How Cherry has fallen

User avatar
HaaTa
Master Kiibohd Hunter

04 Dec 2016, 06:37

I'm going to leave the analysis to the readers. But really, Cherry, you were so. damn. good. at linear switches.

Image


This is what we have now for MX Black switches. Notice how it's a bit different. Perhaps this is enough to convince everyone that there really is a difference between older and newer MX Black switches.

Image


But that's not what this post is about. Take a look, this is a NIB Cherry MX Red RGB switch.

Image


What's going on in Cherry's factories?

User avatar
HaaTa
Master Kiibohd Hunter

04 Dec 2016, 06:44

Here's a NIB standard Cherry MX Red for comparison.

Image

User avatar
nafsasp

04 Dec 2016, 07:07

Hi HaaTa,

Excellent work as always. I have a couple quick questions. First, what has your experience with vintage MX blacks been? I ask because they are my favorite Cherry switch, but I've found a bit of a range in smoothness to them (though my approach is much less scientific than yours), including smooth feeling MX blacks in boards much newer than what people normally consider "vintage".

Second, have you had a chance to test a Cherry MX Silent (Pink) switch yet? I'd be curious to see how they compare to the other Cherry linear switches you've tested. I tried one of these switches for the first time today (at the Washington, DC keyboard meetup I hosted), and thought it felt smoother than a typical new Cherry switch (though I wasn't crazy about the dampening employed on the stem).

User avatar
HaaTa
Master Kiibohd Hunter

04 Dec 2016, 07:35

Hmm. In terms of all the Cherry MX Black switches I've tried over the years. They tend to vary a bit, some smoother than others.
To me what's sad, is that there's this wide deviation. I'm also not surprised by your comment that some later MX Blacks were quite smooth (could've been new tooling for all we know, or added more lubricant in that run of switches).

I have not had a chance to test Cherry MX Slient (Pink) switches yet. If anyone is willing to send me some (1-2), I'll happily test them.

User avatar
Wodan
ISO Advocate

04 Dec 2016, 08:03

One thing that I always wondered about ... how do "vintage" Cherry sliders perform in brand new Cherry casings?

The major source of "scratch" in Cherry switches is the point between the contact leaf and the slider. When a Cherry switch is assembled, that contact leaf is permanently bent back in an open state. This might "wear out" the contact leaf just by age and not just by heavy use. Ever considered that? A worn out contact leaf could reduce the impression of fraction during keypress.

User avatar
DanielT
Un petit village gaulois d'Armorique…

04 Dec 2016, 09:22

This is pure slander, all MX switches are born equal, they just get better with age and usage. I know nobody will agree with me but I don't care :lol:

User avatar
Chyros

04 Dec 2016, 09:41

Can I ask, what's the margin of error in the measurement of each of these data points?

axtran

04 Dec 2016, 09:56

I've put Vintage stems into modern casings. The stems matter more than the housings--it is what the cap is connected to. There's a tiny amount of variance and wobble area where the stem may "rub" on different parts of the bottom housing, however, it will always be directly connected to the stem itself.

Also, I have NIB Vintage WYSE donors... the quality of the switch is definitely there, being zero wear on NOS boards.


Sent from my iPhone using Tapatalk

User avatar
seebart
Offtopicthority Instigator

04 Dec 2016, 09:58

Does not suprise me one bit in fact this confirmes my opinion that vintage MX blacks are the best switch Cherry ever came out with. This of course includes the insanely overhyped nixies (which feel about 10-20 % better at most than regular vitage MX blacks). The great thing is regular vintage MX blacks are not really rare at all.

Good work HaaTa!

User avatar
Wodan
ISO Advocate

04 Dec 2016, 10:04

A little while back I did an accustic based attempt at measuring scratch and compared Vintage Blacks, Nixies, Gateron Red and brand new MX Reds. No doubt the MX red hat the most scratch and the two Vintage Blacks (Nixie, MX) had no measurable difference. The Gateron Red could also be measured to be smoother than both Vintage blacks. But when swapping the Gateron slider with the Vintage Black slider, both casings with the new slider gave the same measurement as with their original slider. My conclusion was that Gateron and Vintage MX sliders are similair but Gateron did something to the crosspoint contacts to further reduce friction.

User avatar
HaaTa
Master Kiibohd Hunter

04 Dec 2016, 10:14

Chyros wrote: Can I ask, what's the margin of error in the measurement of each of these data points?
I haven't gone far enough to determine what this is yet. However, if you click on the graph, then select additional presses, you'll notice that it's quite consistent between presses.

