New Cherry Keymodule information available online

[robin13867]

Hi folks,

Just learned from our lovely marketing department that there's some 'new' keymodule info available on the Cherry site - click here for details: http://cherry.co.uk/cid/b2b_keymodules.htm?

The information has been available in shortened form on our switch website, but they've now expanded the MX datasheets and separated them into individual module types, so it's a little easier (!) to understand. Also, they've added the ML information for both ML1A and ML1B variants, so updates all round then...

Enjoy!

Robin.

[Daniel Beardsmore]

I've never seen that ML variant before.

Does anyone in Cherry UK (as opposed to ZF in the US) have any knowledge of any switch variants that are not on sale? Few people here are in any doubt about current switches (apart from this "new" ML), but there is a huge mystery surrounding all the non-standard MX variants such as [wiki]Cherry MX Orange[/wiki], Alps-mount Cherry MX, and Cherry MX Yellow; the difference between [wiki]Cherry M6[/wiki] and [wiki]Cherry M7[/wiki] (as you can see, I'm speculating that it was the change in the contact plate), and some of the seemingly undocumented Cherry M8 variants.

We only very recently confirmed what M9 looks like outside of catalogues (first post on this page links to the pics):

http://deskthority.net/deskthority-wiki ... 19-30.html

Now if you could take part in this, you really would make a lot of us very happy :)

[robin13867]

I'd love to add to this discussion Daniel, but the info that I have is limited to say the least.

MX Orange - never heard of it in the UK.
Alps mount Cherry MX - nothing from our side in any documentation I have
MX Yellow - again, never heard of it as an 'official' product - I have seen a fake....
To my knowledge, MX series models have only ever come in certain configurations with certain colours.
Black / Grey/ White/ Clear/ Blue/ Green/ Brown (and now red)

Anything else is either:

A) A fake module - there were reports of MX switches with Cherry logo being produced in China some years ago, but I can't say for certain whether this was correct or not. I did once see a fake yellow somewhere from memory though...

B) A module that would have been produced specifically for a large volume customer. This wouldn't have been common knowledge throughout the organisation and so the module would never have appeared on datasheets or any communication between offices. If the customer wanted a specific MX module with a certain feel or attribute, then we would have, no doubt, produced it in a different stem colour to the rest of the range in order to distinguish it from anything 'standard'. Why anyone would do this though is a mystery. With the amount of configurations/colours available, it seems a step too far, but I could be wrong. In 17 years of being here, I've only ever heard of custom keyboards, not custom modules.

There is provision in the part number list for 'special' versions in two places, but I've never seen them used before:

MX-
-1, 2, 3 or 4
A (SPST), B, C(Special), D-
0 (Special), 1, 2, 3, 4, A, B, C, D, E, F, G-
1
N, R, G, Y, D, J
N, W

M7 - what do you want to know? I don't have an official image, but I do have a datasheet
M8 & M9 - same information in the same booklet, along with information on the keycaps.
I have nothing for M6 though
The datasheet/booklet that I have is circa 1982-ish and includes info on various defunct models of product - G80-0225/G80-0246/B70-05AB/B70-4753/B80-3766 and more....

Hope this helps....if you need any more info and I have it, happy to post it on here, not an issue. However, do bear in mind that if there are 'non catalogued' parts, then it's unlikely that ANYONE outside of Germany will hold this information. In the case of switches, all sales/enquiries would be channeled through the factory in Germany and they would keep the information on any custom model within their archives only. Neither us, the US, France, Italy, Australia or any distributors would keep this info on file...sorry!

[Daniel Beardsmore]

Hello Robin, thanks for replying. I'll address your reply by switch, though it will take a few weeks to read and digest this all :)

Cherry M6

There is no "M6" product brand. Prior to the M7 range, Cherry sold externally identical (according to the catalogues) switches with part numbers beginning with M6 (M6x-xxxx). (I think the earliest catalogue we have is from 1973, with M6x in.)

The difference between "M6" and M7 is not immediately apparent. Several of us have really old switches from Electronic Surplus of no known part number¹ and they have a static contact plate of a different shape to switches that are assumed to be M7 (those switches with newer shell designs are assumed to be M7, for example). This difference is depicted on the wiki as linked above, and I am guessing that this is at least one difference between M6 and M7.

