Factual updates and other interesting changes are reported here. Adjustments and tweaks are not listed.
Follow the updates RSS feed:
Updates for 2019
Tuesday, 31st December
Further inspection of Cherry MX switch contacts suggest that Cherry had at least three production lines of MX switches:
- Germany (with “M” style movable contacts)
- Japan (with “A” style movable contacts)
- United States (with “A” style movable contacts)
The “linear clear” MX switches that cherry-jade found are anomalous compared to German and Japanese MX switches, with “A” style movable contacts, but German-style gold-plated wire contact points. The metal surface is also mottled, as though heat treated. It turns out that the contacts are a match for those in the US-made Cherry KXN3-8451 keyboard. This implies a third production line, which may well have been in the US. Manufacturing at Cherry’s US facilities ended in the late 80s, which would be one reason why these switches are so rarely encountered.
The question now is: was “M” style ever used in any German keyboards? “A” style contacts are so rarely observed, that it is hard to be sure.
Sunday, 29th December
Inspection of the switch contacts in the NOS Hirose MX1A-0NNN shows them to be A type (see under MX switch contacts). This suggests that the switches in the US-made KXN3-8451 were sourced from Hirose Cherry and not Cherry Germany. More data is required to understand this more clearly.
Saturday, 28th December
Dr.Click AKA Dr.Dick sent some switches by way of UncleFan. This included something very peculiar: “FD”-branded blue Alps. These are not the ones that show up on Taobao (that are simply SKBM Grey under blueish light): these are proper SKCMAG switches branded “FD” instead of “ALPS” on the bottom, almost the same as the “F D” branding on later SKBL/SKBM switches.
They are grey switchplate and pine, which means that they are neither late-issue nor early issue. They were found in a scrapped keyboard with no case, and Dr.Click threw away the PCB before realising that the switches were unusual. They are interspersed with Alps-branded blue Alps switches, and all of them have a particularly dark shade of blue pigment.
Thursday, 26th December
It has been previously noted on the Deskthority forum that Futaba sealed linear is internally identical to Futaba MD switches. (And indeed, if “MD” indicates a sealed contact type, they may well be MD switches.)
For perhaps the first time, the internals of one are now available to see, thanks to a set of switches from a scrapped Rockwell AIM-65 keyboard. The main difference between the two types is that the sealed switches contain a jumper that also serves as additional securing points for the switch.
Friday, 20th December
The Interdata Master Part Number List from the 26th of January 1979 listed some curious part numbers for what seemed to be 2SD Series keycaps, for Micro Switch SD switches. These are now confirmed to be true, per 2SD Series Chart 30 M supplied by Honeywell. The same part number was used the Interdata listing for three colours and three legends appears, which seemed incorrect, but from the chart, appears to indicate that the part numbers in the Interdata part number list are incomplete. 2SD339 (as listed) does not include the legend text or colours; these are additional characteristics beyond the basic catalogue listing, which only indicates a single-level, single-unit keycap with integrated red LED.
Monday, 16th December
The Micro Switch KB page has received a major update, based on the charts from Honeywell and the promotional literature scanned in by the Computer History Museum. The charts indicate that the reed switches provided both momentary and alternate action, as well as audible feedback in a single model, 7A1HW. Sadly, the specifics of the audible feedback arrangement are not available, as the relevant charts appear not to have survived.
The promotional literature (previously uploaded) also details the method by which KB provides self-encoding key stations: all switches are wired to a common bus, and the bus connections define an 8-bit identity for each switch.
The trombone slide below some switches is found to be a connection from the plunger through to the bailing and lockout mechanisms that can be installed on a lower row in the assembly. This arrangement is supported by the reed switches too, although this has yet to be seen.
As a note, the difference between black and grey shells for KB reed switches is now clear. Black shells are SPST momentary only, while grey shells support up to four poles, bail and lockout, alternate action and audible feedback. Although four poles are physically supported, the limit in all documented types is two poles.
The symbols on the shells seem to be the date codes for the reed capsules themselves.
Monday, 9th December
Theodor from Honeywell has delivered another batch of charts. Most are for 7A1H and 7A1M switches from Micro Switch KB but a few are for Micro Switch SD Series, including a replacement 2 M chart. From the latter, we now learn:
- Keystems S and T denote switches with a small ridge around the plunger aperture (the 0.05″ extra bearing surface mistakenly conflated with the 0.05″ taller plungers in the black-plunger types). The plunger is drawn the same as for 4 A and 4 B, but each plunger type has a different part number than for the standard switches, so they may still be different.
- The flat-stepped plunger style is “N” not “H”. This is clearly readable on the new high-quality copy of chart 2 M, but on the fax-resolution copy from Al Kossow, I misread it as “H”.
- The are three custom plungers for NCR: C, D and H. Chart 4 A does not say which is which: C is listed first but depicted in the centre. All three are quite different. Just as how Datanetics made a custom DC-50 type for NCR, it seems that Micro Switch made various custom SD types for them also.
The 1001SD charts PDF is re-uploaded (now with “M” in the filename) with a replacement chart 1 and the A charts deleted; this PDF is now M charts only. The available A charts are placed into a separate PDF.
Details of 7A1H and 7A1M will be posted once I have had time to digest the charts, some of which have suffered in quality during archival at Honeywell. One curiosity is that the split between “quick connect” H and solder-terminal M is not perfect, with one H type seemingly having solder terminals. Also, KB reed switches supported quad pole.
