What is a good way to make highly stable freely moving armatures?

I'm thinking about freely moving armatures in general. I'm wondering what the best way is to make an armature that won't have play in it.

As a basic type of armature, let's say that we have two rods, six inches in length, with H and W both being 1/4". Let's say that in each rod, there is a hole that is drilled width-wise, 1/4" from the end. The rods are now aligned so that the hole of one rod is on top of the hole of another. A clevis pin with a head is inserted through the top rod, then through the bottom rod. Where the clevis pin exits the bottom rod, an e-clip retaining ring is slipped into a groove in the pin, snug up against the bottom rod.

In the above scenario, let's not assume any inherent properties of the assembly, only that it exists as described. In such an assembly, it seems that there are some tradeoffs. The less extra space there is between the rods and the pin, the less play there will be in the assembly. However, the less clearance there is between the rod and the bars, the more difficult the rods will be to rotate about the pin with respect to each other.

Similarly, the more snug the clevis pin is against the rod, the less play there will be, but the rods will be more difficult to turn.

As another variation, the rods could have a Width of 1" instead of 1/4". All else being equal, as a result of increased contact surface between the rods, there would be less play.

What I've described above is one of the most basic types of armatures I can think of. Variations can be made. Bearing raceways could be used. However, if this is done, wouldn't this add radial play to the joint?

Yet another option is, with or without the bearing,to make a ringed depression around the hole in one rod on the side facing the other rod. The second rod will have a ringed extension around the hole on it. This way, when the rods are put together, the ring on the one rod will extend into the depression of the other. It seems that this will limit radial play.

For general purpose building making armatures of a few inches, are there commercial solutions to this problem? If not, what are your thoughts on the methods that I mentioned? Are there any better ideas? Are there any guides that say, "For "x" type of metal, which expands "x" amount per "x" degree Fahrenheit,"x" amount of clearance should be used,"?

This isn't for a specific project. I'm just wondering about this in general.

Thanks.

If you are considering round rods rather than rectangular, then you want to build up the contact surfaces in some way, so that you've got plane mating surfaces.

Your e-clips should be sprung to take up assembly tolerances (equivalently, provide a bit of preload). Or stack a small spring washer (belleville, wave spring, etc) in there if you have to use flat e-clips.

If I were doing this, I would control radial play by fitting the clevis to the cross-holes, not attempting radial extensions and troughs. If I were attempting radial extensions and troughs, I would give them conical or V-shaped surfaces rather than cylindrical.

Well-made radial bearings should not affect radial play adversely. In fact, they would make it easier to take some play out, because you can make the rods, bearings, and clevis light interference fits, rather than requiring a slip/free fit between the clevis and the rod cross-holes. Also, they would wear better, so the initial play would last longer. Similarly, hardened-and-ground faces (or cheap sold-for-purpose thrust washers) and a thrust bearing would let you add some preload along the clevis axis, and thereby remove some play.

As for "x amount of clearance", consult standard engineering fit tables for direct clevis/cross-hole clearances. First, pick your desired class of fit, which can range from super-sloppy running fits to hard interference fits. In the ANSI world, this would be an RC1 through RC4 class of fit, depending on your application. If this is a sculpture armature, to be occasionally adjusted by hand, then RC1 or RC2 might be appropriate. If this is a four-bar mechanism to be constantly driven, then RC4 class of fit might be more appropriate. Then look up your desired diameter and read the clearance off the table. For face-to-face clearances, use a comparable value. If you are using bearings, use the bearing maker's recommendation for clearance (which will probably be negative, i.e., a slight preload or interference fit).

This sort of thing is considered a well-solved problem in precision mechanisms, and there are lots of off-the-shelf parts to help. If you want any sort of non-planar motion (which your OP more-or-less ruled out), spherical rod ends are the go-to class of part. For purely planar motion, you might consider pressing a miniature cam follower into one rod, and bolting the cam follower to the other rod. Again, pretty standard components.

If you are looking for a bearing that must have absolutely no play and near zero friction with control of degrees of freedom then an arrangement of flexures might be worth considering. These are used in some metrology instruments where that list of properies is very important. They are not usually suited to arrangements with large forces, but they can be made fairly robust. It is not possible to tell from your description whether they would be useful in your application.

Don't forget seemingly weird combinations of hinges at right angles to each other. these can provide very smooth operation and precision under high loads because the spread of the pins in the joints allows good load capacity with reasonable manufacturing tolerances and the 2 plane arrangement keeps the resulting motion perfectly linear.



Bosch 1-Inch Dual-Bevel Glide Miter Saw CM1GD Review | Workshop Addict - Tool Reviews

Thanks, I appreciate it.

How about an example of what your trying to build ?