According to the Imada website (https://imada.com/products/ds2-digital-force-gauge/) they claim +/- 0.2% FS +/- 1 LSD, which is about 1.1 gf, if I'm reading it correctly.
I use the analog output directly, so I might be a bit better (or worse).
Last edited by HaaTa on 04 Dec 2016, 10:22, edited 1 time in total.

User avatar
HaaTa
Master Kiibohd Hunter

04 Dec 2016, 10:18

Also, I don't have any Nixdorf cherry switches. I'd be happy to measure some if I can get a couple.

User avatar
seebart
Offtopicthority Instigator

04 Dec 2016, 10:22

HaaTa wrote: Also, I don't have any Nixdorf cherry switches. I'd be happy to measure some if I can get a couple.
The comparison would be interesting in finally putting some hard facts toward the intense nixie hype.

User avatar
Chyros

04 Dec 2016, 10:24

HaaTa wrote:
Chyros wrote: Can I ask, what's the margin of error in the measurement of each of these data points?
I haven't gone far enough to determine what this is yet. However, if you click on the graph, then select additional presses, you'll notice that it's quite consistent between presses.
Yes, and that's exactly why I was asking.

0.1 gf is large enough that it would invalidate these conclusions, but still, the results are compelling, and like you said, fairly consistent, including in their apparent margin of error. Interesting...

As for Nixies, I think they're pretty much the same as vintage blacks. I think they were just "discovered" earlier because they look different while vintage MX blacks don't.

akryl9296

04 Dec 2016, 11:14

Any chance you could measure Zealios switches?

User avatar
Alkhar

04 Dec 2016, 12:21

What should a newbie see on the graph ?

User avatar
nafsasp

04 Dec 2016, 14:46

Here's something I've been thinking about but don't know the answer to. Does anyone know what material is used for the top housing on Nixies (and other milky switches). Testing milk gateron blacks against clear top gateron blacks here seems to be a slight different in sound quality to them.

HaaTa, I'll try to see if someone can send you a Cherry MX Pink switch. I know a few people who have harvested them.

User avatar
Ail

04 Dec 2016, 15:36

I'm guessing the actuation force on that MX Red RGB is not supposed to be that high at all. Thanks for sharing these, HaaTa.

lepidus

04 Dec 2016, 16:35

Do you know if there is any difference between new and older cherry browns? I could swear the cherry browns from an old compac board I had were way more tactile than current browns. I'm not sure if its due to switch difference, of just because pretty much every board nowadays place the switches upside down.

Btw, looking at your Plotly, there are two different greetech browns, one way more tactile than the other. What is their difference?

User avatar
livingspeedbump
Not what they seem

04 Dec 2016, 17:31

So, there is finally some hard data to back up that vintage MX Blacks truly are superior than modern ones.

User avatar
DanielT
Un petit village gaulois d'Armorique…

04 Dec 2016, 19:13

livingspeedbump wrote: So, there is finally some hard data to back up that vintage MX Blacks truly are superior than modern ones.
Yes and no...
No, because there is no study to take into consideration, material aging, wear etc...
Yes, because we have data that shows that a 30year switch (I assume a NIB 30 year switch) is better than a 1 or 2 year old NIB switch.

User avatar
HaaTa
Master Kiibohd Hunter

04 Dec 2016, 19:28

lepidus wrote: Do you know if there is any difference between new and older cherry browns? I could swear the cherry browns from an old compac board I had were way more tactile than current browns. I'm not sure if its due to switch difference, of just because pretty much every board nowadays place the switches upside down.
Hmm, I'd need some low wear (or NIB) older MX Browns. The only ones I have are quite worn.
lepidus wrote: Btw, looking at your Plotly, there are two different greetech browns, one way more tactile than the other. What is their difference?
They are different switches. The through-hole RGB switch is newer and likely uses a different slider (possibly different contact leaf as well). The older one looks more like a standard Cherry switch.
Alkhar wrote: What should a newbie see on the graph ?
An ideal force curve is a straight/curved line. The bumpiness shows evidence of a number of quality issues (e.g. spring binding, slider imperfections, housing imperfections, dust/dirt, etc.). Using this testing method it's pretty hard to determine what's causing the deviations, but we do know they are there.
akryl9296 wrote: Any chance you could measure Zealios switches?


I have one measured already https://plot.ly/~haata/146
Sadly, I only have a single weighting of Zealios so that's all I can measure for now.

m8max

04 Dec 2016, 20:19

Can it be the plastic? maybe they had to start using a different type for the switch to be more durable?

kami_sama

04 Dec 2016, 22:27

Great work, Haata.
Wonder what could have caused the change, Cherry has now more competition than ever in the mech space (even if it's a small part of their business), so things like this are puzzling.

