Topre rubber dome stiffening

[RiGS]

I lubricated the keys with best lubricant I could get that is safe for plastics and rubber -- Dupont Krytox.

I consder KG8 as dry PTFE lubricant as it is PTFE powder + volatile solvent. I put the disassembled the sliders in a plastic bag and poured some KG8 so that the powder is applied. As noted, the lubrication effect didn't last long although the performance was pretty impressive. Please note that the solvent used in KG8 melts ABS plastic which is used in housing of HHKB.

Have you received your RO-59? It kicks ass.

[limmy]

Have you received your RO-59? It kicks ass.

Yes, I received it a while ago, but didn't use it until yesterday. Yesterday I opened up my HHKB and finally applied RO-59. The stuff is pretty slippery to touch. One thing I like about RO-59 is that it is water based, so I do not have to worry about over applying as it will dry eventually. When I opened up my Krytox lubed HHKB, I noticed there were small puddle of Krytox from over applying. Oh well.. I shouldn't have poured the lube on the sliders hoping for the best. Moreover, I should have learned from previous lesson that I should disassemble completely in order to lubricate correctly.

I painted Krytox over dried RO-59. I like it so far. Krytox tend to bead up a little on plastic surface, but when applied over RO-59 it disperses nicely. I guess RO-59 make the applied surface to have higher surface energy.

[limmy]

I did the same test again after properly lubing the switches with RO-59 + Krytox GPL100 after disassembling. I did three things differently from previous lubrication with Krytox :
1. used RO-59 before applying Krytox (applied two coats of RO-59. I waited about 10 minutes for the first coat to dry before applying second coat. I let it cure overnight after the second coat.)
2. lubed the little groves where two feets of the sliders go in
3. avoided over applying the lubricant by using thin flat brush

Here are the results along with the numbers from the previous test:
60g actuates 10 keys (0 in previous test)
62.5g actuates 14 keys (6 in previous test)
65g actuates 25 keys (19 in previous test)
67.5g actuates all of the tested keys (same result as the previous test)

Since I used both Krytox GPL100 and RO-59, I cannot tell which of the lubricant works better, but I conjecture using them together creates synergy effect. The keys are lubricated just yesterday, so I think the lubricant performance is at the maximum right now. I will update how the numbers change over time as I use them.

Today, I learned doing things properly takes time and effort. (This is what ripster do from time to time. Today I learned stuff.)

Here are my measurements of my 5-year old HHKB which was used daily since I bought them. I measured only the letter part of the keyboard. That is..
QWERTYUIOP[]
ASDFGHJKL;'
ZXCVBNM,./

Most of my keys support 60g without actuating.
When I put 62.5g, only 6 keys actuate. The keys that actuated at 62.5g were mostly non-alphabet keys.
Putting on 65g actuate 19 keys out of 33 keys that I test.
67.5g actuate all of the 33 tested keys.

[Jim66]

So... So far we can tentatively suggest from this thread that...

1). There appears to be quite a substantial variance in the force required to actuate a topre switch. What really bothers me is the number of HHKBs actuating at almost 70g?
2). There is to be some kind of positive relationship between age of keyboard and actuation force (I'm not convinced this would be stat sig though). This effect appears to result from some combination of; a) aging rubber and, b) drying out of slider lubricant. It would nice to collect lots of data (e.g., hours of use, hours of storage/rest, amount of lube present [a difficult one]) and conduct a multiple regression analysis to find the main predictor variable (assuming there is one).

EDIT: This is always going to be a contentious issue due to the expensive nature of topre boards; no one wants to spend 400 on a type-s to find out that it is going to feel significantly worse in a few years. Topre fans (including myself) don't really want to hear this! You might get the same type of effect with Cherry or even ALPS switches, but the main difference here is you can easily replace the offending part in the switch. Their slightly more 'mechanical nature' nature makes them easier to maintain; it's slightly more difficult to replace the rubber on a topre board!

I already own two realforce boards, I'm a massive fan of ergonomic boards and I was thinking about buying the uTron at some point; is this putting me off? As I said, I'm skeptical that any effect would be statistically significant; purely because of the amount of unsystematic variance in the initial actuation ratings. So in this sense, no, I don't think is it. It is a niggling feeling though when you are spending 700 + USD on a board.

Dare I say it, Ripster is right; McRip effect.

[limmy]

So... So far we can tentatively suggest from this thread that...
1). There appears to be quite a substantial variance in the force required to actuate a topre switch. What really bothers me is the number of HHKBs actuating at almost 70g?
2). There is to be some kind of positive relationship between age of keyboard and actuation force (I'm not convinced this would be stat sig though). This effect appears to result from some combination of; a) aging rubber and, b) drying out of slider lubricant. It would nice to collect lots of data (e.g., hours of use, hours of storage/rest, amount of lube present [a difficult one]) and conduct a multiple regression analysis to find the main predictor variable (assuming there is one).

Hi, welcome to the discussion.

Here are my comments on the two points you raised.

1) How many times do you see old HHKB or Realforce? Scarcity of the boards could be suggesting that they wear out and become not attractive. (I am conjecturing here. You might as argue that back than Topre boards were not as popular.) But from my experience, 5 years of daily use is probably enough for a Topre board to wear out. Topre says that their switches can handle 30 million actuation. Hence, it will wear out when it does, it wears out on multiple parts so that in order to restore the feel of a new board replacing one or two parts will not be enough. I will discuss this more in the following point. However, Cherry MX switches are better in this sense as you mentioned.

I would think there is very little incentive to publish negative information on a product that cost $300, especially when it affects resale value. I might be an odd ball there as I have no plans to resell my boards.

All I can say from my comparison is that my 5-year-old HHKB is significantly tiring when compared to like-new HHKB. This is only one observation and as you have questioned, it may not represent the true distribution of the 5-year-old HHKB. However, there is no other observation available, at least I couldn't find any. Combined with the incentive to conceal negative information, I think it is understandable that there is no such information. Unless there are fresh observations that says Topre boards are just fine after 5 years of daily use, I am lead to believe from my ONE observation that Topre boards will develop friction and rubbers will stiffen.

2) When a Topre board gets old, there are multiple parts that age as you mentioned plus some more.
a) rubber loose elasticity
b) sliders get scratched
c) slider housing is scratched
So, even if you get a hold of replacement part for aging rubber sheet, you may not be able to restore the feeling of new Topre board. I think it is pretty similar to car engines. Mercedes(high-end and expensive supposedly more durable) gets old and wears out just like other cars, and even if you replace certain components of the engine the it will not feel same as that of a new Mercedes.

Dare I say it, Ripster is right; McRip effect.

The McRip effect is so vague and subjective. Quoting ripster,
"Some hobbyist forum complaints are mere illusions and MANY are way overblown!"
this is definition of what he calls McRip effect.

