RepStrap Tommelise

I've decided to engage in a design study to determine if Forrest's tentative plan for Tommelise 2.0 could be modified to be a viable RepStrap. If so, it would save significantly on cost and probably on assembly time as well, compared to some of the alternatives.

I think that one of the largest modifications that would need to be made to such a design would be changing the drive system from Forrest's signature gearmotors to stepper motors in order to make them RepRap-electronics-compatible. Currently the design relies on X and Y stages made of 3/4" chipboard driven along aluminum tracks by a threaded rod [farm2.static.flickr.com]. My question is: does anyone have an idea of what kind of torque requirements a stepper motor would have to have to drive those stages (particularly the bottom one)? If it's too high, I suppose I'll have to modify the design to use linear bearings [www.igus.de], but those can get pricey. Does anyone know of a cheaper alternative?

Of course, a REALLY convenient solution would be to modify the RepRap electronics/firmware to be gearmotor compatible. But I'm guessing that's much much easier said than done.

I've been building something that's fairly similar, just uglier. I used drawer slides in place of linear bearings. Two per axis at $5 each. It takes very little force to drive -- I put a 17" monitor on top of my X axis (before adding Y) and it didn't seem to notice. That said, you'll get a fair amount of friction if your threaded rod is slightly bent or if the thrust bearing is mounted at a slight angle. The cheapo steppers I'm using can still deal with the extra friction from a slightly bent rod, but it may set an upper limit on how fast I can drive it.

I'm still several weeks from printing anything so there may be problems with this approach, but the movement feels pretty smooth so far.

Kyle Corbitt Wrote:
-------------------------------------------------------
> Currently the
> design relies on X and Y stages made of 3/4"
> chipboard driven along aluminum tracks by a
> threaded rod

Um, that's not quite true. I've never intended to put the chipboard directly on to those tracks. Instead, I'd planned on putting plugs of grooved HDPE or PTFE to act as bushings. I'm using wood bushings on 1/4 inch steel rod in Tommelise 1.0. My gearmotors can put out 40 oz-in of torque (about what a good bipolar NEMA 18 stepper will put out). Running the y-axis on Tommelise 1.0 might use 5-10% of the available torque and the wood to wood x-axis maybe 2-3 times that. Neither gearmotor gets anywhere near the 60 oz-in clutch limit. They don't even get warm after having been run continuously for a day or so. Now the gearmotor that drives the extruder DOES pull every bit of torque that I can get it to put out.

> My question is: does anyone have an
> idea of what kind of torque requirements a stepper
> motor would have to have to drive those stages
> (particularly the bottom one)?

I personally think that you could get by with not very much torque at all, certainly much less than my gearmotors are capable of delivering.

The main problem with stepper motors and studding is speed. To get steppers to go fast enough you need a high voltage constant current drive and low voltage high current motors (low inductance). The RepRap electronics is constant voltage drive into relatively high voltage motors so its not going to be very fast. Threaded rod drives use much bigger thread pitches than standard studding but are expensive. That's why Darwin uses belts.

---

[www.hydraraptor.blogspot.com]

nophead Wrote:
-------------------------------------------------------

> Threaded rod drives use much bigger
> thread pitches than standard studding but are
> expensive. That's why Darwin uses belts.


I haven't ruled out using belts on this thing if they prove to have a significant advantage. Would that change the torque requirements of the motor?

Kyle

Yes it would need more torque with a belt. The studding gives a higher gear ratio so it requires less motor torque at a higher speed. Having said that stepper motor torque falls off with speed so more torque is available when using the belt, as the motor does not have to turn so fast. I would guess that frictional losses would be a bit higher with a studding drive.

I think Forrest published some torque calculations for screw threads a long time ago in the main blog. To get an idea of the force required you need to measure or calculate the friction of the axes. To calculate that you need to know the weight of the axis plus workpiece and the coefficient of friction between PTFE and aluminium.

The only downside I can see to the belt drive is that it struggles to achieve a step size of 0.1mm. The RepRap design needs such small pulleys that they are not much bigger than the motor shaft. It also uses 0.9 degree motors which are unusual but you can get the same step size with more common 1.8 degree motors using half stepping which is a one line change to the firmware I think.

---

[www.hydraraptor.blogspot.com]

I've rather wondered if Darwin could be made to work with this sort of chain and sprocket system. Robot Objects keep a rather wide variety of gear, sprocket and etc range and are happy to special order just about anything that suits your fancy.

[www.robotobjects.com]

Edited 1 time(s). Last edit at 08/04/2007 10:19PM by Forrest Higgs.

nophead Wrote:
-------------------------------------------------------
> The main problem with stepper motors and studding
> is speed. To get steppers to go fast enough you
> need a high voltage constant current drive and low
> voltage high current motors (low inductance).

What would it take in stepper motors, board redesign and power supply to make that happen? It this vastly more complicated and expensive than what they have now?

The electronics isn't much more complicated you just need a more modern driver chip that does it all. If you go too modern then you run into leadless surface mount chips which gets you into solder paste and ovens.

The main problem is that you probably need to go to a higher voltage supply so the cheap, off the shelf, PC supply is no longer an option as is the idea of running it from a single car battery.

You can't really have people making their own mains power supplies as it is far too dangerous. You could I suppose have a fairly simple boost converter stepping 12V up to say 42V which is the highest voltage considered safe without double insulation etc, but starting to get over complicated.

I think going down your route with small, cheap, light DC motors is probably the best path. I think it could be done with the UCB and some firmware changes. Actually the UCB has enough I/O actually to do two axes with DC motors if we scrap the sync lines.

Alternatively a new board which drives all three motors like yours but which talks snap and emulates 3 UCBs is probably the easiest way to allow people to RepStrap using Tommelise2.

---

[www.hydraraptor.blogspot.com]

nophead Wrote:
-------------------------------------------------------
> The main problem with stepper motors and studding
> is speed. To get steppers to go fast enough you
> need a high voltage constant current drive and low
> voltage high current motors (low inductance).

Another thing on my long list of experiments to try is to just use 7.5 deg/step motor (or even 15 deg/step?) instead of the usual 1.8 or .9 degrees per sec. It's a bit harder to find larger motors with big steps, but I've seen a few that might work. That should give you close to a 4x or 8x speed boost without fancier electronics.

All the 7.5 deg motors I have are 12V. I don't think they go much faster than two or three revs per second with CV drive. I have never come across low inductance ones but if they do exist that might be an idea.

The type I have are known as tin can construction and are relatively cheap compared to the stacked lamination type. They have less torque and less accurate pole piece alignment.

---

[www.hydraraptor.blogspot.com]

Kyle:

There is a guy named Arvin who has already build a half-scale Tommelise and is running his off of an Atmel processor. He just posted over at my forum.

[3dreplicators.com]-

You two might want to get acquainted.