Author Topic: You should be skeptical of "progressive" springs - some force curve measurements  (Read 3484 times)

0 Members and 2 Guests are viewing this topic.

Offline Pylon

  • Thread Starter
  • Posts: 807
Hi all,

As many of you may know, SPRiT, Thic Thock, and others have advertised and sold progressive springs for a while. These springs claim to be nonlinear, having a lower spring constant before the approximate actuation point, and an increasing spring constant after, to provide a more cushioned bottom out. SPRiT, on his website, provides the following image for what the force curve for his progressive and complex springs should roughly look like:



These springs are wound tighter on one side than the other. SPRiT's springs look like this, for instance (image from his site):



Nonlinear springs are known to exist - for example, this article from an industrial spring manufacturer describes some of them, including buckling springs, conical springs, springs of varying diameter, and "dual rate" springs that have a tighter winding on one end. So far the "progressive"/"complex" springs sold by SPRiT, Thic Thock, and others have all been of the last variety, which have a constant diameter but a varying winding pitch on one end of the spring. In theory, these springs work by the tighter end compressing first, and the coils on the tighter end bumping up against each other, which provides nonlinearity. A seller of industrial springs has a nice animation of how these work.

However, as implemented in a keyswitch these don't really seem to be very nonlinear.

As some of you may know I recently designed and built an open-source force curve meter, the files which are published on Github here. The curves of various switches, frankenswitches, and springs I've measured so far are published in a directory on the Github, here.

Windows Dump (I don't believe he's on Geekhack, but he goes by that on Discord) recently lent me a gBoards.ca SPRiT spring sampler, and also various Thic Thock DL and MP springs. I measured the 63.5g SPRiT complex and progressive springs from the sampler, and the Thic Thock 68g MP springs. The SPRiT springs I placed in a Gateron Yellow (one test with the leaf, and another test with the leaf removed), and the Thic Thock spring I placed in a JWK Alpaca "V2" / "B" that Slowshi lent me (the Thic Thock spring is a bit wide diameter wise, and tended to catch on the stem in the Gateron Yellow stem, so I used the Alpaca instead). I did the leafless Alpaca tests using a "V2" / "B" bottom housing from a JWK T1 from which I removed the leaf.

The results are here - I compared the SPRiT springs to the stock Gateron Yellow spring (which is nominally 63.5g), and I compared the Thic Thock spring to a Durock 67g spring when installed in an Alpaca switch. For the leafless measurements, I also plotted them against a theoretical perfectly linear line, which I made by plotting the measured values at x=0.1mm and x=3.8mm and drawing a straight line between them.






Obviously these don't look very nonlinear. The Thick Thock 68g MP pretty much looks linear and very similar to the Durock 67g spring when installed in an Alpaca switch.  The SPRiT springs deviate about 1-2gf from linear, and also don't seem to behave nonlinearly by very much compared to a regular Gateron Yellow spring. The SPRiT springs are also much lighter than the 63.5g claimed.

I don't know if this is an implementation issue with these particular "progressive" springs (the tighter wound sections on the SPRiT and TT springs are relatively short compared to the regularly-wound portion of the spring, and seemingly too short to have much of an effect), or if keyswitches just inherently don't provide enough travel and compression for progressive springs to really enter their nonlinear regime. I also haven't had a chance to test other progressive springs, and it is possible that some of them in fact behave nonlinearly as expected.  But in any case, you should probably be skeptical about purportedly nonlinear "progressive" springs for the time being, until you see measurements that validate their claims.

(edit - fixed a typo on a URL)
« Last Edit: Thu, 22 April 2021, 00:20:16 by Pylon »

Offline Ranger_Trivette

  • Posts: 88
  • Location: Italy
i've designed some springs for work. the progressive spring exist.

if you create a spring with two different pitches you will have a graph with 2 angles.
(you can have the same effect using two different springs, with they own pitches, stacked)
if you use a variable pitch, you really get a non-linear spring.

in both cases, if you want to feel the differences, the variation of the pitch must be HUGE.
there isn't a linear correlation between pitch and strength.

here is an example that clarify.
when you fully compress the upper part of the spring, you will start to compress the lower part, and the graph angle changes.