Saturday, 7th December
I have now started a wishlist page for parts and information.
Monday, 2nd December
The meanings of two Micro Switch SD Series switch variants are now known (by way of 1001SD Series Chart 13 A, which depicts the assembly of SD18 tactile), although both variants are ones not previously encountered.
Successful disassembly of SD tactile model 1001SD11A1A (with only scuffing to the shell) shows that the plastic pin that engages with the shell is a mere 3.7 mm long and 1.6 mm in diameter, and the spring it presses against is 3.1 mm long and 1.55 mm in diameter. This is only marginally larger than the latching parts in RAFI alternate action switches.
Wednesday, 27th November
One of the 80 pages of Fujitsu literature that Kiyoto shared with me contained a curious diagram, depicting three “snap-action” switches side by side in cross-section view. From left to right, these are:
- A tall switch captioned “富士通” (“Fujitsu”) which is their FES-5 clicky reed switch
- IBM beam spring drawn at half its actual height
- A mystery switch
The captions for the other illustrations are illegible due to the photocopy quality. However, the third one bore a surprising resemblance to Micro Switch SD Series alternate action (per one of the charts that Honeywell sent me). This implied that the tactile SD types use a sprung pin arrangement. As I have acquired a couple of 11A1A tactile switches from Radwell, I inspected them with a torch to see just what was inside. Previously I had not observed anything of interest, but now I could clearly see a small brown pin (seemingly plastic) that retracts when the tactile point is overcome.
SD Series is an early tactile type, with the chart date being 1976-09-20. It has already been shown that Fujitsu FES-8 reed/Hall switches offered a tactile leaf as early as 1976, and it is now clear that SD Series did not offer such a design. Micro Switch SW Series also had a rare tactile type (made for IBM) and the internals of that are not yet understood.
(And yes, it is now fully confirmed that FES-8 offered Hall sensing as well as reed, with the idea being that normal keys would be reed to save power and offer high reliability, while “functional keys” would be Hall in order to drive ICs directly, presumably by virtue of being bounce-free.)
Monday, 25th November
Based on the information that Kiyoto has sent so far, the Fujitsu reed keyboard switches page is now updated with the basic details so far of the known types: FES-1/2/3 (metal plunger), FES-4 (cross-reed), FES-5, FES-8 (tall) and FES-9 (less tall). The series names of all discovered reed types are now known. The big mystery that remains is where the term “Peerless” came from.
Sunday, 24th November
Kiyoto has turned up a treasure trove of old Fujitsu literature, which has provided some important data:
- The tall switches found by UncleFan in keyboard model N860-1131-T010 are now confirmed to be FES-1 just as we suspected. These are double-pole-capable reed switches (part of Fujitsu’s reed switch product history). FES-1 is the non-illuminated series, while FES-2 is illuminated; both types were available in momentary and alternate action forms. FES-3 also exists, but we don’t know what its characteristics are.
- Fujitsu “lever action” is FES-370 series, from their スーパーエルゴノミクスキーボード: Super Ergonomics Keyboard. This is numerically one step above FES-360, the discrete leaf-spring switches based on FES-300 and FES-301.
Additionally, a new Alps SKCL switch has shown up, the “ベラボースイッチ” or “Bravo Switch”. Details on this are scarce; it is either a double action switch, or a keystroke speed sensing switch. It was used in cabinets for the arcade game Bravoman (ベラボーマン, amusingly rendered in English as “Bella Bowman”, which is actually written ベラ・ボウマン). No such switch has knowingly been disassembled, and the only description of its operation is in a photo on Erwin’s blog, a site that has been defunct since 2013. From this description, it is not clear whether the second stage is triggered by a high speed impact or a high force impact. What is curious is that the plunger is amber, as with the click switches used in later Apple IIc computers.
Saturday, 23rd November
If anyone can read Japanese and is interested in Alps switches, let me know. I will provide you a long list of Alps patent and utility model numbers, including those for Texas Instruments fat Alps, KFL (momentary and alternate action), SKCP double action and KCC alternate action. The scan quality at J-PlatPat is reasonable, but not sufficient to be able to transcribe all the kanji unless you are fluent in Japanese.
Wednesday, 20th November
Back in 2016, I came across a couple of Korean websites (ePartsHub and ECPlaza) that provided specifications and photographs for Futaba/Sejin switch subseries (MD-4P, MD-7, MR-6C, ML-88, MA41/42 and MA71/72). This allowed most series to be given names that are meaningful even if not fully accurate. This data is all gone from ECPlaza but it remains available on ePartsHub under Push Button Type Switches. (Repeated attempts to contact both sites for further details turned up lots of dead and ignored e-mail addresses.)
What surprised me the most was the rated lifetimes: a mere three million for the little unreliable switches in the BBC Master, ten million for the high-end switches, and yet 30 million for the cheap switches in the BBC Micro. This always seemed implausible, but there was a clear photograph to prove it.
With the discovery of limited amount of official documentation, it is now clear just why that 30 million figure seemed so unlikely: the photograph of MR-6C shows the wrong switch. Comparison of the diagram of MR-6 in the catalogue against the diagram of KBR in the GRI catalogue shows that the two are a close match, while the diagram of ML-41 is a close match for the switches we have been calling MR-6C. The specifications for MR-6/6C also match up with those of KBR. This means that MR-6 is the missing reed type from US patent 4041427 (MR-6C may just be the type with a cruciform mount, but this is not certain).
