Which Types of Keyboard Switches are the Least Prone to Chattering?

[Hypersphere]

As defined by the DT Wiki, "Key chattering or simply chattering, is an undesirable phenomenon that can affect certain types of keyboard switches. The effect of this problem is that a single press of the key results in multiple registered inputs. As an example, typing the word 'Deskthority' on a keyboard with a defective 'K' key suffering from chatter may yield the result 'Deskkthority'."

Cast your vote for up to 3 types of keyboard switches that in your experience or from your research are the least prone to chattering. This is obviously not an exhaustive list; feel free to make comments and cite other types of switches.

Note that this poll is not [necessarily] about selecting your favorite switch(es); instead, it is about selecting the switch(es) that tend not to exhibit chatter.

[XMIT]

Electrically, the various capsense mechanisms - beam spring, Model F, foam and foil, and Topre - should be combined.

[Hypersphere]

Electrically, the various capsense mechanisms - beam spring, Model F, foam and foil, and Topre - should be combined.

Excellent point. However, elecromechanically, there are different implementations of capsense, and a person responding to the poll might have experience with only one or two of the subtypes of switches.

When a capsense switch "goes bad", does it tend to stop registering rather than exhibit chattering? For example, although foam and foil switches can be highly reliable when the foam is new, the foam is notorious for degrading. Does this simply cause the switch to stop responding or perhaps to respond intermittently?

[ohaimark]

I'd like to note that optical switches, though I don't have any experience with them, would undergo complete failure instead of chattering.

[Findecanor]

Chattering can also occur from other sources than the switch itself, for instance if a switch has a bad solder joint.

[HuBandiT]

the foam is notorious for degrading. Does this simply cause the switch to stop responding or perhaps to respond intermittently?

In my experience the foam broke, the foil fell on the board, so the keyboard registered a continuous keypress.

What is your goal with this vote/research? What is the question really asked here?

[Hypersphere]

the foam is notorious for degrading. Does this simply cause the switch to stop responding or perhaps to respond intermittently?

In my experience the foam broke, the foil fell on the board, so the keyboard registered a continuous keypress.

What is your goal with this vote/research? What is the question really asked here?

Thanks for your comments and perceptive scientific questions. I suppose my goal is to satisfy my curiosity and to see if my own notions are supported or refuted by the keyboard community. It would have been nice to have been able to run two scales simultaneously -- going for the least prone and the most prone to chattering.

[Daniel Beardsmore]

I am not qualified to vote. I can only say that, with the limited number of keyboards I've owned, the problem ones were:

2009/2010 Filco Majestouch, with MX Brown — that keyboard was a disaster. The numeric keypad hyphen-minus wasn't seated properly prior to soldering, and the solder joints failed twice, both times leading to chatter; after the second failure, Keyboardco simply wired the switch with flying leads, as the PCB was just too far gone. A while later, a whole load of keys developed chatter and I just put it aside and gave up with it. I don't know if it's more solder joints, or dirt, or what, but I can't be bothered with the wretched thing any more. It's lying around somewhere. One key in my other Majestouch (MX Blue) developed chatter, which was resolved with WD40 (probably WD40 switch cleaner) and that keyboard is still fine.

I also have a ca. 1989 Monterey K102 with one blue Alps switch (numeric keypad 3) that started to go a bit flaky. That's the keyboard I've mercilessly butchered.

Otherwise, it's not a problem I've seen, with anything.

You have the critical question: when you buy second-hand keyboards, are you only buying the ones that survived? If 99% of the keyboards had failed switches or failed printing, those would likely be tossed instead of sold. So if we have the last 1% of old keyboards—that seem to have flawless pad printing and reliable switches—we're not working with useful data.

When I was looking to buy some new USB pen drives, pretty much every brand/model you looked up, had alarming numbers of bad reviews. Basically whatever you decide to buy, it's guaranteed to die on you! Of course, they didn't, but what are the actual failure rates?

Both Brian from EliteKeyboards and Bruce from Keyboardco confirm that the v1 Majestouch was a lemon, and Brian advised Diatec how to make the stupid things properly, and both confirm that v2 has far lower failure rates (and according to Bruce the v2 feels nicer). There are theories that the failure rates of Cherry switches aren't due to Cherry, but poor keyboard manufacturing processes (although Brian also blames the apparent poor design of MX switches). Brian is a busy man, very busy, which is a shame as he clearly has a lot that he could teach us!

How many Kailh PG1511 switches really had keystems break off? Sometimes you can demonise something over one single failure!

Getting a handle on real failure rates isn't easy, and isn't even a question for vendors, as most of the switch types in your poll aren't made or sold any more!

Obviously you won't get contact faults with electromagnetic switches, but how about reed switches? Those are immune to dirt and in theory should be very reliable too.