Offline Leslieann

  • * Elevated Elder
  • Posts: 3796
If you want to test springs, it needs to be done outside the switch.

No matter how much you think you're keeping the results fair there's just too many variables you can't control when trying to nail down the spring variables which will be subtle.
Novelkeys NK65AE w/62g Zilents/39g springs
More
62g Zilents/lubed/Novelkeys 39g springs, HK Gaming Thick PBT caps, Netdot Gen10 magnetic cable, pic
| Filco MJ2 L.E. Vortex Case, Jailhouse Blues, heavily customized
More
Vortex case squared up/blasted finish removed/custom feet/paint/winkey blockoff plate, HID Liberator, stainless steel universal plate, 3d printed adapters, Type C, Netdot Gen10 magnetic cable, foam sound dampened, HK Gaming Thick PBT caps (o-ringed), Cherry Jailhouse Blues w/lubed/clipped Cherry light springs, 40g actuation
| GMMK TKL
More
w/ Kailh Purple Pros/lubed/Novelkeys 39g springs, HK Gaming Thick PBT caps, Netdot Gen10 Magnetic cable
| PF65 3d printed 65% w/LCD and hot swap
More
Box Jades, Interchangeable trim, mini lcd, QMK, underglow, HK Gaming Thick PBT caps, O-rings, Netdot Gen10 magnetic cable, in progress link
| Magicforce 68
More
MF68 pcb, Outemu Blues, in progress
| YMDK75 Jail Housed Gateron Blues
More
J-spacers, YMDK Thick PBT, O-rings, SIP sockets
| KBT Race S L.E.
More
Ergo Clears, custom WASD caps
| Das Pro
More
Costar model with browns
| GH60
More
Cherry Blacks, custom 3d printed case
| Logitech Illumininated | IBM Model M (x2)

Offline Pylon

  • Thread Starter
  • Posts: 807
If you want to test springs, it needs to be done outside the switch.

No matter how much you think you're keeping the results fair there's just too many variables you can't control when trying to nail down the spring variables which will be subtle.

Why test springs that are marketed and sold for installation inside keyswitches, marketed as making a difference while installed in keyswitches, without installing them inside keyswitches? Why would that be useful?

Here's some other spring measurements. Here's a comparison between a 12mm long ~90g MX Clear spring, 13nn long 89g Novelkeys spring, and a 15mm Kailh Speed Pale Blue spring - the difference in spring rate and initial preload is pretty apparent:


Here's a measurement of an MX Clear switch with the Kailh and Cherry springs, with the spring measurement (done without the leaf) overlaid on top:


Here's also a chart I did of a Gateron Yellow with and without its leaf and using its stock spring - the leafless measurement looks pretty dead straight to me:


I don't see issues with measuring springs inside a leafless switch.

Offline Pylon

  • Thread Starter
  • Posts: 807
i've designed some springs for work. the progressive spring exist.

if you create a spring with two different pitches you will have a graph with 2 angles.
(you can have the same effect using two different springs, with they own pitches, stacked)
if you use a variable pitch, you really get a non-linear spring.

in both cases, if you want to feel the differences, the variation of the pitch must be HUGE.
there isn't a linear correlation between pitch and strength.

here is an example that clarify.
when you fully compress the upper part of the spring, you will start to compress the lower part, and the graph angle changes.

Show Image
Show Image


Makes sense - thanks for this. The difference in pitch needs to be sufficient that the tighter wound coils "bottom out" against each other right? The change in spacing in the coils of the SPRiT and TT springs isn't particularly dramatic (nowhere near as dramatic as in the spring you posted), and only around 25% of the spring has the tighter winding. I suppose that's why they appear to behave mostly linearly when installed in a switch.

Offline AJM

  • Posts: 73
  • Location: Germany
Thanks for these tests Pylon!
Very interesting and it confirms, what I have been suspecting.

Offline saltimate

  • Posts: 49
ngl i was pretty disappointed when i tried some out...
was it always just a scam or do they really believe if its noticeable in high precision instruments (or in theory? idk how they evaluate their springs...) it will have a meaningful impact on switch feel?