The catalogue is still sloppy, inaccurate, incomplete and confusing, but it does clear up a major misunderstanding. It also shows that ML switches are not even defined in terms of size, and suggests that instead they are defined in terms of switch contact design, with both sizes being rated for three million cycles each. This single low-rated lifetime makes sense from the perspective of how they are designed, but not specifically why the 2.5 mm version seems to have much lower reliability than the 3.1 mm version. One would also like to believe that MR being reed would be signficant enough to warrant special mention, and it may well be that MR is not reed, but all indications are that it is.
The classification now seems to be as follows:
- Reed, 30 million cycles lifetime, rated for 24 V DC at 5 mA; standard height with 3.1 mm travel
- Sealed metal contact, 10 million cycles lifetime (for momentary non-illuminated; there are strange discrepancies with the other types), rated for 24 V DC at 1 mA; standard height with 3.1 mm travel
- Metal foil contact, 3 million cycles lifetime, rated for 24 V DC at 1 mA; available in standard height with 3.1 mm travel and low profile with 2.5 mm travel
The sealed type and a newly-reported tag terminal type remain without series names. Since ML-81 seems to be a replacement for ML-3, one could argue that the type found in the BBC Micro is going to have a subseries such as ML-1 or ML-2.
Sunday, 17th November
Study of Omron’s discontinuation notices has allowed further details to be found about their reed and Hall effect switches (B2C, B2G, B2H, B2A and B2R), but nonetheless very little is known about them. B2A and B2C have still yet to be seen, and there is no guarantee that all of B2R or B2H shares the same form factor.
Friday, 8th November
The hypothetical Maxi-Switch series 4000 has an official name, as mentioned in Electronic Design magazine in 1976: Captron.
Thursday, 7th November
Electronic Design have given their blessing for me to upload the Focus on Keyboards articles that I have had scanned in by the Computer History Museum and Linda Hall Library. These are placed onto the new references page.
Tuesday, 5th November
I placed a photocopy request from the Linda Hall Library for the 1972 article “Focus on Keyboards” in Electronic Design magazine (ED1972-FOK), which I received a couple of days later. As is always the case with such endeavours, there is no huge breakthrough, but only little morsels of progress. Sadly most of the series covered are not named, including Licon’s ferrite core switches (depicted as the original tall type, advertised as Series 550).
However, there were some surprises. Two Stackpole switch series are mentioned. One is a “flying magnet” type referred to both as “Stackpole Magsat” and “Stackpole/Magsat”. Although it is depicted as having a “magnetic satellite”, Magsat filed the patents (GB1370447 and CA936900) and they also patented a torsion spring type similar to a Stackpole type. The other Stackpole type is a calculator switch called “LO-PRO”, which is likely the name of Stackpole torsion spring whose first patent was filed in October 1972.
The article also names Controls Research Corp’s concentric spring switch—supposedly the switch in the MCM/70—as the “Bi-Pac”.
Some other series names, such as those from Maxi-Switch and Oak, are already known from advertisements of the time.
A little bit of progress. Nowhere near as much as expected, but a tiny step forwards.
Monday, 4th November
I am now starting to think that at least some RAFI full-travel key switch stamping numbers are the date code, which would make sense for the samples they sent me. In the case of the orange diode type, marked something like “258 2”, the date code part makes no sense, but “2” suggests 3.13.005.002, and I found that this part number exists.
I have now posted a huge list of apparent RS 74 M, 76 M and 76 C part numbers, showing how large the range seemed to be.
Sunday, 3rd November
UncleFan recently drew my attention to an unusual Fujitsu keyboard, model N860-1131-T010. (He showed me that it was for sale on Taobao Idle Fish, and Kiyoto found where the owner had previously posted additional photographs.)
The switches inside have not been explored. Tentatively we are calling them “FES-1” for the moment, as both the keyboard series and switch part numbers begin “1”. Whether or not they are reed switches remains a mystery. The keyboard has already sold.
Thursday, 31st October
The Computer History Museum scanned in for me (for a fee, $1/page) an assortment of Micro Switch Series KB literature (allegedly 400 DPI, but it is low resolution and out of focus). Still nothing on the keyboard reed switches, but there is a lot of detail on the other weird switches in the series, including the centipede and trombone types. The former are encoding switches, where the ASCII or scan codes are generated purely electrically without the need for a controller.
Wednesday, 30th October
Some people will have noticed that certain Micro Switch switches, in particular Series KB, can be found marked with pairs of symbols, either in addition to (with Series KB) or in place of (with SM Series) the date code.
I have now made a chart of symbol codes. So far, no meaning can be ascribed to them. Suggestions include calibration markings and date codes, but they do not appear to work as either. If they are batch markings, this was either a specialist or very short-lived process, as these symbols are seldom seen.
Monday, 28th October
Re-examination of unidentified Maxi-Switch keyboard 3101-090-02 reveals that the switches have a patent number ending in “10,060”. US patent 3710060 fits the switch description well … but it is a Bunker Ramo patent! That was one switch type that looked like Maxi-Switch originated it, but possibly not. However, this does appear to confirm that the keyboard, and PCB codes beginning “31”, are Series 3100 as expected.
This raises the possibility that Maxi-Switch licenced the design of “vintage linear” (possibly Series 6300) off SMK instead of copying it without permission.
Tuesday, 22nd October
Back in August 2018 I proposed an explanation for the origin of M1 (or M11): that it was created by Cherry for HP as a drop-in replacement for Datanetics DC-60 switches, based on a comment by Findecanor in 2014 followed by a comparison of dates and specifications.