My HHKB actuates at on average 64g. You might or want to call this overblown or illusion. But to me this is pretty solid. And deviation from original designed actuation force of 45 to me is huge(42% more!!). Elitekeyboards came along a couple of days ago and said they are within range of +/- 15g. Topre is selling 55g boards, and if 45g boards can actually be normal when they actuate at 60g, I am saying something is seriously wrong with their standards.

Also, minor (or sometimes undetectable by normal people or non-experts) difference are important for people who care. Professional violinists using only Stradivarius might as well be called McRip effect. What SEEMS to be an illusion for a normal people is NOT an illusion for people who can discern differences and have preference over this little difference. I would respect their choices rather than calling it some effect. Of course not all violinist need Stradivarius, so ripster's point is valid in some sense, but I think ripster's application of McRip effect is too much of a generalization.

[limmy]

http://www.guardian.co.uk/music/musicblog/2012/jan/03/stradivarius-v-modern-violins-study

my iPad will never get stiff

Your iPad will be outdated and not be used in two years, just like an old laptop.

I guess there were a lot of ripsters out there doing blind tests on Stradivarius. What I meant to convey from the Stradivarius example was that for non-experts it is hard to discern difference between 10K and 1K violins, while experts can spot the differences. I guess there are more than a couple studies on this issue and Stradivarius is not significantly preferred to modern ones. However, the modern ones are still expensive and top-of-the-line, around 30K.

Edit: I said compared modern violins cost around 10K, but I was mistaken. Three modern violins used for comparison seem to cost around 100K according to this article

The three old instruments had a combined value of $10m, a hundred times that of the modern violins.

[itlnstln]

Of course I was a moron when I said my HHKB felt heavy.  

Whatever.

Don't pay any attention to the horse with more **** than brains.

[limmy]

Of course I was a moron when I said my HHKB felt heavy.  

Whatever.

Don't pay any attention to the horse with more **** than brains.

Are you still using the HHKB? Would you provide actuation force measurements?

From about three years of use, I had a feeling that my HHKB was not performing as well as it was new, but at the moment I didn't have a reference point to compare the old HHKB. It was about that time when I began researching lubricants for HHKB. Now that I have another HHKB, I compare them side by side and I can certainly say used HHKB feels different when compared to like-new HHKB. Old one is way more tiring than the new one. The feeling of tiredness seems to be coming from increased friction as well as stiffening rubber domes.

[itlnstln]

Unfortunately, I sold my HHKB.  I still have my RF, however.  Variation doesn't surprise me, though. Your dealing with rubber and plastic; there's always going to be some variation.

[Jim66]

My HHKB actuates at on average 64g. You might or want to call this overblown or illusion. But to me this is pretty solid. And deviation from original designed actuation force of 45 to me is huge(42% more!!). Elitekeyboards came along a couple of days ago and said they are within range of +/- 15g. Topre is selling 55g boards, and if 45g boards can actually be normal when they actuate at 60g, I am saying something is seriously wrong with their standards.

I completely agree with you, a deviation of +/- 15g is half of the force required to actuate one of the 30g pinky switches!

"Sorry Sir, you ordered an all 45g board and you think you have a variable weighted board? Never mind Sir, I'll let you in on a little secret; due to the way the switches are made, there is a chance that every board we make is variable weighted... lolololol."

Unless there are fresh observations that says Topre boards are just fine after 5 years of daily use, I am lead to believe from my ONE observation that Topre boards will develop friction and rubbers will stiffen.

Ouch, that's a dangerous game you're playing. What happens if you are the freakishly bad board? What happens if the fresh HHKBs after five years are the freakishly good boards? Who is right? Topre switches may feel very different after 5 years use, but (and it is a crucial but); the point I'm trying to make is... It is highly unlikely to be a real effect in the population! Due to the seemingly lax tolerances in the switches (see above), this or any other scientific test of difference (see below):

[ Attachment Invalid Or Does Not Exist ] 44986[/ATTACH]

Is higher unlikely to be able to detect any 'real' difference; it would be unlikely to matter how many HHKBs you tested, there would be too much random variance floating around. Any difference which you tested would always be the chance finding (because, there isn't a real effect in the population)! So in this sense, yes, there may be a difference, but, perhaps it wouldn't be big enough to be mathematically or rather statistically, 'real'.

It's that black or white (for me at least).

[limmy]

Ouch, that's a dangerous game your playing. What happens if you are the freakishly bad board? What happens if the individual with the the fresh HHKB after five years is the freakishly good board? Who is right? Topre switches may feel very different after 5 years use, but (and it is a crucial but); the point I'm trying to make is... It is highly unlikely to be a real effect in the population! Due to the seemingly lax tolerances in the switches (see above), this or any other scientific test of difference (see below):

(Attachment Link) 44986[/ATTACH]

Is higher unlikely to be able to detect any 'real' difference; it would be unlikely to matter how many HHKBs you tested, there would be too much random variance floating around. Any difference which you tested would always be the chance finding (because, there isn't a real effect in the population)! So in this sense, yes, there may be a difference, but, perhaps it wouldn't be big enough to be mathematically or rather statistically, 'real'.

It's that black or white (for me at least).

What I am doing here is similar to what statisticians call Bayesian updating. One observation certainly carries a lot of weight when there is only one observation available. Some statisticians hate the idea, but I think it makes more than enough sense in real life situations.

As you noted, it is impossible to control for everything and get precise testing if you narrow down your hypothesis. But if you are fine with lax hypothesis, I am guessing you could get some results. It would be nice to control for everything, but if it is impossible to do so, you could at least say somethings although not sharp. For instance, I could make my hypothesis dull and test the following. Does Topre board change actuation force over time on average? (a board may be used or not, kept in cool and dry place or not)

I think the biggest problem here is that there is simply no sample available. Even if observations are available, due to skewed incentives to conceal negative information the observations will be biased. However, the bias problem is secondary. Putting in reasonable assumptions on the initial actuation force, I think you could get some statistical results. The reasonable assumption would be that on average Topre get their actuation force correct with some variance. This would complicate the t-test a bit (since you didn't observe the initial state), but even if you use standard t-test of difference from null (that is, under null HHKB actuates at 45g for sure) it would be meaningful. You could check robustness using simulated observations drawn from a distribution, but this is already getting pretty complicated..

You are saying there is too much variation, but without proper samples of measurements, how would you assess that such large variation exists? How can you distinguish  cross-sectional variance(variance across HHKB produced at one time) from time-series variance(variances across HHKB with different manufacturing dates)?

I am guessing you would know any statistical or mathematical results are coming from assumptions or axioms. So, my putting in assumptions will not be so problematic, although you may disagree on specifics of my assumptions.

I disagree more on one particular point you raised:

it would be unlikely to matter how many HHKBs you tested, there would be too much random variance floating around.