Offline N8N

  • Posts: 787
For what it's worth, I am familiar with progressive springs in an automotive application.  However, I learned over xmas break that BMW does it differently - when I replaced the struts in my car I found that the springs were very soft (relatively speaking) and extended quite a bit when removed from the struts.  The car actually sits on the bump stops at rest, meaning that the bump stop (basically a piece of closed cell foam slid over the strut shaft) is acting as an additional spring.  I'm not sure why they do it this way but I suspect it is because it offers more "progressiveness" than a single coil springs with coils that go into bind when the increased rate is desired.

Unfortunately, they don't last, and often need to be replaced before the struts wear out - I thought I could let this go and hit a bad pothole one night, I now have a chunk of my RF fender missing (it's plastic) and the RF strut was completely blown when I replaced it.
Filco Majestouch-2 with Cherry Corp. doubleshot keys - Leopold Tenkeyless Tactile Force with Wyse doubleshots - Silicon Graphics 9500900 - WASD V1 - IBM Model M 52G9658 - Noppoo Choc Pro with Cherry lasered PBT keycaps - Wyse 900866-01 - Cherry G80-8200LPBUS/07 - Dell AT101W - several Cherry G81s (future doubleshot donors) (order of current preference) (dang I have too many keyboards, I really only need two)

Offline HungerMechanic

  • Posts: 821
I never expected much variance between Progressive and regular springs.

But the difference between 68 P and regular 68 G springs on Ergo Clears seems pretty significant to me.

I can barely stand typing on regular 68 G stainless in MX Clear housing. But with 68P, it actuates around 45 G, near MX-Brown level, and bottoms-out closer to 68 G, allowing good functioning. 68 P felt bouncier and less rigid than 68 S.

My 63.5 P Ergo Clears are crazy-light at actuation, maybe lighter than MX Brown, but they seem to function fine with Clears.

Offline Pylon

  • Thread Starter
  • Posts: 807
My 63.5 P Ergo Clears are crazy-light at actuation, maybe lighter than MX Brown, but they seem to function fine with Clears.

Have you tested the bottom out weights of each by any chance(using nickels for instance, or the fairly crude "kiss test"?). The 63.5 P I got is bottoming out at 50g on a Gateron Yellow, which has the full 4mm travel. It is possible that gboards.ca included the wrong springs by accident in the sampler, but it visibly has the varying coil spacing, and it's  also possible that they're just much lighter so that they appear to feel different.


Offline HungerMechanic

  • Posts: 821
Sadly, I don't have the ability to test.

What are you describing, 50 G bottom-out, probably shouldn't even work with an MX Clear! And yet I have a board full of these. Very interesting.

Incidentally, a thread on KeebTalk may have reached similar conclusions:

https://www.keebtalk.com/t/sprit-progressive-springs/4055/4

Offline Pylon

  • Thread Starter
  • Posts: 807
Actually I just counted the number of coils on the 63.5 P I have on hand and it matches SPRiT's pictures.

Here are the pictures of the springs I have:


Offline Pylon

  • Thread Starter
  • Posts: 807

What are you describing, 50 G bottom-out, probably shouldn't even work with an MX Clear! And yet I have a board full of these. Very interesting.


Huh, I just put an ergo clear together with the 63.5 P spring and it returns fine. I also tried the 45g spring out of the gboards.ca sampler and put it in an ergo clear and it returns just fine (I did not measure the 45g in my force curve meter, but I can)

Offline Leslieann

  • * Elevated Elder
  • Posts: 3796
Why test springs that are marketed and sold for installation inside keyswitches, marketed as making a difference while installed in keyswitches, without installing them inside keyswitches? Why would that be useful?

I don't see issues with measuring springs inside a leafless switch.
Yes, they are sold for use in switches but every switch is different, even the same switch from the same batch, even re-using the same switch introduces variables.

You don't compare engines in cars, valve springs in heads, and suspension springs in the car. They do test that way but it's (usually) done for determining overall performance, not the performance of a particular item to compare one specific part against another specific part. And yes, this is a switch, it's simple but simple also means it's simple to mess it up.

