←back to thread

90 points amichail | 9 comments | | HN request time: 1.023s | source | bottom
1. eviks ◴[] No.42203948[source]
Analogue controls are indeed an awesome concept for a keyboard. Has anyone shipped any "revolutionary" default setups, e.g., from tiny things like making your pinky suffer less by having lower actuation point to making shallow actuation type lowercase and deeper actuation type uppercase or longer holds on arrow keys accelerating the movement?

Pity, though, the progress is still stalled on the actual layout - the ergonomic splits and other improvements are still a tiny niche

replies(4): >>42204161 #>>42205996 #>>42207327 #>>42207688 #
2. mkj ◴[] No.42204161[source]
I wonder if with a sufficiently fast ADC measuring each switch's voltage transition, you might be able to do velocity detection with normal switches? I guess might need an ADC wired to each switch rather than row/column matrix scan.
replies(3): >>42204504 #>>42204505 #>>42206563 #
3. Eisenstein ◴[] No.42204504[source]
Considering that mechanical switches have to be debounced, how would that work?
replies(1): >>42204610 #
4. bluGill ◴[] No.42204505[source]
HE sensors measure distance, if you measure often enough get velocity out of that. Linear switches should be able to do the same thing (I know almost nothing about linear switches so I might be wrong)

Edit: HE sensors not switches. HE naturally measures distance, but there are switches that have electronics that provide an on/off signal thus making the term switch not useful. If you buy a HE switch make sure you know which style you are getting.

5. mkj ◴[] No.42204610{3}[source]
I guess you'd record the curve of the first transition, and the after the debounce period has elapsed you'd use that curve to emit the appropriate key signal. Or alternatively you could use the whole analogue sequence as input for the key detection (assuming no limits on compute power etc). Neural net powered keystrokes!
6. sevensor ◴[] No.42205996[source]
Would be a fun way to implement a modal text editor. Full press for normal mode, light press for insert mode. Could be unintentionally hilarious though.
7. evoke4908 ◴[] No.42206563[source]
No, not with any useful degree of granularity. The electrical contacts aren't directly driven by the mechanical switch, they mostly rely on their own spring tension to make the connection. The mechanical part simply separates them.

By measuring contact bounce, you can probably detect the difference between fast and slow presses, but not much more. Maybe three or four levels total

8. delecti ◴[] No.42207327[source]
The brand "Das Keyboard" did that almost 20 years ago on a non-mechanical keyboard. The activation forces were tuned in exactly that way across the keyboard. The keycaps also didn't have any legends on them, so it targeted the nerdy/elitist audience a bit. Their later keyboards seem to have gone for uniform mechanical switches though.

NYT article about it from back then: https://archive.is/ANcdu and a picture of it from wikipedia showing the lack of legends on the keycaps: https://en.wikipedia.org/wiki/Das_Keyboard#/media/File:DasKe...

These days, it wouldn't be that hard to make yourself. Various brands of mechanical keyboard switches come in a variety of activation forces. Get a keyboard with swappable switches, and tune the weights to your own preference. For example, on my keyboard I use these switches with a 67g weight, but could get some 62g and 65g ones for the pinky keys.

https://zealpc.net/products/zilent?variant=5894832324646

9. gigaflop ◴[] No.42207688[source]
"Endgamed" keyboard nerd here, and I was able to find different boards that were better for certain roles, until I got to WFH full time. These newfangled magnets aren't convincing enough for me to ditch my workhorses.

I had an HHKB-looking mechanical 60%, and just remapped a few things to be nicer for programming. Function key instead of caps lock, and suddenly so much was at my fingertips. Something like Fn-IJKL for arrows, and pageup/dn, home/end, located somewhere reasonable relative to the arrows. And, a bright red anodized aluminum case, which had a built-in angle, making it very nice to use. More pinky movements than usual, but I made sure to have Fn keys on both sides, to be able to give one or the other a break.

The Preonic '50%' is also pretty neat, and I was pretty productive with the default layout. Hand size was a bit of an issue, but once again, everything I could need was under my fingers with a layer change or two, or some 2-key-combo layer.

Gaming was a nonstarter on either of those boards. I need to at least have a TKL or 65+%, and have it be a sturdy tank. I love my NK87, and use it with Kailh Crystal Box Pink switches. Used to use Box Jades, but these Pinks are crispy.

If someone wanted to get crazy about their mech board and individual key strength, properties, etc, there's a dead simple option of "Buy the stiffer/softer switch as well, install in desired position". Hot-swap sockets are on plenty of keyboards these days, and I've totally heard of people using stiffer switches on the spacebar.