If large enough observations are made, random variances will be canceled out and you will be able to figure out population parameter. See law of large numbers. Of course some weak assumptions has to be made for what I just said to be true.

Edit:
Wikipedia listing of law of large numbers is dumb-down version of the theorem, and less general than most general version of the theorem. I am adding this note in case you argue that the conditions of the theorem is not satisfied. The conditions of most general version of the theorem are so easy to satisfy that it would be note worthily special if they are not satisfied. Of course this is theory and mileage from actual application may vary case by case, but I am mentioning this since you raised a theoretical point. Personally, I think the variations you mention here is random and independent enough that the theorem works just fine.

[fossala]

My 87U is still 50g this morning.

How old is the model? how much time have you done on it?

[Aranair]

and my 2.5 week old hhkb is still 60g odd on the 37 keys

[Aranair]

i wish the radiation softened the rubber instead.

[Aranair]

well /flails arm around. maybe god is telling me to use it to train my fingers.

[limmy]

Two years old.  I'd say a good 4 months of constant use.  I can tell because the spacebar is NICE AND ****ING SHINY!

I like Cherry MX better.

I'm almost certain that it is identical to the day I took it out of the box.  It's just my fingers take a week to fully adjust.

The McRip Effect.

You seem to be a victim of McRip effect.

Your perception of Topre boards will not change its feel over time or your eagerness to say some findings are McRip effect is blinding yourself from hard fact from your own observation that the switches actuated at 52g while they could support 50g without actuating. As I noted numerous times, your dithering method is under measuring the actuation force. If certain keys are sticky, they require more force to actuate. You can not just ignore the stickiness by tapping your fingers three times. I personally think sticky keys are one of the major reasons why a certain keyboard is tiring.

Perhaps you are suffering from Golden McRip effect syndrome(aka Maslow's Hammer, Golden hammer, The Law Of The McRip Effect, Ripster's McRip Effect).

Also, it is funny that you can be almost certain that the switches are identical to the day you took it out of the box. Any comparison over time relying on sensory memory is no where near accurate. I would be skeptical of a comparison relying on memory if comparison was done couple of hours apart, but two years?? ripster, come on...

[RiGS]

The McRip Effect is pseudoscience and useless like most of ripster's other theories. The guy usually doesn't care about accuracy, and he prefers to ignore valid arguments, because he can't prove them wrong, so instead he will just overwhelmingly repeat his self-advertising propaganda relying on your ignorance while addressing ad hominem attacks against the messenger. This is his secret sure-fire way to gain him fame among noobs and ignorant geekhackers on the intertubes.
Repetition is the key here, as his primary target is to contaminate new geekhackers mind with his false ideas, while making the counter-arguments disappear due to the artificial noise he created in the process. He has been doing this for so long, hence the over 54k posts.
The worst part is that his strategy works very well, since most knowledgeable senior geekhackers neither care or have the time, tenacity and diligence to correct him all the time, and report him for spamming. All the noob votes in his favour in the rank #1 poll just prove my point even further.
This is the exact attitude that renders most of his content dangerously misleading and wrong. I warned you.

Edit: If you argree with this feel free to add a link to this post in your sig.

[Aranair]

i don't know if my data will be of any use but I was bored and decided to do a more systematic test like limmy's a few posts back.

(test only involves 33 keys in the middle, and for future references, my hhkb is manufactured in 2011/nov, and date today is 2012/march: actual use is only 2 weeks though)

at 45g: 3 keys
at 52.5g, 4 additional (7)
at 55g, 2 additional (9)
at 57.5g, 14 additional  (23)
at 60g, 10 additional keys (33)

Also, interestingly, for some of the keys, if i leave the coins there and just wait for a few more seconds (without tapping on the table), they slowly collapse and actuates. It almost feels like the coins are JUST enough, so it takes awhile for the rubber domes to collapse or it might just be overcoming the friction, no idea.

In comparison, on my Filco cherry browns, almost every key actuates at 45g or lower and very selected few at 50g.

[limmy]

The McRip Effect is pseudoscience and useless like most of ripster's other theories. The guy usually doesn't care about accuracy, and he prefers to ignore valid arguments, because he can't prove them wrong, so instead he will just overwhelmingly repeat his self-advertising propaganda relying on your ignorance while addressing ad hominem attacks against the messenger. This is his secret sure-fire way to gain him fame among noobs and ignorant geekhackers on the intertubes.
Repetition is the key here, as his primary target is to contaminate new geekhackers mind with his false ideas, while making the counter-arguments disappear due to the artificial noise he created in the process. He has been doing this for so long, hence the over 54k posts.
The worst part is that his strategy works very well, since most knowledgeable senior geekhackers neither care or have the time, tenacity and diligence to correct him all the time, and report him for spamming. All the noob votes in his favour in the rank #1 poll just prove my point even further.
This is the exact attitude that renders most of his content dangerously misleading and wrong. I warned you.

Edit: If you argree with this feel free to add a link to this post in your sig.

I think McRip Effect has own merit although it is effective to only part of the population. If I understood McRip Effect correctly it is about focusing on little effects and differences and making it a big deal. Yes, little differences may not really matter to non-enthusiasts, but for some people even smallest difference is important. I recently bought a flashlight from ebay for 8 bucks and didn't really care about different mods what kind of LED used, but there seem to people paying attention to littlest details. I do not consider myself as a flashlight enthusiast and pretty much everything that flashlight enthusiasts say wouldn't really be persuasive unless I dig deeper and become one.

ripster is sometimes silent about certain things that I point out, e.g. Golden McRip Effect syndrome. I think he agrees to some degree when he is silent. Or maybe he didn't really understand or read what I was saying...

I understand your discontent RiGS, but I must admit for new comers ripster's information and documentation is very helpful. Also, he likes to document and write stuff and that is helpful for many people. I don't really value and understand his jokes, but I would accept them as one of side effects of the benefits ripster provide to this community.

[limmy]

i don't know if my data will be of any use but I was bored and decided to do a more systematic test like limmy's a few posts back.

(test only involves 33 keys in the middle, and for future references, my hhkb is manufactured in 2011/nov, and date today is 2012/march: actual use is only 2 weeks though)

at 45g: 3 keys
at 52.5g, 4 additional (7)
at 55g, 2 additional (9)
at 57.5g, 14 additional  (23)
at 60g, 10 additional keys (33)

Also, interestingly, for some of the keys, if i leave the coins there and just wait for a few more seconds (without tapping on the table), they slowly collapse and actuates. It almost feels like the coins are JUST enough, so it takes awhile for the rubber domes to collapse or it might just be overcoming the friction, no idea.

In comparison, on my Filco cherry browns, almost every key actuates at 45g or lower and very selected few at 50g.

Thanks for the data! It is not as bad as I thought was.

My like-new HHKB produced in 2009 has following results.