The poll I fear is worthless though, as the electromagnetic types are surely just going to dominate. The real question for me, is, how do you make bare metal contact switches that don't suck? For example, the Romer-G has redundant contacts and should prove fairly reliable, but it's widely panned! All that effort for a switch that hardly anyone seems to like, which is a real shame as it's quite an ingenious design. ITT ETL18 is so rare, it's impossible to judge it, but it feels like little else on earth.

(What is a "random linear", anyway?)

[alh84001]

My guess would be that Alps plate spring would not give easily to chatter, but it's just a guess.

[codemonkeymike]

Maybe something like the "Wooting One"[1] will have a good answer to key chatter, I joined the kickstarter a while back at the hope that the switches would be smoother then cherries. I have had quite a bit of problems with chattering with Gateron browns, the Cherry browns seem to hold up better. Only had the privilege to use a HHKB for a couple days and had no chattering at all although I think that is too little of time to judge.


[1] https://wooting.nl

[Hypersphere]

<snip> ....

(What is a "random linear", anyway?)

When I was just starting out on DT, I had some old terminal keyboards that I could not identify and I asked what kind of switches they were likely to have. The reply I got was, "Probably Random-Ass Linear (RAL)". This colorful moniker stuck in my mind, but I shortened the name in the poll to make it more suitable for polite company.

[czarek]

I think anything capacitive or magnetic (hall effect) should be immune to chattering.
In my experience with keyboards, the only switches I saw chattering were very old vintage MX Blacks (couple of them) and quite few complicated white Alps (those were in very bad condition - scratchy and sticky as hell).
I never worn any modern keyboards to the point switches started to chatter though, and I do type a lot - I can shine brand new Filco keycaps in a week, and Realforce in about half year

[Hypersphere]

Probably I should have set the poll up the other way around -- asking for opinions on the switches that are most prone to chattering. In my experience, this prize goes to Alps switches. Nevertheless, Alps are among my favorite switch types -- when they are working well, they are delightful.

[Chyros]

Probably I should have set the poll up the other way around -- asking for opinions on the switches that are most prone to chattering.

Yeah, I agree, because of lot of the "higher" sensing mechanisms appear to be inherently immune to chattering. I haven't had chattering very often with Alps switches, if they go wrong it's more often refusal to actuate properly, in my experience. The worst chattering experience I've seen was with that Lingbao (Gateron, supposedly Outemu) that chattered like hell right out of the factory, yet didn't register double keypresses, so I assume they simply programmed the controller to use an especially long delay to prevent it.

[Daniel Beardsmore]

The worst chattering experience I've seen was with that Lingbao (Gateron, supposedly Outemu) that chattered like hell right out of the factory, yet didn't register double keypresses, so I assume they simply programmed the controller to use an especially long delay to prevent it.

Wait, how did you detect the chattering then?

[Chyros]

The worst chattering experience I've seen was with that Lingbao (Gateron, supposedly Outemu) that chattered like hell right out of the factory, yet didn't register double keypresses, so I assume they simply programmed the controller to use an especially long delay to prevent it.

Wait, how did you detect the chattering then?

Switch Hitter.

[Daniel Beardsmore]

So, the computer can't see double keypresses, but the computer can see double keypresses?

[Chyros]

So, the computer can't see double keypresses, but the computer can see double keypresses?

Switch Hitter flagged it as a chattering issue, presumably due to unnaturally short contact times.

[Daniel Beardsmore]

So what you're saying, is that the keyboard waits so long to confirm a stable switch, that by the time you let go of the key, it's only just registered? I'm guessing that this keyboard is one that uses 1000 Hz polling?

[Chyros]

Ha, not sure actually, might be. It is marketed towards gamers.

[lot_lizard]

the foam is notorious for degrading. Does this simply cause the switch to stop responding or perhaps to respond intermittently?

In my experience the foam broke, the foil fell on the board, so the keyboard registered a continuous keypress.

To me, I have always just equated this term as any intermittent errand key press. Doubled key, occasionally missed (not consistent), or some key registering I didn't even strike. So my definition is obviously too broad.


Even broad, my chatter experiences are highly rare, and almost always related to the controller. As an example, if set too sensitive, the xWhatsit on a beamspring with the solenoid on would win the most "chatter" prize hands down. So much so that keys I didn't even press are registering constantly due to solenoid vibrations (pad leaves the PCB, it's a register for beamers).

[Hypersphere]

the foam is notorious for degrading. Does this simply cause the switch to stop responding or perhaps to respond intermittently?

In my experience the foam broke, the foil fell on the board, so the keyboard registered a continuous keypress.

To me, I have always just equated this term as any intermittent errand key press. Doubled key, occasionally missed (not consistent), or some key registering I didn't even strike. So my definition is obviously too broad.


Even broad, my chatter experiences are highly rare, and almost always related to the controller. As an example, if set too sensitive, the xWhatsit on a beamspring with the solenoid on would win the most "chatter" prize hands down. So much so that keys I didn't even press are registering constantly due to solenoid vibrations (pad leaves the PCB, it's a register for beamers).

So, a beam spring on its own connected to a computer for a long enough time would eventually generate all the works of Shakespeare!