Hey, no worries. Glad people are getting more out of it. I've already got some great ideas from everyone's input.
Right now I'm leaning towards getting the X-keys 52 key pad and something like a Filco tenkeyless. Seems like the combo (with the X-keys pad to the left of the keyboard) would take up about the same amount of desk real estate as a terminal style board, but I wouldn't be wasting any of it with the numpad that I don't often use. Function keys to the left of the keyboard are far more valuable to me.
So what you're looking for is a modular Filco fortytwokeyful...
Seriously though, if numerous extra keys in a particular position are highly desirable, that's probably by far the best approach; forget layouts, get a nice at least compact-ish keyboard and a key pad.
Question for you dw, as you seem to be the authority on terminals.
I might be rambling a bit about them, but terminals made by IBM (never mind countless other companies', often long gone, taking their documentation with them...) are part of a whole other world of computing, somewhat alien to the current world of standardised PCs and the World Wide Web, that I've barely scratched the surface of, never having being "inside" it myself. My informal research has been mostly just with regard to the history and variation of IBM keyboards, via the Model M and backwards and sideways, and so inevitably getting caught up in terminals...
According to this site, the Model M terminal keyboards are totally incompatible with modern PCs. Is the 1397000 a special case? When was it produced?
Is the pun unintended?
From the persepctive of a buyer who wants a single keyboard (hahaha) to "just work" on their PC, this advice is sound. Alas though, this sensible line of thinking means that many, many otherwise good working Models F and M will have been destroyed, even now in these days when we have websites that celebrate them. Unfortunately in this case, most people buy keyboards to use them, not to "save" them, nor simply because the keyboard is interesting in itself...
The intruiging thing is, electronically Model M keyboards for terminals are not necessarily that different to their AT and PS/2 counterparts. Here is
a look at IBM keyboard part number 1390876 for the 3192 terminal, which the communication method it uses is straightforwardly explained by Peregrine, except that
the page is originally in Japanese. I will try to summarise the main points.
- The plug is different from AT and PS/2 plugs, but the pinout shows that it has exactly the same pins with exactly the same uses as an AT or PS/2 keyboard.
- The data transmission protocol appears to be exactly the same as the AT and PS/2 keyboard's.
- However, it uses scan code set 3. [Scan codes, so called because the keyboard is constantly scanning for keypresses, are the data sent by the keyboard to make your finger movements come true on the screen! Sets of scan codes provide standardised defintions for how keyboards should interact.] Now set 3 is a bit of a pain. It was used with IBM PS/2 computers and so is supported by PS/2 101/102 key IBM Model Ms of that time, but it did not catch on in the wider PC industry, and so today scan code set 2 remains the industry standard somewhat. I've read anecdotally that there is however scattered support for set 3 in modern devices, but it is a bad idea to rely on it being there.
There are also the issues of the "wrong" keyboard ID being reported and required in response, and the lack of LEDs, but that seems to be about it. Remapping to personal requirements can be done in software, and keycaps can be replaced or their legends ignored.
In theory then, all that would need to be changed to "make it work on a PC" is to change the connector and a firmware patch. That's all.
In practice, because the source is not available for the firmware, you can't do that. Translating the protocols is a good alternative, because you don't have to worry about internal variations in the same type of keyboard, you don't even need to modify anything in a keyboard at all, and the resultant translator could then be used for any keyboard using the same protocols.
Anyway, I think I've been subconsciously assuming that everyone knows what a terminal emulator keyboard is. Here's some wordy context...
Historically, the IBM PC (probably not the first PC, but that's a huge debate...) appeared in a world where there were many competing standards in almost every respect, in some cases major aspects of technology being unique to certain companies, and not much in the way of universal standardisation. Sure, there was at least ASCII, but not everyone supported it for everything, certainly not IBM. Dumb terminals were no exception, having been around for a while, with their array of different capabilities and for different machines, again particularly so within IBM with their incompatible 3270 and 5250 ranges, never mind everyone else. (I should also note that the 122 keys on a terminal keyboard is by no means a standard, even on IBM terminals with which they are most associated.)
So logically then, there surely wouldn't have been much expectation that IBM terminal keyboards should be compatible with the IBM PC, and perhaps significantly so because the IBM PC was at least in principle intended for an entirely new market of budding microcomputer users who couldn't afford or didn't want a big IBM system with dumb terminals.