A leafless switch is better, but housings don't always  align the exact same every time, lube and dust gets wiped and added, scars develop, springs don't settle the same, especially depending on lube. You can say all these things are minor but they matter, particularly if you're trying to find a subtlety in a benchmark.
Novelkeys NK65AE w/62g Zilents/39g springs
More
62g Zilents/lubed/Novelkeys 39g springs, HK Gaming Thick PBT caps, Netdot Gen10 magnetic cable, pic
| Filco MJ2 L.E. Vortex Case, Jailhouse Blues, heavily customized
More
Vortex case squared up/blasted finish removed/custom feet/paint/winkey blockoff plate, HID Liberator, stainless steel universal plate, 3d printed adapters, Type C, Netdot Gen10 magnetic cable, foam sound dampened, HK Gaming Thick PBT caps (o-ringed), Cherry Jailhouse Blues w/lubed/clipped Cherry light springs, 40g actuation
| GMMK TKL
More
w/ Kailh Purple Pros/lubed/Novelkeys 39g springs, HK Gaming Thick PBT caps, Netdot Gen10 Magnetic cable
| PF65 3d printed 65% w/LCD and hot swap
More
Box Jades, Interchangeable trim, mini lcd, QMK, underglow, HK Gaming Thick PBT caps, O-rings, Netdot Gen10 magnetic cable, in progress link
| Magicforce 68
More
MF68 pcb, Outemu Blues, in progress
| YMDK75 Jail Housed Gateron Blues
More
J-spacers, YMDK Thick PBT, O-rings, SIP sockets
| KBT Race S L.E.
More
Ergo Clears, custom WASD caps
| Das Pro
More
Costar model with browns
| GH60
More
Cherry Blacks, custom 3d printed case
| Logitech Illumininated | IBM Model M (x2)

Offline HungerMechanic

  • Posts: 821
Quote
Huh, I just put an ergo clear together with the 63.5 P spring and it returns fine. I also tried the 45g spring out of the gboards.ca sampler and put it in an ergo clear and it returns just fine (I did not measure the 45g in my force curve meter, but I can)


Interesting. The guides always said best not to go below 55 G bottom-out [they meant the linear springs], and even that's iffy and may require careful lubing. So the actuation weight of 63.5 P would be close to that metric.

Of course, they are thinking of spacebars and other large modifiers, and a high desire for consistency in the switches.

63.5 P will work fine with Clears, I have 64 of them in a board. It's just lower in the actuation than you would want, ideally. Clears really want about 50 G actuation, and I doubt 63.5 P provides that. [In fact, I think it actuates around what a 55 G linear spring would. Which I guess makes it borderline for Ergo Clears.]

With regard to the original discussion in this thread, I think it's true that sub 72 G Progressive springs aren't nearly as 'ramping-up' as MX Clear springs. However, there is a difference. Specifically, lower actuation than the regular version of the spring, and it does slightly ramp up towards the bottom-out weight at the end.

This effect is particularly noticeable among tactiles, in my experience, from Ergo Clears to Holy Pandas and derivatives.
« Last Edit: Thu, 22 April 2021, 19:33:01 by HungerMechanic »

Offline Leopard223

  • Posts: 107
This is why I love this site, thank you for this post!

Offline Leslieann

  • * Elevated Elder
  • Posts: 3796
Interesting. The guides always said best not to go below 55 G bottom-out [they meant the linear springs], and even that's iffy and may require careful lubing. So the actuation weight of 63.5 P would be close to that metric.

55g is just the point things stop being easy.