At some stage I need to scrutinise the catalogue specifications and see if anything in the spec changed that would indicate the reason for changing the numbering from M6x to M7x. According to ZF in the US, there's no-one left at Cherry who still remembers their 1970s switches.

¹ Very few switches put the range or part number on the shell. The only known examples of this are the Omron B3G-S and, formerly, Kaihua PG1511.

Cherry MX

Here is our reference point — a Chinese collector who has the most complete set of allegedly Cherry MX switches in the world:

http://deskthority.net/photos-videos-f8 ... t2149.html

They all have the Cherry logo on them, and they all bear an exact resemblance to genuine Cherry switches. There are also an awful lot of them!

There are plenty of clones which are unbranded (or in the case of Kaihua, are sometimes branded "Kailh" or the series name of "PG1511"), and none of them are the same shape as a real Cherry switch. I have found very little evidence to date to suggest the creation of pirate fakes, i.e. those intended to truly resemble the genuine article. The Cherry MX cloners always produced switches with very distinct differences in the design, usually simplifications. I don't tend to use the term "fake", as it's always so obvious that it's a cheap, unbranded copy.

I did recently raise the notion of genuine trademark violation, in relation to this topic: http://geekhack.org/index.php?topic=20740.0

That is an "Alps" keyboard from Focus, who we know shipped keyboards with some very strange Alps clone switches. The switches depicted in that topic are Alps-branded, but unlike any Alps switch I've ever seen, there are some obvious discrepancies: the mould numbering does not have recessed circles around the numbers, the "ALPS" branding is wonky and distorted, and the numbering itself looks like a child scratched it in with a stick.

Generally, clone manufacturers do not go so far as to outright violate trademarks; rather, they just make similar-looking compatible parts for less money. Did Alps sell off a load of reject batches to Focus? Maybe an inside deal between the Focus and Alps factories in Taiwan to get cheap parts.

We know that Focus used switches that resembled Alps switches on the outside (but without branding) and cheap clone mechanics inside, adapted to fit. The critical factor with the potential Alps fakes above is seeing the switch mechanics. Sadly, the keyboard has been sold off, so we cannot tell, and we may never know. Vintage Alps switches are horrendously complicated inside (up to 13 parts per switch, typically 11) and were never truly cloned (only the overall appearance, with a simple mechanism inside); by comparison, Cherry's mechanisms were always very simple, and the cloners just copy the metal parts outright, to the extent that the parts are interchangeable with real Cherry switches (in the case of Yali/Aristotle, at least, and I imagine also with Kaihua).

There's not a lot more to go on, as "fake" switches are uncommon in Western products; the only widespread "fakes" were the Aristotle (Yali) clones that Chicony bought for a while. There are no other brands I am aware of to have been "faked"; KPT could be considered a combination of clone Omron and clone SMK, but they did clearly brand their switches. (Then you have what could be fake KPTs … at that point you get into a real mess of one-keyboard companies in Taiwan with weird KPT-like switches.) Omron itself is pretty much an Alps clone anyway. SMK gets weird, too. (Prepare to have your mind blown: http://kbtalking.cool3c.com/article/8366 — this page is so intense, even their forum software can't handle it all.)

Yes, these mystery Cherry MX switches are probably all custom orders; it would be nice if you had access to records that would provide more info, but it sounds like we will never find out.

Only Cherry MX Orange has a known keyboard, from NCR Japan, so it's most likely to be legitimate. We have no idea where the other switches came from. Asian keyboard fans are a phenomenal bunch, but the language barrier makes sharing of knowledge very difficult and generally we just peer into their world in wonder. (Though now people are starting to buy into their custom keyboards.) Japanese to English machine translation just does not work at all (it comes back with mostly random words), and Chinese to English isn't a lot better.

[robin13867]

Hi Daniel,

Phew, that's a whole heap of info to plough through!! I've read through some of it and seen the images of the switches from the far-eastern chap - in all honesty, most of us would never have either a) noticed subtle differences or b) been informed that there were changes, so the forums are clearly more informed than the employees of any variations in switch versions. The kbtalking site is way too much for me!!

