Good feeling of oneness with carbon-coated conductive rubber

[002]

So it turns out that Topre don't only make coiled-spring capacitive keyboards.

This morning I busted out a random keyboard with the intention of popping a switch out for this little activity.

As I was undoing the 400 screws that hold the the Topre PCB to it's mounting plate, I could see strange squarish shapes through the PCB where you'd normally see the tell-tale circular capacitive contact pads. Once I got it apart, well...see for yourself! ( full album here: http://imgur.com/a/sCdxL ):







PCB flipped

[Halverson]

Very interesting! Can you tell there are no capacitive springs just from typing on it?

[002]

I was talking to some of the guys on IRC before where I said that I am sure I mentioned in the past (maybe on here or in a PM to someone asking about Sony editing boards) that this particular keyboard felt like shit to type on.

Back then I wrote it off as being heavily abused, considering it's not really that old (2004 I think). It was really dusty when I got it and has a big gouge-mark along 3 or 4 keys. You can see in this image of it where the scratch starts near the F6 key: http://deskthority.net/w/images/a/a9/So ... _Front.JPG

Anyway, it definitely doesn't feel as good as a normal Topre switch. It feels much mushier and the keys don't snap back up nicely either.

[Broadmonkey]

This reminds me alot of how the buttons on a SNES gamepad works.

[Muirium]

So Topre has tried going down market before…

As for SNES pads: so true. Great layout, poor internals. They were surely designed to break with only a year or two's use, as we churned through them like junk food when I was all over that console. Sweaty handed peers (my kid brother especially) had to supply their own, as they just rotted those poor things. Expensive to replace, of course. And the cheaper third party ones weren't ever quite the same.

[002]

Urgh god the Honeywell or Honeybee ones? (can't remember the name). D-pads were thumb shredders on those horrible things.

[Broadmonkey]

Did they call the SNES game pad for honeywell/honeybee?
But the SNES pads actually lasted quite a long time, as long as you took good care of them. If you on the other hand were the kind of person that mashed the buttons like a craze, it would eat up the "contact pad" on the PCB, and fixing said pad is a pain in the ass. It's an okay solution in theory for a gamepad where you hold down the buttons a lot of the time, but using it for a keyboard?! Nasty I say!

[Daniel Beardsmore]

Obligatory nitpick ;-)

That's not membrane. Normal membrane keyboards use three very thin, flexible, transparent layers of plastic (which is what a membrane is: a thin, flexible sheet): lower with circuit traces on the top, middle with holes for each key, and upper with traces on the bottom. The rubber dome (most keyboards), helical (e.g. Alps spring over membrane) or leaf spring (e.g. Cherry MY) or plastic hammer (IBM Model M and Acer) squeezes the upper and lower layers together through the holes in the middle layer. Electric current then passes between the upper and lower membrane sheets where they make contact. The holes in the middle separating layer ensure that only the intended switch is actuated.

Lousy photos here:

[wiki]AppleDesign Keyboard[/wiki]

Keyboards that use conductive rubber domes and similar conductive pads (e.g. Mitsumi "rubber dome on a stick") normally use a single membrane layer, but some older keyboards simply used a PCB like the BKE-2011. I felt so sure that my old beige Dell keyboard (1995 or later) used traces over a beige PCB, but it could have just been a membrane sheet. (Typically there's a thin metal sheet under the bottom membrane sheet.)

Conventional single-membrane style:

[wiki]Apple Keyboard II[/wiki]

Topre Realforce is rubber dome but not membrane. IBM Model M is membrane but not rubber dome. The tendency for people to refer to rubber dome keyboards as "membrane keyboards" is correct (it is for example why they cannot support NKRO, while the PCB of the BKE-2011 keyboard will take the required diodes), but leads to the conflation of "rubber dome" (actuator) and "membrane" (electrical contacts), when both have been widely used independently of the other.

Knowledge of rubber dome and membrane is so poor, that even I didn't understand how they worked until I dismantled a few to see what was on the inside. Hence all the aforementioned shoddy photos :)

[002]

I was hoping this thread might catch your attention. Hasu already called me out on this not being membrane, which is fine.
All I am really interested in now is why Topre even bothered with the plastic sheet in the middle. Hasu suggested that maybe it's just an insulating layer?

