Author Topic: Variable weight Filco Linear Tenkeyless  (Read 102013 times)

0 Members and 1 Guest are viewing this topic.

Offline treigle

  • Thread Starter
  • Posts: 44
Variable weight Filco Linear Tenkeyless
« on: Sat, 31 March 2012, 02:40:25 »
[ Guests cannot view attachments ] 46543[/ATTACH]
Having been impressed by the variable weighting on a Realforce 86U, but finding I preferred the action of Cherry switches, I got a Leopold w/ reds. I rather liked it, but sometimes wished for a bit more resistance, so I figured I'd give blacks another try and traded for a Filco /w blacks. Blacks were definitely too stiff for my taste, so I figured it was time to do some modding, and a good excuse to upgrade my soldering iron and skills. If I were going to go to the trouble of [de]soldering 87 keyswitches, why not make them variable weighted like my 86U had been?

[ Guests cannot view attachments ] 46541[/ATTACH]
Each Cherry black spring has 3 coils butted up against each other at either end, the outer 2 of which may be excised while retaining a relatively flat bearing surface, and all 3 of which may be excised without effecting the spring rate (only the preload). By only reducing preload, one may predictably alter the spring's behavior using Hooke's Law. A black switch provides 40CN of  initial resistance, increasing to 80CN at 4mm compression, for a rate of 10CN per mm of compression. Note that this spring has ~4mm of preload already applied within the switch housing. Thus, if one excised 1mm of preload coils, the switch would go from 30CN to 70CN instead. This only applies to chopping off the outer, butted together coils; once you start chopping up the compression coils, other things happen. Someone who actually knows physics can chime in to address that part... ;)

[ Guests cannot view attachments ] 46542[/ATTACH]
Anyhow, I snipped two coils from either end (for a total of 4) of the springs used in those keys pressed by the index fingers, as well as function keys/escape. Middle finger keys got a total of 5 coils snipped from their springs, ring finger keys got all 6 outer coils snipped, and pinky switches additionally got 80% of another revolution of [compression] coil snipped. Arrow keys, navigation cluster, and print screen/scroll lock/pause keys all got variable weighting as specified above (unlike a Realforce).

I didn't measure actuation weights, but I got approximately 52g pinky, 60g ring, 65g middle, and 70g index weight to bottom out.

I rather liked the results initially, and it felt great while typing at a good clip, but in long sessions, I missed the lightness of red switches. Furthermore, this experiment brought to light some subtle things I never noticed before, such as my left-hand fingers being noticeably weaker than my right-hand fingers, calling into question just how useful symmetrical variable weighting really is.

The cut down springs no doubt slightly accelerate wear on the plastic switch housings/plungers, and seemed the resonate slightly more noticeably. I think variable weighting works best with variable spring constants, not just different preloads. In the end, I scored some brown donor switches from another geekhacker, and turned this into a "normal" ghetto red board, which I am quite a bit happier with, but it was a fun experiment, and I got a lot of practice [de]soldering. For those of you with no soldering experience, this sort of thing might take a few hours, but it's not nearly as hard as you might think....

[ Guests cannot view attachments ] 46549[/ATTACH]
Cherry picker inspired by ishtob's guitar pick switch opener. This will save you a ton of time!

[ Guests cannot view attachments ] 46544[/ATTACH]
[ Guests cannot view attachments ] 46545[/ATTACH]
[ Guests cannot view attachments ] 46546[/ATTACH]
[ Guests cannot view attachments ] 46547[/ATTACH]
[ Guests cannot view attachments ] 46548[/ATTACH]
All done! :)
Kinesis Advantage, Truly Ergonomic (ANSI), Filco 87