After UncleFan came across the internals of a Cherry-made HP 9835 keyboard with M1 switches and a customer part number almost identical to that of the M1 space bar switch, I spent some time investigating, and discovered that the customer part number (Kundennummer or Kd.Nr) in the M11-0101 drawing is actually the HP part number assigned to the Datanetics DC-60 switch used for space bar on the HP 9825 computer, i.e. HP assigned an existing part number (for the switch to be copied) to Cherry when commissioning its replacement. HP for the most part pretended that they made all the switches and keycaps, so there are no Cherry or Datanetics part numbers up for grabs, but the DC-60 switch and Datanetics-made keyboard are clearly illustrated.
For the most part, in that the HP 85 service manual distinguishes between the Hi-Tek and Stackpole switches: Hi-Tek has one solid contact and one fingered contact surrounded by a white plunger, while the Stackpole switches use yellow plungers and a single contact type. This confirms for one that HP did buy from both companies, as believed.
Saturday, 19th October
Additional Maxi-Switch details and examples are up. As is usually the case, there are more examples out there than I realised that I had forgotten or never seen, but nowhere near as many as required to fully understand the product range. The “vintage linear” switches are the most common non-rubber type.
Friday, 18th October
Brief Maxi-Switch page added, just with a table of keyboard examples for the minute.
The suggestion is that “vintage linear” might be Series 6300, depending on how the PCB codes are to be interpreted.
The first example of the older Series 3100 mechanical switches appears to have shown up also.
Tuesday, 15th October
Deepak Kandepet has taken a series of photos of the Alps SCK variant that appeared to be double pole–capable. This has throw up a few surprises. Firstly, it is single pole: the extra terminal holes are seemingly not usable.
More interestingly, it appears to be a lost motion design that offers hysteresis, apparently based on Datanetics DC-50. It may also be the switch type that was classified KCA.
Sunday, 13th October
- Until such time that I hear from Honeywell again, I have now posted the 1SW, 1SW100, 11SN, 101SN and 201SN charts that I have thus far, on the SW Series and SN Series pages. These are internal documents, and the engineering notes are redacted per Honeywell’s request; this does remove some useful information regarding identifying the various types. (I am still hoping that they have older revisions of these documents with less internal redactions, to fill in the details of more of the modes.)
- Sometime around the end of August last year I had an x-ray taken of the two fully-assembled Datanetics DC-60 switches that Meryl sent me: one perfectly linear and one highly tactile. Disappointingly, both are identical internally, despite feeling very different. I have finally posted the x-ray image on a new makeshift Datanetics DC-60 page.
Wednesday, 9th October
Further rummaging has led to the rediscovery of an ITW patent that helps differentiate two of the early ferrite core types. Re-reading the patents page has led to matching up the patent that goes with the double-action ferrite core switches in the HP 9845 workstation keyboard. Understanding of Licon/Cortron/ITW’s ferrite core product range is getting slowly clearer.
Tuesday, 8th October
It seems that I forgot to post a link to the utility model for Omron “Alps-style” switches. The utility model was filed in 1988, giving a rough idea of when these switches were introduced. Sadly there is still no sign of a series name for them.
Sunday, 6th October
At the same time that Bitsavers uploaded the Electronic Engineer magazine keyboard guide, they also uploaded two Cherry data entry keyboard brochures from 1971, which are now the oldest documents to date on Cherry keyboards. These documents offer two very interesting findings:
- It was already known that Cherry filed a reed switch patent in 1970. These Cherry reed switches are now confirmed from official documentation. Only two models are known: 201-0100 (SPST-NO) and 202-0100 (DPST-NO).
- The documents also confirm that Cherry M6 started off with different part numbers. UncleFan already found a Cherry sample card demonstrating this, and this too is now confirmed from official literature. The reed switches were assigned 20x codes, while the gold crosspoint switches were assigned 26x codes, where x denotes the number of poles. Within two years, the “2” would change to “M” for the gold crosspoint switches, and the reed switches would disappear.
Saturday, 5th October
I discovered on Friday that, in July, Bitsavers scanned in a guide to keyboard technology, published by Electronic Engineer magazine in 1971 in conjunction with Micro Switch. This guide has some interesting definitions of switch types.
Depicted in that document is a mercury switch, that is clearly a Mercutronic type from Mechanical Enterprises. After discovering that the all the archive scans of Electronics Australia were removed from americanradiohistory.com for copyright reasons, I went looking to see where it had gone, and found one more mention of Mercutron(ic)¹ that I missed: a depiction in Computerworld—also from 1971—of the exact same switch (complete with the external return spring and base missing from the other diagram), along with its model: MC-210. This implies MC series.
¹ both “Mercutron” and “Mercutronic” can appear together …
Saturday, 28th September
Deepak Kandepet found an Alps calculator keypad on eBay, with PCB code CH34149B. The switches are the type that resembles spring bridge but with a metal top, but the PCB code begins “CH”. This demonstrates that “CH” PCB codes are not specific to the series that became KCC. This suggests that “SCH” is not an older name for KCC, but instead a classification code for mechanical switches. This new theory all depends on what contact type is inside SCK switches, because the theory would indicate that these are neither mechanical nor reed.