Code: [Select]

50g   - 2       -           []
52.5g - 3(+1)   -           []           '            
55g   - 9(+6)   -   ert     []  s        '   c  n      
57.5g - 28(+19) - qwertyu   [] asdfghjkl;' zxcvbnm,  
60g   - 33(+5)  - qwertyuiop[] asdfghjkl;' zxcvbnm,./
Average actuation force of the tested keys is 56.81g. Comparing it with your results, I think it is pretty similar with my like-new HHKB, which was produced 3 years ago (manufacturing date was 2009-02).

My five-year-old HHKB's actuation force results are..

Code: [Select]

55g   -  1      -                        '
57.5g -  3(+2)  -           [            '       m
60g   -  7(+4)  -           []        k  '      nm,
62.5g - 12(+5)  -           []      hjk ;'      nm,./
65g   - 26(+14) - qw rtyuiop[] asd ghjkl;'      nm,./
67.5g - 33(+7)  - qwertyuiop[] asdfghjkl;' zxcvbnm,./

Average actuation force of tested keys are 63.78. Difference between the averages is 7g. Which is not huge in number, but it is significant enough for me that I find the new one to be much more comfortable to type on. The old one give much more tactile feed back, however. I have been using HHKB as my daily driver for 5 years, so I am pretty used to type on these. However, the increased required force is bad enough that I get tired after for a prolonged use. My fingers didn't get any stronger it seems..

[Aranair]

wow,  and there I was thinking my set was behaving like a 5 year old. I was hoping that some day I'll find a "normal" HHKB after having sold that. Guess thats not happening, is it? D:

And yeah, I would agree that 7g definitely makes a difference to some people (me inclusive)... probably more so after spending entire days on the keyboards. Fatigue does build up imo..

[limmy]

I will attempt to RipOmeter my iPad when I get a chance.

My guesstimate is 10 g.

Sent From Brother Ripster's iPad

Your ripometer has to be capacitive to work. It will require no force just need to be close to the screen. Try it with your gold ring.

You need to know theory before using empirical methods.

[fossala]

Somebody willing to trade a White Switch?
http://geekhack.org/showthread.php?22259-Ripsters-Novelty-Key-Trade-Thread

I have a white switch if you want one.

[limmy]

The Golden McRip Effect?  What?  I just post the link and it has lots of pics backing up my opinion -The McRip Effect.  If there is something wrong in the theorem post it in the theorem thread or ignore the link.  It's like me saying I like Filcos.  Not everybody HAS to like Filcos.

Okay. I am more sympathetic(?) if calling something McRip Effect is expression of your opinion. I read the post and I commented on it.

The Golden McRip Effect is the term that I made up inspired by Maslow's hammer or law of the instrument which you introduced to me the other day. You said I was suffering from law of the instrument, but didn't tell me in what context. I am kidding for the most part though. I am not a psychology guy.

Also, you are saying that "it is still 50g". What does that mean? Do all of your keys support 50g without being depressed? Or does it mean that you put 50g and tap your keyboard tray 3 times with your finger then the weight drops? Some of keys on my like-new HHKB can support 57.5g without actuation, don't you have any of those switches? I appreciate your measurements and participation in this discussion.

[limmy]

I thought yo understood what I was saying when I said that one needs to know theory behind in order to make correct use of empirical methodology?

iPad has capacitative screen and it doesn't require force to register. So, using coins to measure something doesn't mean much. Try all you can trying to press the screen with wooden stick, it will not register. I believe it is the amount of finger surface(or any other material that is conducive) that is touching the screen. that is at least how my phone works.

[SidusNare]

iPad has capacitative screen

or any other material that is conducive

You need a material that is capacitive, like this --> http://bit.ly/GSJBFM . Conductive wont get you there unless there is something capacitive behind it.

[limmy]

I think it is about time to summarize the measurements and findings so far.

My hypothesis is Topre keys stiffen over time and use.

I measured actuation forces of 33 keys "QWERTYUIOP[]ASDFGHJKL;'ZXCVBNM,./" on my HHKB.
Here is how you read my results:
weight tested - number of keys actuated(number of additional keys actuated - list of keys actuated

Results from 5-year-old(manufactured on Aug 2006) HHKB used for 5 years. Lubricated using RO-59 and Krytox GPL100.

Code: [Select]

55g   -  1      -                        '
57.5g -  3(+2)  -           [            '       m
60g   -  7(+4)  -           []        k  '      nm,
62.5g - 12(+5)  -           []      hjk ;'      nm,./
65g   - 26(+14) - qw rtyuiop[] asd ghjkl;'      nm,./
67.5g - 33(+7)  - qwertyuiop[] asdfghjkl;' zxcvbnm,./
Average actuation force 63.78g

Results from three-year-old(manufactured on Feb 2009 HHKB) in like new condition:

Code: [Select]

50g   - 2       -           []
52.5g - 3(+1)   -           []           '            
55g   - 9(+6)   -   ert     []  s        '   c  n      
57.5g - 28(+19) - qwertyu   [] asdfghjkl;' zxcvbnm,  
60g   - 33(+5)  - qwertyuiop[] asdfghjkl;' zxcvbnm,./
Average actuation force is 56.81g

This is the weight system I used.
[ Attachment Invalid Or Does Not Exist ] 45592[/ATTACH] [ Attachment Invalid Or Does Not Exist ] 45593[/ATTACH]
Whenever I increase the weight by 2.5g(this is how much each US penny weighs), I test the 33 keys. I put the weight two to three times on different points of key face if it doesn't drop in the first try. So this measurement is somewhat downward biased, but I chose to introduce such small bias in exchange of reducing uncertainties on the measurement. Also this downward bias work against my hypothesis of stiffening of keys.

Aranair was kind enough to post his measurment on his new HHKB (produced on Nov 2011): source

Code: [Select]

45g   - 3
52.5g - 7(+4)
55g   - 9(+2)
57.5g - 23(+14)
60g   - 33(+10)
Average actuation force is 56.36g. This is pretty similar to my like-new HHKB.


Ripster provided some measurements, but I am not really sure what he means by measurement of 50g. I initially thought he meant putting 50g on a key would actuate the key, but his own picture here shows otherwise. All theee keys in the picture is not actuated when he puts on 50g.

This is his measurement on 14 Mar 2012, which shows 50g was NOT enough to actuate but 52g was enough.
[ Attachment Invalid Or Does Not Exist ] 45594[/ATTACH]
measurements and photos are from here

Although Ripster keeps saying that
1. he doesn't feel anything changed over time for this Realforce 87U keyboard
2. he gets same measurement
this is in contrast to his own finding. When a self-proclaimed scientist relies on feeling rather than measurements and even misinterpret measurements, I say something is fishy.


