[Nails]

It seems the pick and place activity has died down, so maybe this will spark some interest. While designing my pick and place, I came across tubular linear motors. They seem very well suited for this application.

Check out: [ame=http://www.youtube.com/watch?v=MUcU6xLQQwU&feature=results_video&playnext =1&list=PL7D5F41665DA52085]Linearmotor Modules XY ThrustTube Parkem - YouTube[/ame]

After endless researching on methods of constructing them, it seems like the best and easiest way is stacking the magnets in a stainless steel tube with the same polarity facing each other WITH NO GAP. South/North-North/South-South/North-North/South, etc. This does pose some construction issues since the magnets repel each other, but that can be solved by making some additional tooling for putting the magnets in the tube and trapping them until the next magnet gets inserted.

There are also some articles about using SMC (Soft Metallic Composite) discs between the coils and forming pickup shoes for the magnetic flux. This started to discourage me, but it seems that Copley's ServoTube "Forcer" simply uses coils stacked together without these discs. It is pretty hard to find a good picture on the net as they are pretty low resolution.

Anyone know what kind of "varnish" is used to coat motor coils?

As far as I can tell, all that is needed to make a forcer is the coils spaced at 1/3 the magnetic length (at least 6 coils - a positive/negative coil for each of the 3 phases) and 3 hall sensors positioned appropriately. Controlling it would be any dumb trapezoidal torque amp, and motion controller. Of course, you would also need linear rails and guides to keep the forcer over the magnet tube, a base to mount the linear rail and end-support for the tube, and a linear encoder for position feedback to the motion controller.

I've got much more info if anyone wants to ask questions.

[keebler303]

You could easily make a pipe and plunger arrangement to compress the magnets into the tube, use a pin to hold them in and then have a threaded end with a cap to screw on and hold them in place. It is important to leave no air space as the magnets could actually move around as the coil pulls on them. It might be best to have a heavy spring under the cap so the magnets stay tight while accommodating thermal expansion.

The varnish is "transformer varnish". Google it and you will find lots of info. It is important to anchor your coils well as they are subjected to the full force of the motor. That is what is doing the moving after all.

On coil design, I think you need overlapping coils and very accurate winding if you want the best performance. Torque ripple and linearity are the biggies. Those are a direct result of coil geometry and magnet accuracy. Almost any old coil will give you movement but I'd think you need fairly high bandwidth control in a PnP machine so they might be important factors.

You may find it more effective to just buy some surplus linear motors.

Good luck
Matt

[Nails]

Transformer varnish.....Thanks.

[dgojanov]

What kind of force was you able to get on your linear machine?

How many coils did you put in tubular moving part.

What is purpose of hall sensor

Thanks for your reply.
Darko

[Nails]

Unfortunately, I hadn't been able to get around to this. It's on the back-burner, as I have a full-time job again.

The hall sensors are needed to sense the polarity of the magnets relative to the forcer position. This provides the information to the motor amplifier as to which coils need to be energized and at what polarity in order to move in the desired direction. The direction to travel is indicated by the sign of the analog +/- 10V torque command input. The magnitude of the voltage in the torque command input corresponds to the amperage to send through the coil.

As originally stated, a motion controller and linear encoder are necessary to close the position loop.

The number of coils, and gauge of wire to use would have to be done experimentally. The smaller the wire, the more turns you have closer to interact with the magnet which equals more force per amp per coil. But smaller wire also has a lower current capacity and higher inductance. Higher inductance inhibits change in current, so it ultimately limits your maximum speed. If you increase the number of coils in the forcer, you get more coils to interact with the magnets and effectively increase the force proportionally.

It would be a balancing act specific to the application needs. As the video show, it is possible to get very good speed, torque, and positional accuracy.

[dgojanov]

Thank you very much for your kind answer!

Best regards,
Darko

do you have any relevant links - pointers how to build this

[dgojanov]

how to align magnets in tube with coils.

coils are likely much wider than each magnet.

The magnets I was looking and ordered are 2 mm wide.

Any comment in possible disproportions in widness between magnets and coils.

REgards,
Darko

[Nails]

Re-read the first post. All your questions are answers there. Coils would be 1/3 the magnet length. I would start off trying 12mm magnets.

Unless you already have a motion controller, linear encoder, and torque mode amplifier you won't be able to build anything that is able to position itself.

>>Any comment in possible disproportions in widness between magnets and coils.

Only the obvious-the smaller the air gap, the better the interaction between the coil and the magnet. The effect is a squared mathmatical relationship. Therefore you want a somewhat thin-walled stainless tube to house the magnets, with as little difference between the inner diameter and the magnet diameter as possible. Like approximately .01 inch difference in diameters.

Look at Copley's brochures on the web. The pictures are not very detailed, but it does give you some idea of how it is constructed.