Monday, August 08, 2005


Screw thread drive preliminary experiment

I've recovered enough from a bout of the 'flu to write a few things up. Having got control of some small stepper motors, we thought we'd try out a screw thread drive instead of belt drive. So Ash and I coupled a section of M5 studding to a stepper motor using some 2mm and 4mm PVC hydroponics tubing. For bearings we used the capstain bearings from old 5 1/4" floppy drives, and trapped them in Meccano holes as shown below:

By using small slivers of plastic tube, we were able to securely grip the M5 studding in the bearings with nuts without fouling the bearing itself.

The motor is not strong enough to drag the Polymorph sliders, so we've built a small wheeled cart while we come up with something better. A pair of angled long-nose pliers held shut with rubber bands grips an M5 nut on the studding, and pushes the cart back and forth:

We've been testing the Allegro UDN2559B buffer/driver part with good results on our stepper motors (the ULN2003 is in "last orders please" mode). It can sink up to 700mA, which we have discovered is enough to blow the output port on a PIC. It also survives being short-circuited and springs back into life when it cools down. Only 4 ports per chip, but as with the ULN2003 no other parts are required.

We've got rid of the old DC motor on the turntable now, and will be using steppers in the future except possibly for the extruder. We're considering using a separate stepper for vertical motion control as the drivers we're using make for very simple circuitry. It also cuts out the long wait between printing an artifact and waiting for the turntable to crank itself back up again in preparation for the next test.

A shortage of plain old 0.1uF capacitors and the 'flu delayed plans to get the MAX202 serial interface chip connected up. When I'm out and about we'll see if we can't control things through the USB/serial interface (blue plug in the bottom left of the control panel).

Thanks, will do.

We're still considering belt drive powered by a larger stepper, but right now we have some steppers that work and we need to learn about driving screw threads - so this is an obvious move and one I hope to learn a lot from.

I'm also suffering from that well-known condition whereby when you have a hammer, all problems start to look nail-shaped...

Vik :v)
If I've followed what you're doing correctly you're going for the Fused Deposition Modeling approach. From what I've seen systems using that approach keep the production stage in a box kept at a temperature just below the melting point of the plastic being worked. I know that with polyethlyene unless the surface being built on was very hot the extruded thread of new material wouldn't adhere. The thermal resistance of the polyethlyene was too high for enough energy from to new thread to melt the surface it was being deposited on so that a good bond was created.

Is this how you plan on going at it?


I've found that the hammer effect can get quite extreme when one realizes that a stepper motor is not quite a box that will rotate a shaft in specific increments, but rather a ring of coils and a ring of magnets that will rotate the magnets relative to the ring in specified increments that is generally put into a box. Or vice versa, for that matter.

As for the comment concerning heating the space where the part is being made, in the preliminary experiments the opposite was true - the parts tried using EVA needed to be cooled in order to avoid making a mangled puddle.
I also believe that there's some contemplation of a hot-air assembly attached to the extrusion nozzle, to pre-warm the region of the part that is to be added to if necessary.
Absolutely. The problem is getting the heat out. When I increased the deposition rate by 60%, no amount of fans could cool the workpiece sufficiently rapidly, and the whole thing always turned into a puddle.

As the plastic is deposited in a very thin layer, it conducts sufficient heat through from the nozzle to melt the underlying material. We're not just passively laying down a noodle of plastic here.

Vik :v)
Allegro appears to be phasing the ULN2003 out, but it seems worth note that STMicroelectronics appears to be producing them as well, and I can't see anything saying they're stopping. Also, they have a part ULN2803 which is roughly the same thing, but has an additional pair, for eight channels - which ends up meaning it could handle either two 500 mA motors or run outputs in paralell and drive a 1A motor.
They are also running out the 2803 :(

I'm using UDN2559B's and ULN2003's side by side in the latest prototype circuitry. They're even more impressive than the 2803's in as much as they sink 700mA per channel and have nice protection features. But they only have 4 channels.

If anyone knows of a ready alternative to 2003's or 2803's please squeak up!

Vik :v)
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to
Posts [Atom]