Did you know Alps designed, patented, and produced their own variant of a Buckling Spring keyboard? In 1988, 2 patents were filed by Alps Electric Co., Ltd. in Japan, detailing their design and implementation of their own Buckling Spring variant of Key Switches. However, it was not a widely adopted design, and as such there is only a handful of known boards using it:
- Alps SM-101 (US-ANSI Layout, 101 keys)
- RM Nimbus 286/386 (ISO Layout, but no actual label depicting the exact model, 102 keys but described by some as a SM-101)
- C.ITOH CIT324/CIT324E (ISO Layout, again no specific label depicting the exact model other than a sticker saying ‘BRITISH’)
- Sega TeraDrive HTR-2106 (Japan only market, and highly sought after by collectors)
There are a few discussed theories knocking around that the Alps buckling spring design was, in part, due to influence from IBM, as they were partnered with to develop the Sega TeraDrive. IBM would also contract out to Alps for manufacturing keyboards in the Japanese markets which would be included with systems they distributed there, so this idea does make some sense. The Alps design in some ways is very similar to that of the IBM Model M, in that it has a steel backplate, moulded barrel plate for the springs & hammers, and then the membrane sandwiched in between them. However, unlike the Model M, the biggest changes were that the keyboard uses a slider mechanism atop of the spring, and has a much shorter barrel producing a shorter profile which ultimately would remove much of the bulk of what makes a Model M. This then required some design changes around how the buckling spring mechanism works, as with the IBM, given the length of the spring and distance of travel of the key itself, the spring would naturally buckle at a certain point. But with the Alps design, as the spring is both shorter, with reduced travel of the key mechanism too, the spring would require a ‘nudge’ to allow it to completely buckle, so unlike the Model M, the Alps flipper would have an actuator which would push the spring laterally to allow it to buckle.
IBM(left) and Alps(right) Buckling Spring Designs from their retrospective Patent applications. Note that on the Alps, you can see the additional actuator(50) which would ‘nudge’ the spring causing it to buckle.
The other big difference with this, is that instead of the keycap itself having a locator nub inside it for the spring, which is actuated directly from the key itself, the Alps design implements a slider, which has a standard Alps mount for keys, and would keep manufacturing costs to a minimum as it would be compatible with all current Alps keycap designs. But what is interesting, unlike the multitude of other Alps keyboards of the time, the keycaps for these were thinner ABS plastic keycaps with pad-printed legends. This is a stark contrast to the thick PBT keycaps with dye-sublimated legends that were common on other Alps keyboards of the time.
But what about this one?
Unlike other keyboards mentioned above, there are no labels, or any other identifying marks to help determine the exact model number of this keyboard. I don’t feel just labelling this as an SM-101 is right, and let me tell you why. Manufacturers commonly use a numbering of the model to depict more so the number of keys across different markets, and in the case of the Alps-made Acer KB-101A, within the UK, the designation of this would be the KB-102A (signifying the fact it is a 102key board). But this pattern doesn’t seem to follow suit with Alps themselves. I have another Alps keyboard to hand which is a 102-key ISO layout, but the model number is AT 101-102 (FCC ID: GYIMIDAS101-102). Alps do manufacture keyboards under a generic model/ID as an off-the-shelf component for other brands to use, so simply looking specifically for FCC IDs against the known Model number for Alps Buckling Spring keyboards (KFNLEA), only 4 are found:
CWTKFNLEA901 has been matched to an existing SM-101 keyboard with the ANSI/101-key Layout. However, due to labels on the other ISO-layout keyboards not being present, we can’t really confirm the same FCCID would apply to these. Likewise, the HTR-2106 has no matching FCCID on any of the labels on those keyboards (especially true given they are Japanese market only). So, given the only examples of the SM-101 where it labels it as that model don’t specifically say 101-102, and with regards to the TeraDrive HTR-2106, in that it has 106 keys, it could well be assumed that the UK ISO variant of the Alps Buckling Spring keyboard would in fact be an SM-102? Just under a different moniker…
But I could be wrong… I may have just had one-too-many coffees today and my brain is running away with me.
I digress…
This specific board is (I assume going from the date codes on the controller IC) a 1990-built SM-101. Overall, the condition of it isn’t too bad, but as you can see it has yellowed quite considerably over the years, to the point you could mistake it as being the same colour as Custard. This is, by far, the most ‘Yellow’ keyboard I own. They keycaps themselves though don’t really show much sign of wear other than the discolouration, the texture of the caps isn’t too shiny and still holds the original texture. The space bar obviously does show more wear than the others, but annoyingly the F2 key itself is broken off. The contrast in colour though with the case and keys from exposure to the warm sun through a window is very noticeable!
Testing the keyboard itself, the switches all (for the most part) feel consistent with each other, the exception being that the S key doesn’t register or buckle (we’ll come back to that shortly), and the Question mark key doesn’t always register consistently, even though it does buckle. The sound they make though is quite different to that of my Model M, in that it’s a much sharper and higher pitched sound, The distance you press the keys is also quite noticeably different as mentioned earlier due to the reduced travel & shorter profile of the switches. The keys are also quite light to press to with not much force needed to press through the actuation point.
