[IC] sorbothane (ULTRA-soft polyurethane) landing pads!

[Melvang]

How is the friction on this stuff?  I was thinking about using this stuff under the feet on my brushed behemoth.  The one problem is there is only going to be 2 square inches at most across all 4 feet, and with as much as this thing is going to weigh.

[mkawa]

your board will actually damage 1/10" or 1mm sheeting. you're going to want much thicker sheet. all duros (30, 40, 70 OO) are sticky and readily attach to smooth surfaces via stiction. they can then be held in place with small retainers on the corners, etc. however, you might want to look at the factory-produced pedestal mount parts, which have thicknesses of 1/4" up to > 1" in cylindrical form, instead of a pad. for your application they will be not only easier to mount, but they will probably be the only way to mount your board on sorbo without damaging it via compression set.

[Merloobi]

Interested! 

[MVPness]

I'm definitely interested in this.

[mkawa]

I’m curious whether sorbothane landing pads will work for Alps switches, and if so whether their effect is pleasant. So I bought a couple sheets of 1mm sorbothane from ebay. We'll see what can be done with the stuff cutting out shapes w/ a utility knife, in a few days.

this is basically what i'm doing at the moment. sorbothane is very visco-elastic. this is like putting more resilient and way denser memory foam in your keycaps. one has to be careful in designing the solid for the load and frequency or else you're either going to damage the sorbothane and render it useless, or it's just going to feel like crap, because it's always compressed and not absorbing any of the power you're putting into the keycap (which is the whole point here)

[jacobolus]

I bought a couple sheets of 1mm sorbothane from ebay. We'll see what can be done with the stuff cutting out shapes w/ a utility knife, [...]

this is basically what i'm doing at the moment. sorbothane is very visco-elastic. this is like putting more resilient and way denser memory foam in your keycaps. one has to be careful in designing the solid for the load and frequency [...]

What kind of shape do you cut, and where do you put it? I was thinking of trying to stick it on the top of the switch housing, around the slider, so that as a keycap is depressed it hits that (instead of the keycap hitting the housing top, or the slider hitting the housing bottom). Different keycaps are shaped differently on the inside/bottom, so it may work better with some keycaps than others. I’ll have to experiment a bit.

[mkawa]

if you read the design guidelines and ask some questions, what you'll find out is that you want as much surface area as possible that can "squish out". this means a couple things i hadn't expected:

1) you don't want a hole that's a close diameter to the MX stem. you actually want to make the inner hole significantly wider than the stem.

2) you want to use jagged cuts to your advantage.

3) you don't want to force the edges of the design into the edges of the inner keycap.

i've been playing around with pieces that go in the keycap as well as ones that go around the stem. there isn't a huge difference.. if the keycap internal surface is ribbed, you're just going to get a much different experience than if the keycap isn't ribbed. that's really where the difference is. this is true regardless of whether it's a stem mount or a keycap mount.

i do have an idea that may have some merit that is similar to the "trampoline" thing some people are playing around with, but haven't tried it yet. need some pin punches hmmmmmmm

[n0rvig]

if you read the design guidelines and ask some questions, what you'll find out is that you want as much surface area as possible that can "squish out". this means a couple things i hadn't expected:

[..redacted..]

i do have an idea that may have some merit that is similar to the "trampoline" thing some people are playing around with, but haven't tried it yet. need some pin punches hmmmmmmm

Yes! Keep up the crazy mad scientist sobothane research Kawa!

[mkawa]

anyone have a meat grinder or knowledge of general properties of threshing tools for soft materials?

[jacobolus]

mkawa: do you have any pictures of the shapes you’ve tried, w/ the sorbothane landing pads, ideally showing where/how you installed them, and some description of the effect?

[mkawa]

no, in fact i just tossed all my experimental parts into the trash yesterday. this followed a trip through the vitamix, in which the stuff just kind of humorously bounced around. the closest thing i've gotten to feeling reasonable is a 0.1x0.5x0.5" square with a hole in the middle and the edges sliced way down, all with craft knives and a steel probe.

the problem with this unit was that the helper spring kicked in almost immediately, as one can't make it any thinner than 2.5mm. it's ok if you're fine with almost scissor-switch like travel and then a very progressive damper at the bottom, but i wasn't personally crazy about it.

the 1mm thick material is almost impossible to design with. with thin material, what you want is really granular particles, almost like activated carbon. hence the blender experiment yesterday. this is because the guiding metric behind sorbo designs is, taking total surface area, loaded area (take one, cross-section, either top or bottom of the part) divided by unloaded area. the 1mm flat sheet, if cut straight, has 1xn mm unloaded area where n is the periphery of the part. so, if we design a standard washer with anything more than 1mm between ID and OD, our shape ratios (aforementioned guiding metric) become extremely high extremely quickly. high shape ratios are bad. low shape ratios are also bad. you ideally want between 1 and 1.3.

second problem. die cutting washers this small is extremely difficult. you're asking your diecutter for 1 thousandths tolerance basically, and that's like... most waterjets can't even do that, and your die is going to be really expensive on top of getting the die-cutter to hold the machine tolerance of the die.

a thresher that can process material and give it an average RMS diameter of like 100 mics (and we don't even care about tolerance as long as the particles are sufficiently granulated) will give you something you can place inside a switch, and the loose fill will keep the shape ratios of each particle about right. (again RMS average)

a meat grinder is a thresher with variable sized dies. you will need a strong traction motor (ie, your elbows + grease), but you should be able to machine a nice small diameter small particle die so that you can make a whole bunch of particulate matter in a few hours. that said, i have no other use of a meat grinder, and i have enough doubt that this will work to make this kind of ehhhh to me.