Saturday, 31st August
Curiously, it seems that there were at least three designs of Clare-Pendar S820 switch. Characteristically dreadful photos of the third type to be discovered are now posted. After revisiting the S820 and S880 catalogue page provided by Terry Trumbull of Electro-Mech, it seems that this third design is that which is depicted in the photograph in the catalogue. As the catalogue page was scanned in 1 bit per pixel, all detail is lost from the photograph, but it does appear to have the diagonally-placed retention clips of this third design.
Although S820 is nominally 2.5 oz, the example in question is significantly lower in force than a Micro Switch SD Series 2.5 oz switch. As expected for a reed switch, it is smooth, and noticeably smoother than SD Series.
Tuesday, 27th August
It seems that the plunger colours for Micro Switch SW Series non-illuminated switches may finally be understood. The original plungers were grey (ca. 1968–1971) followed by black when the plungers were redesigned. Around the mid-70s, the momentary switches were then colour-coded by sensor type: red for sink level, blue for sink pulse, and seemingly green for source level. The last one based on the assumption that “C”-type SN switches are source level, thus green-plunger SW switches with the same sensor type (per the switch bills of materials) should also be source level.
The pale red switches sometimes seen are just red switches, which for some reason can either be a vivid red, or a soft red.
Wednesday, 14th August
Series 7 is not Series 7 after all.
With the help of the fantastic staff at the Smithsonian Libraries (Christine and Trina), we have traced the reed switches to Series KB. For purposes of maximum historical confusion, these switches have an alphabetic series name but part numbers as though the series name was numeric.
By the 1973 Manual Switches catalogue, details on the keyboard switches were very cursory, likely because in keyboard applications they were long superseded by SW Series keyboards and switches.
As for SW Series, one of the last remaining anomalies is 1SW12-BL, which has been M demonstrated to be tactile. That must be what “BL” denotes. Sadly this is not in any of the recovered 1SW Series charts.
Saturday, 10th August
A new, dedicated page on Micro Switch 1SW200 Series, with the first photos of the SW Series bi-level (double action) switches.
Thursday, 8th August
Finally my Micro Switch SW Series models 1SW201-R and 1SW204-R switches arrived from the US. These are exactly what I hoped they would be: the elusive “bi-level” (double-action types). Double action in SW Series is achieved not with two sensors, but with a larger sensor IC with two Hall elements integrated. The stage two springs are attached to the plunger and protrude down from it, making contact with the switch mounting frame.
The sensor in these switches is type “U”.
Tuesday, 6th August
All evidence now indicates that the switches used in Micro Switch RW Series keyboards are
Micro Switch Series 7 . Micro Switch produced a number of switch series with numeric series names, with Series 2, Series 3, Series 4 and Series 6 being confirmed. The catalogue listing numbers of all these series follow the same pattern: the series number, then a classification letter, followed by additional letters and numbers. Series 7 appears to be a derivative of Series 3; both series use the same rail mount system. Like Series 7, Series 3 comprises both reed and non-reed types. Series 7 switches came in a variety of configurations, of which not all were suitable for keyboards, pointing to the series not being specifically intended for data entry keyboards.
Additionally, the keycaps for Micro Switch SW and SN Series and Micro Switch SD Series have been found to be 2SW Series. Although not confirmed by literature, all keycap part numbers appear to fit into this subseries, including those of illuminated switches.
Monday, 5th August
Two years after receiving it in 2017, I have finally scanned in the complete George Risk Industries keyboards and switches catalogue that they sent me via Chaldon International in the UK.
Sunday, 28th July
Micro Switch PB8* keyboard switches are starting to make sense now. The difference between 1PB877 and 1PB878 (as found in Univac keyboards) is now clear: the former is one-shot and the latter is continuous. This exactly corresponds to the use of pulse and level switches in SW and SD Hall effect keyboards. While some PB types are electronically one-shot using an attached PCB, one-shot PB8* keyboard switches mechanically release the internal microswitch immediately after operation. The reason is not known, but it may be the same as with the Hall effect keyboards: a way to achieve N-key rollover.
Friday, 26th July
It turns out that there were at least three series of Digitran metal leaf capacitive switches. These are now collated on a new Digitran page. It also appears that Digitran’s metal leaf capacitive keyboards preceded those of Cortron by several years.
Saturday, 20th July
- As part of the process of documenting Micro Switch and Honeywell, some brief notes are now up on RX Series and WN Series rubber dome over membrane keyboards. Very little is known about the internals of these.
- Understanding of S-type three-letter Alps series names and model numbers is either clarified or muddied—depending on interpretation—by Alps SCB1A163, which has a “CB” PCB code, matching the model series of SCB. This casts doubt on the assumption that the Olympia CD 102 keypad PCB code begins “CH”.
Wednesday, 10th July
The ITW, Cortron and Licon page is updated with a list of Cortron and Licon keyboards and their part numbers, which clearly follow patterns.
Friday, 5th July
The National Museum of American History Library were so kind as to provided scanned copies of the two Micro Switch SW documents in their collection:
- Micro Switch Product Brochure SW: Solid State Keyboards, possibly March 1973
- Micro Switch Product Sheet 51SW12-1: Typewriter Keyboard, possibly August 1973
As a result of the growing amount of data on SW and SN Series, the SW/SN page is now split up:
- Micro Switch SW and SN Series overview
- Micro Switch SW Series keyboards and switches (includes the aforementioned documents)
- Micro Switch SN Series
The National Museum of American History Library has been inducted into the newly-opened Keyboard Hall of Fame.
Monday, 24th June
Edward Walsh is now consigned to the Keyboard Hall of Shame. All the sites run by people who don’t permit you to contact them get tiresome in themselves, but even more so when they hold knowledge that was never made available properly and they don’t let you enquire about it.
