Oddness in MX Clears with varying springs

[QuadGMoto]

Thanks to the recent group buy via Sprit ( ), I now have additional strength springs to play with in MX Clear switches. As part of that testing, I played with the nickel test to check the force required to activate the clear with each of the the various springs in them. Since each nickel weighs 5 grams, I paid close attention to how easily the switch would activate or nearly activate as each nickel was added. That allowed me to estimate how closely the switch activated to the actual weight of the coins and get a finer result than simply a 5 gram range. (Yes, still not excellent, but bear with me.)

So here are my results:

Stock spring: 14 nickels, estimated 67g
65g spring: 13 nickels, estimated 61g
62g spring: 12 nickels, estimated 56g
Blue spring: 12 nickels, estimated 56g
60g spring: 12 nickels, estimated 58g
55g spring: 11 nickels, estimated 54g

This is pretty much as I expected given that the spring the slider rides against to give tactility adds force of its own, with the exception of the 60g spring. Look at the list again and compare the 60g spring to the 62g spring. The activation force is actually higher for the 60g spring than the 62g. That, I did not expect!

I thought maybe I had mixed up the springs, but I was really careful in keeping track and labelling which springs I used. By testing how much force it takes to settle the slider just prior to activation, I was able to confirm that I did not mix up the springs (and neither did Sprit). I also tested multiple switches in both weights in case it was variances within the assemblies. The result was consistent.

Along the way, I also learned that a penny weighs 2.5g and a dime weighs 2.268g, which adds a little more precision to the testing. Using this information, I was able to confirm that the MX Clear with a 60g spring requires a higher force to activate than the same switch with a 62g spring. Both can hold 11 nickels without activating. Both activate with 11 nickels plus a penny. But the 60g will hold 11 nickels plus a dime, where the 62 g version will always activate.

That's a rather counter-intuitive result. Any ideas why?

[ideus]

I suggest you to run an experimental design as follow:



Spring   Switch
55g   Clear
55g   Brown
55g   Red
60g   Clear
60g   Brown
60g   Red
Blue   Clear
Blue   Brown
Blue   Red


This DoE will allow us to get data on the effect of large bump, small bump, and no bump in the switch construction, by testing clear, brown, and red ones. It will also tell us the actual effect of the springs as well of course.


For each combination set at least 2 switches each, all on the top row of the board, or function keys, and use the flatter keycap profile you have, like DSA, i.e. 18 switches total should be set. you will need 2 boards to set up nine each. And for each one, run the test 5 times, record each result in number of coins, it does not matter if it comes the same, that will get 90 results. I can make the statistical analysis if you share the final table with me.


If you want to reduce the amount of switches it can be run with only one switch of each type, that will get half of everything, less information, but at the end we can get some results, with less work.


Use the same coins always, in the same sequence each time. Mark them with a number each to do not get mixed.


What are we gonna get? We should be able to tell the actual effect of the switch design, and the spring module, and their interactions.


Keyboard mad science project  .

[tricheboars]

maybe springs we buy arent quite as accurate as we would like to believe.  how many nickels for stock clears (its what i am using now)?

[laffindude]

That's a rather counter-intuitive result. Any ideas why?

tolerance.

[HPE1000]

I will say I tried the whole coin thing a while back and it was far from accurate from my testing. I was trying coins on 80g originative springs in blues, stock blues, buckling springs and other switches and I ended up getting frustrated and quitting after awhile.

(Did this awhile ago, memory not the best so numbers are going to be wrong/made up because it really doesn't matter in this case)
The thing that made me quit messing around was when I put something like 10 quarters on the 80g blue switch and it would actuate, but I would try it again right after and it took something like 15 quarters before it ended up actuating.

Did you use a bunch of clear switches, or just a single clear switch and keep swapping out the springs?

Maybe try it with a black or red switch? It could be more accurate testing this on a linear switch vs a tactile or clicky switch.

[QuadGMoto]

I haven't had time to do any more testing. I have 10 clears to test with, but only one brown and one red.


