Cherry mx board 6.0 Chatter problem

[45fds68e]

I've bought Cherry mx board 6.0 like 2 weeks ago and at the beginning there were no problems at all (at least I didn't know) but 2-3 days ago I realized when I tap q it sometimes chatters and types more then 1, few minutes ago I realized it only happens when the keyboard is in performance mode which closes windows key, alt f4 etc. But when it is in office mode it doesn't chatter at all, can someone tell me why is this happening, is this a feature or is it just my keyboard being silly? It's a little bit annoying.

[Blue_Moon]

i can only guess as i've never used this board, but it sounds like the keys aren't being debounced in performance mode

When pressed again, the Cherry key turns red and Performance mode toggles on. This is where Cherry's RealKey (RK) technology comes in. It is separate from the scan rate over USB, wherein some keyboards have the option to have it vary from 200-1000 Hz (5-1 ms respectively). This means every 1-5 ms, the keyboard controller checks for whether a key is pressed. If a key press is recognized, different software routines within the keyboard are processed (i.e. de-bouncing, anti-ghosting, and so on). Normally, this has it take ~ 20 ms before the keystroke is output to the PC and then the display. Cherry's RealKey technology uses an analogue controller which detects current flow from a key press, and the signal processing takes ~1 ms instead of the 20 ms most others take. This is a part of the journey from key press to key stroke as seen on the screen and is different from USB vs. PS/2 latency, key switch actuation, display input lag, etc. In theory, this can help reduce total keyboard input latency from the keystroke to the display effect, although with an irreproducible means to quantify this, I can not say much more. Refer to this for more on Cherry's RealKey technology, and also note that no other keyboard since, including some newer keyboards from Cherry, have adopted this. Take that for what you will.

https://www.techpowerup.com/reviews/Cherry/MX_Board_6.0/5.html

this excerpt from a review on the board indicates this is the case, saying that checks such as debouncing cause latency of about 20 ms, so these checks aren't performed in performance mode. it kinda? makes sense, because you shouldn't notice the effects of key chatter when playing a game, but i dunno. i never gamed with key chatter

[45fds68e]

I think I understood it kinda but I still don't understand why does it only happens when I press the q key but not another? Maybe software problem?

[JianYang]

i can only guess as i've never used this board, but it sounds like the keys aren't being debounced in performance mode

When pressed again, the Cherry key turns red and Performance mode toggles on. This is where Cherry's RealKey (RK) technology comes in. It is separate from the scan rate over USB, wherein some keyboards have the option to have it vary from 200-1000 Hz (5-1 ms respectively). This means every 1-5 ms, the keyboard controller checks for whether a key is pressed. If a key press is recognized, different software routines within the keyboard are processed (i.e. de-bouncing, anti-ghosting, and so on). Normally, this has it take ~ 20 ms before the keystroke is output to the PC and then the display. Cherry's RealKey technology uses an analogue controller which detects current flow from a key press, and the signal processing takes ~1 ms instead of the 20 ms most others take. This is a part of the journey from key press to key stroke as seen on the screen and is different from USB vs. PS/2 latency, key switch actuation, display input lag, etc. In theory, this can help reduce total keyboard input latency from the keystroke to the display effect, although with an irreproducible means to quantify this, I can not say much more. Refer to this for more on Cherry's RealKey technology, and also note that no other keyboard since, including some newer keyboards from Cherry, have adopted this. Take that for what you will.

https://www.techpowerup.com/reviews/Cherry/MX_Board_6.0/5.html

this excerpt from a review on the board indicates this is the case, saying that checks such as debouncing cause latency of about 20 ms, so these checks aren't performed in performance mode. it kinda? makes sense, because you shouldn't notice the effects of key chatter when playing a game, but i dunno. i never gamed with key chatter

Debouncing can be implemented without adding any more latency on one of the edges. So you could implement it so that there is no latency on press, but some on the release (in the order of 5ms for cherry). But it is a lot simpler to implement debounce with latency on both edges, which is what most controllers do.

I think I understood it kinda but I still don't understand why does it only happens when I press the q key but not another? Maybe software problem?

Then replace the q switch - it is probably a switch problem if it is only one

[Findecanor]

Chattering shouldn't happen, even in performance mode. Nonetheless, you see posts about it in forums from time to time. The cause could be either of:
1) The individual key switch may has a slight error in manufacturing. Every switch chatters during a short time, usually within 5 ms. This is handled by "debouncing": usually a small delay of 5 ms or more that masks the chatter. However, a bad switch could chatter for a longer period.
Some keyboard manufacturers use a long enough debouncing period to work even with these switches — so that they would need to handle as many returns. Those keyboards do not have any performance mode.
2) The switch has a bad solder joint to the PCB. This would in effect act as a second contact which would have its own bounce.

Fast and accurate debouncing is a selling feature of the MX-Board 6.0. It is supposed to use some special method for debouncing that Cherry advertises as being faster than others.
Therefore you should have a reason to be able to return it under warranty.

[45fds68e]

You get it wrong, chattering only happens in "performance mode" but it doesn't happen in "office mode" and today I saw a reddit post about this problem and that guy was having the same problem as I'm having. This is his post btw: When I enable “Performance Mode”, I get consistent debouncing. Anyone with this keyboard notice the same? Enabling “Office Mode” resolves this issue.

I don't think this is a switch error. Maybe software? I emailed the support and am waiting for their response and I also asked the guy if he found a solution for it.

[Findecanor]

I'm sorry that my previous post was a bit unclear. What I meant to say but missed to was that I believe that Office mode could be using a longer debouncing period than Performance mode.
Thus, the longer period would mask the flaw in the switch.
Anyway, I don't think that the community has figured out exactly how the Cherry MX-board 6.0 works, as it is supposed to use some unusual fancy way of doing debouncing in performance mode.

[45fds68e]

Update : Replaced the keyboard with a new one. If you are having the same problem just change it under warranty because THERE IS NO SOLUTION FOR IT