Switch Progression Alps/Datacomp Keyboards

[Polecat]

Don't know if this is of any use to anyone, but I found it interesting to see the changes in switches in a particular keyboard model over a few years. Specifically for the Datacomp DFK-121 and DFK-191 keyboards, of which I have six.

First I had to figure out the date codes, to know which came first. The six I have are in two layouts - without or with the Windows keys on the bottom row. I have three of each layout. No brand name on the front of any of them, but one is branded "Super" on the back, the other five are branded "Ancer" (not Acer).

Date codes are part of the serial number, and I believe they switched format somewhere between the Windowskey-less and Windowskey-having boards. The first three are serial 10931127, 10941509, and blank. The last three are 960205851, 960806939, and 97053078. Presumably the 93 and 94 on the first three are the year codes, and the 10 is the month. On the last three the 96 and 97 would be the year codes, and the 02, 08, and 05 would be the month. This makes sense, because the Windows keys started being used when Windows95 came out (~1995). With me so far?

For switches, 10931127 has Futaba white sliders with inverted crosses. That's the oddball, which I bought new at Fry's in 1993 or 1994. The rest were all used when I got them. 10941509 has Alps-branded white simplified. Housings marked "1x" with x being a letter. The undated one has white unbranded Alps clones with four mounting tabs. Housings marked "Byy" with y being a number. The last three have unbranded clones, marked "1xyy" where x is a letter and y is a number. 960205851 and 960806939 have white sliders; 97053078 has cream sliders. All are clicky.

I'm guessing the DFK-121 version had Futaba switches and the DFK-191 had Alps/clones, so that was probably not a chronological change. But the changes in the Alps/clone switches on the other five was interesting to me. White ALPS simplified to white unbranded clones to cream unbranded clones over the period of 1994 to 1997.

[Daniel Beardsmore]

The idea that 10931127 and 10941509 were both made in October seems a little suspect.

Housings marked "1x" are not simplified Alps. Those have the single letter and single-digit number written separately. Also, I'm sceptical about simplified Alps as early as 1994, although they seem to have been introduced in 1995, not 1996 as previously thought.

The change from white to cream sliders is curious: I thought this change occurred a earlier than 1996–7, more like 1994–5. White sliders suggest OA2 switches, and cream suggests T1, but I thought T1 came in earlier.

This is the sort of information that would go in Keycombo if it were ever to be somehow resurrected. (Hence why there are fields for serial numbers and PCB codes etc; [wiki]Keyboards and switches by year[/wiki] is nowhere detailed enough.)

However, it would require one switch from each Alps clone keyboard to be opened for inspection. Also, there is the matter of where to store the required photos of each type — every data point (each keyboard example) needs a permanent set of photos to be used as evidence and for future reinspection). Up to a certain point, examples can be added to the relevant wiki pages, but the amount of examples needed to make Keycombo work would overload wiki pages. Images could be stored in the keyboard database website directly (anywhere where they will be secure) or they could be stored in MediaWiki sub-pages.


In essence though, for me the objective would be to just record as much data as possible, and make derivations from it later.

[Polecat]

Thank you for the info and corrections! I'm still new to this, but I have a pretty good collection of boards that I saved back in the '80s and '90s that I'd be happy to document.

The example here has a serial number of 44920028:

photos-f62/unitron-branded-datacomp-key ... 12387.html

So you're probably correct about the "10" not being the month. I guessed that because none of mine had a number larger than 12, but more likely it's the week, as it is on chips and some other components.

The letters after the model number on the Datacomp boards seem to indicate specific versions. For example there are a couple on ebay with foreign characters that were used on Hunter wheel alignment machines. Those have some custom key caps specific to those machines. And that's apparently the reason so many skin overlays pop up when you search this model. It looks like Datacomp is still building versions of this keyboard.

I have at least one Focus board with cream sliders, so I can check the date on that one. I'll also check the mold dates on the Datacomps.

[Daniel Beardsmore]

Datacomp are pain … they're mysterious and shadowy. I just wish I could find out from them what's going on and what went on in the past. For example, I'm still anxious to prove who made those clones. I'm not sure that I was aware that Datacomp ever used clone switches (doesn't ring a bell), but since they did, then they may still remember who they bought them from.

Monterey gave me the name Hua-Jie, but at the time it would have been Himake (now owned by Hua-Jie Group), but of all the other companies known to use standard, non-Tai-Hao¹ Alps clones—chiefly Ortek, Strong Man, Nan Tan and Focus—they're either gone or won't talk. Datacomp won't talk either. (I used to list ELSA, but I realised the other day that the ELSA keyboard in question is just an Ortek.)