Saturday, 22nd June
While digging around for more Micro Switch information, I came across an article in Modern Data from December 1970 for Licon Series 550 keyboards and switches. This was the first time observing an official term for one of their products. Cortron themselves suggested some series names (although they have yet confirmed them from their literature), one being “55”. Further digging shows FC2500 to be another official term, which appears to be their DIN-compliant ferrite core keyboards, that eventually transferred to Devlin.
A brief page is now up on ITW, Corton and Licon with all notes so far.
Monday, 10th June
I am still working my way through Micro Switch’s keyswitch product range with the help of Honeywell. It’s now clear that SN Series comprised three subseries: 11SN (snap-in panel mount, both illuminated and non-illuminated), 101SN (self-adhesive PCB mount) and 201SN (illuminated version of 101SN). A list of 11SN Series types is up, with charts to follow once I have had time to redact them, as they are all engineering charts. These details may offer some insight into SW Series keyboard switches, which were intimately related to SN Series.
Sunday, 2nd June
UncleFan has rescued me from a blunder of Japanese translation. I misinterpreted Omron PRDN 389 as being a list of switch models, thinking that “キーボード形” meant “switch types suitable for keyboards” on the basis that all the part numbers began B2C, B2H or B2R.
He has since found Omron B2R-84E2L-602, and it is a keyboard, not a switch. Therefore, PRDN 389 is a list of discontinued keyboards, and now all the codes listed make far more sense! All the strange numbers in the middle are of course just the number of keys, analogous to how Micro Switch numbered their keyboards, using the same series names for switches and keyboards.
Saturday, 1st June
Based on all the charts from Honeywell, the Micro Switch SD Series page is extensively updated with all new details on the series.
Thursday, 30th May
Another jackpot! This time, Honeywell have turned up 13 charts from SD Series, dating the series all the way back to 1975 (although the official start of sale date is not confirmed). The charts do not answer every question, but they shed a huge light on the series. All 13 charts are posted both in a collated PDF and as a zip file of all the separate high-quality TIFF images.
Tuesday, 28th May
Following the unexpected discovery of an older schema of Micro Switch SD Series switches, it seemed worth giving Honeywell another try to see if they could locate any details on their old switches.
Details are scarce with the series having been discontinued nearly 20 years ago, but they were able to turn up a couple of drawings, one each for SW and SN Series, which has helped clarify a number of historical details. Most interestingly it shows that SW and SN appear to have been companion series from at least the period of 1995 to 1999, rather than SN having replaced SW. With luck, they may able to recover more documenation.
Monday, 27th May
I thought I read somewhere that Micro Switch SD Series was DIN-compliant, as highly unlikely as that seems. That would point to an introduction date of around 1982. Very little practical data gathering has taken place with SD Series, so off-hand it was hard to be sure. However, it is now clear that SD Series was in production as far back as 1976, as demonstrated by model 125SD12-1.
This keyboard uses an older part number schema, with one fewer position in the identification number, and seemingly at least one position with a different interpretation to the 1001SD Series schema found from 1998. One possibility is that the old switches were 1SD Series, and that 1001SD was chosen when revising the schema. This remains pure speculation.
Saturday, 11th May
UncleFan found an eBay listing for the Yamaha QX1 in which a broken Alps KCC switch could be seen. On Geekhack, the QX1 is noted to have “a linear switch found on the NRC f020 Yamaha qx3, and qx1”, which is Hirose MX1A-0NNN. As such, this machine seemed not to be worth further investigation, as the Cherry switch part number could be retrieved from the readily-available QX3 service manual.
However, having now purchased the QX1 service manual from Mauritron (to avoid hassling Yamaha further), I can now see that indeed the QX1 uses switch KCC10903, which is the model number I speculated it would be upon seeing the eBay listing. The switch can be seen more clearly on the Introvert’s Yamaha QX1 Retro Review, and is a standard KCC green momentary switch.
This now means that KCC lower-weight momentary uses old part number KCC10903 and new part number SKCCBK (as previously obtained from Eleport in Japan). This marks the very first Alps data entry switch type where we have both the old and new part number.
Tuesday, 7th May
Based on an unfortunately barely-legible copy of two pages of the May 1985 Fujitsu magazine (which cannot be posted online), the Fujitsu reed switches page is updated with more details on each of FES-5, FES-8, FES-9 and FES-4.
Saturday, 27th April
- Sandy has permitted me to have a copy of the 1994 Alps keyboard catalogue. The Alps series names and model numbers page is now updated accordingly. In particular, “KF” is found to denote “full keyboard”, completing understanding of the series split from three-letter series names (containing both keyboards and switches) to four-letter switch and six-letter keyboard series names. Membrane switches use KF codes instead of SK codes, since they are not discrete units. While this catalogue may not be published online, the 1993 catalogue has also been found, and once this is available, everyone will be able to review this information. (Tentatively it appears that the two catalogues contain the same keyboard switches, except for one additional membrane switch type in the 1993 catalogue.)
- Fujitsu reed switches are now moved to a separate page; very little is known so far, but cross reed is now confirmed to be FES-4 series from a blurry photo of a page from a Fujitsu magazine. More details will be forthcoming.
Tuesday, 23rd April
With the forthcoming 1993 Alps catalogue scan due from lsoelxi, I figured that I should ask Sandy to confirm whether the 1994 Alps catalogue really did give the series names of “CM” and “CL”, as this would revert a change to “SKCL” and “SKCM” by the time of the 1994 catalogue.
