Does anyone else obscess over switch weight?

[jcoffin1981]

I like to replace the springs on a new build.  I am opening the switches anyway, and replacing the springs they are more consistent from switch to switch and you choose the spring weight.  My current build is a DROP Shift, Gateron RGB, DROP skylight keycaps, and 63g aftermarket springs.   I typically get 60 or 62g springs and it is difficult to tell the difference between them.  My most recent order they were out of 60g and 62g so I chose 63g springs. 

After assembling it and using it they are noticeably heavier.  While a change is welcome (I'm sure many of us switch between boards for a change), I find my fingers are fatiguing fast.  Using nickels and pennies and a gram scale I can measure the force it takes to actuate the switches, and the previous one (same switches but 60g springs) take 50 grams of force to actuate the switches, while these take 53 grams of force.  You would not think that such a subtle difference would be noticeable, but it most definitely is. 

Does anybody else obsess over small details like this?  Do you find that your fingers adjust to the added weight?  I have always preferred lighter switches, typically "brown" tactiles.  This keyboard is for work and I spend several hours a day typing.  My current work board has Cherry Browns with 60g springs.  Fortunately with this board the connections are "hot swap," so I don't have to break out the soldering iron and I can swap them out quicker.

[Leslieann]

You will adjust, it will take a few days to a few weeks depending on how steep the change is.


As for obsessing over switch weights, yes and no...

Force curves matter more than the bottom out rating itself.
(Note these numbers are not accurate, except where noted near the end. (This took too long as it was, I didn't need to go looking for charts)

Let's start with a linear because it will show what I mean the most.
You can have two identical switches  but two different 62g bottom out springs, one may start at 25g the other may start at 30g, this means you could have one switch with a 45g actuation point and another with a 40g actuation point. Same switch, same spring rating, just a different curve and some have a wilder curve than this,

So start/actuation/bottom out could look like:
25g - 40g - 62g
35g - 45g - 62g


Now take a 60g spring and a 63g spring, both with different curves and the difference at actuation point could be as much a 10-15g difference.
start/activation/bottom out could look like:

23g - 38g - 60g
35g - 48g - 63g

While the spring only shows a 3g difference, your fingers are seeing a 5-10g difference where it matters.


Now start with a very soft 60g spring and a rather flat 63g spring
So start/activation/bottom out could look like:
20g - 35g - 60g
40g - 55g - 63g

Now you're hit with a 20g difference at activation.
With one nickle being 5grams, 20g is a massive change.


Now throw tactile bumps into the mix...
How a switch is built decides how the tactile bump changes with the spring, most switches have a set tactile bump, for example browns have about a 5g bump over spring pressure at activation, doesn't matter if it's a 50g spring or an 62g spring the bump will be (about) 5g over whatever the spring pressure is at the actuation point.

So start/bump/falloff/bottom out could look like:
25g - 35g - 30g - 50g
30g - 45g - 35g - 62g

Others switches have a set force, for example 62g Zeals, the tactile bump is relatively locked at 45g (give or take a few grams). So using a lighter the spring, in this case a 39g spring results in the bump growing relative to the spring pressure at activation, instead of the default 8g bump(?) it had with a 62g spring it now has a 25g bump.

So start/bump/falloff/bottom out would look like (note these are semi-accurate numbers!):
30g - 45g - 35g - 62g
15g - 45g - 20g - 39g

In this case, the tactile bump goes from hitting a speed bump to hitting a brick wall and then falling off a cliff.*
I haven't tested it, but I suspect the opposite would happen with a heavier spring, the bump gets washed out more and more.



*Warning
While an interesting combo, do not copy this, it's an example of extremism and it will bite you. While I do successfully use this same NovelkeysOriginative/Sprit 39g spring on several other switches (with varying degrees of effort) the Zeal 62g bump is too sharp for the spring to handle and even though they're lubed and I spent tons of effort to break them in as well as over a year (or 2?) use some switches still fail to reset on slow release. Even just a tiny misalignment of the plate can cause them to bind, seriously, I've spent a ton of time and effort trying to make them function perfect with only some success. I wouldn't recommend going below 45g (would be very difficult still) and would even recommend a 50g minimum spring rate to be reliable, even then you will want an extremely light grease or just go with an oil.


Edit: I said they were Novelkeys springs when they were actually Originative.

[StefanVoda]

Are these springs longer than the other ones you have used in the past?

[jcoffin1981]

You will adjust, it will take a few days to a few weeks depending on how steep the change is.


As for obsessing over switch weights, yes and no...

Force curves matter more than the bottom out rating itself.
(Note these numbers are not accurate, except where noted near the end. (This took too long as it was, I didn't need to go looking for charts)

Let's start with a linear because it will show what I mean the most.
You can have two identical switches  but two different 62g bottom out springs, one may start at 25g the other may start at 30g, this means you could have one switch with a 45g actuation point and another with a 40g actuation point. Same switch, same spring rating, just a different curve and some have a wilder curve than this,

So start/actuation/bottom out could look like:
25g - 40g - 62g
35g - 45g - 62g


Now take a 60g spring and a 63g spring, both with different curves and the difference at actuation point could be as much a 10-15g difference.
start/activation/bottom out could look like:

23g - 38g - 60g
35g - 48g - 63g

While the spring only shows a 3g difference, your fingers are seeing a 5-10g difference where it matters.


Now start with a very soft 60g spring and a rather flat 63g spring
So start/activation/bottom out could look like:
20g - 35g - 60g
40g - 55g - 63g

Now you're hit with a 20g difference at activation.
With one nickle being 5grams, 20g is a massive change.


Now throw tactile bumps into the mix...
How a switch is built decides how the tactile bump changes with the spring, most switches have a set tactile bump, for example browns have about a 5g bump over spring pressure at activation, doesn't matter if it's a 50g spring or an 62g spring the bump will be (about) 5g over whatever the spring pressure is at the actuation point.

So start/bump/falloff/bottom out could look like:
25g - 35g - 30g - 50g
30g - 45g - 35g - 62g

Others switches have a set force, for example 62g Zeals, the tactile bump is relatively locked at 45g (give or take a few grams). So using a lighter the spring, in this case a 39g spring results in the bump growing relative to the spring pressure at activation, instead of the default 8g bump(?) it had with a 62g spring it now has a 25g bump.

So start/bump/falloff/bottom out would look like (note these are semi-accurate numbers!):
30g - 45g - 35g - 62g
15g - 45g - 20g - 39g

In this case, the tactile bump goes from hitting a speed bump to hitting a brick wall and then falling off a cliff.*
I haven't tested it, but I suspect the opposite would happen with a heavier spring, the bump gets washed out more and more.



*Warning
While an interesting combo, do not copy this, it's an example of extremism and it will bite you. While I do successfully use this same Novelkeys/Sprit 39g spring on several other switches (with varying degrees of effort) the Zeal 62g bump is too sharp for the spring to handle and even though they're lubed and I spent tons of effort to break them in as well as over a year (or 2?) use some switches still fail to reset on slow release. Even just a tiny misalignment of the plate can cause them to bind, seriously, I've spent a ton of time and effort trying to make them function perfect with only some success. I wouldn't recommend going below 45g (would be very difficult still) and would even recommend a 50g minimum spring rate to be reliable, even then you will want an extremely light grease or just go with an oil.

Thank you for that detailed reply.  I had not even considered the entire force curve.  These springs are Upgrade Keyboards branded (not the previously used Sprit) springs and I just assumed the curve would be the same since they are supposed to be same as stock MX springs.   I kind of thought that it was +3 grams over the entire force curve, but this definitely not the case.  It is not just fatigue, but I have pains in my wrist and forearms and have to stop typing.  So as time consuming as it is I have to swap out the springs again. 

Edit: trying these out side by side, there is a considerable amount of difference in the resistance of that first 1.5- 2mm up until the tactile event.  There is a LOT more force required to reach it with the 63g springs vs the 60 and 62g springs.  You are correct Leslieann. 

[jcoffin1981]

Are these springs longer than the other ones you have used in the past?

The previous Sprit springs are 14mm.  These are 15mm.  The stock Gateron Brown springs they replace are also 15mm.

[Leslieann]

Thank you for that detailed reply.

Glad it helped and that you found the culprit.
Luckily, like you said, it is hot swap.

Gotta watch those force curves or they will bite you, companies really need to publish that info.
I don't really care what the bottom out is, it's bottomed out, while handy for identifying springs it only tells you the pressure at the absolutely most useless point in travel.

[jcoffin1981]

After using the heavier springs for a couple days and then returning to the lighter ones they feel incredibly light.  Like, too light; and it felt good to get back on the heavier board.  It's kind of like when you swing a heavy baseball bat and go back to the lighter one they feel even lighter.  Some baseball players do this to warm up.  I do think that they are in fact too heavy.  But, I do want to experiment with heavier springs.  The problem is that it gets expensive as well as time consuming to just try different springs.  I'd like to try springs that are just a little heavier.  Obviously comparing brands can be apples to oranges.  Where is it that these force curves are published?  This might be able to help me make a decision- like you said the bottom out force alone is pretty pointless.  If I could add just a few grams to the entire force curve that would be perfect. 

I think that my typing may actually be a little more accurate with the heavier springs.

[Leslieann]

A correction
I said they were Novelkeys springs I use when they were in fact Originative Sprit. It's corrected above as well.


I don't know of any that publish the force charts, only some switch makers seem to. Sprit explains their force curves, but no actual numbers.

As for prices, they're all over the place and it's easy to overpay.
Novelkeys has about the cheapest I've seen lately but very limited selection. I'd go this route for anyone just wanting "stiffer" or "softer", but avoid if you want to match something you have.
Originative has a decent selection of (mostly basic) Sprit for about the cheapest you'll find. This is my go-to, I don't care to get too fancy.
Sprit has a list of other vendors, some with massive selections, but expect to pay more. Note that Sprit doesn't like selling direct, so unless you want tons forget buying direct from them.

I try to wait for Black Friday sales.

BTW, I too seem to type better on stiffer when I go back but I think it's a temporary effect due to the extra force slowing you down, once you get adjusted your accuracy goes right back to what it was. Also on some switch/spring combos I needed stiffer springs in the modifiers to keep from activating them on accident but this is only on super light springs and light tactility.

[Rhienfo]

Force curves matter more than the bottom out rating itself.

You can have two identical switches  but two different 62g bottom out springs, one may start at 25g the other may start at 30g, this means you could have one switch with a 45g actuation point and another with a 40g actuation point. Same switch, same spring rating, just a different curve and some have a wilder curve than this,

So start/actuation/bottom out could look like:
25g - 40g - 62g
35g - 45g - 62g

So does that mean that when companies like tx are talking about consistency, are they refering to actuation along side the force that bottoms out the switches (like a 62g spring having +-1g of variance)

Also this is brought up in the thread but I'll mention it here. Sprit is a scammer who was (and I'm pretty sure still is) banned from selling on geekhack. If you do truly want sprit springs for whatever reason (if your very particular about weighting of a spring that sprit has and no other company does or you prefer "slow springs" or things like that as I know that sprit is the only person who does springs like that, which are also consistent) only buy from a reputable vendor as they should have them in stock.

That's the reason why I don't by any sprit products personally. I mainly buy from TX who from my knowledge have had a pretty good track record when it comes to delivering projects on time and their springs are good and consistent from my end. I've also heard good things about geon springs as well.

[jcoffin1981]

Force curves matter more than the bottom out rating itself.

You can have two identical switches  but two different 62g bottom out springs, one may start at 25g the other may start at 30g, this means you could have one switch with a 45g actuation point and another with a 40g actuation point. Same switch, same spring rating, just a different curve and some have a wilder curve than this,

So start/actuation/bottom out could look like:
25g - 40g - 62g
35g - 45g - 62g

So does that mean that when companies like tx are talking about consistency, are they refering to actuation along side the force that bottoms out the switches (like a 62g spring having +-1g of variance)

Also this is brought up in the thread but I'll mention it here. Sprit is a scammer who was (and I'm pretty sure still is) banned from selling on geekhack. If you do truly want sprit springs for whatever reason (if your very particular about weighting of a spring that sprit has and no other company does or you prefer "slow springs" or things like that as I know that sprit is the only person who does springs like that, which are also consistent) only buy from a reputable vendor as they should have them in stock.

That's the reason why I don't by any sprit products personally. I mainly buy from TX who from my knowledge have had a pretty good track record when it comes to delivering projects on time and their springs are good and consistent from my end. I've also heard good things about geon springs as well.

I have read about Sprit's unethical business actions and don't want to support him/them.  The fact is though I like their springs.  I have purchased other springs (perfect example here) and been unhappy with them.  Purchases I make are from reputable vendors. 
If you have a recommendation for a spring to purchase I will check it out.

As far as "variance," I don't know what part of the force curve they are referring to- but I think it's bottom out force.  They are quoting +/- 0.3 g variance in their most recent spring springs.  They also have their economy line that are less money, but +/- 1g. Many switch makers will quote +/- 5 g variance between switches.  How much of that is in the switch vs the spring I don't know, but they are mass producing springs so I'm sure close to half of that could be from the spring.

I'm not sure how much the length plays a part in this.  The original Sprit Spring was 14mm and 62 bottom out force.  The other Upgrade spring that was a standard MX spring at 63mm was 15mm in length.  I don't know how much that additional mm equates to pre-load at the very top of the switch press making it feel much heavier.

[Leslieann]

So does that mean that when companies like tx are talking about consistency, are they refering to actuation along side the force that bottoms out the switches (like a 62g spring having +-1g of variance)

I suspect that has nothing to do with the curve but the primary rating (bottom out) is within 1%.

Also this is brought up in the thread but I'll mention it here. Sprit is a scammer who was (and I'm pretty sure still is) banned from selling on geekhack.

While buying them from Originative means they get my money, they get less and a vendor is far more likely going to bring legal action than an individual user keeping them at least somewhat in check.
Not excusing it, but if I stop buying from every company that's ticked me off at any time in the past I wouldn't be able to buy much of anything at all.


I always forget TX springs and that's probably because none of my usual vendors carry them (also no 39g!).
Seems they can be found even cheaper which is nice but now I gotta find a vendor I can trust I guess.

[Rhienfo]

I have read about Sprit's unethical business actions and don't want to support him/them.  The fact is though I like their springs.  I have purchased other springs (perfect example here) and been unhappy with them.  Purchases I make are from reputable vendors. 
If you have a recommendation for a spring to purchase I will check it out.

Yeah I understand that, as I said there are valid reasons for buying from them, like I can't blame you for that. As I said I hear geon springs are good (and I think they might be the cheapest "consistent" springs which is cool) but sadly there aren't tons of options for good aftermarket springs.

Many switch makers will quote +/- 5 g variance between switches.  How much of that is in the switch vs the spring I don't know, but they are mass producing springs so I'm sure close to half of that could be from the spring.

It's still noticeable, but they are quite useable, just not really consistent like sprit or tx in terms of bottom out. It's definitely better than stock cherry, kalih and jwk switches, which have really inconsistent springs (they sometimes feel like +/- 15g variance)

I'm not sure how much the length plays a part in this.  The original Sprit Spring was 14mm and 62 bottom out force.  The other Upgrade spring that was a standard MX spring at 63mm was 15mm in length.  I don't know how much that additional mm equates to pre-load at the very top of the switch press making it feel much heavier.

Yeah length does make a decently big difference, I'm not too well verse in spring physics, but it wouldn't be as massive (unless we're talking 18-22mm which some brands sell) but would be noticeable to the point where you prefer a length. Tx do sell 14mm springs so if you haven't tried they might be much better for you (geon is also 14.5mm so that might be a turn off for you)

I always forget TX springs and that's probably because none of my usual vendors carry them (also no 39g!).
Seems they can be found even cheaper which is nice but now I gotta find a vendor I can trust I guess.

I'm lucky that daily clack sells them in Australia, the only issue that they are usually sold out in the weight and length that that I really like (62g 15mm), but I do trust dc and from my experience I had no issues that weren't resolved very quickly.

[Leslieann]

Many switch makers will quote +/- 5 g variance between switches.  How much of that is in the switch vs the spring I don't know, but they are mass producing springs so I'm sure close to half of that could be from the spring.

5 grams is a massive amount and on the average switch that represents an 8-12% variable, that is an embarrassing amount of variance in such a precision product, especially one with so few parts.

Springs are (relatively) simple to make, yes, you have metallurgy involved but that's not a big deal these days if your material comes from a reliable source, after that it's just math. Switches on the other hand can come from multiple molds for each part, each part can be from different plastic batches but we can control this pretty well really and you have tolerances just for that reason.  All of this is automated and easily tested (same for springs), you can monitor how molds wear, you can easily test a sample from a batch and get a good idea about how things stand so this should be pretty consistent, it may be a consistent downward slide as molds wear but barring a serious problem/failure it's going to be pretty consistent from batch to batch and switch to switch.