About three years after the initial release of the original Type 5150 PC, IBM did however announce a version of the PC that could function as a 3270 series terminal. This is the Type 5271 "3270 Personal Computer" mentioned above that Sandy has the keyboard for. Adding this facility required a special monitor as well as the keyboard, and used almost all of the expansion facility in the PC. So a theoretically viable business use that would require using a terminal keyboard with a PC, proved complicated and required considerable extra expense just to meet the hardware requirements. However this, nor its relative unpopularity, did not prevent updated versions of the 3270 PC appearing.
At this point I was held up for hours trying to find out what happened to various product lines [I just cannot work out what product that P/N
1393656 would have been used with, or even what IBM officially named it...], so I will leave that for now.
Basically, with more powerful PCs and generally advancing technology, it became viable, and in the longer run probably much cheaper too, to emulate dumb terminals as software to run on the PC rather than with additional hardware. Consequently it made sense to use a keyboard that would talk to the PC in the usual fashion and handle the inconsistencies in software.
And this - yes, finally - is the magic point where you get the Terminal Emulator keyboards, or as IBM called them "Host Connected" keyboards. They use exactly the same electrical connections, signalling protocol and scan codes as their 101/102 counterparts, so can be treated as generic PC keyboards if so desired. However, either a driver, specific application, or more likely in our case, remapping software is needed in order to do anything with the extra keys. I would suggest autohotkey, I talked about this earlier.
Yes, this I believe is what the 1397000 is, a terminal emulator keyboard. However, all my searching through the IBM product announcement letters just doesn't show up that part number, even though lots of parts sellers seem to have it. I can confirm the existence of the 1397003 German version, I'm using it right now, and this is the keyboard that lowpoly refers to:
I think one terminal keyboard on clickykeyboards was supposed to work on ps/2 and it was the German version. They appear regularly here on ebay.de and many seem to work on a modern PC.
I should point out that as far as I've been able to find, all these Host Connected terminal emulator 'boards are built in the later 122 key housing, here's a typical example of the housing on a "real" terminal keyboard:
and here's the current production Unicomp terminal emulator version (note the addition of the LED panel):
While I couldn't find that darn part number 1397000 or 13970003 at IBM's site, I have found numerous other part number references for the Host Connected - ie. 122 key PS/2 compatible terminal emulator keyboards, so try searching for these:
IBM SELECT-A-KEYBOARD AND IBM PS/2 HOST CONNECTED KEYBOARD
Announcement Letter Number 191-086 dated June 11, 1991:
Host Connected Keyboard
Belgium 1397501
Canadian French 1396900
Cyrillic 1397515
French 1397506
German/Austrian 1397500
Italian 1397507
LA Spanish 1396901
Swiss French 1397511
Swiss German 1397512
English (U.K.) 1397513
English (U.S.) 1396400
IBM Personal System/2 Host Connected Keyboard
IBM Announcement Letter No. ZG91-0245 dated June 11, 1991:[this is the same letter as above but for the EMEA [Europe, Middle East, Africa] region instead of the US]
Austrian German 1396902
Belgium 1396903
Cyrillic 1396916
Danish 1396904
Dutch 1396905
French 1396990
Greece 1396917
Iceland 1396919
Italian 1396908
Norway 1396909
Portugal 1396910
Spain 1396911
Sweden/Finland 1396906
Swiss/French 1396912
Swiss/German 1396913
Turkey 1396921
Yugoslavia 1396920
UK 1396914
US (EMEA only) 1397025
Finally, an alternative part number to look for for that very same US English "Host Connect Kybd (Select-a-Keybd)": 1397050
[from the
IBM Electronic Pocket Reference Manual, 4 Jan 2000 revision]
Of course, don't forget the part numbers 1397000 or 1397003.
Or, you can just
buy a new from Unicomp...
EDIT:
snipped picture of IBM part number picture of 1387033
Just wanted to say that this layout is awesome. Very ideal for me.
It is a shame that there's as good as no chance at all of this format being made again; the housing (not made by Unicomp AFAIK), Model F action (not supported by Unicomp), the terminal type (ancient and uncommon) and the layout (122 and 101/2/3 'board more readily available and will do the same thing), and so justifying the existence of an emulator version would be impossible without specific demand for a very large quantity...
You should be able to swap controllers, assuming that the terminal space saver uses the same controller as the full size one.
In this case, I don't think this is possible because the 1397003 is a Model M and the 1387033 is a Model F, which presumably has the controller attached to the same PCB as the key contacts, and even you work around that the contact system is different, and the matrix could be a problem [matrix differences could be a problem for any controller swap...]. On top of that, Unicomp probably couldn't help, alas... "Cracking" or finding out the protocol and scan codes would probably be more straightforward.
I do apologise for being so long winded, but hopefully something I said is of some use.