Mythbusters - The RipOmeter Tool

[o2dazone]

in for the prosperity of this wiki article

[itlnstln]

Strong work, ripster.  Good info here.

[wellington1869]

rippy - this is really helpful and interesting!  

I knew the browns were extraordinarily light. M's are pretty heavy.

I'd love to know what black cherries are at since subjectively I felt they were among the heaviest ever.

And topres?

and worn-out-brown cherries vs relatively new?  (I felt the worn out ones were even lighter).

it's widely reported but the Unicomp is 65gm or 5gm lighter than the IBM.

is this verified? I couldnt tell from your pic if you had 65gm on a unicomp?  If its true, it would back up my subjective feeling that unicomps were lighter than the M.

[wellington1869]

one question:
when you place the nickels on the F key, do you continue adding nickels until the switch depresses? OR - do you depress the switch and REMOVE nickels until it rises?

I ask because -- my very-light dell rubber dome is at 90gm and still hasnt depressed the key on its own.  ON the other hand, using method #2 (removing nickels), it measures a more reasonable 50-55gm (ie, starting at the down-pressed position, at 55gm, the key stays down, when remove one more nickel, at 50gm, it rises).  Which seems more realistic a range than 90gm!

[itlnstln]

one question:
when you place the nickels on the F key, do you continue adding nickels until the switch depresses? OR - do you depress the switch and REMOVE nickels until it rises?

I would think it would have to be the former.  On many switches, most of the resistance occurs before a buckle/tactile point.  If you were to depress the switch, the up force would be much less on these switches since you would have buckled/pressed the switch past is its strongest point.

[wellington1869]

I would think it would have to be the former.  On many switches, most of the resistance occurs before a buckle/tactile point.  If you were to depress the switch, the up force would be much less on these switches since you would have buckled/pressed the switch past is its strongest point.

but in that case my rubber dome dell is registering 90gm and that cant possibly be accurate   There might be some 'tactile bump dynamics' at play here (where a "downstrike" only needs around 50gm, but a "static resting force" (like nickels) can be upheld by the switch by up to 90gm.)

[huha]

Measuring it like this is not really useful. Keyswitches have quite strange force-displacement diagrams, so just stacking weights on a key will give you just a rough estimate of the maximum force required.
Doing decent measurements is not really easy and involves a lot of patience. I did three rounds some time back with different keyboards, but haven't gotten around analyzing the results and making fancy diagrams yet, so it didn't seem worthwile to publish it. Also, this doesn't really make sense until I've further analyzed the data to determine the error of my measurements, which unfortunately should be quite high.

There's an incredibly high error in measurments with hacked-together devices, so it's not really usable as a reference. From my measurements, however, activation forces of BS are between 70 cN and 80 cN, which is in the dimension you'd expect. These tend to vary somewhat between measurements, so really measuring forces that small is quite a pain due to incredibly large error.

Doing this stuff properly takes hours per switch; time I currently just don't have. The measurements get more refined every time I do a new batch from what I can tell, but the error is still too high (probably, I haven't even calculated it yet) to take this as a reference. I didn't have any linear switches at hand, so this was not linearized or otherwise cleaned up properly, which really should be done before publishing these things. But seeing the topic came up, it was really tempting. So, without further ado, one of my measurements.

Done using completely hacked-together equipment, so please don't take it seriously. The diagram doesn't even look decent. Labels are in German, but it shouldn't be too hard to figure it out.




-huha

[wellington1869]

Measuring it like this is not really useful.

critique sandwich, huha, critique sandwich.
;-)

[huha]

FAIL. Thou art confusing the tactile feeling and graph with the activation force. Activation is just a single point. Ripsters method works for tactile switches, but not for linear switches.

The measured activation force depends on how you measure it. If you put down weights on the switch until it gives in (which wouldn't be the activation force, but rather the tactile point's position, but that's just a detail), you'll get the forward tactile point. If you put weights on the switch and take them away, you'll get the reverse tactile point. On most switches, they don't coincide. On clicky switches, they're way off. On rubber domes, tactile and activation points are totally different.
The problem in keyswitch measurement lies in the force-vs-displacement graph: It's non-monotone, so if you apply a force, you won't get anything lower than the current force applied. If you get to the tactile point, you don't know how much extra travel the switch needs to activate, because it will just drop through that point.

Activation force just isn't enough to describe a switch. The feel is in the force-displacement diagram, not in the activation force. Getting to know the activation force isn't even really helpful to determine any switch properties (e.g. a switch with a fairly shallow force-displacement diagram and high rise for the tactile point should feel snappier, but lighter than a switch with a linear rise to the same tactile point).

That said, no measurement is of no use, but the error of measurements like these is incredibly high (who would have thought that?). When you put down a coin, you will invariably apply some extra force, which could trigger the switch too early.

-huha

[wellington1869]

FAIL. Thou art confusing the tactile feeling and graph with the activation force.

critique sandwich, webwit, critique sandw...  oh never mind.

[Hak Foo]

FWIW, the nickel as we know it wasn't introduced until 1866-- when the world was beginning to think "Hey, metric seems sensible" all over.  Before that, we had an insubstantial, delicate silver 'half dime' instead.

As I understand it, the current Japanese one-yen coin is aluminium and masses exactly one gram, so if you want to do really refined testing, it's a good model.  You can probably get them in junk boxes at coin shops for 10-25 cents each.

[Hak Foo]

Well, my thought was that you get four 1-yen coins (estimated outlay: $1, or four cents if you can harass a friend going on holiday)-- and then put it on top of a stack of nickels to get the desired weight.

[ch_123]

What about the Model F? I'd recommend a different measuring device:

I tested mine out. It was hard to get a single consistent result, but on average it took 15-16 10c euro coins to depress the switch, which would correspond to about 60-65g. I'd say it would be more towards the 60 end as it doesn't feel as stiff as my Unicomp did, and that's rated at 65g.

[ch_123]

Yeah, the Model F I have is one I got off eBay that definitely looked like it had been well used when I got it. I don't have my Unicomp to hand (I sold it to one of my cousins, but I'll be vising him during the week so I can check for definite) but I do remember the Unicomp feeling somewhat stiff to type on, even when I had "broken it in". (Not enough that it become uncomfortable to use, but enough to make me think that I wouldn't want to use anything more stiff than it) Typing on this Model F feels effortless by comparison.

I don't know whether the difference is down to some difference with the springs, the fact that it's using a capactive switch under the spring, the aging, or just the build quality differences, but I can tell you that this thing has a much nicer action (in terms of how easy it feels to type on it, and the sharpness of the tactility) than the Unicomp I had, and a 1993-vintage 1391406 I used once.

[wellington1869]

Welly, you listening??  Sounds like you're going down the wrong path with Unicomp petition.  You should be having them revive the Model F with capacitive switches!!

hmmmm, do those come in ps2? I think there's an AT model right?

[ch_123]

Mine is AT, but it works with a simple AT-PS/2 adapter. It even supports n-key rollover...

[wellington1869]

Mine is AT, but it works with a simple AT-PS/2 adapter. It even supports n-key rollover...

I should look into it. I had one way back in the mid-80's but I dont remember it too well. I just remember it was loud as heck and very bouncy.  Oh great, another keyboard to buy.

[wellington1869]

ya age does seem to make a difference.

Anyway I just remembered why I never got the F model -- the spacebar is way too big and the alt keys are too far apart for my thumbs (I use the alt keys with my thumbs as control keys for my humungous Autohotkey file and its hotkeys).

I did see one model (cant tell if its at or xt) on ebay that had a shorter spacebar, but then the board itself is humongous (with an extra 24 function keys along the top that i'd never use).

So i'd have layout issues with it.  

So ya, the only solution here is for unicomp to produce a modern version

[ch_123]

I would maybe PM some of the other Model F users like Webwit first - what's going on here may just be years of loosening up the springs as much as anything.

Lam/Laurie got a new in box one a few weeks ago. From a post of his from another forum on which I sometimes lurk -

I just got myself a Model F origonal AT IBM keyboard!
Born on 05/06/85 It has never done a days work so is like new.
The sound and feel of the keys is like a Model M but lighter and snappier!
I love it.
It has the buckling spring mechanism but unlike the Model M boards that have the hammer hit a membrane this board uses capacitive switching like a Happy hacker! Or realforce.
Now it may not look as snazzy but I paid £11 for this compared to the £200+ each I paid for the HH and realforce.
If anyone can find one I would strongly recommend it!

the spacebar is way too big

Agreed, probably the main weak-point of the design (along with the less than desirable layout). You do get used to it though.

[wellington1869]

Lam/Laurie got a new in box one a few weeks ago. Whenever he gets out of hospital, you should ask him to see what a brand new one is like. I remember him posting favourable comments about it though.



Agreed, probably the main weak-point of the design (along with the less than desirable layout). You do get used to it though.

he's out of the hospital now. Oh lammy....?

[ch_123]

Oops, hadn't read that thread since yesterday!

[ch_123]

Well, unless he also has a NIB Model F, the above would probably be the more objective test. ^.^

[ch_123]

I wonder how this impressive looking contraption would compare against them... If there was some easy way of converting it to PC compatibility without building adapts etc, I'd buy it.

[wellington1869]

a 'lighter and snappier' bs switch does sound pretty ideal.  too bad about the crummy layout.
Maybe we have to wait for an enterprising geekhacker to hack an F board into an M layout.

[wellington1869]

clearly I need one of those.

Worlds Heaviest keyboard - The Olivetti ANK28-106J 80g Ooof! Actually, I'm sure some Ebay Coke drenched keyboards are significantly higher

wow, an 80g keyboard -- more like a "fistboard".

[itlnstln]

wow, an 80g keyboard -- more like a "fistboard".

Crude jokes aside, at least you wouldn't have problems with accidently pressing the A or ; keys.

[o2dazone]

Crude jokes aside, at least you wouldn't have problems with accidently pressing the A or ; keys.

What crude jokes?

You mean like "I just got done furiously fisting my keyboard for a deadline project"

[wellington1869]

lol, I just meant you'd need to pound on the keys.  man, its impossible to talk about keyboards without sexual innuendo.


But now that I think about it, I suppose I could have said "80g! I'd have to fist that board rather than finger it!"

And that, my friends, would have been a crude joke.

[ch_123]

Has anyone tried this method on Black Alps switches? I've never seen a force rating for them.

[timw4mail]

I was just looking them up to see if I'd like a AT101W.

They are in the middle of the page here.  Almost as stiff as a Buckling Spring.

Show Image

That's really a horrid-looking response curve to me...it looks like it would be really easy to get confused as to if you've pressed the key or not...

[ch_123]

I wonder what the two different lines correspond to, but I'd say the top one is probably the right one - they are a rather stiff switch, on the other hand, the journey down isn't as stiff, so the second one looks more correct overall. There's something about them I don't like which you can probably see from the graphs - most of the force is concentrated near the top of the switch, and once you press it it just bottoms out. That means that you apply a lot of force to get over the actuation bump and then bottom out straight after, contrasted with a buckling spring where the force builds up more gradually and you can train your fingers not to bottom out as easily. They're pleasant to type on, but because I keep bottoming out on them I can't type quickly on them.

That said, if you find any $5 NIB specials, what have you go to lose?

That's really a horrid-looking response curve to me...it looks like it would be really easy to get confused as to if you've pressed the key or not...

Definitely. That's becomes a problem when you try to not bottom out. Because the noise the switch makes is from depressing the key as opposed to a click on the point of actuation, you can think you've pressed it properly from the noise, but in reality you haven't. As I said, not my favourite switch in the world.

[JBert]

Seriously, I'm awestruck at your patience. Doesn't this play out like Jenga?

Anyway, I would like to comment on the "randomness": isn't this caused by friction? Basically, once the key is moving it will have less friction than when it was still in its up position. So when you add another coin, it could be that you slightly nudged it to move and hence buckle under the weight.

Still, an intuitive way to measure keyboards on the go.

[qso]

I was shocked to find out that the low profile keyboard I usually use (which I thought was very light), has 70g resistance.

[keyb_gr]

I was shocked to find out that the low profile keyboard I usually use (which I thought was very light), has 70g resistance.

With the weight method, you only determine the height of exactly one point in the force displacement diagram - sort of a keyhole view. Subjective feeling, however, is influenced by the whole force displacement characteristic. In this case it's the short throw and thus narrow tactile peak (which means not so much actual work) that saves your scissor switch board.

When your finger hits a key, a certain amount of work is easily performed due to plain ol' interia. The key's moving mass is accelerated, and the switch is moved some distance. It gets interesting afterwards, when you have to apply extra force to depress the key even further. That's most of what you consciously register. Due to the inertia part, there'll always be an (individually varying) offset in perception.

[clickclack]

Another myth BUSTED.  Clickykeyboards used to say this (source):
It now says this on the home page:

haahaa, thats awesome!

If you want to REALLY do the whole keyforce thing right you need to plot the entire force curve.  This is the rig that is used on this site for a variety of keyboards.

Show Image

hmmmm.... yup, I definately like it! Gimme, I wanna play with it![

Worlds Heaviest keyboard - The Olivetti ANK28-106J  80g  Ooof!   Actually, I'm sure some Ebay Coke drenched keyboards are significantly higher.

I have found "Board Chow" that could support a house!!!!!!!!!!!!!!
YUCK!

I really like this "RipOmeter" thingy. And I have found it to be very useful and accurate to desribe the general stiffness to comparable keyboards. I have also found it to be rather consistent if certain things are kept in mind and practiced.

-First, consistently being within a few grams to me IS very consistent with regards to feel and preference.

-Second, I have noticed that virtually all keyboard keys are softer if hit dead center and close to floating in the key stem chamber/shaft. Regardless of it being a mechanical type or rubber dome type.

-Third, I have found that if I stack the coins close to where it will colapse/activate/click and just lightly tap the sides of the stack, just enough to let it roll around and find the path of least resistance you can usually find where the key is hanging up. (if only slightly)

I have a mechanical keyboard that will not depress at all (within reason) when struck off center, while I have noticed many Dells have a similar problem but not as severe.

majorly dirty keyboards seem to be more finicky about how accurate the key presses are.

Just some observations that I thought were interesting.
:)

[didjamatic]

That's interesting their coins are 1g exact.   Good find.

I use a Didjometer.  (Sorry, I couldn't resist)  There are many manufacturers of these tension/push/force gauges, just be sure you get a 100g version.  Preferably 10-100g.  I have 2 of them and they are very reliable and always accurate.

Here's an ebay search you can save that will show you some good ones.  It's pretty broad so it will show a lot of things you don't want but most things you do should be there.  If you look often you'll find they pop up in the $5-20 range frequently.  And those are the Swiss/German/French gauges that are a few hundred retail.  There are some new ones from China that go for $45.




EDIT - Here's a 10-100g Sherr Tumico like one of my gauges and an Arpo that looks good.

[sixty]

Sixty really needs to get his company to calibrate his postal scale so we can get a good measurement of the Leopold.

Yeah yeah.. I keep forgetting to re-weight it. I have a theory what I might have done wrong the first time, though.

[clickclack]

I will update this wherever needed, but maybe I can save sixty the trouble.

These figures are from my stock unmodified keyboards.

Archiss AS-KB87T - 944.5g
Brown cherry
Black with legends
(with cord attached and cable tie)

Filco FKBN87MC/EB - 947.5g
Blue cherry
Black with legends
(with cord and cable tie)

Filco FXBN87M/EWEK - 954.5g
Brown cherry
White without legends
(with cord and cable tie)

I know this doesn't belong here but I wanted to post it before I forget to.
I will update it elsewhere as needed soon.

=)

[sixty]

But the Leopold keys are awesomer nonetheless.

If you do not sweat acid, at least.

[NeeGo]

My IBM Model M Silver Label takes 11 nickels and about 10-11 pennies so that's about 80.0-82.5g. I need more nickels to be more consistent so I'd have to check that again. And I have got to get me a younger Model M for my fingers. :sad:

Oh I see a Futaba on the wanted list and I think my BSR has that so I'll check the switch later today.

[ricercar]

Hey ripster, my Apple Adjustable ripOmeter scores 70cn (or 14n, depending on how the count is recognized). Is this number confirmed anywhere else? This seems high, even for an ALPS switch.

[keyboardlover]

Over 5K posts and I only TODAY learned how the ripOmeter works.

Results: I like it! it gets my KL stamp of approval.



It's fun learning how stuff works.




Here's some results I got today:

55g - ergo clears
35g - mx reds
75g - mx green

Are black springs the same as gray or green springs? I'd like to try ripOmetering Panda clears.

[keyboardlover]

I think expecting me to glue my nickels together it a bit much to ask.

[sordna]

Here's a result for you: My Comfort Keyboard System extra-soft-touch version rip-o-meters at 30-35g. I dug it out of the storage. It also has cherry MX compatible keycaps, costar style stabilizers, and is programmable/remappable... it even treats the 2 spacebars separately so you can make one of them backspace.

[ Guests cannot view attachments ] 22737[/ATTACH]

[sordna]

The regular CKS (Comfort Keyboard System) is 55g. There is a soft-touch version at 45g and an extra-soft-touch at 35g, which is the one I have. You might as well list all three. Also I'd list the extra-soft touch as 35g, it's only the numpad area that seems a tiny bit softer. On average my keyboard activates at 35g. Here is a pic of a switch and cherry-compatible keycap, it's interesting, after the initial bottom-out and actuation, the switch can be depressed further 2mm with a little extra force. Feels like linear switch, with a soft-landing pad installed. It's actually quite comfortable. Also including a pic of the left shift stabilizer.  What do you think ?

[ Guests cannot view attachments ] 22824[/ATTACH] [ Guests cannot view attachments ] 22823[/ATTACH]

[Input Nirvana]

2 years ago I bought a beige ErgoMagic for $!5. I gave it away but afterwards thought I should have kept the adjustable stand for the Kinesis Split project. The keyboard was pleasant to type with, and obviously super adjustable, that was some inspiration for the Kinesis mod.

[hcry4]

Envy.

[woody]

Added RipOmeter Cherry ML published versus measured results.

