Jump to page content

Updates

Factual updates and other interesting changes are reported here. Adjustments and tweaks are not listed.

Follow the updates RSS feed:

Updates for 2022

July

Sunday, 24th July

I just learned late last night of the unfortunate and untimely death of Michael Schmid, Cherry’s head of MX marketing. According to the article, Michael died in May; I only found by chance. I copied Michael in on an e-mail to Cherry on Thursday and immediately received a bounce for his address. LinkedIn showed that he still worked for Cherry, as did Cherry’s own website, but something seemed amiss, so after not hearing back from anyone at Cherry on Friday, I had a look last night to see if there was any other clue as to Michael’s whereabouts, and found that he had died.

I have Michael to thank for some of the Cherry source material on this site, including an old MX datasheet, and the M8 part number schema. Michael seemed passionate about Cherry; it’s tragic that he was taken from the world too soon.

Tuesday, 19th July

Another Micro Switch mystery is finally solved. Thanks to discoveries by “fricked” and Deepak Kandepet, we finally know how timed repeat works. Ever since discovering the existence of timed repeat switches (in SD Series, and now in the original SW Series), we have been faced with the question of how a (presumably) simple four-terminal switch could accommodate auto-repeat timing. Such keyboard are fairly unusual, and it has taken until now for someone to trace the circuit. Timed repeat switches are much simpler than expected: they are modified sink pulse switches where one output terminal sits at an intermediate voltage while the key is held, invisible to the encoder but able to be detected by a custom signal generator IC.

Details are given on the SW Series page under timed repeat.

Sunday, 3rd July

“fricked” has found the first example of a wired inductive Licon keyboard, part number 55-000084, listed for sale on eBay for a paltry $599.99. This example is the type described in US patent 3714611A and, from the part number, will be a low-profile Series 550 keyboard, and the first Series 550 keyboard ever seen. Strangely, despite the use of encoding switches, the logic board is the same size as the keyboard itself, with a total of 49 logic chips! Unfortunately eBay photos are too small to allow anyone to read off the chip codes. (See the ITW page for a link to the eBay listing, since at some point it will die and–as whoever is buying all these rare items squanders them and they vanish without a trace—the link will have to be replaced with one pointing to my archived photos as the only thing keep alive the item’s existence.)

The model I forgot to archive the photos from (55-100007, which the new owner has predictably kept hidden from the world) was probably an original Super Switch keyboard, considering that the switches were soldered to a PCB instead of threaded. Conceivably these PCB-mounted models were created to allow matrix wiring and thus N-key rollover, and may have been classified as Series 551.

The switches in 55-000084 are a lot like the “Style A” switches from Series 555, which implies that the patents filed in 1973 for tactile feel and alternate action were more likely intended for either Series 550 low profile or Super Switch. (The patents, uncharacteristically, omit the details of the switch itself.)

June

Monday, 13th June

A new Olympia page briefly summarises the majority of the discovered switch usage to date, now that I have finally got around to adding some of Steve Kuterescz’s electronic typewriter photos that he’s sent over the last couple of years. At present, there is no detailed breakdown of chronological information, as so little such data is available at present.

This page is similar in function to the recently-added and less comprehensive Olivetti page.

May

Sunday, 1st May

Sometime between 1975 and 1977, US keyboard manufacturer Collimation was acquired by Applied Dynamics International (one of the two ADIs, the other being Advanced Datum Information Corp). Applied Dynamics International, founded in 1957, is still in business, and they reported a year ago that “there are those who remember and participated in ADI’s illustrious past” but have since refused to be forthcoming.

The Applied Dynamics International page is now updated according to material collected by Marcin Wichary. Due to the often poor quality of that material and its poor stewardship (no indications of origin and sometimes not even an indication of source document or year) I have tended to instead use a higher-quality Bitsavers scan or, in at least one instance, obtain a higher quality scan from a library. In this case, the most detailed material is from Electronic Design’s Gold Book series, and unlike the rival Electronic Engineers Master series, it is not possible to read the Gold Book volumes online as they have not been archived. It’s quite possible that there are some fascinating gems awaiting discovery there.