There were only three people who participated and actually post their measurements. I obtained full measurements of 3 HHKBs(one heavily used and two like-new), one suspicious measurement from ripster. This lack of providing measurement could be due to incentive for users conceal negative information on the products they use or love. They are incentivised to act that way for possibly two reasons. 1. they fear the resale value is negatively affected, 2. they don't like to hear bad things about what they love. However, I cannot really rule out that they don't really care.


So far, from what I observe, I tentatively conclude that  
1. using HHKB for a while will stiffen the switches.
2. storing it in cool and dry place and not using it will not change key feel that much.
3. Topre's specification of 45g +/- 15g generally works but we only see measurement towards +15g. (This is shame on Topre, because they are selling 45g and 55g boards separately, but it turns out 45g board will be likely to actuate close to 55g range.)

This conclusion is from a small number of observations which may or may not represent true population of HHKB or Topre switches. If you have any other information that says otherwise, I am open to revise my conclusion and say I was unlucky one.

[Aranair]

eh, ripster you actually measured the topres with ripometer on the day you got it?!

[hasu]

This is measurement of my moded HHKB.  It is given variable rubber sheet, silencing moded and lubed with RO-59.
Backslash, backquote and Fn key still have HHKB rubber cup while other alpha and numeric key part have variable rubber sheet.

I'll measure my other HHKBs as well later if I have time.

Code: [Select]


Weights to register(actuate) 2012/03/23
HHKB pro(PD-KB300 2003-05) with variable weighted rubber sheet from Realforce 101(2007 maybe), silencing mod and lubricated by RO-59.
,-----------------------------------------------------------.
|43 |34 |40 |47 |45 |48 |50 |49 |48 |43 |35 |35 |36 |55 |55 |
|-----------------------------------------------------------|
|47   |33 |41 |51 |49 |50 |49 |50 |51 |43 |36 |36 |37 |44   |
|-----------------------------------------------------------|
|49    |34 |41 |50 |51 |52 |51 |51 |50 |44 |37 |37 |41      |
|-----------------------------------------------------------|
|48      |35 |42 |50 |53 |52 |51 |51 |50 |45 |36 |45    |61 |
`-----------------------------------------------------------'
      |54 |50   |        40             |50   |58 |
      `-------------------------------------------'


[limmy]

This is measurement of my moded HHKB.  It is given variable rubber sheet, silencing moded and lubed with RO-59.
Backslash, backquote and Fn key still have HHKB rubber cup while other alpha and numeric key part have variable rubber sheet.

I'll measure my other HHKBs as well later if I have time.

Code: [Select]

Weights to register(actuate) 2012/03/23
HHKB pro(PD-KB300 2003-05) with variable weighted rubber sheet from Realforce 101(2007 maybe), silencing mod and lubricated by RO-59.
,-----------------------------------------------------------.
|43 |34 |40 |47 |45 |48 |50 |49 |48 |43 |35 |35 |36 |55 |55 |
|-----------------------------------------------------------|
|47   |33 |41 |51 |49 |50 |49 |50 |51 |43 |36 |36 |37 |44   |
|-----------------------------------------------------------|
|49    |34 |41 |50 |51 |52 |51 |51 |50 |44 |37 |37 |41      |
|-----------------------------------------------------------|
|48      |35 |42 |50 |53 |52 |51 |51 |50 |45 |36 |45    |61 |
`-----------------------------------------------------------'
      |54 |50   |        40             |50   |58 |
      `-------------------------------------------'


Thanks for the measurement, hasu. This is pretty impressive and thorough.

I didn't have serious problems on switches that uses a single rubber dome, such as control, shift, tab, etc. That is why I focused on the middle part, where two large pieces of rubber sheet is used.

There is one possible factor that could have negatively affected(reduced) the actuation force, namely the silencing mod. In my old HHKB, if I depress the switch slightly(not pressing anymore but let my finger stay in place) and then put on weights, the measured actuation force tend to be smaller than what I measured before. I am not sure how significant of effect this mod created, because I do not have measurements on how much key travel distance is reduced by the mod, but I would like to try the mod sometime in an effort to reduce the actuation force.

So it seems from measurements from hasu and ripster that Realforce variables are likely to be within spec after some time.

This is pretty similar to mine.  The gradient decline from 45/35g is a gradual one.

Details in the Topre Wiki.

I also would suspect RO-59 doesn't affect RipOmeter measurements all that much.

Yes, I saw that information on the wiki, how the actuation force declines gradually on 35/45g border. However, when I searched for the basis of the claim, I didn't really locate actual measurement, but only one post by wellington1869 that simply stated without mentioning any measurement that it is 40g (see here) I might have missed the discussion somewhere else, but any claim has to be backed up by measurement and proper sources.

However, looking at the same result, I have a different opinion on RO-59. I think it preserves the original smoothness over time. As I noted before, the sliders when unlubed get scratched. So, I conjecture that if the sliders are properly lubed, it may extend its life considerably. (I heard from a mechanical engineer that anything that moves it is critical that proper lubrication is present.)

[hasu]

This is result of my another moded HHKB pro(+silencing mod + RO-59 but original rubber sheet) and two HHKB pro2s without mod.

Code: [Select]


Weights to register(actuate) 2012/03/26
HHKB pro(PD-KB300 2004-11) with silencing mod and lubricated by RO-59.
,-----------------------------------------------------------.
|48 |52 |54 |53 |53 |55 |52 |56 |58 |54 |52 |52 |49 |53 |53 |
|-----------------------------------------------------------|
|50   |53 |54 |56 |54 |55 |54 |53 |53 |52 |50 |52 |53 |46   |
|-----------------------------------------------------------|
|52    |53 |52 |54 |54 |52 |53 |47 |52 |51 |49 |46 |47      |
|-----------------------------------------------------------|
|50      |51 |52 |54 |54 |53 |53 |53 |51 |53 |51 |52    |53 |
`-----------------------------------------------------------'
      |52 |52   |        59/49 *1       |45   |51 |
      `-------------------------------------------'
*1 with/without spring under space bar


Code: [Select]


Weights to register(actuate) 2012/03/26
HHKB pro2(PD-KB400W 2007-10) I bought this used HHKB for spare and I myself hardly use it.
,-----------------------------------------------------------.
|48 |51 |52 |51 |49 |50 |56 |57 |57 |58 |56 |54 |52 |56 |51 |
|-----------------------------------------------------------|
|50   |55 |54 |54 |55 |54 |57 |57 |60 |60 |58 |51 |53 |52   |
|-----------------------------------------------------------|
|48    |56 |56 |58 |56 |54 |53 |55 |56 |57 |56 |57 |43      |
|-----------------------------------------------------------|
|50      |57 |59 |58 |56 |58 |54 |57 |55 |53 |55 |50    |53 |
`-----------------------------------------------------------'
      |47 |55   |        58/47 *1       |47   |47 |
      `-------------------------------------------'
*1 with/without spring under space bar

 

Code: [Select]


Weights to register(actuate) 2012/03/26
HHKB pro2(PD-KB400W 2010-08) I bought this used HHKB for spare and I myself hardly use it.
,-----------------------------------------------------------.
|39 |46 |49 |53 |57 |54 |54 |53 |54 |54 |53 |49 |47 |47 |45 |
|-----------------------------------------------------------|
|46   |50 |58 |58 |61 |61 |59 |56 |61 |59 |58 |47 |49 |46   |
|-----------------------------------------------------------|
|48    |54 |58 |63 |62 |61 |54 |56 |60 |59 |59 |45 |49      |
|-----------------------------------------------------------|
|42      |56 |57 |55 |56 |56 |56 |56 |57 |62 |60 |49    |46 |
`-----------------------------------------------------------'
      |44 |42   |        53/44 *1       |47   |45 |
      `-------------------------------------------'
*1 with/without spring under space bar


[Aranair]

lol. man, the variance is amazing. I would think they would have it under more control at those kind of prices.