These rivets are possibly overkill for what they need to be, but it certainly adds some reassurance to its longevity!
Inside, the main key mechanism is again very similar to that of a Model M, in that it has a plastic moulded top barrel plate, steel back plate, and sandwiched between is the membrane. Interestingly, the rivets on this are much thicker and stronger to that of the Model M, and it makes me think were Alps aware of this being a weakness on IBM keyboards already? Or was it just over-engineered to last longer? It’s nice though opening something and not seeing a load of rivets rattling around inside the case! Certainly, don’t see a need to bolt mod this one!
Other than the C.ITOH controller, details around the genuine OEM Alps one are sparse…
The controller PCB does have Alps branding! Unlike the previously mentioned C.ITOH branded version which seemed to have its own made controller, this seems to have just an Alps made one. The controller has an Intel P8052AH microcontroller, and a 74LS06N. The Intel MCU is also tagged Alps. I’ll need to check my other Alps keyboard to see if this is a common component between the two.
So, back to the 5 key… This just simply refused to buckle like any of the others, and only when being pressed with a fair bit of force would sometimes register. Taking off the caps to compare the switch to others, visibly I couldn’t see anything wrong, looking doen the sides of the slider as best I could didn’t really show anything broken or out of place from what the others looked like, it just simply would not buckle. However, one thing I did notice, when pressing the switch, it doesn’t quite travel as far as the others do, it’s literally only just less than 1mm in total difference, but if you look it’s there. The only way to get to the bottom of this then would be to take the switch apart, but it seems no one has ever done (or documented) this before, the only information I could find which went into some degree of detail was the information around the C.ITOH branded board. This had pictures from a teardown they did to troubleshoot and replace a broken controller, so it gave me some insight into how the board may be pieced together, but it didn’t show the switch itself being disassembled, only the underside of the barrel plate where you could see all the individual hammers. However, that didn’t seem to give any idea as how to go about disassembling the switch itself. The only other information I could find which detailed the switch to some degree was the details in one of the two patents which were filed for this (US5012055A), as this showed more of a teardown of the switch assembly. Great, something to start with!
Exploded schematic of the Alps Buckling Spring Mechanism from their Patent Application
According to the diagram then at least, the slider is secured into the assembly using two small clips at either side, and from the pictures of the teardown, and going from the drawings in the patent documentation, the switch appears to be assembled by pressing the slider into the barrel itself from above, with the hammer being inserted prior to this. This made sense going from the teardown pictures, but how do I remove the slider? I don’t want to end up splitting the whole thing and then having to bolt mod it to gain access from the bottom, but then it also doesn’t look like there is enough room to push the slider up from beneath it, as the hammer is in the way. Conversely, trying to slide something down between the slider and the barrel to press in or release the clips wouldn’t be an option either as there just isn’t enough room, so the only thing I can think of is to pull the slider out from the top. So, thick rubber band over the top and some wide nosed pliers, gently, but firmly, grip the top of the slider using the rubber band for extra grip and to stop the pliers biting into the slider, then just wiggle it gently back and forth while pulling up. Took a bit of force in the end, but it did come out cleanly, and with no visible damage either to the slider or barrel. Success! And as for the hammer and spring, this just simply lifted out without any issue.
So here we are, possibly the first actual pictures of what the spring and hammer assembly looks like beyond just the drawings in the patent! But what’s causing the issue of it not registering? The hammer itself looks fine, nothing appears broken, missing or bent. The spring doesn’t appear stretched or damaged, so what’s going on? A quick look into the barrel made things clear.
As you can see, there is what appears to be a small piece of dirt resting in the lower-right, like a little stone. This could have gotten in there at any point in its history. I don’t think this would have been something from the F2 key being broken, as with the way the plates are all sandwiched together, there is no way for something to move freely about in the internals, maybe someone had the S cap off before, or at some point in its life it was just left covered in all sorts of junk like old tech is before it’s discarded, but either way, this is what’s causing it.
Reassembling the switch is easy too, simply drop the hammer and spring back into the barrel, hold the keyboard so the right side is raised slightly that the spring leans towards the left side of the keyboard, then press the slider back into the barrel. The 5 key now works again!
As for the Question mark key, this seems to have just been a bit of dirt or dust but did clear after a few presses and now works fine.
So where do we go from here? My other Alps board has PBT caps with dye sub legends, so the contrast between those and the ones from this is like night and day. I could replace all the caps on this one with those, but then what do I do with that one? I could also look at sourcing replacement caps for this, but complete Alps boards don’t come up too often, and certainly not at a decent price. If another AT-101-102 board did come up (given it’s also a Dome with Slider variant, so not truly mechanical), that may be an option? Other choices are a new set of Tai-Hao keycaps for it, even though they’re ABS, they are at least double-shot caps and look very nice.
These are both the ABS keys and PBT keys from both of my Alps keyboards. I didn’t realise until properly looking at it, but the profile is also ever so slightly different? Perhaps this is normal between the two types…
I’d also need to do something about the yellowing of the case too, otherwise I’ll be craving a bowl of custard every time I look at it.
Leave a Reply