Lube on the other hand can be used to control or fine tune costs and is easily variable, need to make a few extra pennies use less, want a more refined feel, use more. Running low, stretch it out. How much and how it's applied can change a lot about a switch and it's precisely why on a high end board you rip them apart and apply it yourself. We do this with precision so that you do have consistency and not just some randomly placed blob you hope will eventually reach all the right places.

[jcoffin1981]

Many switch makers will quote +/- 5 g variance between switches.  How much of that is in the switch vs the spring I don't know, but they are mass producing springs so I'm sure close to half of that could be from the spring.

5 grams is a massive amount and on the average switch that represents an 8-12% variable, that is an embarrassing amount of variance in such a precision product, especially one with so few parts.

Springs are (relatively) simple to make, yes, you have metallurgy involved but that's not a big deal these days if your material comes from a reliable source, after that it's just math. Switches on the other hand can come from multiple molds for each part, each part can be from different plastic batches but we can control this pretty well really and you have tolerances just for that reason.  All of this is automated and easily tested (same for springs), you can monitor how molds wear, you can easily test a sample from a batch and get a good idea about how things stand so this should be pretty consistent, it may be a consistent downward slide as molds wear but barring a serious problem/failure it's going to be pretty consistent from batch to batch and switch to switch.

Lube on the other hand can be used to control or fine tune costs and is easily variable, need to make a few extra pennies use less, want a more refined feel, use more. Running low, stretch it out. How much and how it's applied can change a lot about a switch and it's precisely why on a high end board you rip them apart and apply it yourself. We do this with precision so that you do have consistency and not just some randomly placed blob you hope will eventually reach all the right places.

I think that metal leaves are a big part of this variance.  The switches get knocked around quite a bit, and even the slightest twist or bend can seriouisly change the tactility and resistance of the switch. 

[Leslieann]

I think that metal leaves are a big part of this variance.  The switches get knocked around quite a bit, and even the slightest twist or bend can seriouisly change the tactility and resistance of the switch.

I've seen that on Matias, but nothing else.

I still think that's an insane amount of variance.

[Rhienfo]

Many switch makers will quote +/- 5 g variance between switches.  How much of that is in the switch vs the spring I don't know, but they are mass producing springs so I'm sure close to half of that could be from the spring.

5 grams is a massive amount and on the average switch that represents an 8-12% variable, that is an embarrassing amount of variance in such a precision product, especially one with so few parts.

Springs are (relatively) simple to make, yes, you have metallurgy involved but that's not a big deal these days if your material comes from a reliable source, after that it's just math. Switches on the other hand can come from multiple molds for each part, each part can be from different plastic batches but we can control this pretty well really and you have tolerances just for that reason.  All of this is automated and easily tested (same for springs), you can monitor how molds wear, you can easily test a sample from a batch and get a good idea about how things stand so this should be pretty consistent, it may be a consistent downward slide as molds wear but barring a serious problem/failure it's going to be pretty consistent from batch to batch and switch to switch.

Lube on the other hand can be used to control or fine tune costs and is easily variable, need to make a few extra pennies use less, want a more refined feel, use more. Running low, stretch it out. How much and how it's applied can change a lot about a switch and it's precisely why on a high end board you rip them apart and apply it yourself. We do this with precision so that you do have consistency and not just some randomly placed blob you hope will eventually reach all the right places.

I think that metal leaves are a big part of this variance.  The switches get knocked around quite a bit, and even the slightest twist or bend can seriouisly change the tactility and resistance of the switch.

From my experience when it comes to mx tactiles, the leaves probably played a very minor role when it comes to changing the tactility. When I fixed up some of the tactiles to make the bump consistent across all of the switches, the two things that I saw impacted the most was the inconsistency of the springs (I changed the stock springs with tx springs and it was better) and inconsistent lubing (these were some of the first switches I lubed, so some of the legs were lubed, which made the tactility inconsistent so I tried to get rid of it and worked fine as well)

[mathisart]

I'm interested in switch weight. Also travel distance. Also switch spacing. I think 35gf is too much weight. (19.05mm spacing is also way too much btw). I still don't know what weight I prefer.