I have some tactiles that took 39g springs with only minimal effort and medium lube while another tactile simply does not work properly below 45g regardless of the type or viscosity of the lube you use.
Novelkeys NK65AE w/62g Zilents/39g springs
More
62g Zilents/lubed/Novelkeys 39g springs, HK Gaming Thick PBT caps, Netdot Gen10 magnetic cable, pic
| Filco MJ2 L.E. Vortex Case, Jailhouse Blues, heavily customized
More
Vortex case squared up/blasted finish removed/custom feet/paint/winkey blockoff plate, HID Liberator, stainless steel universal plate, 3d printed adapters, Type C, Netdot Gen10 magnetic cable, foam sound dampened, HK Gaming Thick PBT caps (o-ringed), Cherry Jailhouse Blues w/lubed/clipped Cherry light springs, 40g actuation
| GMMK TKL
More
w/ Kailh Purple Pros/lubed/Novelkeys 39g springs, HK Gaming Thick PBT caps, Netdot Gen10 Magnetic cable
| PF65 3d printed 65% w/LCD and hot swap
More
Box Jades, Interchangeable trim, mini lcd, QMK, underglow, HK Gaming Thick PBT caps, O-rings, Netdot Gen10 magnetic cable, in progress link
| Magicforce 68
More
MF68 pcb, Outemu Blues, in progress
| YMDK75 Jail Housed Gateron Blues
More
J-spacers, YMDK Thick PBT, O-rings, SIP sockets
| KBT Race S L.E.
More
Ergo Clears, custom WASD caps
| Das Pro
More
Costar model with browns
| GH60
More
Cherry Blacks, custom 3d printed case
| Logitech Illumininated | IBM Model M (x2)

Offline tp4tissue

  • * Destiny Supporter
  • Posts: 13374
  • Location: Official Geekhack Public Defender..
  • OmniExpert of: Rice, Top-Ramen, Ergodox, n Females
it's also not always the case that the outer fingers are weaker.  for example, it's possible that depending on your technique with an axe, your pinky and ring finger gets stronger than middle/index

Offline Faceman76

  • Posts: 76
  • Location: Long Island
it's also not always the case that the outer fingers are weaker.  for example, it's possible that depending on your technique with an axe, your pinky and ring finger gets stronger than middle/index
It's safe to assume, that for most of us, the inner two fingers are stronger. 

I've played with various Captains of Crush grippers over the years,  my middle and index fingers are still the strongest. 

Sent from my SM-G975U using Tapatalk

TKC 1800, Crystal Box Navy, Sprit 100g

Offline Ranger_Trivette

  • Posts: 88
  • Location: Italy
Makes sense - thanks for this. The difference in pitch needs to be sufficient that the tighter wound coils "bottom out" against each other right?
correct.

The change in spacing in the coils of the SPRiT and TT springs isn't particularly dramatic (nowhere near as dramatic as in the spring you posted), and only around 25% of the spring has the tighter winding. I suppose that's why they appear to behave mostly linearly when installed in a switch.
if you don't modify the pitch in a "dramatic" way, you will never see a difference in the graph.

in the spring the pitch move from 6,5 to 15.5.
the graph change the inclination from 20° to 28°.
and the difference is barely visible.

i think you will never see any difference in a graph if you test the springs you have posted before.

Offline Pylon

  • Thread Starter
  • Posts: 807
Ok, so here's my hypothesis as to why the "progressive" springs I tested aren't  behaving nonlinearly, or are very minimally nonlinear.

The equation for the spring constant of a coil spring is the following:


k = spring constant (higher = stiffer spring)
G = shear modulus of the material the spring is made out of
d = wire diameter
D = Mean diameter of the spring (outer diameter - wire diameter)
na = number of active coils

Most springs are made of steel so G (shear modulus) tends to be fairly constant, and this is also fairly consistent across different alloys and tempers of steel.The overall diameter of the spring (D) tends to be constant for all MX-format springs, though if you had a conical spring (e.g. Topre) it would vary over the course of the travel. d (wire diameter) varies between different springs, but is constant for a given spring.

What a progressive spring is trying to do is to lower na over the course of the travel, by causing coils to bottom out against the inactive coils at the end of the spring over the course of the travel. As individuals coils "bottom out" against the end, they become inactive coils, and so na should go down, increasing the k of the spring as you push the switch down. This is done by making some of the coils spaced closer to each other so that they bottom out first.

The main issue that springs are already compressed when you install them in an MX switch. A typical MX-format spring is ~15mm long naturally, but becomes ~8mm long when installed in a switch. When the switch is pushed all the way down to bottom out, the spring is shortened to 4mm long (typical MX switch has ~4mm of travel). So the entire action of an MX happens between spring lengths 4mm to 8mm.

