deskthority

I ask because it seems Beamspings work the opposite of Topre and the Model F where the conductive rubber is always on the capacitive sensor while Topre and Model F have their conductive mechanism come down onto the capacitive sensor.

I'm really not sure what you mean by "action". They work on the same principle: a moving conductive object changes the capacitance of a capacitor when a key is pressed. Topre and model F switches move this object towards the fixed capacitor plates when a key is pressed, increasing the capacitance. Beam spring switches separate the plates more when a key is pressed, decreasing the capacitance. If you define "action" as whether the plates are moved together or apart, then yes, they have a different action. Don't let this fool you into thinking there is a fundamental difference between them though! A properly-designed controller for capacitive keyboards could work for either "action" by flipping a single bit, as really the only difference is whether high or low capacitance represents a pressed key.

The principle is pretty much identical just that beamspring is active low and the rest are active high.

Exactly.

Beamspring's an oddball because it pulls the target up and away from the sense pads on actuation, like this:



Turns out that's not a great way to produce a clear signal. Much better to move the target a longer distance, and that's much easier when it's going in the same direction as the slider, instead of working against!

Interesting, so if I put a Beamspring switch over a Model F/Topre board it should be able to activate a key press?

In theory: yes. In practice: Not likely. Beamspring has such a weak signal. But the other way around should work. Hell, I've even detected just my fingers on a Topre PCB. Capsense is Capsense.

Muirium wrote: In theory: yes. In practice: Not likely. Beamspring has such a weak signal. But the other way around should work. Hell, I've even detected just my fingers on a Topre PCB. Capsense is Capsense.

It's hard to swallow that such a huge switch produces a weak signal.

y11971alex wrote:

Muirium wrote: In theory: yes. In practice: Not likely. Beamspring has such a weak signal. But the other way around should work. Hell, I've even detected just my fingers on a Topre PCB. Capsense is Capsense.

It's hard to swallow that such a huge switch produces a weak signal.

It's a more refined mechanism, in a way it makes sense .

I still haven't tried beam spring, I think it should be pretty cool though .

It is awesome. My favourite ever switch.

But IBM definitely made several improvements with the complete redesign they did for buckling spring. Capsense is great, but Beamspring offered it a surprisingly small target. They really overhauled that with Model F. Similar area, but much more movement.

Then of course there's the staggering difference in switch height. Yikes! Beamsprings are deep, man, deep.

Topre is analogue. Well... all of the switches produces a rise/drop in capacitance when actuated. But the capacitance of the Topre switch is supposedly practically linear in relation to key travel -- because of the conical shape of the actuating spring -- where as other switches produce a sharper difference. It is explained better in more detail in one of Topre's patents.

This means that Topre switches do not need different operating modes above and below the actuation point like other capacitative switches. Actuation is triggered simply when capacitance has reached a threshold.
Topre displays a "Realforce RGB" keyboard on Computex 2016 right now, which allows the user to change this threshold between three presets. At last year's Computex they showed a keyboard that could emulate an analogue gamepad and a MIDI clavier.

Chyros wrote:

y11971alex wrote:

Muirium wrote: In theory: yes. In practice: Not likely. Beamspring has such a weak signal. But the other way around should work. Hell, I've even detected just my fingers on a Topre PCB. Capsense is Capsense.

It's hard to swallow that such a huge switch produces a weak signal.

It's a more refined mechanism, in a way it makes sense .

I still haven't tried beam spring, I think it should be pretty cool though .

We heard on your reviews that the Model F would have cost around $900 in today's money. How much would a beamspring keyboard cost now?

Where'd he get that number? iSuppli?

Muirium wrote: Where'd he get that number? iSuppli?

I looked up the prices for an M and an F from somewhere in the late 80s iirc and converted them to 2016 $.

y11971alex wrote: We heard on your reviews that the Model F would have cost around $900 in today's money. How much would a beamspring keyboard cost now?

Not sure you could buy it standalone tbh Oo .

Chyros wrote:

Muirium wrote: Where'd he get that number? iSuppli?

