Rolamite!

September 12, 2024 12:05 AM

Whatever happened to Rolamite? Invented in 1966 by Donald F. Wilkes, it looks like a simple gadget made with two rollers and a steel band, but it's much more. As basic as the wheel, the lever, or the hinge, it is the only simple machine discovered this century. Its use will be widespread --- in everything from switches, thermostats, and valves to pumps and clutches, and as almost frictionless bearings. More details. In 1969 the "First Symposium on Rolamite" was held (186 pages). How to build one. Two videos. Maybe rolamites' promise didn't live up the hype, but here's a scale, and from 2023: How To Clean a Mass Flow Controller Rolamite Valve.
posted by ShooBoo (14 comments total)

That's super interesting,, it feels like I could do fun stuff that otherwise requires axes ad stepper motors, I kinda want to build a 3d printer out of these. If part of the lack of uptake the wear and tear on the ribbon do you think?
posted by Iteki at 1:41 AM on September 12

The ribbon doesn't wear because there's no sliding contact with it, and doesn't tear because its composition, thickness and bend radii are chosen so as to keep all of the required deformation elastic rather than plastic. You probably could engineer in some planned obsolescence by using something woven instead of a thin sheet.

These are similar considerations to those associated with any other kind of flexure. Many compliant mechanisms actually end up outlasting those based on traditional mechanical joints.
posted by flabdablet at 4:28 AM on September 12 [4 favorites]

That just blew my damn mind.

Also, cried a bit inside seeing Physics Girl back when she was in better health
posted by JustSayNoDawg at 5:52 AM on September 12 [2 favorites]

I suspect they're not widely used because conventional bearings and slides are good enough enough for the majority of cases. You talk about the 3D printer case, and all of the extant systems are based on ball bearings, either in circular races on an tracked wheel, or recirculating races on a linear bearing on a rod. The motors involved are much stronger than the friction forces of fairly tight bearings, the number of moves over the lifetime of the device is relatively low so bearing wear is not important, and the need for precision movement with low runout is high. Commodity parts answer these design questions just fine.

I think today, especially with the availability of one-shot molded UHMW plastic bearings and flexures, there's not a lot of call for something like this. Which isn't to say they wouldn't be useful at all, it's just why we don't see them in our most common machines.

It is cool, though.
posted by seanmpuckett at 6:14 AM on September 12 [2 favorites]

Just to follow up, I've built a few linear motion machines -- pen plotters -- and working on a design that would use a rolamite for a linear bearing is like ... that just makes the machine more complicated. You need two very rigid parallel surfaces within which the rolamite can operate, that's going to have to be machined steel or extruded aluminum with axial clamping, you need a wide spring steel band, you need to capture the band so it doesn't fall out, you need rollers, you need a groove for the rollers so you don't get runout (or an extremely wide band??), you need a cage for the rollers so they don't fall out or get misaligned, and the cage has to attach to both sides of the rollers to prevent misalignment, you need to attach the rollers to the cage which requires a rotary bearing (ball bearing) anyway, and only now can you bolt the cage to your gantry and at this point I'm like. Fuck this.

I can stick a UHMW linear bearing on a rod and clamp it to my gantry or I can put ball bearings in a wheel in a v-slot and bolt the axle to my gantry. I don't need anything more complex than that.

It may be a simple machine, but there are simpler ones that provide adequate solutions in most cases.
posted by seanmpuckett at 6:29 AM on September 12 [2 favorites]

> ... you need a cage for the rollers so they don't fall out or get misaligned, and the cage has to attach to both sides of the rollers to prevent misalignment, you need to attach the rollers to the cage which requires a rotary bearing (ball bearing) anyway, and only now can you bolt the cage to your gantry and at this point I'm like. Fuck this.


"Simple machine" that needs to be embedded in a lot of expensive complexity to properly do it's thing.
posted by Aardvark Cheeselog at 7:32 AM on September 12 [1 favorite]

Though I should probably be on Team Rolamite (the patent is exactly the same age as I am), the internal rollers need to be quite large and the tape/ribbon relatively very thin to keep stresses low. Thus it will be quite bulky for a limited load, and simple slides can be smaller/stronger.