User avatar
Zeal

04 Dec 2016, 22:51

Thanks for sharing the graphs HaaTa. :)
HaaTa wrote: Sadly, I only have a single weighting of Zealios so that's all I can measure for now.
Let me know your address and I'll send over some stock R4 samplers. Will probably throw in a few "Zealiostotles" (Aristotle stem + Zealio housing) so you can measure the click/tactile response on those!

User avatar
nafsasp

05 Dec 2016, 01:18

HaaTa wrote: Hmm. In terms of all the Cherry MX Black switches I've tried over the years. They tend to vary a bit, some smoother than others.
To me what's sad, is that there's this wide deviation. I'm also not surprised by your comment that some later MX Blacks were quite smooth (could've been new tooling for all we know, or added more lubricant in that run of switches).

I have not had a chance to test Cherry MX Slient (Pink) switches yet. If anyone is willing to send me some (1-2), I'll happily test them.
Hey HaaTa,

I have another follow up question/observation. A while ago I compared the springs from six or so different MX blacks from various years, and found a wide variety of spring sizes/shapes. Some were longer or shorter, and others had a larger or smaller diameter, with different numbers of coils. They were all close in weight, tested very unscientifically, but I would guess all six of them were within 5-7g of each other.

My question is, aside from weight, are there characteristics of springs that affect how smooth a switch feels?

User avatar
HaaTa
Master Kiibohd Hunter

05 Dec 2016, 04:35

So, something I've talked about for quite a while that affects most keyboard switches is "spring friction". Or said another way spring binding.

When springs are compressed, they are held into place in switches. The coils of the spring rub against the side of the housing causing additional resistance against the movement of the slider. The material, plating, lubricant and corrosion can have a lot to do with this to varying degrees depending on the switch design.

The number of coils may have a positive or negative affect as well. Few coils means less contact points, but each contact point applies more friction. More coils means more contact points and more chances to get caught on housing imperfections.

In terms of length of springs. this mainly matters for switch pre-load force. Otherwise it doesn't really matter.

Something don't see talked about very much (if at all) is vibrations/force transfer through the spring. When you compress a spring, it doesn't compress uniformly when the coils are being resisted by friction due to the housing. This causes force impulses to be transferred through-out the spring. I suspect this is one of the sources of noise on my force graphs, though it's pretty hard to validate without some sort of highspeed camera and the switch cut open.

Anyways, yep. Springs can affect how a switch feels. But I think the design of the switch itself is more important.

User avatar
Chyros

05 Dec 2016, 13:10

HaaTa wrote: So, something I've talked about for quite a while that affects most keyboard switches is "spring friction". Or said another way spring binding.

When springs are compressed, they are held into place in switches. The coils of the spring rub against the side of the housing causing additional resistance against the movement of the slider. The material, plating, lubricant and corrosion can have a lot to do with this to varying degrees depending on the switch design.

The number of coils may have a positive or negative affect as well. Few coils means less contact points, but each contact point applies more friction. More coils means more contact points and more chances to get caught on housing imperfections.

In terms of length of springs. this mainly matters for switch pre-load force. Otherwise it doesn't really matter.

Something don't see talked about very much (if at all) is vibrations/force transfer through the spring. When you compress a spring, it doesn't compress uniformly when the coils are being resisted by friction due to the housing. This causes force impulses to be transferred through-out the spring. I suspect this is one of the sources of noise on my force graphs, though it's pretty hard to validate without some sort of highspeed camera and the switch cut open.

Anyways, yep. Springs can affect how a switch feels. But I think the design of the switch itself is more important.
I have thought about this, but I don't know manyt switch designs in which is spring is so exposed that it would clearly rub against stationary parts. Could you give an example of a design in which this is very prominent?

Hking0036

05 Dec 2016, 15:21

Super interesting! It's interesting to see some substantial info to back up the vintage blacks, they are smoother after all. That drop in smoothness is crazy :shock:
Have you done gateron blacks yet? I remember people hailing them as the best linears when they first came out and I'd be interested to see how they compare.

I'm also interested to see how a spring-swapped clear compares to a brown or a zealio but that's something else entirely.

dlightningd

05 Dec 2016, 16:51

Great analysis. I've heard a lot of people complaining about new Cherry switches, but attributed most of that to nostalgia for older boards. Fascinating to see the data.

Post Reply

Return to “Gallery”