Encoder placements (more wild thoughts bout encoreds)

[Konstantin]

So if the encoder will register actual table/head movement instead of ballscrew/motor spindle movement will it give more pricise positioning?

Imagine the encoder pressed against the table with a rubber/whatever wheel.
Something like this

[HuFlungDung]

The problem with that is the loss of resolution, that is if you can reliably obtain an absolutely slip free surface to surface connection.

If the encoder is on the screw, the overall accuracy is a factor of the pitch of the thread. But, running direct on a surface, let's say a 1" diameter wheel would travel Pi inches per revolution, so each division of a 1000 line encoder would pan out to be .003". If the encoder was mounted on a 5 tpi screw, the resolution would be .0002.

Also, the inertia of the spinning disk of the encoder could itself could lead to loss of position, if movements are fast.

[Chagrin]

Cut a timing belt and lay it flat on the surface and put a timing wheel (pulley) on the encoder and you'll have more precision.

[Bloy2004]

I agree with Huffed..
You'd be better off using the "protected" rack and pinion from a long mechanical calipers than the unreliable friction wheel.....assuming good contact, one bit of debris passing under the wheel and right away you're off. And then there's the compression factor of the wheel causing constant deviation and multiple calibrations. Seems like too much work to set up when so many replaceable "cheap" digital scales are available.
I guess it all depends on how much accuracy/precision and total travel distance you require.

[Bloy2004]

I wild thought I often entertain is the use of a laser pen type system. The beam bouncing back to a receiver that locates according to the time lapse interval ....but the receiver would have to measure the speed of light!
Please elaborate on this by shooting it down so I can clear my mind of such entertaining thoughts.

[Stevie]

This is the same as the old travadial system; it's not new; and believe me if there is enough pressure against the surface it will not slip; the travadial i still have has a knurled wheel to grip the base surface; most trava's were always setup to contact the side of the bed way; never did see one on the cross-slide

[arvidb]

Originally posted by Bloy2004 I wild thought I often entertain is the use of a laser pen type system. The beam bouncing back to a receiver that locates according to the time lapse interval ....but the receiver would have to measure the speed of light! Please elaborate on this by shooting it down so I can clear my mind of such entertaining thoughts.

I just shot it down in page 2 of this thread: Wild thoughts about encoders

Sorry

[arvidb]

Originally posted by Chagrin Cut a timing belt and lay it flat on the surface and put a timing wheel (pulley) on the encoder and you'll have more precision.

This is a really good idea!

"Long" ago, when I had just started thinking about building a CNC machine, I thought about making my own encoders. It's possible to buy cheap IR sensors that can "see" a white paper that's held close to it, but does not react to a black paper. So I thought, why not print my own encoder wheels? It should be quite easy to align two sensors so that they output quadrature signals.

Then next problem appeared: resolution. A 600 dpi laser printer is not very good at printing really fine resolution circular patterns like an encoder disc (I tried it). What about scaling up the disc and copying it to a smaller size? No, a copyer can't handle it either.

So my final idea was to print out a large disc and connect it via an axle to a smaller "roller" wheel that would roll against the underside of the table - just like the picture above.

But a big disc that is strong enough to support itself has a not negligible rotary inertia, and I couldn't convince myself that the slip would be small enough to even give it a try. A timing pulley riding on a timing belt would have solved that.

And now I think I'm too lasy to try it ... but wanted to share my idea anyway! I might be able to scrounge up the encoder wheel printing program I made if anyone's interested.

Arvid

[balsaman]

Any encoder (linear or rotary) attached to the table instead of the motor can work but it is not a solution to the backlash problem. Any backlash in the system will result in wild oscilation of the motor as it hunts back and forth to try to position the table "just right", overshooting each time. There needs to be NO backlash in this configuration.

E