Sadly, like with so many other advertisements, the ADI advertisements in the Gold Book leave unsolved mysteries. The specificatons for Series 50 and Series 55 seem to be akwardly merged. The electronic interlocking in the photoelectric encoder keyboards is not properly explained. There are no depictions of the switch type used in the Series 55 mechanical contact switches. Even the series names kept changing without explanation, and it’s difficult if not impossible to be sure whether series were renamed or represent slightly different products, and the lack of information about Collimation’s product range means that we don’t know what was on offer before ADI’s acquisition of Collimation.

As is almost always the case, we have as many new questions as we have answers to old ones.

April

Thursday, 28th April

There seems to be no limit to the ingenuity of keyboard engineers. Photoelectric encoder keyboards for the most part have two-key rollover at most, in common with other self-encoding keyboards, because the electronics are unable to read a useful value with more than one key engaged at once. However, some clever Germans devised a way to beat this limitation: translucent shutters. The photoelectric encoder keyboards page is updated with details.

Wednesday, 27th April

As more information has slowly come to light on photoelectric encoder keyboards, the time has come to give them a page of their own.

Tuesday, 19th April

Details on the history of Collimation keyboards is starting to become clearer. Although the manufacturer (Applied Dynamics International) is still around, and claimed that there are “those who remember and participated in ADI’s illustrious past”, they proved to be less than forthcoming. However, examination of the various advertisements shows that ADI acquired Collimation sometime between 1975 and 1977 and dropped the Collimation name around 1977. The “Western Digital Systems” to whom the original design was patented is not the Western Digital: storage manufacturer Western Digital Corporation (WD) was founded in California in 1970, while the keyboard maker Western Digital Systems was founded two years later in Texas and traded under the Collimation name.

February

Monday, 28th February

Just when it looked like February was going to be a washout compared to the success of January, persistence with contacting vendors paid off with the acquisition of the 1986 ITT catalogue entry for the elusive ETL 18 family. Many questions still remain, but this is a good first step in understanding these switches. The question of their unfortunate scarcity may lie in the DIN ergonomics standards of the early 80s, which ETL 18 appears not to comply with. This may have limited its application to standard keyboards. One important question that remains is whether the switch could be re-implemented in a DIN-compliant manner by a company such as Kailh, considering that it is one of the finest examples of a tactile switch ever constructed.

January

Sunday, 30th January

Yesterday brought a surprising discovery: the “intermediate” Cortron ferrite core switches (AKA “ITW magnetic valve intermediate”) apparently really are Cortron SS3, which had been ultimately ruled out as it does not match the magazine description. The identity of SS3 was discovered not through any change of heart by Cortron themselves, or by any discovery of Cortron material, but rather a Celanese advertisement for Celanex 2012 and Celcon plastic resins, both of which were chosen by ITW for the SS3 switch, that is depicted in the advertisement. Curiously, the advertised switch looks to be taller than the finished product, as well as being a slightly different design.

Nonetheless, magazine articles for SS3 clearly disagree, so either the stated dimensions for SS3 are incorrect or deliberately misleading, or Cortron gave Celanese either the wrong product name or the wrong photograph.

Friday, 28th January

The miniature elastic contact switches from Maxi-Switch are almost certainly 8500 Series: the description of this series fits perfectly, and 8500 was one of the two theorised names for the series.
The strange alternating 0°/180° Alps reed switches are now determined to be a derivative of AKS-5 series, by way of a magazine article in JEE from 1978. AKS-5 is the same overall design, but the later type has been revised. Possibly these are KGF and KGL.