I looked up the prices for an M and an F from somewhere in the late 80s iirc and converted them to 2016 $.

y11971alex wrote: We heard on your reviews that the Model F would have cost around $900 in today's money. How much would a beamspring keyboard cost now?

Not sure you could buy it standalone tbh Oo .

That would be a tragedy, since IBM charged something like $280,000for their display terminals.

Lost your keyboard? Buy a whole new terminal!

Well since I imagine these were available (and affordable) really only to large companies and organizations, I'd think replacement components could have been purchased directly from IBM.

emdude wrote: Well since I imagine these were available (and affordable) really only to large companies and organizations, I'd think replacement components could have been purchased directly from IBM.

The 3101 costed only $1295.

Well, okay, but did they have wide consumer adoption like the IBM PC? And you just said yourself that IBM's early display terminals cost 280k.

So in the end they are all just based on the threshold of voltage over the matrix. Beamspring and Model F try to spike voltage, and Topre gradually raises it. Is this correct?

The video you posetd the other day y11971alex is a nice excerpt of the Beamspring switch in action. Of course most of it is not visible here but it's a nice demo:

emdude wrote: Well, okay, but did they have wide consumer adoption like the IBM PC? And you just said yourself that IBM's early display terminals cost 280k.

280k was inclusive of the controller; not sure about the standalone cost of the terminal, which probably could have been bought alone, since Wiki mentioned that the up to four 2250s could be connected to a controller.

I'm not so sure about the 3101, but I do know that 3270 terminals were proliferate in many cities, since they were the only terminals that could be attached to the S/370 mainframes, which were a staple for business computing (for the businesses that had the bank accounts and floor space to accommodate them, that is). In fact, one of my relatives recalled that rows of 3270s used to be in his office in Taipei, but the office was since demolished and the terminals never seen or heard of again. They were probably scrapped for metal, given their prodigious weight. The 2250, however, he never saw before. The 3270s were probably less expensive than the 2250s, which, of all things, are (light pen) touch screen terminals.

The 2250s remind me of IBM POS machines... wonder if convenience shops in the '60s used 2250s as POS units?

seebart wrote: The video you posetd the other day y11971alex is a nice excerpt of the Beamspring switch in action. Of course most of it is not visible here but it's a nice demo:

I should like to give it an English title, perhaps.

Looks like the size of the capacitive part is roughly the same.

Anyone has beamspring _and_ model F keyboards around Seattle?
Or a psoc5lp kit?
We now have a luxury of direcly measuring capacitance thru the signal amplitude.
Can even do it simultaneously. Need to press all the keys in the column (and couple of adjacent columns, just to see if there's any effect. On model F, at least, it's negligible.)

When I send you the FEXT (103-key) assembly, I'll send you the assembly for a 5251 as well. I would love to have someone do these tests on a beamspring. Additional shipping is minimal in USPS flat rate, so why not. That is with the assumption you are up for shipping back

I will ship it back, no question about that. I just feel uneasy about handing it to USPS (or UPS. Or fedex for that matter.) after both my model F ended up like this after shipping. They weren't properly packed, though.

DMA wrote: I will ship it back, no question about that. I just feel uneasy about handing it to USPS (or UPS. Or fedex for that matter.) after both my model F ended up like this after shipping. They weren't properly packed, though.

Ick... that's brutal. I will pack well, and trust you. Just use the same setup returning. We can knock out two birds with one shipment this way

If you end up in NYC with your equipment I could help. I don't think I could send my 20lb beamspring across the states though

codemonkeymike wrote: If you end up in NYC with your equipment I could help. I don't think I could send my 20lb beamspring across the states though

Well.. last year I actually have driven there. Oh wait, no, that was 2014. Don't really know if I'll be able to repeat that though. I'd like to, but The Baby is starting going to daycare now so the car is needed at home every day. And I'm _really_ trying not to get a second car until I get my own home
So far I even manage to do that, living 4 miles from work and 1 mile from the daycare
Not sure it will last after wife starts working though..