Blur Busters Forums

A4Tech answers Cherry's Realkey (RK) with Lightstrike (LK)
A4Tech Promotional Video: https://www.youtube.com/watch?v=hScU5v_ ... e=youtu.be
2:10 we see response times .2-2.8ms

Cherry's Promotional Video: https://www.youtube.com/watch?v=cLWUTOrifXk

ugh. I should build a sub 1ms keyboard using traditional diode matrix, just to spite these people.

http://www.bloody.tw/cn/ProductsKEY.php?pid=11
Translator needed - what's all the writing about 6mm/4mm/3mm about? Looks like they're trying to sell the 6mm highlighted area is a mount that helps stabilize the switch... 1.2mm is the actuation distance.

Sparky wrote:ugh. I should build a sub 1ms keyboard using traditional diode matrix, just to spite these people.

isn't the problem in the switch itself and not the layout or coding of the electrical components?

I stole this from geekhack. [The worst case on Black cherry(with Finger flick)]


Wouldn't your measurement of less then 1ms be destroyed by the "fake press" that happens in the middle?

lexlazootin wrote:

Sparky wrote:ugh. I should build a sub 1ms keyboard using traditional diode matrix, just to spite these people.

isn't the problem in the switch itself and not the layout or coding of the electrical components?

I stole this from geekhack. [The worst case on Black cherry(with Finger flick)]


Wouldn't your measurement of less then 1ms be destroyed by the "fake press" that happens in the middle?

No, you just need to listen continuously for the longest possible single bounce, not the settle time. Looking at that scope shot, the settle time is about 1.6ms, but at no point does the switch stay open for more than about 50 µs. So you sample each set of keys for maybe 60µs, getting through a 10 row matrix in 600 microseconds using the minimal pincount for a 104 key keyboard. Add in the USB poll time and that's about 1.54ms maximum delay between pressing a key and the keyboard actually sending it to the computer, 1.6ms maximum for key release. (if you lay out your PCB carefully and align the sample timing with the USB poll, you can get lower latency on high priority keys, like WASD, space, arrow keys, etc. You just sample those last, right before the usb poll is expected.)

The spite/overkill option would be using a much more expensive USB 3 microcontroller like this: http://www.cypress.com/?docID=46736

so your USB poll interval is 0.125 ms, and you have enough IO pins to just use 2 rows, so you can sample the whole keyboard and report it in under .25 ms, including USB poll interval.

By the way, how fast do optical switches usually recognize a movement change?

Ghilbi wrote:By the way, how fast do optical switches usually recognize a movement change?

probs as fast as you program them?

Can someone give me some code to compare SteelSeries 6Gv2 (with PS/2) and Logitech G910 (USB)?

Q83Ia7ta wrote:Can someone give me some code to compare SteelSeries 6Gv2 (with PS/2) and Logitech G910 (USB)?

Sparky probably knows. Looking forward to the results. I've got a A4Tech B640 if anyone wants to have it shipped to test out PS/2 vs USB using an arduino.