How would you class this switch based on the descriptions here?: http://deskthority.net/wiki/Contact_mechanism

[Daniel Beardsmore]

Honestly I didn't notice it was a plastic sheet, I thought it was a solid sheet of rubber (but obviously the domes are discrete). Doh.

The sheet on the face of it appears to just provide the conductive pads. Normally the bottom of the dome is carbon-coated, or you have a carbon-impregnated foot. If this is the case, you need something to hold the conductive pads away from the PCB.

I would hazard a guess that it is in fact a membrane board, and that the plastic sheet is two membrane sheets: lower one with holes, and upper one with conductive pads.

This may have been a way to re-use standard Topre domes without needing to devise a carbon-coating process for them. Just put the carbon pads on an membrane layer.

Membrane-over-PCB? Looks like Topre have found a whole new way to use membranes that still requires a PCB. You get the NKRO diodes, but also no need for expensive metal contact switches. It's not anything I've ever seen before. No membrane is necessary to use a PCB with domes, but it will be if you don't make the domes conductive.

[Muirium]

I'm no expert. But could this still be a capacitative matrix?

Note the interlocking pattern on the PCB:

That could well be to maximise the ease of making electric contact when the membrane is squeezed. But it makes me think of a design to maximise sensitivity to capacitance, too. Perhaps?

What does Topre's usual capacitative sensor look like?

Diodes suggest it's not capacitative, I suppose. Could just be barking up the wrong tree!

[Daniel Beardsmore]

Those trace ladders are for conduction.

Capacitive keyboards look completely different. Generally two very large PCB pads under each switch. Topre's are two semicircles forming a single circle, while IBM's were rectangular.

None of this is covered by the wiki though :)

[Muirium]

Nudge nudge!

I'm keen on the idea that we're getting closer to rolling our own Model M membranes and even building capacitative switch keyboards from custom designs someday. The better documented this stuff is, the more I'll know why we're not there yet!

Capacitance intrigues me as it, presumably, has shades of grey between simple ON/OFF. I've vague ambitions for velocity sensitive keyboards with enhanced accidental key press awareness and better hints for automated spelling correction on the host.

[002]

What does Topre's usual capacitative sensor look like?

http://mineko.fc2web.com/box/kb-room/it ... rce89.html
Some images of the PCB near the bottom which should give you an idea.

No pictures of the silent switch today sorry - got a little side-tracked when I found this

[Muirium]

Different indeed:

Just like Daniel said.

[Daniel Beardsmore]

Right, and those domes are different to the ones in the BKE-2011, so Topre are not re-using domes.

What threw me is that the pressure points on the domes themselves look to be carbon-coated, so why do you need another sheet of carbon pads?

Needs a very careful examination of that plastic sheet. Also, it wouldn't hurt to verify with a multimeter that the domes actually are conductive: they may be a darker colour in the middle for some completely different reason.

[002]

I'll have a look tomorrow. It might also be worth pointing out that Topre might've done a very similar thing with Sony in the past: http://matsusaka.ne.jp/~ueno/techical/W ... -1200.html No sign of a plastic sheet in those images though.

Info courtesy of hasu

[Broadmonkey]

If these works as a SNES game pad (which I guess it does) the grey circle inside the dome is there to complete the trace and thus actuating the "switch". On the SNES game pad the grey dot is what is used to complete the trace, but I think that on this keyboard, they use it to press down on the sheet which will then complete the trace. Why have the additional sheet? maybe because it helps decrease the wear on the PCB traces (like you use a sock to to minimize wear on your feet by creating a middle layer) as I wrote earlier the SNES pads had a problem with wearing down because the grey dot inside the dome was too hard on the PCB contact.

on the image the shell is facing downward while the PCB is facing upwards

[Muirium]

The shoulder buttons were always the first to go. Street Fighter II was the cause of their existence, and their curse.

[Daniel Beardsmore]

This is my guess:

Topre guess.png (24.14 KiB) Viewed 8756 times

(Scale is kind of out, though …)

[002]

I think that's pretty accurate. The membrane is comprised of two very well stuck together sheets. The only thing that's missing is the carbon pad on the dome.

Some more pics (not that great). You can see where I scratched the pads with the multimeter probes on one of the images:

[Daniel Beardsmore]

OK, so that's weird …

The last two photos show the membranes from the top, I take it?

If I understand this correctly, the keyboard should function correctly without the membrane sheets (since the domes already are conductive themselves), but they added conductive membrane sheets as well for some unknown reason? (I didn't draw a carbon coating on the dome in the diagram as I wasn't sure about what I was looking at, and it didn't, and still doesn't make sense as a design.)

By the way, if you furnish me with all the dimensions, you can have a scale diagram for the wiki.

[002]

Yep the last two photos show the side that the domes make contact with.
A friend from work suggested that perhaps the membrane was added to make the contact with the PCB 'flatter' in the event that the dome didn't collapse evenly. Maybe without the membrane, keypresses weren't registering well enough?

Another friend from work has a set of Vernier callipers that I can borrow so I'll get some measurements on the weekend and update this thread. Thanks for offering to update the image - it'll be good to have it on the wiki.

[Daniel Beardsmore]

From what I can tell, the design appears to be a last-minute hack, since they left all the domes as-is. I don't know whether making the domes conductive could serve any other purpose, other than to signify that the product was hastily redesigned.

It would be interesting to discover whether later production runs moved over to plain domes. Also, is Hasu's membrane-free version a predecessor or successor to the BKE-2011?

You've certainly found a strange keyboard!

[002]

The Sony NEWS laptop that hasu told me about was almost certainly made before the BKE-2011. I'll see if I can find any other info on these conductive dome Topre boards. Lots of wiki editing to do this weekend -_-

Top


Bottom


Side (I think the top to bottom measurement of 7.48mm might be a bit out. It's probably closer to 7.4)

[Daniel Beardsmore]

Also I need the dimensions of the membranes and PCB. Some height measurements of the inside (top and bottom) with the caliper's tail end would be handy.

I was wondering if you'd say that — in other words, it would imply that Topre have found an issue with the the plain dome implementation and added a membrane layer (perhaps as suggested, inadequate contact with the PCB). That would explain the re-use of a dome-making line that includes carbon-coated domes.

[002]

Ok the internal measurements were a bit hard to get.
Distance from the bottom surface of the carbon pad on the dome to the bottom of the whole dome unit is ~2.40mm
Distance from the top of the whole dome unit to bottom of the internal "well" (as seen in the first image) is ~3.40mm

PCB thickness: 1.60mm
Membrane thickness 0.18mm
Carbon pad on membrane diameter: 5mm
Black ring width on dome contact side of the membrane thickness: 0.55mm

Edit:
Carbon pad on the dome is 0.80mm tall
Peg in the middle of the well is ~3.30mm tall

Sorry about no pics this time. Let me know if anything is unclear.

[Daniel Beardsmore]

Quick back-of-an-envelope sketch, complete with annoying smudge. Dimensions x, y and z are missing. The dimensions of the dome don't add up.

[002]

Ok so:
y is roughly 0.5mm high. Note that these sections don't go completely up against the walls of the dome - this image illustrates what I mean:

Spoiler:

x is 5.3mm wide and ~4.9mm high. The circular points in the top left and bottom right are ~1.5mm in diameter
z I am not really sure what I'm supposed to be measuring?

Some other things:
- Where you got 0.28mm in the top right with an exclamation point, shouldn't the equation be simply 7.48 - (3.4 + 2.4)?
- To explain the other exclamation point (13.4), the thickness of the dome itself is not 0.80mm. It's not really possible for me to measure it with any degree of accuracy but I can tell you now it's not that thick. If I were to give it a very rough guess...I'd say it's probably about 0.325mm

[Daniel Beardsmore]

So you mean that the outer wall of the upper well is 0.8 mm, while the outer wall of the lower well is only 0.325 mm?

z is the diameter of the holes in the bottom membrane, but for some reason I got mixed up between bottom and top.