The issue with the progressive springs is that when you install them in a switch, virtually all of the "tight" coils have already compressed simply from installing the spring in the switch. When you go push the switch, the remaining coils left that area all evenly spaced, and so you  have little to no non-linear "progressive" effect over the course of the travel of the switch.

See this picture of the Thic Thock MP 68g spring comparing its natural state, and compressed to 8mm (which is what happens when you install it in a switch). When compressed to 8mm all of the tight coils have already bottomed out, and all of the remaining active coils are evenly spaced apart, so there's no tighter-wound coils left to provide nonlinearity when you go push the switch.

267600-0

« Last Edit: Fri, 30 April 2021, 12:23:49 by Pylon »

Offline tp4tissue

  • * Destiny Supporter
  • Posts: 13374
  • Location: Official Geekhack Public Defender..
  • OmniExpert of: Rice, Top-Ramen, Ergodox, n Females
it's also not always the case that the outer fingers are weaker.  for example, it's possible that depending on your technique with an axe, your pinky and ring finger gets stronger than middle/index
It's safe to assume, that for most of us, the inner two fingers are stronger. 

I've played with various Captains of Crush grippers over the years,  my middle and index fingers are still the strongest. 

I must emphasize, DO NOT play with captain of crush grippers. The hand ligaments are very delicate and we have evolved very far from the days where extreme grip force is a necessity in our lives. Exercising grip can cause carpel tunnel. It is an excess with no purpose and high probability of damage.

Offline Faceman76

  • Posts: 76
  • Location: Long Island
it's also not always the case that the outer fingers are weaker.  for example, it's possible that depending on your technique with an axe, your pinky and ring finger gets stronger than middle/index
It's safe to assume, that for most of us, the inner two fingers are stronger. 

I've played with various Captains of Crush grippers over the years,  my middle and index fingers are still the strongest. 

I must emphasize, DO NOT play with captain of crush grippers. The hand ligaments are very delicate and we have evolved very far from the days where extreme grip force is a necessity in our lives. Exercising grip can cause carpel tunnel. It is an excess with no purpose and high probability of damage.

Was up to the 195lbs gripper for a bit, but backed down to the 140 for the reasons mentioned above. Been 15 years+ and no issues so far.

Sent from my SM-G975U using Tapatalk

TKC 1800, Crystal Box Navy, Sprit 100g

Offline The0rigina1

  • Posts: 5
  • fumoposting
I never expected much variance between Progressive and regular springs.

But the difference between 68 P and regular 68 G springs on Ergo Clears seems pretty significant to me.

I can barely stand typing on regular 68 G stainless in MX Clear housing. But with 68P, it actuates around 45 G, near MX-Brown level, and bottoms-out closer to 68 G, allowing good functioning. 68 P felt bouncier and less rigid than 68 S.

My 63.5 P Ergo Clears are crazy-light at actuation, maybe lighter than MX Brown, but they seem to function fine with Clears.
Are you using 63.5p from sprit? According to my samples, they are only 13mm, one mm shorter than normal springs.
So the actuation point is almost at 1mm, which rendered them unusable for me.

Offline HungerMechanic

  • Posts: 821
Yes, so what I did in that case was purchased Ergo Clears made by someone else using 63.5 G Progressive Spirit springs, switches lightly lubed with 3203.

They work surprisingly well compared with what I expected, although they are light.

My keyboard is in the shop right now, but when it comes back I'll check the actuation feel. I don't recall having any functioning problems or other issues with the switches [aside from noise from insufficient lubing].

63.5P Spirit so far seems like a viable spring for very-light Ergo Clears. I'm going to continue testing them, they are like alternate Browns that could be made to sound better.

Offline Pylon

  • Thread Starter
  • Posts: 807
So I was sent samples of a spring that hasn't been released yet. I didn't post it out of confidentiality, but the person that sent me it ended up posting the redacted force curve anyways on a public Discord server, so I'll post it here. This springs has an ~4gf deviation from linear. I'm not sure if I'm at liberty to discuss the technical aspects of the spring so I won't for now. But it seems that progressives are possible.



This was measured in a Gateron Yellow with the leaf removed.
« Last Edit: Thu, 20 May 2021, 19:31:49 by Pylon »

Offline Ranger_Trivette

  • Posts: 88
  • Location: Italy
it doesn't seem soo progressive to me...


