IBM 5281/5282/5285/5286 Data Station Typewriter Keyboard (P/N 2658984)
Quick summary:
- 528X-83 type Model B keyboard
- 83-key Italian layout
- IBM beam spring clicky key-switches
- Double-shot SAN keycaps
- Made by IBM Italia
- Some call this the "beamfoot" as it resembles the later IBM 5291/5292 Model F keyboard but in Model B form
- The speaker present was likely used to simulate what a solenoid did for other Model B keyboards or provide audible cues from the system
That's a lot of numbers in the name... I hear you. This is what I call a 528X-83 type Model B keyboard! 528x family, 83 keys. I call it so because the habit of calling most IBM Model B(eamspring) keyboards by simply "5281", "5251", "3276", etc. isn't really accurate, in my opinion and experience. As per the cable tab from the speaker, this keyboard must've been made in the first half of 1981.
The IBM 5280 Distributed Data System was announced in January 1980 and was a 'diskette-based system' for distributed processing and data entry. There were several data stations within the family:
- IBM 5281 Data Station
- IBM 5282 Dual Data Station
- IBM 5285 Programmable Data Stationn
- IBM 5286 Dual Programmable Data Station
The stations themselves resembled the more common 5250 series terminals such as the 5251 and 5252. 5281 and 5285 were single-operator stations, 5282 and 5286 were dual-operator stations. 5285 and 5286 were additionally "programmable" stations. My understanding is that thee dual-operator ones had a single CRT inside that's optically split to either user. A 66-key data entry version of this keyboard also exists.
I also have a New Model M. It uses the same matrix Unicomp has used for years with their Ultra Classic and EnduraPro keyboards. As such, you should be able to observe 2KRO by pressing any 3 WASD keys together (and see my matrix simulator to find more examples).
I don't mean to sound like an arsehole, but it's not a matter of debate; a conductive membrane using a key-matrix is practically limited to 2KRO. You'd either need to stick diodes in (which is harder and not cost-effective to do with membranes versus PCBs) to prevent back current from ghosting keys, or you'd need dedicated traces for each switch position instead of using a matrix (which is costly in traces and the design of controller you'd need for that). Because neither is a practical option, the controller must block keys susceptible to ghosting because not sending all intended keys is preferable to reporting keys not even being pressed. In a key-matrix without diodes, 2 is the maximum number of any given keys guaranteed to be reliably sensed and reported together, thus Model Ms (like most membrane keyboards) are 2KRO. That doesn't mean more keys can be pressed under certain situations - the matrix is like a table of traces, having columns and rows and what combinations are being used matters - but it doesn't change that key rollover means the maximum guarantee across all keys and the number in this case is 2. Not the maximum possible key combination under any condition. I think that part is what confuses a lot of people (which is understandable, I've seen it misrepresented for marketing and used as a buzzword). But, Deskthority Wiki describes as such which is probably generally easier to understand:
I'd be surprised if this is not the case with your keyboard. With the Unicomp Mini M, they improved supported combinations by making the matrix larger (from 16x8 for New M to 16x12) and then using maths you can find out where in the matrix the most commonly used keys should go, but you can still observe 2KRO on a Mini M by trying to press any 3 arrow keys down.