If a customer orders a specific module in the UK, then yes, we'd know about it and it would have been contained in our systems at one time or another. However, with the rules on companies keeping records extending to only a few years (7 I believe) then most of the older records for past sales have been destroyed - certainly from a UK perspective anyway. We moved offices 7-8 years ago and a lot was thrown away then (along with quite a few 'collectible' keyboards at the same time...d'oh!).

If a custom module was made in Germany, we're not necessarily privy to the information - it's customer specific/product specific and the only people that would know are the chaps in Germany and the customer themselves. I will ask one of my colleagues though if some investigation can be done - I don't hold out a lot of hope, in fairness, but may be worth a try. You have to understand that between offices around the world, there wasn't an integrated PC system at all (still isn't) - back in the 70's/80's, there was no PC system, so everything's on paper and then when PC systems appeared, each country was using a different one. Here in the UK we used a 'Bull' order processing system, which was useless, then the Bull system coupled with a PC 'customer database' system called Tracker, then Goldmine, then SAP for a while (hopefully never again!) and now Maginus. Germany continue to use SAP and we have a portal, but it's cleansed periodically, so a lot of older info disappears.....

As far as clone switches are concerned, there were various rumours circulating a few years back. We were supplying one customer with large quantities of MX's, only for them to stop and rumours were rife that they were buying a clone product from the far-east as they were still using things like space-bar mechanisms/levelling mechanisms bought in the UK. Nothing ever proven though and we didn't get to see the switches, so it could all just be hearsay rather than anything concrete. But rumours of clones have been around probably for longer than I've been here - certainly I remember talk of it not long after I started and it comes up from time-to-time, but little is done about it.

Also, really not sure about the colour differences in some of the switches. I would have thought that variations in shade are purely down to material used during our manufacturing process, rather than them being separate switches with different properties. As I'm sure you can imagine, different shades of colour would be very difficult to manage, whereas a complete change of colour is quite easy. Again, I'll ask the Germans to see if there's some historic info available.

I'll be back...!! Oh, FYI literally just received EOL notification for all M8 series switches - they're not being produced any more.....

[Daniel Beardsmore]

The current major clone is [wiki]Kaihua PG1511 series[/wiki]. This is found with "PG1511" or "Kailh" branding.

Chicony bought [wiki]Aristotle Cherry MX clone[/wiki] back in the day; Aristotle confirmed to me what I learnt from Chicony, that this product is no longer sold.

There's a few clones of the MX switch, all unbranded apart from Kaihua's, and not a single one looks the same as a real MX switch. yab(lotsa numbers) has a photo somewhere of his MX collection that includes all of his clones. Whether any are still sold besides Kaihua's, I do not know; it took some effort to track down the various major Alps clone manufacturers, and there's still at least one current product from an unknown manufacturer. (The switch is marked "YH-B", which means nothing according to Google.)

[Daniel Beardsmore]

OK Robin, try this one:

http://deskthority.net/keyboards-f2/tre ... ml#p120900

It's fascinating to note that some switches have some kind of mould numbering instead of branding — is this the birth of Cherry MX? Or some sort of weird anomaly with the mould making? I've never seen this before, and there's no reason to believe that they're fake.

The thick plottens.

[Muirium]

Here's the link for the "clones? what clones?" department:

Keyboard Switches.jpg (945.37 KiB) Viewed 8992 times

The gentlemen at Kaihua have a familiar range to sell you!

[Daniel Beardsmore]

I already mentioned the PG1511 range twice :P

See also Ripster Switch Guide: Kailh Cherry MX clones — note that images 13, 14 and 15 are not Kaihua, but in fact the Aristotle (Yali) clone that Chicony used, which is white ("clear"), clicky, and fragile, and has compatible internals to real MX switches, or so I read.

Somewhere, yabb has a better photo of his collection that shows the fakes, but this is a start:

http://deskthority.net/marketplace-f11/ ... tml#p61397

Interestingly, there are ones that are identical on the outside to real Cherry switches, just without any branding.

[Muirium]

No self credit implied, Daniel, as I surely heard of them through you! I'd like to know what Cherry thinks about such clear clones, but I doubt good Robin could tell us either. China is not exactly famed for its intellectual property enforcement.