Sandy confirmed that “CM” and “CL” were his own abbreviations, and that the 1994 catalogue does name the series as “SKCL” and “SKCM” as expected.
This clears up one nasty point of confusion over the naming. Now we only have to figure out “CM(KCM)”, “CH”/“SCH”/“SCK”, and the new problem that the “KF” prefix—denoting “full keyboard”—was used for both keyboards and switches in the full-travel membrane line! This may become clearer when the whole 1993 catalogue is ready. Sadly the catalogue is not old enough to cover SKFN/SKFM series.
Sunday, 21st April
Added some brief details on Tipro and Tipro keyboards. Currently details on the older keyboards are scarce.
Saturday, 20th April
Added some notes on NMB desktop membrane keyboards as well as the Model NOTE85A notebook keyboard.
Wednesday, 10th April
Kiyoto has located a set of Japanese patents relating to Fujitsu leaf spring. I have tentatively reclassified the various leaf spring types according to the suggested chronology from these patents, but there is still a long way to go.
Monday, 8th April
My TESLA Hall effect switch and some TESLA keycaps arrived from 10ko in Bulgaria today. I have no idea what happened to the switch (the plunger has been sliced narrow and now it is quite wobbly) but the keycaps are intact, and these and the switch permit keycap mount comparison between Micro Switch, TESLA and Omron. These three brands have very similar keycap mounts that are almost but not quite compatible, for different reasons.
These types are now detailed under the new keycap mounts section. Unlike the existing keycap compatibility page, this section places each group of similar mounts on a separate page for simplicity and to permit more details to be recorded.
This section (like any other) will be expanded upon over time.
Saturday, 6th April
Kiyoto is on a roll. He has located a large number of Fujitsu datasheets on their site, including the datasheets for Fujitsu FES-360 Series, which is first-generation leaf spring. FES-360 Series is still in production, but only for existing customers. It seems to be a fairly small series, offering 59 cN momentary action only, although the oldest PDF for it is from 2001, so it may be that other models were long since discontinued.
Thursday, 4th April
Kiyoto (聖人) sent me a link to the datasheet (or catalogue entry) for Fujitsu FKB2500 Series. Although it does not offer any clear insight into Fujitsu’s naming and classification system, it does match the suggested naming arrangement listed under Fujitsu keyboard series. It also offers the first specifications on a Fujitsu mechanical switch type.
Monday, 1st April
Posted a list of keyboard types discovered to date in Roland and Yamaha service manuals, along with Yamaha non-keyboard types. The Roland non-keyboard types will be added later when time permits. This page represents some of the raw data already presented under Alps series names and model numbers.
Also included is the first confirmed instance of an Hirose Cherry M83S, specifically the standard part M83S-0N00 as found in the Yamaha QX21 (1.5 mm travel version).
Sunday, 31st March
As noted under wiki errors, US patent 4529849 is incorrectly cited as being the patent for second- and third-generation leaf spring. In fact, it is the patent for the membrane leaf spring types. Membrane leaf spring has been mentioned at least once—as “4th gen Fujitsu leaf spring”—but not depicted.
Kiyoto in Japan has demonstrated that the Fujitsu FMR-30BX has a moulded Fujitsu membrane leaf spring keyboard, as documented in the patent. It is possible that the integrated-membrane type shown in the patent was changed to use a modified form of the existing contact assembly from the original leaf spring switches.
Thursday, 28th March
- Added a brief page on Cherry M8, MD and MJ series, along with the end-of-life notification for M8 and M85.
- Added all the M8 types from the end-of-life notification to the Cherry parts page, along with the custom MX types and the two known M9 types.
- Updated the M8 schema and common schema pages to reflect the fact that German custom M8 types are now confirmed to exist (as given in the end-of-life notice).
Sunday, 24th March
The Fujitsu date codes and Fujitsu keyboard series pages are both updated with additional examples. In the latter case, this includes 3100 and 3700 (ultra-flat leaf spring keyboards), 8000 and 8200 (tall reed), 9200 (reed that might be shorter) and 2300 (third-generation leaf spring). 2700 is removed for now, as it referred to a single photo that did not show the switches.
Thursday, 21st March
Having received two SN Series switches from STRONIC in France (101SN11 and 201SN1B1, with the latter in its original packet), it has become clear that these are identical to SW series, except for having a base to hold them together for PCB mounting.
Wednesday, 20th March
Since the SMK series and part numbers page does not offer any illustrations, a separate SMK series page is now posted with illustrations of the known series, mostly taken from SMK’s catalogue pages. In particular this includes KBS03 Series, the membrane-based alternative to the better known KBG family (which looks similar but uses conductive domes over a PCB).
Sunday, 17th March
Finally written up and illustrated the dual-magnet tactile and click mechanism of the Omron B2H-F7W Hall effect switches that I obtained in September last year.
Sunday, 10th March
Re-inspection of the Hirose M83S, M83A, MD and MJ series datasheet shows that the force figures are not indicating the permissible manufacturing range, but in fact the tolerance (as recently determined for US M4/M5/M6). The Cherry common schema page along with the MD and MJ schema pages are updated accordingly. The range for MD, 50〜80 gf, comes out at 65±15 gf, and 65 g is what the Epson PX-8 Technical Manual gives for its standard weight MD switches.