Thursday, 27th January

JEE on Google Books has offered a few hints about Jelco, the mysterious Japanese keyboard and switch manufacturer whose JKS-91 switches were used in the Microbee computer. The adverts depict JFK-3000 and JKF-6000 keyboards, and JKS-2, JKS-5 and KRS-6 switches. JKS-5 and KRS-6 are momentary and alternate action respectively and appear to be classified together as SKS-5. JFK-6000 uses JKS-2 switches, while the switch type of JFK-3000 is not specified.

Although there is still no mention of JKS-91, for the first time, Jelco is no longer a complete mystery. There may be more to the JFK-3000/6000 advertisement; Google Books does permit access to the remainder of it.

Wednesday, 26th January

The Alps mystery deepens once again. An advertisement from 1981 for KCC that includes an exploded illustration of the KCC switch—that is added to the KCC page—names three switch series heretofore not encountered: KGF, KGL and KCF. Although the K-series names were not restricted to keyboard types, this advertisement appears to concern itself solely with keyboard switch series. KGL and KGF are reed types, of which only KGL was included (as a “Lead switch type”) in a later 1982 advertisement only partially accessible on Google Books, so possibly KGF went end of life around that time.

KCC is described as a “mechanical contact (diaphragm)” type, as is KCF, so KCF might be the switch currently described as SCK. That would leave us without an explanation for KCA, which does not get a mention in these advertisements.

Naturally, there is no other trace of KGF, KGL and KCF, and only KCC is illustrated in the advertisement.

Tuesday, 25th January

The Alps series names and model numbers page contains what until now was my current understanding of that subject. An accidental discovery on Google Books has cast some doubt on this. Switches resembling those that became known as KCC Series were advertised in JEE (Journal of Electronic Engineering) in 1978 and 1979 as either “AKC2C” or “AKC2N”: both types are depicted seemingly identically (the scan quality is poor) and share the same brief description. There is no way to determine whether they are variants within the same series, or two separate series sharing the same housing (which we know to have existed).

Although one could argue that the S-codes (SCB, SCF, SCH, SCK) only represented internal assemblies, SCK is definitely used differently. Roland service literature gives codes SCK41168 and SCK41167, not for switches, but for switches combined with relegendable keycaps. There is also SCK-41097, suggested to be solely a switch.

AKC itself is presently taken to be the original spring bridge type, of which KCC is taken to be the replacement type. It may be that Alps briefly dubbed it AKC2 (there is also the non-keyboard AKC8, which is a miniature TACT switch, with subtypes AKC8S and AKC8N advertised in JEE), before formulating a new series naming system. However, this still offers no further insight into the S-types and how these came to exist in parallel to AKC, AKB and AKM.

This mystery appears to be deepening rather than becoming clearer. None of AKC2, AKC2C or AKC2N can be tracked down, although an archived LG/Zenith parts listing with an entry of “SWITCH AKC8S W/O CAP/SKHCAA20” indicates that AKC8S became SKHC, and the two look the same. (AKC8N seems to have become a different series.) The fact that the AKB-3420 and AKB-3320 literature did not name the switches used deprives us of observing what switches went with SCH-type keyboards. All we know is that AKB and SCH were noted together in the manual for these keyboards and CH was used for the PCB code. With Google, Google Books and the Internet Archive all exhausted, it may be a while before anyone encounters a source of answers.

Saturday, 22nd January

Lucky Austrian “Lukas” has achieved the impossible: he has obtained the catalogue entry for B2H from Omron themselves. It turns out that B2H was always a PCB-mounted type. Thus, the existing B2H information on the B2x page and the separate B2H-F7W page are now combined to form a dedicated B2H page. The Linda Hall Library was predictably unable to scan the B2H article, but not for the predicted reason: they don’t have a copy, despite what their catalogue implies (library catalogues remain in the dark ages with regards clarity). However, although permission to share the catalogue entry has yet been obtained, it does provide a lot of useful data on the series.