I can't yet obtain proof that they were all made by Himake.

¹ Tai-Hao switches are distinct and readily identifiable

[Polecat]

>>>Datacomp are pain … they're mysterious and shadowy>>>

Indeed.

The first two boards have mold dates of 2/93 and 11/93. The other four have no dates.

Funny you mention Strongman because some of the FCC codes and microprocessor labels point to them, rather than Datacomp on four of the boards, even though the keyboards are externally identical. Boards are not the same though, and the FCC codes follow those, no surprise. FBX5E9DFK191F (10941509), FBX5E9DFK121 (10931127), and KM988KKB88191 (the other four). "Ancer" label ID on all but the Futaba switch board, probably the marketing company here in the U.S.

I also have two Mac (ADB) boards using the same basic case. Model is DMK-105US (Techlink)and DMK-105AUS (Datacomp). FCC FBX5E9DMK105 on both. Mold dates 7/1993 and 9/1993. Serial numbers 33930646 and 02943041. White unbranded Alps (clicky) in both.

The Focus (FK-2001) with cream Alps has an upper case date of 9/1994, lower case 12/1994. CPU date code 9541.

[Daniel Beardsmore]

There is a definite crossover — [wiki]Strong Man SMK-8851[/wiki] is Datacomp-style, for example. Part of the difficulty is knowing exactly who has injection moulding capability; Strong Man for example is highly unlikely to have had any production facilities of their own. As I recall, Datacomp bought out Strong Man after Strong Man went bust, and they supply what was [wiki]Strong Man KB-8861[/wiki] to DSI. I do have some slightly more specific details in an e-mail somewhere, although officially this is all confidential ;-)

[Daniel Beardsmore]

You might have some of the rare [wiki]alps.tw Type OA4[/wiki] then, as these are known from "Strong Man KB-8851" (I've yet to find any page with the rest of the photos of that keyboard — alps.tw never linked from one page to another!!)

[subcat]

I have just acquired a Datacomp DFK-101 with SKCM Blue and ALPS Electric doubleshots. Post to come.

[Polecat]

There is a definite crossover — [wiki]Strong Man SMK-8851[/wiki] is Datacomp-style, for example. Part of the difficulty is knowing exactly who has injection moulding capability; Strong Man for example is highly unlikely to have had any production facilities of their own. As I recall, Datacomp bought out Strong Man after Strong Man went bust, and they supply what was [wiki]Strong Man KB-8861[/wiki] to DSI. I do have some slightly more specific details in an e-mail somewhere, although officially this is all confidential

Right you are. I dug deeper and removed the frame/pc boards from the lower cases and found some numbers on the three later (windows key) cases. Looks like these were hand carved into the mold - "KB 8851/8852" and below that "03 18 94". All three have the same markings, including the date, not on the earlier ones. Probably the date the mold was made, rather than when the cases were molded. Late cases are a bit different also - those have three case screws compared to four on the early ones. Switch plates are metal/early, plastic/late. Labels on the back side are all Datacomp style, silver mylar (?) with somewhat crude printing, like the ebay one (and the Mac boards). Again, all but one are branded "Ancer" on the labels, the odd one is "Super".

[Daniel Beardsmore]

That does sound like the date that the mould itself was manufactured — you see that a lot with Chicony keyboards. Many PCBs also have a date on them, and that too seems to be the date that the PCB was designed, instead of the date that it was manufactured. Hi-Tek have PCB manufacture dates, done in 7-segment digits, presumably some sort of optical setup during the photolithography, possibly entirely automated. (Oddly, this is not seen in the Apple keyboards that look like they come from the same PCB OEM …)

At the moment we just need everything documenting on the wiki.

[Polecat]

I'll be happy to provide any info or photos on these or any of the other 50 or so (mostly Alps) keyboards I have here. I've been trying to sort out the Northgates also, and not finding much about those online.

[Daniel Beardsmore]

I look forward to your wiki updates then.

[Polecat]

Ok, now all I need is to buy a good camera, build a photograpy studio, learn to write, learn to program wiki, describe the universe in 500 words or less, and build a working scale model. See you in ten years or so.

[Daniel Beardsmore]

I don't anticipate dying within that period, but as for the state of human civilisation in another ten years … I suspect that keyboards will be less of a concern by then that they are now.

[Polecat]

Well I voted for the Giant Meteor.

(...not quite so...) seriously though, I would be able to provide photos and info on the boards at least, if that would be any help.

[Daniel Beardsmore]

You know what needs doing — just do it!