For the tests I've done so far, I used two clears for each spring strength. Given the consistency shown by all the switches, while tolerances is a possible answer it seems unlikely given the consistency within the pairs.


In using the nickel test I've noticed that you have to be very careful to place it gently on the switch. Even the slightest drop adds motion to the mass and changes the result. For this testing I based the results on the highest number of nickels(+) that I could stack on a switch without activating it under the theory that smaller numbers could be the result of me dropping the stack or adding weight with my fingers.


I've found it easier to test tactile switches than linears because it's very clear how much weight it takes to activate them. (Come to think of it, maybe I should put a meter on the linear switches so I can tell when they activate.) It's possible to get the nickel stack to "settle" on the tactile bump if the stack is light enough. If the stack is heavy enough to activate, it's not possible to get it to settle. That gives me a consistent, if crude number.


I'm also testing the switches mounted on a test plate, not within a keyboard.

[Techno Trousers]

Question: For the 65g springs, how close to 13 nickels was the activation? I mean, was it closer to 12 or to 13? The reason I ask is that I want to mod some clears to be closer to the feel of capacitive buckling spring, and by my own experimentation I've identified that as somewhere right around 12.75 nickels worth of force to activate.

[QuadGMoto]

Question: For the 65g springs, how close to 13 nickels was the activation? I mean, was it closer to 12 or to 13? The reason I ask is that I want to mod some clears to be closer to the feel of capacitive buckling spring, and by my own experimentation I've identified that as somewhere right around 12.75 nickels worth of force to activate.

I had estimated that it was just barely holding 12 nickels last time I tested. Tonight both samples are just barely activating at 13 nickels and able to hold 12.5 nickels. (1 penny is 2.5 grams to the nickel's 5 grams, exactly half the weight.)

[CommonCurt]

Wish I would have gotten a few more sets of the 60g now.

[QuadGMoto]

I did a quick test of the red with its stock spring so I could test testing the activation point using a multimeter. It activates at almost exactly 9.5 nickels, or 47 grams, easily the lightest measurement I've had yet.

[Techno Trousers]

Question: For the 65g springs, how close to 13 nickels was the activation? I mean, was it closer to 12 or to 13? The reason I ask is that I want to mod some clears to be closer to the feel of capacitive buckling spring, and by my own experimentation I've identified that as somewhere right around 12.75 nickels worth of force to activate.

I had estimated that it was just barely holding 12 nickels last time I tested. Tonight both samples are just barely activating at 13 nickels and able to hold 12.5 nickels. (1 penny is 2.5 grams to the nickel's 5 grams, exactly half the weight.)

Cool, thanks! Sounds like the 65g gold springs are just what I'm looking for.

[Zeal]

Well, that's interesting! Can we get some pictures of 60g and 62g both standing / side shots?

[HPE1000]

Well, that's interesting! Can we get some pictures of 60g and 62g both standing / side shots?

I can, I wish it wasn't so rainy outside, the picture will probably suck

[CPTBadAss]

I agree with laffindude. Even the stock switches from Cherry have a huge tolerance band. For example, from the datasheet, Blacks have a actuation force tolerance of .7 oz or 19.84 g. Likewise, the springs that sprit made will have a tolerance band as well. So some springs will be 62g but might register on your test as 63.23g.

To average out these tolerances, you'll need to test multiple springs.

maybe springs we buy arent quite as accurate as we would like to believe.  how many nickels for stock clears (its what i am using now)?

That's a rather counter-intuitive result. Any ideas why?

tolerance.

[HPE1000]

Picture is bad, but here.

[Zeal]

Picture is bad, but here.

Show Image

Well, there's nothing wrong with the coils/size of the spring. It /may/ just be the tolerance...or #GOLD plated shenanigans!

[davkol]

Nickel's are so 2011...

[QuadGMoto]

Nickel's are so 2011...

Maybe so… 


'Course I don't have the $4,000+ needed to buy a proper meter and integrated stand to properly measure the force over distance.    Maybe if you donated it to me…