The article itself has been placed onto a new wishlist page for desired magazine and journal articles.

Wednesday, 19th January

The mystery of IMS International’s lone moving ferrite keyboard may finally be cleared up. The switches are identical to Keytek’s Inductric line, and it is certainly possible that Keytek was the supplier of the IMS keyboard. The inventor may well be Victor Bernin, formerly an engineer at Illinois Tool Works responsible for their inductive switches.

Tuesday, 18th January

Further scouring of Google Books has turned up confirmation that Mechanical Enterprises T16 Series is indeed the conductive rubber switch type. Again, access to the complete article is impossible; it is somewhere inside Electronic Products Magazine, Volume 32, 1989, possibly inside the article Conductive rubber aims to capture switch market. (The page is not yet updated.)

Monday, 17th January

Our understanding of SMK JM-0400 series has taken a step forward with the first confirmed breakdown of the series structure, based on fragments of an advertisement accessible via Google Books. The known subseries are as follows: J-J0019 (momentary sloped reed), J-M0404 (momentary sloped mechanical), J-M0409 (alternate action sloped mechanical), J-M0410 (illuminated sloped mechanical), J-M0432 (momentary stepped mechanical) and J-M0434 (alternate action stepped mechanical). Although the reed switches were advertised as being part of JM-0400 series, the one known reed subseries falls outside of that numbering range. This discovery explains part J-M0409#01 and series JJ0019 that were previously encountered.

JM-0200 however is now known to not be a keyboard switch: it is a series of elastic contact pushbuttons that, while tall, only offer 1 mm of travel.

Sunday, 16th January

Thanks to a “Lukas” from Austria, some interesting information about B2H Hall effect switches and their B2HK keycaps has come to light. B2H and B2KH were described in detail in the JEE, Journal of Electronic Engineering article Hall IC Keyboard Switches Become the Leading Type by Yoshikazu Kitao. Google Books has numerous volumes of this journal scanned—which is what Lukas found—although Google or the submitter (the University of Michigan) have done a monumentally bad job of the process. You can search for terms and see a portion of the page in context, but there is nothing to indicate the issue in which the result was found. For example, there is a “block diagram of a contactless keyboard switch” on “Page 60” somewhere between issue 133 and 144, with no guarantee that these are even the correct page numbers. Instead of treating each issue separately, they appear to have scanned entire volumes as single documents. Although the request seems unlikely to succeed, the Linda Hall Library may be able to scan in the B2H article: this depends entirely on whether they are able pay the appropriate copyright royalty payment to its Japanese publisher.

In the meantime, we now have the switch height (29 mm, same as B2R), an explanation of the model numbers and the different operating forces. This tells us that B2H-F7W is a tactile model (“F”) with open collector outputs (“7”). The “W” suffix is not known to be covered and may indicate a PCB-mount model.

Further, we have some brief explanations of some of the subdivisions of the B2KH keycap series, that has now been moved to its own page. B2H has yet to be split off, as the page still shares material with B2R and B2A, of which not enough is yet known.

Wednesday, 12th January

In December last year, a switch was presented to Deskthority for identification, marked “IEVT” and having SMK JM-0404-like mechanical contacts. I mistook them for being a Far East type in the same vein as the “HXJ-1”, “CS” and “YH-B” types with equally unidentifiable branding. The switches looked similar to old Tipro switches, but did not appear to be sufficiently similar. Consequently, I did not investigate them further, incorrectly assuming that it would be hopeless.

Fortunately, the owner promptly traced the branding to IEVT: Inštitut za Elektroniko in Vakuumsko Tehniko (Institute for Electronics and Vacuum Technology), a Yugoslavian organisation based in Ljubljana, now in Slovenia. More specifically, they located a Serbian seller offering for sale an IEVT switch and keycap brochure. At only 360 dinar (around €3 or $3.50) there is no reason not to snap this up, but my attempts thus far to obtain it have been thwarted. The brochure remains available with no takers, or at least no successful takers.

