deskthority

LOL! Sounds like you're having fun!

XMIT wrote: Time to go get the hammer and a nail...

Mechanical Violence: if at first you don't succeed, use more.

Don't know if that's MIT way, buts it's certainly the Mines way!

Nope, the MIT way is "Brute force is the last resort of the incompetent." http://studentlife.mit.edu/mindandhandb ... es/hacking

Incompetence aside (spoiler: brute force didn't work, the rivet is still in there, no damage done), I have found a procedure for safely removing switches, after only damaging two of them. The trick is to gently lift them using a small flat head screwdriver, just enough to have the bumps on the switches release the keyboard switch plate.

Time for another video!

At last, the back plate is off. A 5/16" drill bit was able to remove the rivets.

There is a diode under every single switch! Fantastic. It supports hardware NKRO.

Photos and videos will come later. (I've been uploading a video to YouTube for like two hours now, it was taking 5 minutes before. What gives?)

NKRO and soft click, very nice!

XMIT wrote:Nope, the MIT way is "Brute force is the last resort of the incompetent." http://studentlife.mit.edu/mindandhandb ... es/hacking

I guess that's why one school specializes in mental pursuits and the other in putting big ass holes in the earth.

I obtain the following measurements:

45-50g force, 2.5mm travel - actuation
70-75g force, 3.0mm travel - click, tactile response
4.0mm travel - bottoming out

Yes indeed, actuation is before the click! How remarkably odd.

I've posted yet another video to demonstrate this.

"It is recommended to not drop the nickels." LOL! Love it!

Pretty cool stuff there, glad you're digging into these things and enjoying them!

Nice, cool video . I like how serious you sound in these xD .

It's good to see you stack the whole thing on top of the switch rather than putting them on one at a time, which is how I suspect some people came up with fairly low weights for keyboard switches (impact of the coin providing more force) or with key caps on. I didn't know a nickel is 5 g; tbh this method does mean that you can only measure in 5 g brackets and that you assume each coin is equally heavy. When I measured forces on my switches, I used a variety of coins to get different weights, and then measure the weights of THOSE coins on an analytical balance from the lab (+/- 0.0001 grams) because I found the coins, particularly the pennies, had very inconsistent weights.

I have to say, actuation before clicking is pretty common. On anything I've looked for it, I've been able to do it, except for buckling springs - try it on other switches sometime, just push them down extremely slowly and you'll see it's not that hard to do. I don't imagine that the difference in force on other switches is QUITE as huge as this, though.

The caliper is a great idea, very nice .

Yeah I don't have an analytic balance at home and don't plan to get one. My little digital scale reads to 0.1 grams. I confirmed all of these weights after the fact to within 0.5g. That's easily error on the scale itself. I've found nickles to be extremely consistent at 5.00 grams.

The fun never ends with this board.

Earlier on, I mentioned how the switch plates can separate from the switches. Two little plastic rivets can break and the switch plate goes loose.

It gets better. Due to the extremely tight tolerances, it is important to match the correct switch plate to the correct switch body, to make sure that the broken rivets line up correctly for correct switch actuation.

The fun begins when you have ten switch plates, ten broken switches, and don't know which one goes to which. This is an O(n^2) problem and it is expected to take (n^2)/2 attempts to match everything again.

Brittle plastics strike again. Aaaargh!

Work continues.

retr0brighting was a success on the top case and a few key caps. I am now whitening the remaining key caps.

The PCB got a nice cleaning with isopropyl alcohol and is much cleaner now. All keys register (tested by shorting connections manually without switches in place).

The switch plate has some rust that I'm too lazy to bother removing. If this were e.g. an F62 I would sand it down and re-paint it.

There are a few broken key switches. The switches are not made from ABS plastic. I am guessing they are POM/Delrin. Krazy Glue (cyanoacrylate) is not effective for repairing switches. Switches do still function if the switch plates are detached from the rest of the case so long as they are soldered down well.

I'm hoping to have it all back together some time tomorrow.

They don't make em like they used to it seems. Lol. But for real, that must be very sturdy.

I never thought I'd see a complete teardown of an NEC APC here. Great work XMIT. Keep it up.

Yup, this is a great thread, keep up the good work dude.

Sorry about the lack of photos. I'm taking lots of them but I kind of want to get the keyboard wrapped up by tomorrow. Still working.