[robin13867]

Hi folks,

Not good news I'm afraid - I've contacted my colleagues in Germany and, as I suspected, they're unwilling to supply any other information to me that I can post on the forum

Basically, because of the risk of clone products being made in the far-east, they're just not prepared to let anything go. Any product that was not previously published on the website or in catalogues is particularly sensitive, so really sorry, nothing more I can glean from them I'm afraid.

Oh well - back to the drawing board!

[Daniel Beardsmore]

You serious?

See this page from the Kaihua website:

http://www.kailh.com/gb/Newsdetail.asp?Newsid=37

It's not their own material, as I understand it: they've republished something from an IT magazine.

However, you can see there exactly how their clone switch compares with yours through detailed disassembly and comparison. From what I can tell (visually, as it's all in Chinese), the metal parts appear to be fully compatible with Cherry MX switches; the key difference is that they don't use gold crosspoint.

These switches are in widespread use in Asia. Cherry MX clones have been on sale for probably 20 years or more now, although Kaihua switches may be more recent. If Germany still thinks that there is a "risk" of cloning, then they're not really paying any attention to the global marketplace. We've known about the clones for years. (The only clones I know of are of the MX series, though; I don't know of clones of ML, MY, M9, M8 etc. I think the cloning began in earnest in the early 80s when a market arose for it, around the time that MX was introduced.) That said, Yali switches are the only ones that have seen any incursion into the West, mostly in Chicony KB-5181/5182 keyboards.

I was sure I read somewhere that gold crosspoint was patented by Cherry, but I can't find that now, only a catalogue that implies that it was a Cherry innovation. I discovered that some RAFI switches are also gold crosspoint (RS 74M, some RS 76M).

Did you see the link above to the Xerox keyboard with the really old MX switches? It would be nice to know why some have stamped numbering instead of the Cherry logo.

[Daniel Beardsmore]

PS I am confused. The specifications are public. The switches are a public product that anyone can buy. What exactly does Germany think they're hiding from us exactly?

I can't see what trade secrets exist in ancient knowledge of products that are no longer made or sold (all the old MX variants), or that nobody in Asia cares about (like M6/7/8/9 — no-one's going to clone those, especially when they've been cloning MX for a couple of decades).

[robin13867]

Hey, Daniel, don't shoot the messenger!

Look, I can actually understand where they're coming from. The specification on some switches may not have been made public knowledge at all - if products have been made specifically for one customer (hence the 'odd' colour schemes), then they aren't in the public domain and so information is kept specifically for that customer. They really don't have to let this information go at all.

MX information on currently available models is public knowledge, hence being published on the website with datasheets available (as there are for all of our switch products too - see www.cherryswitches.com). However, even on this site, it doesn't necessarily tell the whole story. We make completely custom products occasionally for very large clients and if something's modified, this isn't information that's published anywhere, nor will it be. There'll be some form of agreement in place with the customer that we won't make this public knowledge. Yes, anyone can buy our switches, but not ones that are specifically made for one customer - obviously if redundant products come up for sale, then you have the opportunity to buy them and 'reverse engineer' them if necessary, but the info still isn't for public consumption.

What the other manufacturers do, is entirely up to them - putting this info into the public domain is good for the enthusiast, but it does tell your competitors what you're doing. If I wanted to get fired very quickly, I could release information on this forum that will give competitors a heads-up on what's coming in the future, therefore giving them a competitive advantage - after all, you never know who reads this site at the end of the day. Remember, I work for the manufacturer and read it! So, it's, unfortunately, their right to withhold the info that they have, rightly or wrongly and I have no control, sorry.

This is the actual text from the e-mail sent to me by my colleague in Germany:

"Some of the products shown are copies which are not manufactured by us. As you mentioned already, some are old customized products which are no longer produced and we would not have samples of these. Since many Chinese manufacturers try to copy our switches we won’t release any information which hasn't been published on our website. "

I've just re-checked all the 'old' info that I have and there's nothing that hasn't previously been published - good job really, as If I'd scanned and sent it to the forum, I'd be in some serious hot water....and I need this job!!

I also understand your frustration with what's been said. It just feels like withholding information for the sheer hell of it, but this is business at the end of the day. MX switches are extremely popular - so much so, that the production lines are flat-out, 24hrs a day to keep up with demand - and anyone making a 'clone' is potentially damaging our business, so they're just very protective. Make sense?