However, there was a clue in the mention of Ljubljana, as that is where the aforementioned Tipro are located. I enquired with my contact at Tipro and he revealed that Tipro were formed by staff from IEVT, and he found a later version of the brochure in their archives where the Tipro logo had been added. The details from the brochure are now available on the new IEVT Series TC (tall reed switches), Series TY (typical pre-DIN mechanical switches), Series TZ (DIN-compliant switches) and keycap pages.

The parts list for the Galaksija computer gives two possible switch series: TY and TX. The only clear photos show TY, and TX is not in the brochure. Some Galaksija photos hint at a low-profile RAFI-like switch that could be TX, but this remains to be seen.

Huge thanks to Tipro for helping out with this research.

Monday, 3rd January

The mystery of Key Tronic’s “Butterfly” switches can finally be laid to rest … for the most part, at least. One of the staff involved with the design has confirmed my suspicion that that switch depicted in US patent 4209819 “Capacitive keyswitch” is indeed the “Butterfly” type. The “Butterfly” name presumably stems from the shape and motion of the movable capacitor plate. “Butterfly” was to be a fully-discrete capacitive switch, soldered to a PCB, with the variable capacitor entirely contained within the switch housing; the design would have no foam pad, which (in theory at least) would avoid the increase in force associated with compression of the overtravel pad.

The finer details, and the rationale for abandoning the design may never be known. The principal inventor—Ewald E. “Sig” Seignemartin—passed away in 2012, and the person who confirmed the patent is not at liberty to disclose details as he left Key Tronic many years ago. We only know that the “Butterfly” type never entered production.

For whatever reason, Key Tronic decided to continue down the foam pad route, introducing the discrete plate-mounted switch module design instead.

Saturday, 1st January

2021 knowledge round-up

Another strange year has come and gone, with disappointingly no huge breakthrough to match those of 2019 (and Cortron have failed to come through, as it seemed most likely they would: this is something I’ll have to tackle alone). With that said, the collection of outstanding mysteries has been steadily chipped away, and general knowledge is steadily rising.

As part of my ongoing process of documenting the electronics within keyboards and the encoding process in particular, there are now dedicated pages for each of MOS/LSI single-chip keyboard encoders (commenced in December 2020 and expanded significantly in 2021), microcontroller-based encoders and the related but rather different keyboard controllers that go inside the computer instead of the keyboard. The encoding and output page has been continually improved throughout the year as new information comes to light.

Datanetics history has seen a significant boost. A chance discovery of Jamie Fingal, daughter of the late Datanetics president Marshall James Styll, as well as more information from Meryl Miller, has allowed Datanetics history to be expanded considerably. 2021 has seen the discovery of several varieties of their original batch-fabricated array keyboards, and Meryl has explained their origin in NCR’s CRAM 2 memory storage system. The first photograph of a calculator with a batch-fabricated keyboard has come to light, from Meryl’s private collection.

Other highlights of 2021 include:

Four series of Maxi-Switch discrete switches have now been formally identified, as has been the origin of the name “Maxi-Switch” (from maximilian armour). Series 2700 reed switches and Series 3100 mechanical switches are both Amphenol-designed switches, and possibly both part of Amphenol’s 601 series (details on the Amphenol products remains limited, and Amphenol have failed to respond). The “vintage linear” switches—and the keyboards built around them—are 6000 Series (or Series 6000, depending how they felt like writing it), while the extremely rare “integrated dome” types are 8000 Series. A number of 6000 Series switch part numbers have also been recovered from NSN entries by way of HP part numbers.
Despite Corton’s refusal to co-operate, progress in understanding Licon and Cortron keyboards continues. The widely-encountered “full-height” type—which is shorter than the “intermediate” type but has a larger PCB-to-plate distance—is now determined to be Series 54 Superswitch II. A number of switch part numbers for these switches have also been recovered from NSN entries via HP part numbers. General understanding of their product history is improving with time. It’s still a huge shame that whoever bought the 1971 55-100007 keyboard from eBay has squandered it: this is one of many keyboards that, once sold, were deleted from the Internet leaving no trace that they ever existed, and one where I forgot to preserve the photos. Possibly this was an original Superswitch keyboard; details of early 70s Licon keyboards remain extremely scarce.
SMK JM-0400 reed switches, and the corresponding elastic contact types, have now been seen more closely, but the owner has predictably failed to document these properly. Currently there is no information on whether the elastic contact type—which is the basis for Maxi-Switch 8000 Series—is part of JM-0400 series, and the JM-0400 membrane switches remain to be encountered.
There is now tentative evidence that the two styles of Mechanical Enterprises T-15 Series—sealed metal contact and elastic contact—represent two separate series: T-15 sealed contact and T16 elastic contact (always seen written without the hyphen). “t16” is the name that Mechanical Enterprises themselves previously offered to someone else (which I took to be a mistake); this name has since appeared in a keyboard flyer in Bill Buxton’s collection, one that is not yet publicly available (and cannot be shared as he has vanished into oblivion). Datalux (formerly Mechanical Enterprises) no longer respond at all, and no visual depiction of T16 is known, so it may be a while before this matter is cleared up.
Omron B2C has now been split off from the B2x page following the revelation that this is in fact a capacitive series of keyboards, using foam pad switches.
Hirose solid-state capacitive has been found, which was unexpected; naturally, this has not been properly documented anywhere.
While it’s well known that ИЗОТ cloned Cherry M8 (as 0130Д) and M9 (as 0125Д), I don’t recall seeing a clone of Cherry M6/M7 before. A mystery switch found in Romanian keyboards has now been determined to be a copy of M7. It’s not an M7 clone; rather, it simply copies the way that M7 is designed and operates, but the dimensions do not match M7. The manufacturer remains unknown, but FEPER or ICE Felix seem to be likely. ICE Felix also made keyboards with what now seems very likely to be a clone of Stackpole KS-200E interlocking switches (already vaguely known but now more examples have come to light); see the Eastern Bloc switches page for brief details. More information may be forthcoming.
Last but not least, a number of Alps switch series have now been identified by way of an Alps Electric low-profile keyboard switches sample pack. This was originally sold on eBay, and neither the eBay seller nor the buyer ever made any attempt to actually document this discovery. The pictures were later salvaged from Discord after being accidentally rediscovered, and copied to Deskthority where they are now accessible. In some ways it’s remarkable that 2021 has seen the progress that it has with the keyboard community so determined to not recognise, preserve or pass on knowledge. Discoveries are often left trapped inside Discord where they are quickly forgotten.

The following remain my top outstanding mysteries, and maybe somehow 2022 will bring some answers:

Details of the various Mitsumi discrete keyboard switch series (KAM, KCT, KDM, KLM, KLT etc): not a single shred of official information has shown up for any of these types yet
Clearer details of the various Futaba series (MD, ML, MR etc), as details of these products remains fragmentary and vague
A clearer breakdown of SMK JM-0400 and details on the successor DIN-compliant series that remains unidentified
The identity of the “intermediate” Cortron ferrite core switches, and the external appearance of the mysterious Cortron Superswitch 3
The meaning of “WEAB” in Mechanical Enterprises T-5 Series
The manufacturer of the “Katano” switches
Details of Omron’s B2x range of reed and Hall effect keyboards and switches
Clear details of the product range of Pye Electro-Devices AKA PED, in particular the switch series used in their BBC Micro keyboards

There is also a long list of unseen switches: switches known from patents or advertisements that have yet to show up in any keyboard. This list includes a number of exotic types, in particular the final fling of the self-encoding bounce-free school of thought before matrix scan took over the industry completely in the early 70s.