[limmy]

Upon cursory reviewing of the new data, there seems to be quite a variance.
Also, I notice that there seems to be no clear correlation between the age of the board and the actuation force.

There are several things that stand out from measurements from 4 HHKBs that hasu posted:
1. lubed and silence-mod boards tend to have lower actuation force
2. letter keys tend to have bigger actuation force (this is what I feel from my 2 HHKBs too)
3. the space bar spring adds about 10g to the actuation force

I will measure my HHKB again so that it would have similar format as hasu's measurement and complete full measurements of 6 HHKB. (4 from Hasu, and 2 from me)

Most of the boards are used, 5 out of 7 HHKB we saw so far were used. So, it is hard to tell whether the variance we see is out of the factory variance or the variance has built up due to usage. Also, it seems to me that how it is taken care of is an important factor, but it would be very hard to draw conclusions from them. I think lubing and protecting the sliders are important, but that is my conjecture so far and only backed up by a couple of cases.

[fossala]

My HHKB Type-S is nearly new. I only have 1p coins (3.56g) to measure with. Every key is weighing the same 56.96g (16 coins). I would like to find a better way of measuring this as I seem to have one of the only boards that measures the same througout.
I type quite a lot on this board and will update over the months how it wears.

[limmy]

I have measured my HHKB again. My measurement have gone down a little bit, possibly due to tilting the board so that the face of the keys are parallel to my desk. Along with raw data, I drew a chart in Excel for visual representation of the numbers.

Code: [Select]

HHKB pro2 (2009-02) in like-new condition. (03/26/2012)
,-----------------------------------------------------------.
|50 |53 |53 |50 |48 |50 |53 |55 |53 |53 |55 |50 |53 |48 |48 |
|-----------------------------------------------------------|
|43   |55 |55 |55 |53 |55 |53 |53 |58 |58 |60 |48 |45 |48   |
|-----------------------------------------------------------|
|48    |55 |55 |58 |58 |58 |55 |55 |58 |55 |55 |48 |50      |
|-----------------------------------------------------------|
|50      |55 |58 |53 |55 |55 |53 |53 |55 |55 |58 |53    |50 |
`-----------------------------------------------------------'
      |48 |50   |        53             |50   |48 |
      `-------------------------------------------'
NOTE: 43=42.5, 48=47.5, 53=52.5, 58=57.5


Code: [Select]

HHKB pro2 (2006-08) used for 5 years daily. (03/26/2012)
,-----------------------------------------------------------.
|48 |58 |60 |63 |60 |60 |63 |63 |63 |60 |63 |48 |50 |48 |48 |
|-----------------------------------------------------------|
|48   |63 |65 |65 |63 |63 |60 |63 |63 |63 |60 |58 |58 |50   |
|-----------------------------------------------------------|
|55    |60 |63 |65 |65 |65 |58 |60 |58 |58 |58 |53 |50      |
|-----------------------------------------------------------|
|53      |65 |68 |63 |65 |65 |55 |55 |60 |63 |63 |53    |55 |
`-----------------------------------------------------------'
      |55 |55   |        60             |55   |50 |
      `-------------------------------------------'
NOTE: 43=42.5, 48=47.5, 53=52.5, 58=57.5, 63=62.5, 68=67.5

source

Code: [Select]

Weights to register(actuate) 2012/03/23
HHKB pro(PD-KB300 2003-05) with variable weighted rubber sheet from Realforce 101(2007 maybe), silencing mod and lubricated by RO-59.
,-----------------------------------------------------------.
|43 |34 |40 |47 |45 |48 |50 |49 |48 |43 |35 |35 |36 |55 |55 |
|-----------------------------------------------------------|
|47   |33 |41 |51 |49 |50 |49 |50 |51 |43 |36 |36 |37 |44   |
|-----------------------------------------------------------|
|49    |34 |41 |50 |51 |52 |51 |51 |50 |44 |37 |37 |41      |
|-----------------------------------------------------------|
|48      |35 |42 |50 |53 |52 |51 |51 |50 |45 |36 |45    |61 |
`-----------------------------------------------------------'
      |54 |50   |        40             |50   |58 |
      `-------------------------------------------'

source

Code: [Select]

Weights to register(actuate) 03/26/2012
HHKB pro(PD-KB300 2004-11) with silencing mod and lubricated by RO-59.
,-----------------------------------------------------------.
|48 |52 |54 |53 |53 |55 |52 |56 |58 |54 |52 |52 |49 |53 |53 |
|-----------------------------------------------------------|
|50   |53 |54 |56 |54 |55 |54 |53 |53 |52 |50 |52 |53 |46   |
|-----------------------------------------------------------|
|52    |53 |52 |54 |54 |52 |53 |47 |52 |51 |49 |46 |47      |
|-----------------------------------------------------------|
|50      |51 |52 |54 |54 |53 |53 |53 |51 |53 |51 |52    |53 |
`-----------------------------------------------------------'
      |52 |52   |        59/49 *1       |45   |51 |
      `-------------------------------------------'
*1 with/without spring under space bar

source

Code: [Select]

Weights to register(actuate) 03/26/2012
HHKB pro2(PD-KB400W 2007-10) purchased used, but not used recently
,-----------------------------------------------------------.
|48 |51 |52 |51 |49 |50 |56 |57 |57 |58 |56 |54 |52 |56 |51 |
|-----------------------------------------------------------|
|50   |55 |54 |54 |55 |54 |57 |57 |60 |60 |58 |51 |53 |52   |
|-----------------------------------------------------------|
|48    |56 |56 |58 |56 |54 |53 |55 |56 |57 |56 |57 |43      |
|-----------------------------------------------------------|
|50      |57 |59 |58 |56 |58 |54 |57 |55 |53 |55 |50    |53 |
`-----------------------------------------------------------'
      |47 |55   |        58/47 *1       |47   |47 |
      `-------------------------------------------'
*1 with/without spring under space bar

source

Code: [Select]