As the ribbon goes from outside curvature to inside curvature as the Rolamite moves, it also means that the material's surface will be alternating between compression and tension. This is right where fatigue cracking likes to lurk, and I don't see a way out of it through preloading.

I don't know if the Rolamite came out of missile or space research, but there were a whole bunch of inventions and materials that came out of the Space Race that work really well once, but fail badly through fatigue or reuse. Since you don't have to think about reusing most missiles, you can operate right at the edge of material properties. A couple of good material examples are Kevlar (amazingly, almost impossibly strong fibre if always kept in tension, but compress it and it crumples and crumbles) and Kapton (high temperature resistance, but flex it the wrong way and it's dust). Kevlar's crumbliness has become a virtue in "bullet proof" vests, in that it can diffuse a lot of kinetic energy once.
posted by scruss at 7:39 AM on September 12

I have this issue of Popular Science! I found it in my dad's stuff in the attic. 12 year old me was fascinated by the ads in the back: Build a Hover Car! Find Gold! MELT METAL!

I even bought plans for a one-man jet helicopter.

I remember at the time being interested by the Rolomite, but the hype in the article was pretty over the top considering the issue was 19 years old at the time and I'd never seen one of these things.
posted by Horkus at 8:42 AM on September 12

Thanks for the follow-up info, good explanations! I don't care, I love it and I'm gonna make something with it! I think I can make a cheap solonoid/linear actuator out of it!! Like if I capture the one roller it should move the cage right?
posted by Iteki at 12:24 PM on September 12

If a precisely positionable, backlash-free, low-friction, high endurance solenoid replacement is what you need, you could do a lot worse than the stepper motor plus wrapped taut band mechanism that used to be commonly used for head positioning in 5.25" floppy disk drives.

It's pretty much half of a Rolamite: leave out one of the rollers and one of the rails, fix the centre of the ribbon to the remaining roller (actually a capstan driven by the stepper motor) with a screw and Loctite, split the ribbon in two on one side of that fixing screw, and spread the halves to leave a lengthwise gap that the unsplit ribbon on the other side will fit in as it wraps around the capstan.

The circumference of the capstan, less a little bit so that the ribbon fixing screw never bumps into the driven flat rail, sets the full throw of the actuator.
posted by flabdablet at 12:40 PM on September 12 [1 favorite]

you need a groove for the rollers so you don't get runout

If I understand "runout" correctly, I think you could use a crowned pulley shape?
posted by pwnguin at 2:08 PM on September 12

This feels like something Rusty Venture would talk about during a tour of Venture Industries.
posted by RakDaddy at 3:54 PM on September 12 [1 favorite]

I don't think crowned rollers would work in a Rolamite because the ribbon has no stretch in it by design. Even if it did, I think the forces that induce a belt to climb to the crown of a fixed roller, when translated to the Rolamite context, would just end up misaligning the floating rollers and ejecting them sideways. But intuition around belts is tricky and I could easily be wrong.
posted by flabdablet at 1:52 AM on September 13

Worth mentioning perhaps that the wrapped band mechanism (e.g. in the floppy drive) relies on the linear driven part to be constrained in all but one dimension with low friction rails and uhmw bushings so it doesn't skew or flop around, and the capstan is constrained in every way except rotation by two sets of ball bearings in the motor assembly. I'm not saying it's bad, just that these band-based mechanisms need significant external support and constraint to function reliably.

The wrapped band technique is used in some large scale linear motion devices -- with flexible toothed belts. You put your motor on the traveling part, use idler pulleys to trap the toothed belt around a toothed drive pulley on the motor. Actuate the motor and it will pull the machine along the belt. Essentially a rack-and-pinion but the rack is a belt, not a solid part.

(More commonly, the motor is stationary and the toothed belt is most of a loop attached to the moving part. The traveling motor solution is typically only used when the range of motion is large, you don't mind having the weight of your motor coming along, and it's relatively easy to get power to the motor either through a cable chain or on-board power supply.)
posted by seanmpuckett at 5:57 AM on September 13 [2 favorites]