Saturday, 9th March
MEI T-5 subseries T-5D has since been confirmed to exist, and is presently on order from Datalux. Once these arrive (Brexit willing) it will be possible to see what they are.
Friday, 8th March
- Phosphor bronze is added to the materials page, under a new spring materials section.
- There is now a Hi-Tek/NMB section; more will be added later. The details of the liquidation of Minebea keyboard subsidiary Shanghai Shunding Technologies Ltd is moved from now deprecated factual nuggets page.
Wednesday, 6th March
- The MEI T-5 page is updated with a list of currently available parts (all on order from Datalux) and the part number schema.
- Tenite™ is added to the materials page.
Tuesday, 5th March
- There is now a Micro Switch/Honeywell section. At the moment this comprises a single page put there to contain the theory that SD16 series corresponds with switches whose sub-part number begins “16” (as found in MouseFan’s Sun Microsystems SD keypad).
- Vintage ABS notes are moved from the deprecated factual nuggets page to the materials page.
Sunday, 3rd March
- List of confirmed MEI T-5 parts added.
- Added a brief page on Clare-Pendar Series S880, to have somewhere to put the link to the YouTube video showing an excerpt from Electronic Design magazine that mentions it.
- Added also a page for Clare-Pendar Series S820.
- Added a section on “PWB” to the Alps keyboard codes page.
- Posted a brief history of Mechanical Enterprises, along with a list of relevant patents and what they apply to.
- MEI Mercutron page is added, listing the patents and what they appear to correspond to.
Saturday, 2nd March
Added newly-discovered T-5V subseries to the MEI T-5 page.
Friday, 1st March
- LFW is now confirmed to be a series (LFW-CT will be a switch model).
- Mechanical Enterprises page updated with the newly-discovered switch series: Mercutron family (including M-5), LM, and LFW (changed from LFW-CT).
- “WEAB” is added to the mysteries page, as years are passing and it remains persistently unresolved.
Thursday, 28th February
- Added the first mention of MEI LFW-CT.
- Added a mention of Don Lancaster’s 1976 book TV Typewriter Cookbook to the T-5 page also, as “T5” gets a mention there (giving us a low water mark for age).
Tuesday, 26th February
- There are now separate M4/M5/M6 schema and M7 schema pages, for convenience, as these are two separate series.
- Explanation of weight ranges for US M4/M5/M6 switches (they refer to tolerance and approximately match the German tolerances).
Sunday, 24th February
- Added the Clare/GI/Pollak/Electro-Mech CAGE code (97564) to the Clare/Pendar page to clear up the mystery of that number marked on switches.
Friday, 22nd February
- Alps OEM model numbers page uploaded, albeit with just two items for now.
Tuesday, 19th February
- Initial list of Key Tronic product codes posted, in an attempt to understand their structure or meaning.
Saturday, 16th February
- Added Newtech YMW, GoldStar YYWW, Intel YWW and Mitsubishi YYWW to the IC dates page along with a copy of the image grid from the Keycombo dating page.
Wednesday, 13th February
- IC dates page updated following determination that Texas Instruments’ three-digit date format represents YWW as expected.
Tuesday, 12th February
- Added a page on Mitsumi models, derived from the list previously posted to the wiki, and extended to cover KCD-A65YU (found in the Commodore SX-64).
Sunday, 10th February
- HFO added to the IC dates page.
- Alps series names and model numbers page updated to note that CM series may have also been called KCM.
- Added a page on Alps PCB codes; for now this is just a stub, but the Olivetti 11K code incompletely documented at Deskthority is moved to this page.
- Added a very brief page on Oak Series 400.
Saturday, 9th February
- Started a page on IC date information, to help people date ICs and thus keyboards and indeed other equipment. More formats will be added over time.
Sunday, 3rd February
- Travel and hysteresis mentioned on the glossary page, as well as on a new Travel, actuation and hysteresis page (the latter being mostly created just to have somewhere to report on the useful information from the General Instrument S700 Series patent).
- Honeywell mount added to the keycap mounts page (as there is a patent for it).
- Switch design characteristics page updated with a brief mention of buckling rubber sleeves, as there is a 1970 patent for those, too.
- Added a brief guide to reading the header fields of Japanese patents.
Wednesday, 30th January
Tuesday, 29th January
- Added a page on SAGEM keyboards.
Sunday, 27th January
- Added a page on common keyboard switch characteristics to help with lay identification purposes.
Friday, 25th January
- Glossary updated.
Sunday, 20th January
- Added a page on TGL standards.
- Added a page on the HFO B 461 G Hall sensor IC used in RFT and RAFI switches.
- Updated the RFT TSH 19 page with brief details on the two variants.
Monday, 14th January
- Started a page on East German manufacturer RFT.
- Started a page for RFT TSS reed switches.
- Added a stub page for RFT TSH 19 Hall effect switches.
- Use of rhodium as a contact material is now mentioned on the materials page.
Saturday, 12th January
- Materials page gains a recreation of Cherry’s contact material selection chart as well as some additional details on material choice.
Wednesday, 9th January
- Updated the Cherry MX page with the years from Hirose.
- Updated the Cherry M7 page with Hirose’s M5 and M6 years.
Sunday, 6th January
- Added MX Olympia Linear Clear to the MX variants page, and revised the page a little.
- Changed the Cherry MX notes page to be the main Cherry MX page, and added a brief table of historical occurrences relating to MX.
Wednesday, 2nd January
- Updated the Cherry gold crosspoint/M7 design page with illustrations and greatly expanded content.