The great size change
- Identity of the standard
In the early 1980s, the Deutsches Institut für Normung (the German Institute for Standardization) introduced workplace ergonomics standards which included requirements for how keyboards should be designed. Compliance with this standardisation led to widespread redesign of keyboards by many manufacturers.
Curiously, low-profile keyboard switches existed in Germany prior to this standard: Cherry M8 from around 1979, and RAFI RS 74 from around 1975. At the Deskthority forum, OleVoip noted:
Whilst there was no requirement for keyboards to be low-profile in the standards of 1980, there was one in the German safety-at-the-workplace regulations of that era owing to the Workers Union's ergonomics movement that started in the mid-1970s, when RAFI already built keyboards as low as that. As an employer, you may ignore standards but you mayn't ignore safety regulations. Hence, low-profile boards already were 'standard' in Germany when a proper DIN standard eventually was released in late 1984.
Identity of the standard
A Guide to Human Factors and Ergonomics, Second Edition by Martin Helander (2006), page 262, references “DIN 66234” in relation to the introduction of low-profile keyboards, but not the name of the standard or which part of this standard is relevant. He in turn is referencing “Helander and Rupp, 1984”, the definition of which is not directly available via Google Books:
The German DIN 66234 standard had a pervasive effect. All computer manufacturers complied with the standard and manufactured low-profile keyboards (Helander and Rupp, 1984).
Looking at beuth.de, the relevant standard might be DIN 66234-6 “Display work stations - Design of the work station” (drafted May 1982, published December 1984), or DIN 66234-7 “Display work stations; ergonomical design of the work station; lighting and arrangement” (drafted October 1982, published December 1984). Both of these parts are contemporary with all the changes identified.
A Siemens Switches and Pushbuttons Data Book (of unknown date; see under Siemens STB 11) confirms this:
Keyboard height – important for correct working posture – in accordance with DIN 66234, Part 6 < 30 mm (measured from the desk top to the middle key row). This requirement is easily met since the key height is only 17.5 mm (with 4 mm travel) or 16 mm (with 2.5 mm travel) from the upper edge of the PC board to the upper edge of the keytop.
The same document also cites DIN standards for the standard spacing of keys:
Center spacing – important as regards finger width – in accordance with DIN 2112 and 2127 – 19.05 mm spacing for keytops
The precise requirements cannot be known without access to the relevant standard, but as noted above, it is understood that the distance from the desk surface to the top centre of a keycap on the home row must not exceed 30 mm. Amkey’s 1991 EEM entry simply refers to “30mm ergonomic requirements (DIN standard)”.
D’Milo Hallerberg, the son of Hi-Tek’s late founder and a one-time Hi-Tek employee, revealed in the Hi-Tek Corp. History topic at Deskthority:
The 725 refers to the ergonomic standard (Max .725" off the desk surface) that came out during that period that effectively obsoleted the standard series we'd been making for many years prior.
This appears to be an incorrect recollection, as a Hi-Tek (or NMB) catalogue diagram shows that the height is measured between the top of a home row keycap and the bottom of the PCB. This measurement does not take into account the bottom of keyboard enclosure (including non-slip feet) and tilt angle of the keyboard. 0.725″ is 18.42 mm, allowing the bottom of the PCB to be 11.5 mm above the desk surface.
Omron B3G-S was described in the 1986 catalogue as “DIN規格に適した全高18.1mmのロープロタイプ”: “Low-profile type with an overall height of 18.1 mm suitable for DIN standards”. Here, 18.1 mm is the distance been the standoff on the bottom of the switch and the top of the keycap.
The subject is discussed within the article Keyboards downsize to meet DIN requirements (Frank Weghorn; Electronic Products, September 30 1982, pages 90–92 [EP1982-KDMDR]). This article notes another requirement of the standardisation: the slope angle of the keyboard must fall within the range of 5° to 15°, with 12° “found to provide the most compatible operator-machine key layout”.
The standardisation (and the history leading up to it) is also discussed within the Deskthority topic The Alps Trilogy Part 1: Apple M0110A review (Alps SKCC tall cream). In this topic, user “jacobolus” noted:
They insisted that keyboards and keycaps must be of a neutral grayish color, falling between specified light/dark values (I think beige was okay, but certainly nothing brightly colored, white, or black). They were worried about potential eyestrain if there’s too much glare off your white keycaps, or something. If you look at the ways an IBM Model M keyboard differs from a beam spring keyboard, almost all of those differences were mandated by the German standard.
There is no reference to where this information came from (or which standard) and although it is a plausible idea there seems to be no readily accessible corroboration of this claim.
