to a serial port. Since I was about 90% sure that a DB-25 meant TTL-level parallel (and likely ASCII), I supposed it would be fried by RS-232 voltage levels. But, in the end, it is fine and indeed still working.

Features:
  • the space bar and legends on the gray and black keys have yellowed, but the white keys are still bright
  • the overall scheme, while not as colorful as some similar keyboards, is still appealing
  • the nav cluster is quirky
  • CAPS LOCK is stepped, physically locking, and lighted (turns out this is done more or less electrically direct: not even the keyboard's controller is involved)
  • some words are spelled out instead of abbreviated, even though they end up wrapped and squeezed, such as FOR-MAT instead of FMT and CALCU-LATE instead of CALC
  • but then CTL instead of the standard CTRL (yes, that's what it does even though it's way up there)
  • and, finally, ABB-REV
  • the power LED is green, still unusual for the early '80s (and still pretty pale)
  • most switches are 4B3E, that is, 2.5oz.; but ESC and CTL are 4B6E, that is, 6oz.
  • LOCAL is locking, so it's a terminal mode
  • these particular switches are still nice and smooth to type on
  • the case itself is just plastic
The PCB has an 8048 (not socketed -- so not trivial to read), a couple inverters, a demux, and a one-shot.
Honeywell-BKBD-top.jpg
Honeywell-BKBD-top.jpg (1.12 MiB) Viewed 3385 times
Honeywell-BKBD-bottom.jpg
Honeywell-BKBD-bottom.jpg (1.21 MiB) Viewed 3385 times
It also has a full DB-25 connector for the cable instead of some cheaper header. The cable itself appears to be purpose-built, because it omits seven lines that don't matter.

The parallel signals from the DB port of the controller are ordered 6,5,4,3,2,0,1 on the cable. I don't know whether this was due to some constraint of the single-sided PCB layout or because the designer got momentarily confused by the pin mapping of the second half of the 7404 signals (outputs are 2,4,6,8,10,12 for 1,3,5,9,11,13, respectively).

There is a connection for a speaker all the way out to the cable, and what might be a standoff for it on the case, but just a jumper on this keyboard.

The DATA SET READY LED is controlled by a line that also goes (through an inverter) to the external interrupt of the controller. I thought that might mean that not ready (high) would inhibit output, but that doesn't seem to be so. If DSR is toggled after initialization, the keyboard sends some special identification(?) sequence.

Teensy and connector:
Honeywell-BKBD-converter.jpg
Honeywell-BKBD-converter.jpg (385.45 KiB) Viewed 3385 times
The LOCAL key sends a different code for lock and unlock (8B and 8C). These codes and the ident string seem to be the only bytes with the high bit set.

The dark keys that don't correspond to ASCII control characters otherwise send ESC codes. Some are compatible with a VT52, such as A, B, C, D for up, down, right, left. All but EXECUTE (ESC i) change when SHIFT is held down. So, ERASE is ESC K and SHIFT ERASE is ESC J, erase end-of-line and erase end-of-screen on the VT52. But INSERT and DELETE are ESC [ I and ESC [ P, as though ANSI escape sequences, except not matching them.

MMcM

10 Apr 2021, 23:27

Here is a keyboard in the same family with more keys and more color, SD-16604 / 108SD30-4.
Honeywell-BKBD-108-case.jpg
Honeywell-BKBD-108-case.jpg (429.54 KiB) Viewed 2983 times
Honeywell-BKBD-108-back.jpg
Honeywell-BKBD-108-back.jpg (397.25 KiB) Viewed 2983 times
It has the same SD-series feel with a steel plate and a plastic case. The Backspace key is more heavily weighted than other keys, which takes a bit of getting used to.

This one has the red BREAK key. As seen in the earlier topic, some do and some don't. I am not sure what else this correlates with.

The PCB has more or less the same components in a somewhat different arrangement.
108SD30-4-top.jpg
108SD30-4-top.jpg (1.22 MiB) Viewed 2983 times
108SD30-4-bottom.jpg
108SD30-4-bottom.jpg (1.34 MiB) Viewed 2983 times
This board also has a speaker connected to the previously identified signal.

The basic parallel strobe protocol is the same. Except that instead of sending an ASCII code, it has the key number, in straightforward left-to-right, top-to-bottom order: CLEAR RESET INIT is 01 and ATTRB is 6C. Using previously unused signals on the DB-25 connector, CTL and SHIFT as well as BREAK and AUTO LF and LOCAL (which are locking keys) are output directly.

It is possible that there is an ASCII variant of this bigger keyboard and/or one like this of the smaller one.

It is compatible enough that the same converter hardware works for both, after adding a few more wires.
Honeywell-BKBD-converter-2.jpg
Honeywell-BKBD-converter-2.jpg (498.81 KiB) Viewed 2983 times
But the firmware needs to be different and here can come closer to an actual keyboard, but with no up transitions, not quite enough shifts and no flexibility in their behavior.
Last edited by MMcM on 24 May 2021, 16:38, edited 1 time in total.

User avatar
Lalaland124

11 Apr 2021, 00:13

Very pretty keyboard and great project!

User avatar
XMIT
[ XMIT ]

11 Apr 2021, 01:04

MMcM wrote:
10 Apr 2021, 23:27
[...] with no up transitions [...]
So, these have the "pulse" type sensors? I encountered these a while ago here:

viewtopic.php?p=277892#p277892

HaaTa had the idea back then of strobing power. If you get at most one pulses per power on event, you could use this to detect "up" transitions implicitly by a lack of pulse.

I had my chance at one of these lovely boards recently and missed out:
https://www.ebay.com/itm/284204742932

The boards with the red key caps, also seem to have non sculpted keys.

John Doe

11 Apr 2021, 02:34

XMIT wrote:
11 Apr 2021, 01:04
So, these have the "pulse" type sensors?
Do the 3-terminal sensors have "pulse" type? All of them I have seen are "hold" kinds but I'm confused with this now.

MMcM

11 Apr 2021, 18:56

XMIT wrote:
11 Apr 2021, 01:04
MMcM wrote:
10 Apr 2021, 23:27
[...] with no up transitions [...]
So, these have the "pulse" type sensors? I encountered these a while ago here:
I meant no up transitions from the installed firmware. As with the smaller board, most switches here are 4B3E, so there could be with different code / wiring.

Post Reply

Return to “Workshop”