[jacobolus]

Hm. What if you glue a few little 1mm * 3mm * 3mm squares [or something] onto the cap bottom / switch housing top?

Edit: or better would be like a 3mm radius circle I guess (or slightly less). Then that would be top area of pi * 9mm^2, and side area of pi * 6mm^2, so that's 4/3 = ~1.3.

[mkawa]

then the entire surface becomes loaded and your shape ratio goes to infinity. das is bad.

[jacobolus]

I don’t quite follow. What are we trying to look for here? Percentage of the whole 3d shape that’s loaded? I thought you were talking about the ratio of top surface to side surfaces.

Doesn’t the amount of the load matter too?

[mkawa]

my understanding of the shape ratio is that it's a measure of compressive load to surface area, so if you have compressive stress on the object, you want to take the total area of compression (two sided) then divide by two, to be a bit more accurate. one can make the simplifying assumption though (note that shape ratio is totally a first order appx guideline) that all loaded area on one face is matched by loading on the geometrically opposite face. hence, take top of bottom.

because it's non-newtonian everywhere, then by definition there is no part of the stress/strain graph in which it can be linearly approximated, so rough guidelines are the word of the day.

and yah, the load and resonant frequency vs loading frequency are extremely important. the first of the first order calcs though is shape ratio. if your shape ratio is horrible, your design is bad and you're either not going to load the material at all and will see no benefit from it or you're going to destroy the stuff.

[jacobolus]

Did you try cutting shapes out of the 1mm stuff, and using them for landing pads? What was the result? [Or were your tests all inside the switch / with the 2.5mm sheets?]

[mkawa]

i cut a bit of 1mm off and put it in various places but was generally unimpressed by by attempts. the other thing is that some keys are ribbed internally, and actually underload the material. with the 2.5mm, this didn't matter, the sorbo would fill the key regardless. with the 1mm you had a choice between very little springiness and absolutely no springiness :|

[mkawa]

so, i don't see a way to get a pad into the space between the switcha and keycap. there just isn't enough room. however, a pcb mounted board with a large patterned pad that had isolated hemispheres until each key would work quite well.

[mkawa]

if someone is really interested in making this happen, has a copy of illustrator and can draw up a DWG with it against a phantom and gh60 board, i'm willing to make some prototypes. i have a pretty good idea of how to design for the material now, but the upshot is that you need to to take a relatively thick sheet and waterjet cut it so that it sits behind a (preferably pcb-mount) board's pcb. the non-conservative spring-like action will then happen between pcb and case. basically, you'll get very little resistance from the sorbothane as you press down on the keys, but when you bottom out, you will get a large helper spring effect both pushing back on your finger and absorbing the impact of your finger on the bottom of the switch. further, when you release the keyswitch, the sorbothane (with an appropriate amount of preload) will dampen any resonances in the switch/pcb/plate assembly. it will work especially well with PCB mounted switches and with switches plated in aluminum. some careful tuning will need to be done with regard to disc size under each switch, and switch thickness, however, but the effect should be quite interesting, certainly not like any switch-feel modifier people have used to date..

[jacobolus]

What have you actually tried here, mkawa? I’m a bit skeptical that putting something below the pcb (as compared to under the keycap or inside the switch) is going to have a huge impact on tactile feel, though I could definitely see it cutting out the noise.

In about 2 months, once I get back from a trip, I’m planning to join the SF TechShop and try CNC cutting some combined case/plates out of wood, to use with direct wiring. The idea is: take a solid piece of wood, as thick as you’d typically have for plate + pcb + the space in between, and then cut out the shape of the switch, such that it sits on top of a thin bit of wood at the bottom with a couple small holes for the leads, and the whole bottom half of the switch fits snugly encased (the exact 3-D shape to cut out needs to be figured out for each type of switch).

I’m curious if, in such a setup, a bit of sorbothane between the wooden shelf (standing in for a PCB) and the switch would have a useful effect here on noise/vibration damping. Then again, maybe there won’t be so much resonance in the wood slab to begin with.

[mkawa]

hardwood's going to be pretty dead as it is, but it definitely depends on thickness, finish, grain, etc.

done correctly, sorbothane can make things _extremely_ dead as far as vibration is concerned, and it can act as a spring when needed. the problem is that the spring constant (over its tiny tiny linear range) is due to the "bulging" of the material. you're displacing very dense, gel-like polyurethane, then, because it has very strong memory (unlike most polyurethanes), it pushes back. this means that very thin samples of material are pretty much useless, because they can't bulge. it also means you can't just put a mat down under something to dampen vibration unless that mat is much more of a cube than pretty much any mat is. i've worked with quite a bit of sorbothane bulk material. some of it has been effective, some of it has been damaged by me, and some of it has been for applications you normally wouldn't even consider offhand.

for the switches, i tried pretty much every variation on an o-ring or pad with a hole in it that i could think of. again, cubes work great, but they're huge, and cut stroke down to miniscule amounts. at 30 OO with room to bulge, there's still some travel from the spring action of the sorbo, but it was not something that i think anyone would like to type on. however, if you look at the keyboard assembly globally, which has impact frequency and location, things get a lot easier to manage. want room to bulge as a helper spring? make an appropriate series of cuts. want to just dampen? make a different series of cuts.

again, to design parts out of sorbo, you need do a fair amount of thinking about its spatial geometry both stressed and unstressed. the situation inside a keycap or switch is pretty much terrible. outside, things get a lot better.