geekhack
geekhack Community => Keyboards => Topic started by: Hak Foo on Sun, 19 April 2009, 21:36:29
-
I was pondering: Why doesn't a keyswitch mechanism like this exist?
In the attached image, the underlying PCB is yellow.
The blue lead is connected to a metal base,, which has a spring mounted on top. When the key (green) is pressed down, the buckling spring will contact the outer ring (black), which would be metal, to close the switch.
This would provide the feel of a buckling spring with the "if it goes bad, it can be desoldered and replaced" nature of a non-membrane switch.
Or am I missing something and "Oh, that's how the IBM Model Q operated. (Incidentally, what were the models between F and M?)"
-
Hmm... How'd you connect the spring reliably? And how good might a spring-to-metal contact be (and stay in the long run)?
-
Couldn't the spring connect to a little nub of conductive plastic?
I think this is a great idea :D
The original design, for reference:
http://www.geocities.jp/kenjin_keyboard/IBMbuckling.jpg
http://www.teknorapex.com/Library.73/current_category.112/divisions_detail.html
-
Hmm... How'd you connect the spring reliably? And how good might a spring-to-metal contact be (and stay in the long run)?
If that's an issue you could always make it capacitive.
-
I think there are a few problems:
1) You need a metal "pot." Metal is more expensive than plastic, especially when it's shaped. This would be a considerable cost factor.
1.1) To prevent the contact from going bad, the metal pot would have to be made from materials not prone to oxidation or degradation, so normal cast iron is out. You'd need something like stainless steel, aluminium or even gold-plated metal. Expensive, expensive.
1.2) The metal part has to have an extremely intricate shape in order to hold the keycap. If it weren't already expensive, it would be by now.
2) The metal-on-metal sound would freak you out. Seriously.
3) I think debouncing may be a real issue here. Typically, switches are specially optimized to minimize key bounce, as this is a majoy PITA to deal with. The spring is inherently bouncy and when it buckles against the wall, it could potentially bounce a lot.
So why has nobody done this? Easy:
It's expensive and it doesn't offer any benefits compared to other switch technologies.
-huha
-
It's too bad there isn't a cheap way to measure the strain on a spring...
-
A small contact stripe should do as you can give the spring a direction towards it buckles.
I see problems with soldering the spring precisely (stainless?) and bouncing.
When all problems have been worked out it probably won't be as simple anymore.
But I suggest you work on it and make it a self contained switch using an established drill pattern. :smile:
-
Would it help if the springs were made of phosphor bronze?
-
Here's a wild idea:
Buckling springs are loud; we can easily hear the keyclicks, especially if we put an ear to the backplate.
So let's attach three microphones to the backplate, to detect and triangulate keyclicks from anywhere on the board.
Software determines which spring is activated and whether it was pressed or released :)
-
Get a hall effect if you want to spend so much money on such an intricate switch.
Btw anyone have a link to a pic of what they look like inside?
-
Ahh, the mysterious Hall-Effect 'boards. Often spoken of, but never seen. If anyone knows of one, could they please post a link to a pic? Thanks.
-
Ahh, the mysterious Hall-Effect 'boards. Often spoken of, but never seen. If anyone knows of one, could they please post a link to a pic? Thanks.
http://dvice.com/archives/2007/07/maglev_keyboard_as_hightech_as.php
(http://dvice.com/pics/maglev_keyboard.jpg)
Make sure you let me know if you find out where to buy it though, lol :D
-
It's not hall-effect based, but Cherry produces linear reed switch modules for some time. I couldn't find any further information than a patent (http://www.freepatentsonline.com/3644855.html) on those, but basically, a magnet is held in the keycap, which is in turn supported by a spring. When the key is pushed, a small reed switch inside the housing closes as soon as the magnet is close enough.
Seems like a very durable and sturdy design to me, if not somewhat expensive because of the large number of reed switches needed. It might be worth it, as they shouldn't have any bounce at all (maybe they're even mercury-coated).
-huha
-
Some reed switches use Rhodium as contact material.
The hall effect sensor has to be powered to work? That means they will fail not only from usage but also from non-usage. It will take years but it adds another element of failure that a mechanical switch does not have. I think they are overrated.
-
Some reed switches use Rhodium as contact material.
How does this influence key bounce? I always thought in the old days, switch contacts were mercury coated, which effectively eliminates bounce. They could as well be on the reed switches, but this is just speculation; given they need to be encapsulated in glass anyway, I don't see any reason not to coat them--remember, they're old, so mercury toxicity and environmental concerns don't count.
The hall effect sensor has to be powered to work?
Yes, that's how hall-effect sensors work. Basically, you send a current through a block of a special material (I think silver was used before semiconductors became available; today, semiconductors are used because they're better AND cheaper). If a magnetic field penetrates your block perpendicular to the direction of current flow, this leads to a force (Lorentz force) on charge carriers (electrons; if you're using semiconductors, electrons and holes, but this gets somewhat nasty), which are in turn displaced. If you attach some leads perpendicular to the magnetic field and your "feed current," you can then measure a voltage which is dependent on magnetic field strength (the stronger the magnetic field, the stonger the Lorentz force--charge carriers are displaced more severely, leading to a bigger potential difference at the sides of the block), called the Hall Voltage.
So yes, hall effect sensors need to be powered, but they don't really decay. Hall effect doesn't work without auxiliary power, but the sensors themselves are what you'd call "passive." You do need electronics to amplify and interpret the results, though. Hall sensors are also somewhat cheap these days, so they're used in a lot of equipment (as far as I know, standard PC fans use them to synchronize the motor for example)--making a hall-effect based key switch seems to be a somewhat logical decision, while coupling it with a membrane sheet does not.
That means they will fail not only from usage but also from non-usage. It will take years but it adds another element of failure that a mechanical switch does not have. I think they are overrated.
As I wrote, they need to be powered to operate. Hall sensors are basically a chunk of metal (or semiconductor; the effect is present in metals, but the resulting hall voltages are really tiny, so you'll need expensive amplifiers and whatnot to get an acceptable result), so they don't decay over reasonable periods of time. They'll basically last as long as the electronics last, but lifetime most likely seems to be limited by the mechanism to provide tactility anyway, as with mechanical switches. One big advantage of hall sensors is they don't bounce; you'll need electronics for processing the signal anyway, and adding a large enough hysteresis removes bounce effectively.
-huha
-
hall effect sensors
Thanks for explaining this.
-
It's a bit hard to understand without an illustration, but maybe this picture makes it clearer.
(http://geekhack.org/attachment.php?attachmentid=2338&stc=1&d=1240610428)
whereas I is the "feed current", B the magnetic field, U_H the Hall voltage, F_L the Lorentz force.
-huha
-
There's a whole world (http://www.magneticsmagazine.com/conf-2009/mag_conf09_index.php) out there...
Of course anything that increases the number of elements needed for production has no chance compared to simple rubber dome keyboards. Especially if it is durable. Two extra elements for a switch means hundreds for a whole keyboard.
Of course! No company that sells on the consumer level will want their keyboards to last forever, otherwise the market will eventually become saturated and they won't sell anymore.
-
new switches will need some nice new keycaps...
check out the sheet of double-shot keycaps that are half completed :)
http://keycapsdirect.com/custom.php