In his 1982 article [EP1982-KDMDR], author Frank Weghorn claims that “in many cases … key travel has been sacrificed to meet the 30 mm home-row height requirement, affecting the keyboard’s tactile characteristics.” The author also reported that “the shortened keystroke has, in some cases, let to enhanced feedback characteristics in the new models.” He goes on to say, “This has been accomplished by simply increasing the linear key pressure or by totally redesigning the switch mechanism to provide a true tactile differentiation at the point of contact.”
Unfortunately, the article remains somewhat vague about these changes. Some examples are given, but not enough to affirm the claims that were made. Key Tronic for example are said to have reduced their keycap height from 0.546″ to 0.310″ (from 13.9 to 7.8 mm) and retained their 0.150″ travel (3.8 mm). Of course, 3.8 mm was already a shorter throw. By comparison, the travel of Keytek’s Inductric keyboards was reduced from 0.150″ (3.8 mm) by 0.020″, giving 0.130″ (3.3 mm). Keytek appear to be one of the companies who adapted the feel of the switch accordingly, offering the traditional linear feel alongside a new tactile option. Per the article: “The new switch design provides an override or negative force profile at the time of switch contact. This optional feature should eliminate any negative effects associated with the shorter stroke and will be especially useful in applications requiring high input rates.”
The table below (which is not taken from any article) shows a comparison of pre-DIN and DIN travel distances for a number of prominent brands (all figures converted to metric for consistency):
|Alps||KCC||3.5 mm||KCL||3.5 mm|
|Cherry||Early M6||4.8 mm||MX||4 mm|
|M7/later M6||4 mm|
|Futaba||MD||3.1 mm||MA||3.8 mm|
|ML (short)||2.5 mm|
|Hi-Tek||Dovetail Series||3.6 mm||Series 725||3.6 mm|
|Keytek||Inductric||3.8 mm||Inductric||3.3 mm|
|Key Tronic||Unnamed||3.8 mm||Unnamed||3.8 mm|
|Omron||B3G||4 mm||B3G-S||4 mm|
As the table shows, the DIN-compliant designs in some cases offered considerably longer travel than the pre-DIN designs. When Cherry Germany began their G80 series of domestic keyboard designs, they used M7, M8 and M8 switches, and the most prominent form (the only form to have ever been discovered) seems to have been M8-based keyboards with only 2.5 mm travel; this series transitioned to larger MX switches with the full 4 mm travel. Futaba’s proper DIN-compliant M8 switches offered more travel than any of their late 1970s designs.
There is a suggestion that Micro Switch SC and CT suffered a reduction in total travel from 0.150″ (3.8 mm) to 0.125″ (3.2 mm), but the limited available data on these series makes this impossible to prove.
The claims that Key Tronic’s key travel was not reduced may be a little disingenuous however, as the datasheeet for model L1696 from March 1979 (see under Key Tronic keyboards) gives the travel as 0.171″ (4.34 mm) nominal.
The article also notes that the need to retool production lines was able to bring about lower keyboard costs, as some manufacturers took the opportunity to increase the level of automation involved and choose designs that were cheaper to manufacture.
The keyboard makers themselves were seemingly not pushing these changes; the article claims that the “proof is in the orders, with several keyboard manufacturers reporting that prototype requests are weighted four-to-one in favor of the low-profile design.” No comment is made as to the geographic distribution of these requests, but the article covers a number of US manufacturers and it seems safe to say that the effects of the standard were being strongly felt on the far side of the Atlantic.
This overall reduction have ramifications for what type of switch mechanism would be permitted; for example, the extremely smooth and bouncy feel of ITT ETL 18 may not be achievable in a switch reduced to DIN-compliant levels. A community attempt was made to achieve buckling spring with Cherry MX mount, but this would increase the switch height: such a design is not likely to be compliant.
The difficulity in creating the Input Club Silo Beam spring switch raises the question of whether modern switch size expectations are in any way responsible. In a Kickstarter blog post on the Keystone keyboard from August 2021, Input Club wrote:
Unfortunately, there have been some metallurgical issues with the beam spring development: the beam fatigues and breaks too quickly. As we are not material science experts ourselves, we have been working with some other companies to identify some better materials to use for the beam spring itself. Currently we're waiting on our switch factory on whether or not they can source (and work with) the new materials.
In the Geekhack forum topic The Silo Beam Switch - Beam Spring switches for the modern era, user “hvontres” wrote:
I think this is an indication that an MX-sized beam spring is a bridge too far. I am asuming they were already using some type of spring steel in the original design. Given that, there probably isn't that much improvement to be had, since the endurance limit is generally about 50% of ultimate strength and most spring steel is already pretty far up there. My best guess is that given the length of the spring, the relative deflection to get to 4mm of keystroke is causing a larger percentage deflection compared to the original IBM beam springs, which were much taller than what you can fit inside of an MX sized housing.