This is not ML, most likely MY.
ML is a cute little low-travel switch with a very light tactile bump.

http://www.cherrycorp.com/english/switches/key/ml.htm

[enoy21]

Interesting. After reading so much on the brown and red switches being considered so light I came here to look at something else and low and behold.

My new Filco is also at 55g.  So I can confirm the previously linked source.   I had thought for a while that my blues felt a little lighter than these ( See in my review )  and that the browns weren't nearly as light as people had me worried about.  

Sometime ago in some thread on here I mentioned that I had a feeling that 55g would be about perfect for me after typing on the blues..... No damn wonder this feels like my own plot of southern boobs to type on.

Now again they are new. So what is typical break in time ?  And will the keys used the most often like WASD break in faster causing an off balance feel to those keys ?   I would love to hear that these will stay at 55g for years to come but now I wonder if I will get 3 months.


Also , would this attribute to people saying they don't feel the tactility much at all in the beginning then say after adjusting they can ?  


If reds are really 45g to these as 55g then Reds may very well be too light for me.

[enoy21]

Which for me would probably equate to 3 months.    Once BF3 comes out next week I'll put alot more work onto the WASD keys and can retest in a few weeks.

I tested the F key tonight.

[Lenny_Nero]

Cant be bothered to wade thru, so my bad if its already been said.
but just in case no one has ...over here in the UK/GB

1p == 3.5 g or an eighth
2p == 7g or a quarter
A Benson *** or 10 quid (a cockle in local parlance) are a gram.

I have yet to check them on my pointer's scales,
but it was a well known 'street' measures everyone used in case they had no jiggers.

On keyboards use a few 1p's to start the stack then the 2's save a big pile.
Might get round to checking them against my Rabone gauges to see the truth,
but it matters not if we all use the same numbers.

...and a thumb end joint is an inch and the nail is 25 thou good for plug gaps.

[Internetlad]

hey ripster can i buy you coffee sometime?

I probably don't live anywhere near you so i'll just ship you a can of folgers.

[keyboardlover]

GLWNT.

[limmy]

Why not use pennies for more accuracy? They are easier to get and 2 pennies equal one nickel in weight. i.e. one penny = 2.500g
http://www.usmint.gov/about_the_mint/?action=coin_specifications

Your dithering method only reduces friction hence creates downward bias of the measurement. What you want to measure is force required for the keys to actuate and it should include friction. Any method that would reduce friction by slightly shaking / by measuring in a car with running engine(origin of dithering) / by hitting the desk three times with constant force is biased and not representative unless you type in those situation.

[sordna]

I've been using the RipOMeter method on and off, but when I got these arcade buttons, I got the idea to measure the force with precision scale. It's really very easy, you hold the switch upside down, and you press down on the scale with the stem. I tried it with cherry reds too and it works fine (you just have to hook up the leads to an ohm-meter to tell the activacion point for linears). I even tried putting my Poker on the scale, zero-ing it, and measuring the force by simply pushing the H key, and the results were surprisingly accurate! Isn't a using a scale more accurate and even easier than RipOMeter ?

[Input Nirvana]

No respect for The Ripometer
Using a scale...where's the fun in that? Tech guys need to be more right-brained sometimes (reference Steve Jobs).

[sordna]

The RipOmeter has a margin of error of almost 5g. A digital scale with better accuracy/resolution than 5g will provide more accurate results and is easier to use. Modern digital kitchen scales are available with 1g accuracy/resolution.

Methodology:
1. Take a good digital scale and set it to grams.
2. Put the keyboard to be measured on the scale, in a balanced position, and zero the scale (tare button).
3. To validate scale accuracy if you have doubts, put a reference weight on the keyboard, or even some nickels! Each nickel should add 5g, so if 10 nickels on the keyboard show 50g, your scale is good.
4. The scale should show 0 with the keyboard on it and the weights removed. Push a key that falls over the center of the scale slowly, until the activation point. The highest reading you see right before the tactile point is overcome (or the switch registers for linear switches) is the actuation force. Try it a few times to be sure.
5. Measuring loose switches is even easier, you just hold the switch upside down, and push the center of the scale with the stem until the click happens (use an Ohm/continuity meter to determine the activation point for linear switches).
6. If Ripster says the RipOmeter is more accurate, just validate your scale's accuracy with nickels one more time.

[huttala]

Saw that you had Swedish Öre in that list, we don't have any coins for öre anymore.
So you should edit that and put 1kr, 5kr and 10kr in the list instead. They are the only coins we have left.
1Krona = 7grams
5Krona = 9.5grams
10Krona = 6.6grams

Hope this helps.