deskthority

Pretty soon I'll have a significant portion of the sought-after blue switches on the market (SMK Alps, Blue Alps, Cherry MX, NEC Oval). The nickel measurement technique from ripster isn't necessarily perfect, so I'm considering the purchase of a precision force gauge (and the construction of a force testing platform).

I'd also be willing to test other switch types.

One of my main concerns is whether people are interested in the results -- if no one is interested I won't bother. Also, if another user has a better testing setup, I won't bother.

I actually built a switch testing device with a pressure sensor. It's all very interesting... but at the end it's still down to personal preference. The graphs don't help you much understanding if you'll like a switch. That being said, it's a fun project, so why not?

I think HaaTa also owns one. Like matt3o says the numbers are interesting, especially in comparing but the data cannot convey the keyswitch feel and sound. Now for our wiki exact data would be great.

I think force measurements would be extremely helpful. I've done a few measurements of my own with weights, but they aren't super precise. The data is very useful because it allows us to class switches by weight and partly by keyfeel (which has to do with the actuation force compared to the terminal force), as well as force gaps (which is extremely prominent on the NEC oval switch).

Indeed. Well calibrated graphs would let us compare things objectively. It's not a perfect picture of everything about a switch, but it's a hell of a lot better than a link to a thread of our typical touchy feely comparisons, especially when most people involved haven't tried a board of the switch in question and are blindly extrapolating. (Mea culpa!)

There's never truly an alternative to trying switches in person. But we can certainly help make the search more informed with this kind of data.

Also, actual force graphs tell quite a lot from the keyfeel as well. You can tell a lot of the smoothness of a switch from its force curve. It also shows anomalies in the keyfeel, such as the second tactile bump in Matias switches or the tactile bump in linear Alps.

Chyros wrote: Also, actual force graphs tell quite a lot from the keyfeel as well. You can tell a lot of the smoothness of a switch from its force curve. It also shows anomalies in the keyfeel, such as the second tactile bump in Matias switches or the tactile bump in linear Alps.

And if testing is done dynamically, it can help deal with the issues of static friction vs dynamic friction.

Would we be better off sending switches to the users who have them already or coercing those users into posting the data on the wiki?

I planned on measuring smoothness and ringing with a piezoelectric element. The resulting audio graphs show how much friction/ping/etc. occur.

ohaimark wrote: Would we be better off sending switches to the users who have them already or coercing those users into posting the data on the wiki?

I planned on measuring smoothness and ringing with a piezoelectric element. The resulting audio graphs show how much friction/ping/etc. occur.

Oooh! The Model F and Pingmaster must be off the charts on that xD .

Chyros wrote:

ohaimark wrote: Would we be better off sending switches to the users who have them already or coercing those users into posting the data on the wiki?

I planned on measuring smoothness and ringing with a piezoelectric element. The resulting audio graphs show how much friction/ping/etc. occur.

Oooh! The Model F and Pingmaster must be off the charts on that xD .

Yes the IBM 5954339 Alps SKCC but also the NEC APC-H412 blue ovals should be right up there on the "ping" scale.

So the consensus is that force measurements won't really matter (as others have done them and should post them on the wiki , but that you all want to see the measurements from a piezo?

NO measurements do very much matter and we need them for our wiki, piezo and other.

ohaimark wrote: So the consensus is that force measurements won't really matter (as others have done them and should post them on the wiki , but that you all want to see the measurements from a piezo?

I'd post my force measurements, but I don't consider them precise enough. Proper force measurements are definitely very useful.

So... What kind of precision and granularity are we looking for? .01g?

I'd build a mechanical "presser" with something like .1mm increments and hand adjustments. I'd measure the switch at each of those increments and any particular points, like just before and after actuation/click.

I wouldn't be able to make a low-granularity graph of the press in motion, at least not as detailed as the MX Charts and whatnot.

Objective data. Good.

ohaimark wrote: So... What kind of precision and granularity are we looking for? .01g?

I'd build a mechanical "presser" with something like .1mm increments and hand adjustments. I'd measure the switch at each of those increments and any particular points, like just before and after actuation/click.