Weights to register(actuate) 03/26/2012
HHKB pro2(PD-KB400W 2010-08) purchased used, but not used recently
,-----------------------------------------------------------.
|39 |46 |49 |53 |57 |54 |54 |53 |54 |54 |53 |49 |47 |47 |45 |
|-----------------------------------------------------------|
|46   |50 |58 |58 |61 |61 |59 |56 |61 |59 |58 |47 |49 |46   |
|-----------------------------------------------------------|
|48    |54 |58 |63 |62 |61 |54 |56 |60 |59 |59 |45 |49      |
|-----------------------------------------------------------|
|42      |56 |57 |55 |56 |56 |56 |56 |57 |62 |60 |49    |46 |
`-----------------------------------------------------------'
      |44 |42   |        53/44 *1       |47   |45 |
      `-------------------------------------------'
*1 with/without spring under space bar


[ Attachment Invalid Or Does Not Exist ] 45783[/ATTACH]

[limmy]

My HHKB Type-S is nearly new. I only have 1p coins (3.56g) to measure with. Every key is weighing the same 56.96g (16 coins). I would like to find a better way of measuring this as I seem to have one of the only boards that measures the same througout.
I type quite a lot on this board and will update over the months how it wears.

Thank you for your participation. All same key weighting is indeed something new. I noticed that there is inherent error in measuring peak actuation forces in Topre. Webwit once mentioned the randomness too in his measurement. So, I think using 3.56g coins are fine as long as you measure more than a couple keys. Due to this random nature, I think multiple measurement would be desirable. (either it is multiple measurement of a single key or multiple measurement of multiple keys)

Also, I find from the 6 samples I have that the rubber domes that are single(not bound to each other) tend to have lighter actuation forces. I would guess it is the same for you. Did you measure only the alpha keys, or did you measure other keys as well, e.g. alt, tab, etc?

[Tycn]

How and where have you been lubing the switches? I'm not getting the uniform smoothness one expects of a stupidly expensive keyboard, the travel of some keys are slightly frictiony and some are not. Don't have anything quantitative but in the two brand new uniform weighted boards I've had some keys definitely feel heavier than others - arrow keys and tilde the lightest, alphanumeric area the heaviest, modifiers and rest of board somewhere in between.

[hasu]

I just measured some mechanical boards in hand but only checked home row.
From this result the consistency of Cherry switch is impressive.
Though I've never felt inconsistency on my HHKB and don't like key feel on those Pokers.

Code: [Select]

keybord & switch        A  S  D  F  G  H  J  K  L  ;  '  Return
-----------------------------------------------------------------
PokerX Cherry Red       41 42 42 42 42 41 40 42 41 40 41 40
PokerX Cherry Brown     49 48 49 47 48 49 48 48 48 48 48 49
X68000 ALPS Green       52 50 47 50 49 47 51 53 48 51 49 47
G80-3600 Cherry Red     38 37 38 39 38 38 38 39 39 39 39 37
These Pokers are slightly used and X68000 keyboard came from two decades ago.
(X68000: http://geekhack.org/showwiki.php?title=Island:29060)

lol. man, the variance is amazing. I would think they would have it under more control at those kind of prices.

[fossala]

Also, I find from the 6 samples I have that the rubber domes that are single(not bound to each other) tend to have lighter actuation forces. I would guess it is the same for you. Did you measure only the alpha keys, or did you measure other keys as well, e.g. alt, tab, etc?

You are right. They are lighter by around 10g.

[limmy]

How and where have you been lubing the switches? I'm not getting the uniform smoothness one expects of a stupidly expensive keyboard, the travel of some keys are slightly frictiony and some are not. Don't have anything quantitative but in the two brand new uniform weighted boards I've had some keys definitely feel heavier than others - arrow keys and tilde the lightest, alphanumeric area the heaviest, modifiers and rest of board somewhere in between.

If it is new, I hear it gets smoother as you use it. As for the lubing method, you can see the following tip by hasu

I applied cylinder part and foot of plunger, and hole and foot rail of housing. Namely I think you'd better apply all of sliding surface.
And I applied it twice with small brush one by one, it is very PITA work.

Though I don't know that dunk method  I don't recommend to apply PTFE to part not needed.

In order to properly lube HHKB, you would have to disassemble your HKKB completely. If you have rubber and plastic safe lube and want to quickly try it out, you could brush it on the sliders after taking the key caps off. Do not put drops onto the sliders as it would make your sliders sticky(do not over apply). I tried Krytox GPL 100 and I like it so far.(search ebay with keyword krytox 100) If you can get a hold of RO-59, I would recommend that.

I would recommend taking it apart, but if you don't want to put so much time on it, you could try the brush method.
Caution: You would want lubricant that is safe to ABS plastic and EPDM rubber. And do not over apply the lube.

[limmy]

Here are some summary stats of the measurement data we have.

Code: [Select]

Man. dt  Avg    SD    Avg*   SD*  Note
2009-02  52.85  3.64  53.95  3.21 #1
2006-08  58.71  5.47  60.60  4.46 #2
2004-11  52.16  2.62  52.60  2.13 #3
2007-10  54.03  3.65  55.28  2.55 #4
2010-08  53.16  6.02  55.41  4.69 #5
2003-05  45.65  6.81  44.65  6.71 #6


Man. dt = manufacturing date
Avg = average of actuation weight of all keys
SD = standard deviation of actuation weight of all keys
Avg* = average of actuation weight of alphanumeric keys
SD* = standard deviation of actuation weight of alphanumeric keys

Notes
#1 measurement by limmy - HHKB2 like-new condition
#2 measurement by limmy - HHKB2 used daily for 5 years
#3 measurement by hasu - HHKB1 silence mod, lubed with RO-59
#4 measurement by hasu - HHKB2 purchased used, not used after the purchase
#5 measurement by hasu - HHKB2 purchased used, not used after the purchase
#6 measurement by hasu - HHKB1 silence mod, lubed with RO-59, variable-weight rubber dome from Realforce 101(purchased around 2007)

Please see here for raw data and their original sources.

[limmy]

ripster's way of over-generalization is just amazing.

I think one would know what McRip Effect really is once he/she takes a look at his list of theorems.
Here are some examples:
Ripster's KeyAboo Theorem

Geekhack is now 99% Asian or are weeaboos and/or keyaboos.

Ripster's Weight Theorem: Build Quality is Directly Correlated To Weight of Keyboard

Build Quality is Directly Correlated to the Weight of the Keyboard.

Ripster's LATEST Keyboard Theory - The McKEYMouse Effect

Noobs decorate their keyboards with little sense of restraint and color coordination, pay too much, and have too many Interest Checks (GLWIC) but think it's cool because other people are doing it.