Offline Pylon

  • Thread Starter
  • Posts: 807
I mean angle and degrees is not a measure of spring rate, and changes depending on what aspect ratio I make the graph. But you're showing a ~2.4x change in spring rate. It's going from around 4gf/mm to roughly 10gf/mm.

Whether you can feel that is another matter.

Offline Faceman76

  • Posts: 76
  • Location: Long Island
it doesn't seem soo progressive to me...

Show Image

Wouldn't different scaling affect these angles?

Sent from my SM-G975U using Tapatalk

TKC 1800, Crystal Box Navy, Sprit 100g

Offline Ranger_Trivette

  • Posts: 88
  • Location: Italy
it doesn't seem soo progressive to me...

Show Image

Wouldn't different scaling affect these angles?

Sent from my SM-G975U using Tapatalk
Yes. But it doesn’t matter.
A graph with two different angles is not progressive.
Two springs combined, or a spring with two different pitch, give you a graph like that.
Not a progressive spring.

Offline Leslieann

  • * Elevated Elder
  • Posts: 3796
it doesn't seem soo progressive to me...
How much were you expecting on a spring that has only 15-20grams difference from start to finish? There's not a lot of room for a large curve.

A graph with two different angles is not progressive.
Two springs combined, or a spring with two different pitch, give you a graph like that.
Not a progressive spring.
That is exactly what a  progressive spring is.
Two different ratings, they change the angle depending on how close or how far apart the two springs are, the closer they are the more gentle the curve the further apart they are the sharper the curve.

Go look at car/motorcycle/mtn bike progressive springs (or wikipedia or any other encyclopedia about this topic), this is exactly how it works.
Novelkeys NK65AE w/62g Zilents/39g springs
More
62g Zilents/lubed/Novelkeys 39g springs, HK Gaming Thick PBT caps, Netdot Gen10 magnetic cable, pic
| Filco MJ2 L.E. Vortex Case, Jailhouse Blues, heavily customized
More
Vortex case squared up/blasted finish removed/custom feet/paint/winkey blockoff plate, HID Liberator, stainless steel universal plate, 3d printed adapters, Type C, Netdot Gen10 magnetic cable, foam sound dampened, HK Gaming Thick PBT caps (o-ringed), Cherry Jailhouse Blues w/lubed/clipped Cherry light springs, 40g actuation
| GMMK TKL
More
w/ Kailh Purple Pros/lubed/Novelkeys 39g springs, HK Gaming Thick PBT caps, Netdot Gen10 Magnetic cable
| PF65 3d printed 65% w/LCD and hot swap
More
Box Jades, Interchangeable trim, mini lcd, QMK, underglow, HK Gaming Thick PBT caps, O-rings, Netdot Gen10 magnetic cable, in progress link
| Magicforce 68
More
MF68 pcb, Outemu Blues, in progress
| YMDK75 Jail Housed Gateron Blues
More
J-spacers, YMDK Thick PBT, O-rings, SIP sockets
| KBT Race S L.E.
More
Ergo Clears, custom WASD caps
| Das Pro
More
Costar model with browns
| GH60
More
Cherry Blacks, custom 3d printed case
| Logitech Illumininated | IBM Model M (x2)

Offline Ranger_Trivette

  • Posts: 88
  • Location: Italy
a progressive spring has a curve as graph.
this is obtained moving gradually from a pitch P1 to another one, P2.

if you have half spring with a pitch P1 and the other pitch P2, you get two straight lines, with 2 angular coefficients, like the last graph.

(instead of changing the pitch you may use different diameter of the spring, sickness of the wire etc etc)

Offline Pylon

  • Thread Starter
  • Posts: 807
Just measured Gazzew's 55g, 65g, and 72g springs. The 72g is showing decent nonlinearity (~5-6gf maximum deviation from linear). The 55g and 65g are significantly less nonlinear, though it's still there.

Offline Ranger_Trivette

  • Posts: 88
  • Location: Italy
the 72gf look pretty not-linear! :thumb: :thumb: :thumb: