Secrets Of Topre Revealed - Part TWO (mod your key weighting)

[wellington1869]

good stuff. need next installment.

one question: do you think bumping to 45g on home row helps reduce/prevent "aaaaaaaaaaaaa"s occurence? Or not much difference?

[lam47]

This is looking good Ripster!
Can't wait for the next post.

[wellington1869]

btw ripster, what are you using for your macro shots? they're always crystal clear and in focus!  I suppose you have a 'real' camera or something? I gotta get one of those one of these days...  my camera goes out of focus if I get closer than 6 or 8 inches even in so-called 'macro' mode.

I have one of these:


Its a great camera for what it is, but I cant get super-zooms of small objects with it, even in macro mode.

[wellington1869]

where'd ripster go? maybe he's pinned down by a large pile of nickels that toppled onto him?

[wellington1869]

ooooh, looking forward to it! very exciting

(see? I told ya she likes it rough )

[lam47]

Oh my god you have been cutting the rubber!
I do not think I could bring myself to do that. And I did some soldering yesterday with a blow torch because my Iron died (again)
I take my hat off to you.

[wellington1869]

I did adjust somewhat over time but felt in the end that I had extra tension in my hand because of a stupid Topre design decision.

I sorta feel that way too

I stuck them in a tin can with magnet on bottom.

magnet on bottom is a great idea, just a few weeks ago i read about that somewhere and  just did this with a plastic box I use to store screws in. I was a bit concerned about magnetizing the springs though, in case of any electro-magnetic interference with the switches.

I assembled face up, using gravity to keep domes centered then carefully lowered metal frame down.

personally i'd do it face down for this board, it seems the only sure way to see that all domes go into the center of all the keys. At least, I felt better about it doing it face down. Either way this is the hard part in reassembly.

At least I confirmed that the spring's primary function is switch duty, not key feel.

brilliant work. very good to know.

The ol pinkies don't feel much of a difference

very good to know

I haven't had any annoying repeating a's or ;'s yet

excellent. I was worried about that because I occasionally even get "ssssssssssss" and "s" is supposed to be 45g.  Let me know if the improvement holds up over time (ie, the reduction of "aaaaaaa" and ";;;;;;;;;;;;;").


brilliant job rippy! fantastic, and hats off to you for taking the knife-plunge. I'm fairly certain I'm going to follow.
:cheer2: :rockon: :drum:

[wellington1869]

p.s., what do you use for your macro shots?  is it a real camera or just a steady hand?

[wellington1869]

Glad you liked it Welly.  Your threats of pouring goop on the domes made me rush to try this in the morning.  

lol, I think thats still an alternate approach. I might have to try that out on the ol' "Pause key" to see if its effective.

I use a Nikon D-200 with a Macro lens and special flash equipment.

damn you

 My son is really into carniverous plants so he post pics like this in - get this - the International Carniverous Plant Forum.  

good god.

And he likes to come to my office and laugh at me for posting in a keyboard forum!

yea, like the ICPF is really 'normal'

[bsvP585hUO2Y6]

ripster writes:

> Other mystery is what happened to the diodes?? I didn't take closeups
> of the rubber dome side of PCB because I wasn't about to start peeling
> off more domes - but I don't remember seeing a diode. Must be more
> than one way to fix that pesky 6-Key rollover problem.

I don't think the diode trick can be applied here the way it is with
traditional switches.  To repeatedly measure the state of a capacitive
switch you need AC, don't you?  Diodes happen to be rather good at
preventing current to go *both* ways...

[lam47]

Nice work again.
If you guys are getting repeating letters does that not mean you are suffering from the dreaded chattering? I have not once had a key repeat on me.
I was going to say something very rude about Welly and his weird keyboard romp from the other day. Something to do with speed and strength and 30million strokes.
However I think most of it would be censored.

Anyway, could you not complain if you have the chattering? Your boards are from US right. Mine came from Japan so I have no warranty but I would think Majestouch would give you one. Or am I missing the point here?

[keyb_gr]

The big chip is a Fujitsu MB90F337 "16-bit proprietary microcontroller". This is specifically intended for USB applications. Might also do the matrix decoding grunt work - any other chips in sight?

[rdjack21]

The repeating keys are definitely not chattering - you wouldn't get a dozen of them.  It's simply that 35g is a little too light for me and the way I rest my hands on the keyboard.  Other Geekhack members have reported it as well - not just Welly and I.

The other chips are in the big pics attachments - don't remember anything else that interesting but I'm not a chip expert - bunch of little TI guys.

Yea I get the repeating key issue on occasion as well and just like ripster on the a and ; keys mostly. But I've mostly adjusted to it by now and it does not happen nearly as often now. But then again I also use a Filco palm rest with mine which means most of the pressure/weight from resting my hands is now on the palm rest and not the keys.

I think the mod is pretty neat but I really don't want to start cutting mine up yet.

[ch_123]

ripster writes:

> Other mystery is what happened to the diodes?? I didn't take closeups
> of the rubber dome side of PCB because I wasn't about to start peeling
> off more domes - but I don't remember seeing a diode. Must be more
> than one way to fix that pesky 6-Key rollover problem.

I don't think the diode trick can be applied here the way it is with
traditional switches.  To repeatedly measure the state of a capacitive
switch you need AC, don't you?  Diodes happen to be rather good at
preventing current to go *both* ways...

Yeah, when I disassembled my Model F, I don't remember seeing any diodes (unless they were somehow really well hidden) so I'm pretty sure capactive boards have some way around the rollover problem. And considering that AT/PS2/USB only supply 5V DC, I dont think there's any AC floating around there.

[o2dazone]

I have to hand it to you, you've got big brass ones to be putting a razor to a $250 keyboard

[talis]

ripster writes:

> Other mystery is what happened to the diodes?? I didn't take closeups
> of the rubber dome side of PCB because I wasn't about to start peeling
> off more domes - but I don't remember seeing a diode. Must be more
> than one way to fix that pesky 6-Key rollover problem.

I don't think the diode trick can be applied here the way it is with
traditional switches. To repeatedly measure the state of a capacitive
switch you need AC, don't you? Diodes happen to be rather good at
preventing current to go *both* ways...

Yeah, when I disassembled my Model F, I don't remember seeing any diodes (unless they were somehow really well hidden) so I'm pretty sure capactive boards have some way around the rollover problem. And considering that AT/PS2/USB only supply 5V DC, I dont think there's any AC floating around there.

The diodes would indeed prevent the correct measurement of the capacitance of the contacts.  As its a capacitive switch at no point do you make it a dead short, all you're doing is altering the value of the capacitor at each switch.  Because of this, you don't set up a loop in the circuit that would cause ghosting.

Likely they pull one end of the switch to ground (row), and measure the rise time of the capacitive charge through a known value resistor on the col.  you don't need AC, a step DC voltage would be enough.  

Another standard way to measure capacitance is to use the element you're trying to measure in a RC type oscillator, and measure the frequency it produces.  This is a  more stable, but much slower way to measure C, so its unlikely what they use.  The metal in the spring should produce sufficient change in capacitance that a simple charge time measurement should capture it.

[talis]

I was surprised the spring did not make electrical contact with the PCB.  The spring is conductive.  It also is highly attracted to magnets - the spring compressed when I put in my little magnetic tuna tin.  But no continuity to the PCB.

I'm no electrical engineer but I thought there is a + and a - on every capacitor.

Nope, take a look at any ceramic Capacitor, they are completely non-polarized.  In its basic form, the capacitor is a non-polarized device.

A Capacitor is made up of two plates of conductive material, with a non-conductive substance between them.  When you put positive charge on one plate, negative charge is attracted to the other plate.  That allows it to store some charge (since the material between the plates is non-conductive, there is nowhere for the charge to go).  The electrolytic capacitors are basically made up of two strips of tin foil with a sheet of paper in between*, ceramics are similar but have a number of leaves interleaved, with a ceramic material in between.

C = eA/d

The value of the capacitance is based on three things:

A - The size of the pates (area)
d - The distance between the plates
e - The dielectric permittivity(this is based on the material between the plates).

Obviously with the keyboard, d and A are fixed, what changes is the permittivity of the material betwen the plates.  The spring has a different permittivity then air, and therefore when it collapses down near the contacts, it changes the e, and the capacitance.

* You can actually make a capacitor exactly this way, glue a sheet of tin foil to each side of a sheet of paper.  The real electrolytic is a bit more complicated then this, as one sheet has an oxide layer created on it that acts as one terminal, and the other terminal is made up with a mix of the other sheet, and the paper soaked in an electrolytic fluid.  This differential structure is what causes it to be polarized.

[wellington1869]

I was going to say something very rude about Welly and his weird keyboard romp from the other day. Something to do with speed and strength and 30million strokes.
However I think most of it would be censored.

lol, you should have, i'd have enjoyed it

ya its not key chatter, just that its too easy to hit those keys sometimes cuz they're so light.

I'm still surprised they didnt choose to put diff springs for diff resistance and instead built it into the dome thickness. just seems more logical and intuitive to have modded the springs instead. I'm sure they considered it. I'd be interested in knowing why they went with the domes instead. Maybe that membrane is poured in a mold and maybe it was easier to build diff thicknesses into the mold itself. Then just use a standard spring. Still, seems counterintuitive.

It is a pretty nifty key design I have to say. Conical springs - who'd a thunk it.

[ch_123]

Obviously with the keyboard, d and A are fixed, what changes is the permittivity of the material betwen the plates. The spring has a different permittivity then air, and therefore when it collapses down near the contacts, it changes the e, and the capacitance.

So something can change the capacitance of a capacitor by just being near it without having to make contact with it? I was rather confused when I was trying to figure out how the Model F's mechanism worked, but it's start to make sense.

Few parts = Cheaper to make than your old style F capacitance switches

I don't think that there's a huge difference in part count between the F and the Topres, they both have a PCB with loads of capacitive terminals, they both have springs (although they do completely different things) and whereas the Topre has a rubber dome to cover the spring, the F has a plastic hammer that attaches to the underside.

That said, the innards of the Model F are somewhat over-engineered, but they probably could have made a more frugal version without affecting the overall quality of it.

[talis]

So something can change the capacitance of a capacitor by just being near it without having to make contact with it? I was rather confused when I was trying to figure out how the Model F's mechanism worked, but it's start to make sense.

It can, yes.  If you want a good example of this, listen to "Good Vibrations" by the beach boys.  The Theremin they use works on exactly this principal.

Most capacitors this change is too small to measure, unless you go out of your way to design one specifically for this purpose.

[ch_123]

Yeah, that would make sense. On the F, the hammer is positoned just above the two contacts, but when you press the key, the hammer is pivoted forward so that it lies across the two points. I know about the permittivity of the medium affecting capacitance, but I never thought that resting a nonconductive material over the two plates would have an effect (albeit probably a small one as you say)

[wellington1869]

I think i'm going to go whole hog and replace 9 keys altogether.  Basically everything that is currently 35g in the main alphabet cluster (the sole exception being "\" which is so out of the way that it doesnt bother me).

So I'll swap their domes from the arrow cluster (the 9 keys in the top part of the arrow cluster).

At that point it'll be effectively an all 45g board. (I dont care so much about the numbers and the few remaining 35g keys).  Its plenty light already, I'm not expecting a great difference in feel; this is really just to try to reduce the instances of accidental key presses.

Best part is, despite the cutting, this mod is basically functionally reversible (just put the domes back in their original spots).

[wellington1869]

Great link. Its much easier following along the auto-translation having opened up my topre a couple of times already.

"The screw it is many." Indeed.

"As for the 45g key when you compare to 30g, the [takutairu] impression is strong clearly, is."  I disagree, I think the "takutairu" isnt that noticeable between them.

I like the different colors too.

[wellington1869]

Looks like he tried o-rings "to make stroke shallow". I wonder how he came up with those force-resistance graphs.

And look, he sort of had my idea of putting gloop on the keys:

"The key top being too light? It tries pouring in the silicon resin to the key top?

It puts the lotion on its body.

Looks like none of his mods worked out:

"The silicon rubber -> hardening time is long. Formation is difficult.
The optical hardenable resin -> formation is difficult.
The polyester resin -> it becomes feverish. "  

Well at least it didnt get the klappe.

The putty for Plamodel -> hardening time is long.

I never had that proble... nah, too easy.

When it expands, such feeling.

:-|

[wellington1869]

look, in 2007 topre had at least one board with black keys and yellow-ish lettering - it looks great. I wish they'd do this for an english version, I'd love to buy these keytops.

[wellington1869]

sweet! I love that look. The yellow lettering works well on the topre black keys.

[keyb_gr]

It can, yes.  If you want a good example of this, listen to "Good Vibrations" by the beach boys.  The Theremin they use works on exactly this principal.

They used a Theremin in that one? Never noticed. Then again, I don't think I've listened closely to this song in ages.