McRip Effect is just one of his own theories and opinion of things. I once took it seriously looking only at what he claims in his McRip Effect, but I have learned not to take it so seriously from the discussions with him and seeing his other theorems. I think McRip Effect and views therein are only useful for a novice who do not care about relatively small things and improvements.

The on-average 7g difference between my two HHKB is huge and I bet any one who typed them side by side would recognize the difference right away. The numbers from RipOmeter or any of the sort is only representing one facet of the whole picture. Moreover, ripster's measurement here suggests his measurement increased by 7g. And I do not see anywhere that his measurement was indeed 50g when he purchased it, moreover I see many cases that his measurement was 45g. Does that really mean ripster's measurement was the same?

Looking at new data from members here, I learned following new things.
1. Topre stiffening is not as severe as I initially thought it would be. Especially for boards that are kept in storage and not used. Ripster's RF was used only for 4 months and it is probably why his weight change was not significant and not so different from designed peak force of 45g.
2. New Topre board's that are designed to have 45g peak force usually require more force. The average of peak force on alphanumeric keys from two newly produced HHKB was about 55g.

Thanks!

And as I put away my Topre it STILL is RipOmeter measuring the same as the day I bought it.  I will add all this to The McRip Effect when I get a chance.

I'll post pics and summarize when I get a chance here as well.

WHAT A RELIEF to be back on my Beloved Cherry MX.  Those Topres are uninspiring switches.

Namaste.

[limmy]

I'm still with EK!!!!

Sorry.  Too busy to take pics but my RealForce 87U is the same as the day I took it out of the box.  Both two weeks ago and two years ago.

It just FELT lighter after two weeks.

And the whole Cherry MX spring thing being different is ALSO hokey science and due to manufacturing variance.

THIS is how Keyboard Science works.

I still disagree with EK.

You don't have measurement saying actuation force was 50g back then. Period. On the other hand, there are posts that your switches were indeed 45g back then, and also posts that they are 52g a few weeks ago. So, what you are saying doesn't make sense. Please stop saying the same thing over and over again.

If you disagree with me on Cherry MX springs, do your experiments and prove me wrong. I did mine. You didn't do anything except taking some pictures and putting on some nickles. Obviously you cannot tell if two springs are identical by just looking at them!

[limmy]

Yeah well, that was the FIRST time I used the RipOmeter so it makes sense my methodology 2 years ago wasn't as refined.

Anyway I'm right and you're wrong.  Just ask all the people that have done Ghetto Reds and other spring swaps.

The springs are the same.

The Topre rubber domes don't age in 2 years.

Your red stickers don't do ****.

The anal retentive method of painting Cherry MX is a weird Korean past-time that I don't recommend for Americans and other sane people.

Why the digression? First the springs, second the sticker mods, third the lube business? I sense something.... but I cannot describe what it is. Anyway, please stay on the topic.

Why would you have to ask people? Would you believe them if they said otherwise? It is matter of facts not opinions.

I agree Topre rubber domes don't age in 2 years when not used heavily. Your Topre is no where near my daily-used-5-year-old HHKB. Even if a Topre board is very old, well-taken-care-of boards are very much usable. See hasu's examples.

I don't have problems you not recommending something I suggest. However I am a little wary that you have not tried it and thus you don't know what it is. Not knowing what it is and building up an opinion is very close in nature to forming a prejudice. I dare say you have prejudice issues. It can be seen in many of your theorems, so much generalization only observing little part of the whole picture...

[Squelos]

By still, do you mean it measured 50g when you got it ? Mine measures 45g. But then my grammes must be the same as yours.

[urbanus]

(Edit) under the assumption that my like-new HHKB is very similar to my old HHKB 5 years ago.

A dubious assumption, I fear.

and provide proper evidence.

I would challenge this.

You have two groups -- older versus newer -- each with a sample size of 1.  That's not enough to calculate variance within each group, so you can't say that the difference between the groups is statistically significant.

There is no evidence that the difference in weights between these two keyboards is the result of stiffening over time, as opposed to (say) differences in manufacturing processes over the years that Topre has been manufacturing keyboards.

You provide no information about other environmental factors that may have an effect on results: temperature, humidity, etc.

This could be resolved by a longitudinal study: take a sample of keyboards and measure stiffening over time by taking measurements at certain intervals.  However, the evidence presented here is inconclusive.

[limmy]

I am not saying my finding is statistically significant. Why are you guys so into statistics? The OP has only one observation and there cannot be any statistics(none at least in classical statistics, which is what I believe you are referring to) in the first place. I didn't even say one word about statistically significant finding! This documentation is one instance of rubber dome stiffening which is widely perceived but not documented phenomenon. The OP documents it under one assumption that is reasonable.

I feel my HHKB is noticeably stiff and it is almost as stiff as heavily used rubber dome, which is NOT how my HHKB felt when I first got it. The increased actuation force is so obvious to me, but I just didn't measure the actuation force when I got it. If you don't believe it, because you don't see any numbers.. then fine. If you don't believe it because there is no statistics, fine. Just take this as anecdotal evidence. Who would have resources, time, and money to produce the research you want? That is simply impossible for a consumer to do.

As far as the measurement goes, we only have one sample that is measured over time. Ripster's RF measured 45g when he got it two years ago and now it measures 52g. Ripster's board was said to be used about 6 months. My new realforce measures very close to the specified spec only off by +/- 5g so I suspect there is not much of cross-sectional variance within newly produced boards. Besides, Japanese QC is pretty decent.

Here is one post about my stance on the statistics I am using if I am using any.

Here is my summary and tentative conclusion before observing Hasu's measurements.

Here are measurements of 6 HHKB(two of them are modded and one of them is using rubber dome of Realforce variable)

(source)

[limmy]

Please stop saying your RF measures the same. Your own measurement says otherwise. See here. Your measurement increased by 7g. This is the tiebreaker.

Mine feels/ measures the same as the day I got it.

sent  from brother Ripster's 87u



We obviously need tiebreaker here or this could go on for a while.

[limmy]

I think I did it wrong.  It was my first TIME after all.

Other keys are definitely the same.

Don't make nickles and 1 yen coin lie, ripster. It was you who said gravity doesn't lie.

So, you are saying your other keys measure 45g?? You were saying 50g(52g) all along, numerous times! Your dithering method, which is later developed I suppose, always lowers measurement so the first time measurement argument also works against you.

[limmy]

I agree my evidence cannot be generalized and it will probably never will because I don't have time and money to test multiple Topre boards over time. It is only indicative, which is what a consumer can do at most. Besides I didn't put it in any wiki, so you don't need to notify me. Also, it is funny that you take it out because it lacks evidence because most of your wiki entries are collection of anecdotal evidences.

5g differences may happen, but that is not what you said. You said they measure the same, and they don't measure the same! Also, three instances in three different keys are unlikely. The below are your own measurements.

Three keys support 50g.


At 52g, the keys collapse.