I wouldn't be able to make a low-granularity graph of the press in motion, at least not as detailed as the MX Charts and whatnot.

Yeah that should work I guess. Of course, the smaller the increments you can get, the higher the resolution, and the better it'd look like. Still, that should work . At the very least it'd provide high-quality actuation, tactile and botton-out forces, and with some luck it'd make some nice force curves too .

So after the Summer of Ping, now we have the Winter of Ping.

I'll start with the piezoelectric measurements. After I get those done I'll make a few purchases and build the press-tester.

ohaimark wrote: I'll start with the piezoelectric measurements. After I get those done I'll make a few purchases and build the press-tester.

Cool thanks!

Yeah we really need someone with a precision switch tester who can generate and post force curves of all the switches. I have plans for one on the back burner but it could be two years before it is usable. Don't hold your breath.

I think this is a fantastic idea. Please do it! I agree that it boils down to feel, but it'd be nice to have an accurate reference for people to use.

After taking a break for most of 2015 due to work commitments, I came back and wanted to make sense of all the Cherry MX compatible switches out there. I ended up making a spreadsheet but found that there are different weight measurements out there. Manufacturers measure actuation and 3rd party springs and Zeal all use bottom out weights.

The Deskthority Wiki is an amazing resource and adding accurate information like this would only make it better.

Deleted my last post as there was a change of plans... A $500 press force meter with a 1 gram resolution became available for $50.

I struggled to find a way to achieve .1mm movements -- it will require a bit of fabrication, but I think I have a solution.

I plan to mount a fine-threaded nut and bolt on a cheap drill press conversion kit. Moving the bolt should give me the granularity I need when measuring, much like the operation of 3D printer axes.
To measure the distance I've moved, digital vernier calipers (accurate to .03mm) will be mounted somewhere on the drill press. I'll figure out the exact spot later.

I think the total cost will fall somewhere under $125... Still, I could use a bit of support. If U.S. users want to pitch in (I prefer to ship in the CONUS) I'll send them a link to my eBay store. I have a bunch of keyboards and some other assorted junk listed.

I'm going to be shameless and post my eBay store link here. I'm low on cash for this project due to college expenses.

http://www.ebay.com/usr/devrial

There will be a Green Alps keyboard listed soon. I'm holding on to a few keyboards for testing purposes, so if you want more of them in the store funding me wouldn't hurt. Alternately, you can donate -- PM me. But I don't want to seem like a bum.

These are the items I need to make it work:
http://www.ebay.com/itm/171970158990?
http://www.ebay.com/itm/231700202477?

The press tester arrived today. It works like a charm, and I picked up mounting hardware for it. Once I get the whole thing built I'll post a "call for keyboard switches" on Deskthority and, perhaps, Geekhack.

I'm hoping that the community will ship me desoldered switches and, in some cases, whole keyboards to test. I'd be able to help a little bit with shipping costs, but couldn't foot the whole bill. European users probably shouldn't bother unless they have a lot of disposable income.

Just ordered the drill press and calipers. Will start a new thread with build photos sometime next week.

Based in the US and happy to help if I'm able.

It's surprising how hard it is to find accurate standardized information on this, so thank you for doing this Do you have a list of switches that you plan on trying?

Alps (& clones):

• Acer
• SKCM Blue
• SKCL Green
• SKFL Grey
• NEC Blue Oval
• SMK/Monterey Blue
• SKCM White

IBM:

• Buckling Spring, F
• Buckling Spring, M

(All items below will be lent to me by mashby, who is a scholar and a gentleman.)

Cherry:

• Red
• Black
• Grey (linear)
• Blue
• Green
• White
• Brown
• Clear
• Nixdorf (vintage black)

Gateron

• Clear
• Red
• Black
• Yellow
• Brown
• Blue
• Lock

Zealio

• Clear (don't recall the grams)
• Purple (don't recall the grams)

Misc

• Green Razer
• Aristotle (clicky)
• Lots of various Korean springs

No mention of cyan Omron B3G-S?

Here’s Omron, Alps, SMK, Cherry:

I didn't think Omrons were popular. Perhaps I'm mistaken... If you want to lend a switch I'd happily accept.