As of this writing all of the switches are soldered back into place and the keyboard passes testing in Switch Hitter using Soarer's Converter. I had to replace about eleven (11) switches using spares from a donor board. Some chatter, some fail to actuate, some were broken during disassembly, some are missing a tab that makes it impossible to mount a key, some never worked.

I have some intact but unusable freebies if anybody wants a spare switch to play with. You can make a keychain out of it.

I did confirm NKRO using Soarer's Converter and Switch Hitter. I put both palms on the alpha section of the keyboard and saw a whole bunch of keys light up. Neat!

Alright, I need to take another break from keyboard work. Family life calls. This evening or tomorrow morning the retr0brighting of the key caps should be done and I should finally have everything back together. I may keep the broken space bar on this board just as a tribute to being careful. Plus I promise lots of photos. Stay tuned!

I'd definitely be up for one or two freebie ovals to play around with! Could you please keep one to one side for me for a later date?

Sure. I don't think they're going anywhere. Just a couple of switches in a padded envelope to Europe end up being super cheap. Just ask nuum.

XMIT wrote: The switches are not made from ABS plastic.

Could it be PBT? I know some switches are made out of this.

XMIT wrote: The switches are not made from ABS plastic. I am guessing they are POM/Delrin.

Most keyswitch sliders (e.g. Alps, MX, ...) and many of the housings seem to be made from Delrin or some similar plastic.

All done! I'm typing on the keyboard right now.

I keep promising photos. There are loads, trust me. But today is a busy day for me. I probably won't be able to post photos until later this evening or tomorrow. I am borrowing a macro lens from a friend to get some really great switch close-ups.

Everything is working. The switches (which got a bath in isopropyl alcohol) are definitely cleaner and less scratchy. The stabilized keys are smoother and easier to use with some Suer Lube teflon grease. Why even the cracked and welded space bar is holding up okay.

It's been an adventure. I have four more of these in the house and hope to get three more perfect boards.

Photos are here!

https://goo.gl/photos/LK3buGxEJU6pyEsB8

I'll go through and pick some nice ones for the thread.

Hard to tell from the pictures, but how do you feel the retrobright concoction worked?

Worked great! Have a look at the 1! and 2@ keys to see the difference between non-whitened and whitened key caps. But, if doing this yourself, practice on something first until you get the hang of it. The creme that I used makes things much easier, no need to mix up a paste yourself. (Phosphoric acid is safe with ABS plastics.)

(I practiced on a board that I really liked and ended up with poor results. For me this board was going to be a throw-away and it turned out better than expected.)

It helps that I do this indoors and have a UV lamp which gives me controlled conditions. Every few hours I came along and gave everything a spray with some water to keep things from drying out. It took about 12 hours to get those results.

Good to know. I'm debating what to do with my 5181, try and retrobright it or paint it.

And if the results from the first board are any indication you're going to have some great boards to sell in the coming weeks!

Do you mean a Chicony KB-5181? Something like this:
http://coronthica.com/by-uuid/52067671- ... bf8ef609f/
?

If so: this procedure should work quite well. I used:
1/2 cup "40 Volume Creme Developer"
1/16 tsp "Oxi-Clean"

I mixed these with a melamine chopstick, applied with a paint brush, and let sit under a fluorescent UV lamp for 12 hours, spraying with reverse osmosis water every 3-4 hours.

Nice! Good work . I hope I find one someday .

Is the slider on these reversible btw? It only ever shows one side in the pictures.

XMIT wrote: Do you mean a Chicony KB-5181? Something like this:
http://coronthica.com/by-uuid/52067671- ... bf8ef609f/
?

Yep! I can probably practice on the AEKII case I got from you as it will be a casualty to my tinkering

Chyros wrote: Is the slider on these reversible btw? It only ever shows one side in the pictures.

Surprisingly, it is not! Hopefully these macro shots tell the story better.
(f/32 to f/40, ISO 400, exposure as required for +-0EV, Mup mode, camera on table, helping hands holding subject. Overhead and spot lights. No flash. Nikon AF-S Micro NIKKOR 60mm 1:2.8G ED at 1:1.)

Excellent pictures. The non-reversibility makes sense. The front, slanted bit supposedly touches the electrical contacts, and the fact that it's ramped presumably controls the distance, exactly like with Cherry switches. The rear bit I'm assuming is stabilised by an oppositely-shaped guidance gutter?

I don't know what any of the photographic details mean xD . I'm paid extremely little so I can't afford anything near something that has settings like that xD .

Far out! Those pics are fantastic!