Tammy wasn't able to get the switch out of the keyboard, so in the end she just cut the entire end of the guide plate off!
Since the Keyboard II is a membrane keyboard, there is nothing to solder the switch to. Therefore, SMK provided the switch with its very own PCB:
I don't actually know quite what Tammy tried. You can see glue in the photo, and it seems to have been used to glue the PCB prior to soldering. The switch wasn't glued to the guide plate, but I removed it not by desoldering the PCB but by removing the guide plate from around it using sidecutters. The fact that the PCB remained attached is not good for the photos, but it made the switch far, far easier to reassemble. (Just as the [wiki]terminal sealant[/wiki] does for first generation SMK switches.)
Now, I am not the first person to photograph this switch, but hopefully the pictures below will offer more insight into the design.
Just like the first generation design, the latching mechanism uses a spring-loaded pin with the ability to swivel in different directions. This is in comparison with the RAFI design, where the pin can only move backwards and forwards. The positions are reversed: the pin now fits into the slider, and the latch track is the separate part:
With pliers, you can pull the pin out:
Surprisingly the time spent looking that spring was probably less than a minute, so that's a huge improvement over RAFI where one of those springs is lost forever. It's a whole 4 mm long and 1.38 mm in diameter, which is why it has to come second place to the [wiki]RAFI full-travel key switch[/wiki]: the equivalent RS 74 spring is only 2.7 mm long and 0.9 mm in diameter! RAFI latching switches are the most evil switches I have ever encountered.
The latch track block:
The switch itself was always advertised as "finicky" and I found that it was jammed. You could press the slider, and that would open (not close) the contacts, but the slider would never release. I can't explain why pressing the slider opened the contacts, but the movable contact had ended up on the wrong side of the crossbar, blocking the slider from returning to its home position:
That was a bit of a nuisance to fix, but I rearranged the contacts without needing any further disassembly, and with the terminals soldered in securely, it wasn't too much of a headache to get the switch back together. Testing with a meter shows that it's now working perfectly. It's got a sharp click, but it's not got the smoothness of RAFI or Mitsumi.
The switch components:
One day I may desolder it for the sake of some final photos, but at the moment I'm rather glad for it be able to be opened and reassembled with comparative ease for SMK!
I still don't know how the switches in the Keyboard II work, though. Maybe one day someone will come across one and take some decent photos of it. 2016 has been a fantastic year for knowledge, after all.
As a footnote, Tammy also mistakenly referred to Deskthority as "deskauthority.net" -- it seems to be a natural misreading.
Mine had come out of an Apple IIGS keyboard known as the 