EDIT:

Contrary to popular belief, the theremin was not used on the 1966 recording of "Good Vibrations" by The Beach Boys, which featured Paul Tanner's "box", later called the electro-theremin. However, for concert appearances, an oscillator slide-controller was designed and built for Wilson by Robert Moog. Wilson helped to popularize the instrument when he recorded Paul Tanner playing his electro-theremin—for the first time in recorded music history—on the song "I Just Wasn't Made For These Times." The song appeared on The Beach Boys' 1966 album Pet Sounds, considered one of the most influential albums in popular music history.

http://en.wikipedia.org/wiki/Theremin

[rdjack21]

Here ya go - the Dharma Tactical.  Probably have to book a ticket on JAL to buy it.

Interesting think about this OEM version of the board was that it was SUPPOSED to be an all 30g board but someone must have gotten a hold of 6 U.S. nickels and found out it wasn't.

I think that is the first time I've seen a rebadged Realforce board. But I do like those keys I would not mind having those as an option. But I would still want a set of the current keys as well.

[wellington1869]

I think that is the first time I've seen a rebadged Realforce board. But I do like those keys I would not mind having those as an option. But I would still want a set of the current keys as well.

i guess the only way to get those keys right now is to shell out $230 plus shipping on a dharma tactical.

Unless majestouch is listening...?

[nocturn4l3030]

i know this thread is old now but i didnt want to start my own topic..

still debating whether to do this mod ...maybe it would make more sense if i opened it up first but i have some questions

do i need a keycap puller for this?

i noticed when cutting the rubber domes and then switching...the cut is obviously not a precise fit...is there something that holds the dome in place? if not...i can see that trying to center the dome over the spring and then putting the plate on wouod be a pain in the ass because it would move the rubber domes around

i think thats really jus my main concern is that ill fail to put the dome precisely over the spring and then put it back together without misaligning everything...any tips on this? or just gravity -.-

when cutting..is there a precise area that you're trying to cut? i was jus gonna replicate the area that you did but i figure itd be smarter to ask first

has someone outside of ripster and willington done this mod before?

 wanted to ask though, whats holding those rubber

[nocturn4l3030]

thx for the quick response

but oh man jus surface tension?? FML.. duno if i want to go through this now..i can see myself misaligning a rubber dome with every attempt... >.<

i only jus want to replace the A key for gaming purposes too..

no water for me! im scared enough as it is lol

anyway thx again

[vynn]

Since the springs add 5g, has anyone tried simply removing them from an all 55g Topre, effectively making it 50g? I'm guessing no if the spring is necessary to bouncing the keys back up, but it does seem like the rubber dome might be sufficient for that.

[vynn]

oops, just found the patent about the function the spring provides.

http://www.google.com/patents?id=MjEuAAAAEBAJ

[nocturn4l3030]

if anyone is trying to do this mod.. be careful about stripping the screws in the back.. a screw stripped VERY easily and just ruined my whole project ... was excited to do this damn mod and now i'm just pissed off >=[

[nocturn4l3030]

hm i duno.. they were kinda tight and it caused one of the screws to be stripped.. still mad about it !!  argh

[RiGS]

Did you use the correct screwdriver?

[nocturn4l3030]

it worked for all the other screws... /shrug

[nocturn4l3030]

the black screws were fine.. it's the outside silver screws that were screwed in tight and managed to strip

[RiGS]

it worked for all the other screws...

Most of the time you can unscrew a screw with an improper screwdriver, however in some cases they appear "tight".
When you use the correct PH/PZ screwdriver then it fits perfecty, and doesn't take any special effort to unscrew it.

[nocturn4l3030]

Probably M2s.  Pull and replace.

 
what's a M2?

[nocturn4l3030]

Most of the time you can unscrew a screw with an improper screwdriver, however in some cases they appear "tight".
When you use the correct PH/PZ screwdriver then it fits perfecty, and doesn't take any special effort to unscrew it.

 
/shrug maybe i picked the wrong screw then.. i jus have a small screw driver set and i'm pretty sure i picked the "correct" size.. but yea if others didn't strip their screws it's prob somethin i did wrong then

[whizz]

Guys, how many of you did the weighting mod? Any problems with it? I want to buy the keyboard and mod it, so I want to be sure that it is OK and there is no drawbacks (except the warranty of course). If for example the modded keys don't feel any different, like chunky or something. Thank you.

BTW, do you think I can use domes from the realforce numberpad? That would be great, if they are the same.

[hasu]

I exchanged rubber dome sheets each other between HHKB and Realforce.
I have had no problem on this mod except cost.

I'm happy with variable weighted HHKB now and all 45g weighted Realfoce 101 works well, though not impressive for me.