At this point, the best bet will be using a thinner spring (yes, while that may be counterintuitive, in a bending spring like this, the stresses go down for thinner material). However, that may not give the desired force profile, so this will probably turn into a protracted back and forth battle between stiffness and fatigue.
This is only a single user’s speculation, and it may not be correct. All that is known for sure is that 1960s keyboards were able to take all the space that they needed. During the 1970s there was already a drive towards lower keyboard profiles, resulting in series such as Micro Switch SD and ITW’s SS3 that were not DIN-compliant but were nonetheless lower in profile than previous generations of switch. The widespread adoption of laptops has brought even typical desktop keyboards significantly down in height, as many of them now mimic the look and feel of laptop keyboards, to their detriment.
Prior to the widespread redesigns, most switches were designed such that the keycap always remained above the switch. ITW’s SS3 appears to be a counter-example, anticipating the theme of the DIN-era changes without meeting DIN compliance. Plate-mounted switches typically had the mounting plate just below the top of the switch. To meet the significant reduction in height, larger switches were required to allow the keycap to be lowered past the top of the switch. This meant that the width and depth of the switch needed to be small enough that the switch could fit inside the keycap, such as by chamfering the top edges (as with Alps) or tapering the switch body (as with Cherry). As the mounting plate would now be in the way, the position of the plate relative to the switch was also changed, with the plate being much further down, closer to the PCB.
Adaptation for compliance to DIN ergonomic standardisation did not require entirely new switches to be constructed, but the majority of the switch had to be redesigned. Alps had seemingly not long introduced KCC series, and when they created the DIN-compliant KCL series circa 1983 they were able to continue using their “switchplate” contact module design from KCC series. The rest of the switch was a new design. In addition to the redesigned shell, the new plunger allowed the keycap to sit lower down, and for its central post to pass inside the switch when the key is pressed; this is the origin of the famous “Alps mount” that saw widespread adoption in Asia.
Hi-Tek Series 725 required all new parts, although the contact design from the earlier High Profile and Dovetail Series switches was retained, in smaller form and with revisions to the structure. Series 725 was also introduced in 1983. Series 725 featured a radically different design, where the switch has no top at all. The plunger occupies the space where the switch top should be, functioning also as the top of the switch. The keycap then fits directly over the top of the plunger.
Cherry MX, another 1983 introduction, was a ground-up design with DIN compliance as one of a number of uncertain goals. Cherry already had a highly miniaturised switch in the form of M8, but this is not known to support alternate action (although it should be possible considering that RAFI achieved ultra-small alternate action parts) and MX also allowed Cherry to continue using their existing rotary alternate action mechanism. The shell seems to be derived from M9, but the contact system was new. It also coincided with Cherry’s adoption of lower-cost gold contacts, but those were made available for M8 and M9 switches too. Cherry MX retained a protruding stem for keycap attachment, but the stem was placed onto a platform that is lowered inside the switch. This allows the centre post of the keycap to pass inside the switch.
Omron B3G-S was also a wholly new design. It appears to have been inspired by Alps KCL/KCM and shared its keycap mount, and its contact module was based on Alps but used a contact leaf spring rather than the metal foil of KC-type Alps switches. SMK also produced switches with hollow plungers, although the standard design used SMK’s own mount. Alternative versions were produced that accepted Alps and Cherry MX keycaps.
Other product ranges changed around the same time. Somewhere around 1985, Mitsumi standard mechanical was modified similarly. Futaba appear to brought their smaller ML models to market in 1983: notable usage of these includes the Acorn Electron, Memotech MTX series and Atari 600/800XL, all introduced in 1983. The chief difference between the larger and smaller Futaba ML models is the height: this is one design that could be scaled down, although the smaller version is more fragile and not as robust. Futaba followed up with a radical redesign, the MA series, shortly afterwards; this new series appears to have been introduced by 1985, but there is very little clarity on this.
The “D” in Comptec and Signature Plastics keycap family names stands for “DIN”. From the Pimp My Keyboard FAQ:
In the mid 80’s an attempt was made to standardize keycaps to a ‘DIN Standard’. DIN stands for “Deutsches Institut für Normung”, meaning "German institute for standardization". This resulted in a new high profile family being produced, the DSS family, which was a DIN standard, Spherical touch, Sculptured key family.
Comptec introduced three known DIN-compliant families: DSS, DCS (cylindrical sculptured) and DSA (spherical flat). DSS has since been discontinued, while DSA and DCS remain in production.