Believe me, if there are things I can say on here and contribute to, I'm happy to do so, but as with any manufacturer, I can only go so far...

[robin13867]

I was sure I read somewhere that gold crosspoint was patented by Cherry, but I can't find that now, only a catalogue that implies that it was a Cherry innovation. I discovered that some RAFI switches are also gold crosspoint (RS 74M, some RS 76M).

Did you see the link above to the Xerox keyboard with the really old MX switches? It would be nice to know why some have stamped numbering instead of the Cherry logo.

I believe that the Gold Crosspoint was a Cherry innovation (seem to recall a patent somewhere!), but I have no evidence to back this up I'm afraid.

I did see the Xerox keyboard with the old MX switches and I can offer absolutely no explanation for this at all!! I can only think that it's from the dawn of the MX era and some switches may have been manufactured prior to the logo appearing full-time on the module itself. The ones with the number stamp may have been from previous production runs and kept in-stock, but then mixed with the newer logo'd versions. Other than that, stumped!

[Muirium]

I've just re-checked all the 'old' info that I have and there's nothing that hasn't previously been published - good job really, as If I'd scanned and sent it to the forum, I'd be in some serious hot water....and I need this job!!

Of course! But I must chime in with Daniel over the difference between old and new info. As exciting as unreleased inside stuff may be, we are much better off not knowing it. It just ruins the surprise and kicks up trouble.

The info that gets us going is about the old oddities we already have in our own collections! Things so weird that not only is there zero interest from cloners, but a haze of mystery, theories and unanswered questions surrounding them that naturally sends us asking. The most competitively useless stuff is exactly the info we most crave! Well, those of us into vintage and rarities, right here on this thread.

Clones want to be a direct replacement part for something in demand. I wouldn't worry about the prospect of clones showing up for things even Cherry can't be bothered producing any more! Though I do understand, of course, that you're not the head honcho making the call.

I'd love if there was some kind of statute of limitations where obsolete products get the documentation release that they deserve. But there isn't, and someone would have to do the work even if there was.

Thanks for doing what you can Robin.

[robin13867]

Not a problem - happy to provide what I can on here.

Although it may not sound as if it's of any use to anyone else, the point is that ANY information on old/non-available switches has the 'danger' of being re-used to create a brand new version of that product that customers may buy, therefore potentially taking business away from the current batch (basically, as they're different!).

I would have thought that information could be sent out in the public domain after a set period, but this would create a very sticky situation for MX's in particular - as it's a product still widely available and ultimately configurable into any number of different variants, even old models are potentially still possible to manufacture......if you want enough of them, we'd build 'em (we are talking in the millions though, not a few thousand!)

As I say though, if there's info I can tell you all about, I'm happy to, but I have to be very choosy and make sure that I'm not releasing info that I shouldn't.

[Daniel Beardsmore]

I think the one thing I'd like to know more than anything else, is how M6 and M7 differ. Is my hunch correct, or am I completely wrong? Why "M6x" — did M1x to M5x also exist? There was clearly some reason for the part number change, and that for me is the big mystery.

I'd also love to know what those Cherry MX Alps-mount switches were used for. Someone must know, as the switches are out there in wild.

However, it looks like we will never know.

[robin13867]

I'll go back through my old stuff and see if anything falls out that's relevant on the M6/M7, no worries. However, one word of warning.....never, ever go by part numbers with Cherry products!! A bit of an example.

For many, many years we've had G80 (MX), G81 (FTSC), G83 (NTK - New Technology Keyboard - Membrane), G84 (ML), G85 (Scissor), G86 (ATK - Advanced Technology Keyboard - enhanced G83). What the hell happened to G82? It's been a mystery in here for many, many years - especially for one of my colleagues who worked in membrane production back in Harpenden and has been here for 30 years. It wasn't until about 3-4 years ago that we finally got one - the G82-27000/G82-27020 (Initial & Initial for Mac). At some point in the future, we'll see G87, but whether G88 and G89 will ever appear, your guess is as good as mine. Therefore, I wouldn't mind betting that M1x-M5x never saw the light of day....

Part number changes can happen for a thousand different reasons, or (indeed) not at all in some cases. We have keyboard products that have undergone only a few changes over the years, some with none at all even when there's been a significant shift in components. Other times, part numbers are changed for insignificant reasons, which can cause mass confusion with the customers - G83-6105's for instance. G83-6105LPNGB-2, G83-6105LRNGB-2 and G83-6105LPQGB-2......They're all the same part to look at, they all have identical features, they're all PS2, they all come in the same box. None is a customer special, none have any special 'features', they all use the same components inside. So, what's the difference? LPN has a Cherry logo. LRN doesn't. LPN and LRN are in box multiples of 65pcs, LPQ in box multiples of 10. That's it.

The same will apply to any product in our line - the part number can change for many, many different reasons from a change in cable, a change in electronics, a change in box-size, almost anything you can think of. Yet occasionally (as with our G84-4400), a major electronic change takes place - a shift in trackball mechanism from one manufacturer to another and there's no part number difference to tell us when/why/how this has happened.

My personal thoughts on the Cherry/Alps switches? 1) Combined Cherry/Alps project (can't see it!) 2) Customer wanted to use existing Alps caps on our modules (doubtful) 3) They're fake (probable), but unfortunately, no definitive answer....sorry.

[Daniel Beardsmore]

Our current page on [wiki]Cherry article numbers[/wiki] — still a long way from completion, but a useful guide. You're always welcome to fill in any gaps on that page :-P

What we don't have proof of, yet, is fake switches that pirate the branding off another company. So far I only have those weird white Alps as evidence, and there's no proof that they're fake, instead of just being really badly made. (They look like the edge inscription on fake £1 coins!)

[robin13867]

Daniel - won't go poking too far into the Article number page through sheer fear of messing someone's hard work up. One snippet of info though FYI...

The middle letter of some (not all - it's very confusing!) article numbers, relates to whether it will have Windows keys or not. For instance, G84-4100LPAGB-2:

G84 - ML
41xx - small footprint housing
L - laser etch caps
P - PS2 connection
A - 83 key model (no Windows keys)
GB - UK layout
-2 - Black

The 'A' in this instance, denotes any non-Windows variety. Any letter M (or after) denotes a Windows version. In this instance there would be 83 keys. This is also true of most of the German built models (G83-61xx, G80-3000, G81-3000 etc...

In more recent times, the middle letter has also been used for different purposes. The new G80-3000's, for instance, carry a 'C'. All keyboards in this range are now 105 key, so the explanation above doesn't hold any longer. The 'C' is used to denote a combi electronic variant - both USB and PS2 (generally USB with PS2 connector).

However, as with the part number revisions, there's no clearly defined structure any longer. When I started in 1996, it was easy to learn the part codes. Now though, it's a nightmare. I'll give you a good 'for instance' of when it all started to go wrong....Back to the G84-4100. Original models carried a 3 letter suffix - PAG or PPG. They were all 83 key versions (long before Windows keys were popular). By all accounts, you'd think easy - PAG, AT Connector. PPG, PS2 Connector, yes? NO, NO, NO. It's the other way around. Then, when the 86 key versions did appear, it' all moved to some easier understanding - PPMGB (Tampo Print, PS2, 86 key, GB), PAAGB (Tampo Print, AT, 83 key, GB). However, with the addition of black versions, we also had to contend with PRMGB and PRAGB. Quite what the R is for, I've no idea....

Unfortunately, even with the more 'regular' older models, it still isn't easy - there are a thousand different configurations.....!!

PHEW!!

[Muirium]

Oh boy…

This will be the ultimate data entry / destructive test for KBDB! I expect we'll never get to the bottom of it for historical models.

[Daniel Beardsmore]

Actually it may be something KBDB could help with: if every known Cherry keyboard is filed, it will help decipher more of those patterns. Eventually, we won'tneed the Cherry article numbers topic as you could just look up the exact keyboard ;-)

"Tampo" print? Is this the Special Cherry Word for pad printing? :) (there's a lot we don't know about pad printing, too, but that's for another topic/day.)

[Daniel Beardsmore]

Another fake, this time from "Proworld":

http://geekhack.org/index.php?topic=46705

Switch looks OK, but those keycaps are horrendous.

[Daniel Beardsmore]

OK, now this is funny:

http://geekhack.org/index.php?topic=45470.0

Stackpole switches on a US Cherry keyboard! Hard to call "fake" on weird Cherry switches after seeing something this bizarre.