Sunday 26 April 2009

Tiny stepper torques big

Having calculated that the tiny stepper and GM17 gearbox combination should be able to drive a pinch wheel, I made a lash up to test the theory.

When you have a 3D printer "lash up" is probably not the right term as quite sophisticated parts can be made easily.

Here it is pulling a spring balance with a piece of HDPE filament.

It got to 10 Kg and then the coupling from the GM17 to the 4mm shaft of the pinch wheel let go.

Not surprising given the torque involved and the fact that it was made with 25% fill. I made it again with 100% fill. I can't remember the last time I made a solid part.

It is coupled to the shaft with a hexagonal steel insert drilled out to 4mm and tapped M3 for a set screw onto a flat on the shaft.

With the 100% fill coupler it easily pulled the scale to the end, i.e 12.5Kg. The motor was powered from 8V (to stop it getting too hot) and stepped at 200pps. With a step angle of 15°, the GM17 default gear ratio of 228:1 and a 13mm pinch wheel that gives a feed rate of: -
200 × 15 / 360 × 1 / 228 × 13 × π = 1.5 mm/s.
That would give an output rate of 54mm of 0.5mm filament per second. I think that is comparable to the rates Adrian Bowyer has reported from a NEMA17, but it only weighs about 60g whereas a NEMA17 is about 200g. There are a lot more parts to wear out though, so a NEMA17 may be a better option. Darwin can easily throw 200g about and HydraRaptor is moving table, so the head weight has little relevance.

I have some NEMA17's arriving this week. I tried one from an old disc drive but it didn't have much torque. I don't know if that was because it had aged in the 20+ years I have had it or whether modern motors are much better.


  1. Awesome!

    Besides weighing little, your geared stepper also has much finer filament control. So this looks like the best pinch wheel extruder.

  2. A big disadvantage I have just realised is that there is significant backlash. Because the rotation is so slow that means that it will take a signifiant time to reverse the feed, perhaps causing more ooze.

  3. A single stage gear couple, with about a one to six ratio, powered by a NEMA17 motor should give a good compromise of fine filament control and only a bit of backlash.

    I played around with the Gearweaver script in the Art of Illusion Scripts folder in the skeinforge zip file, to get a reasonable size shovel herringbone gear couple. The parameters different from the defaults which I ended up with are the following:

    Number of Teeth for First Gear: 79
    Number of Teeth for First Gear: 13
    Pitch Radius of First Gear (mm): 40
    Minimum Rim Thickness Ratio: 1.0
    Shaft Radius (mm): 2.0
    Shovel Segments (surfaces): 50
    Thickness (mm): 20.0

    I set the shovel segments at 50.0, being 20 mm over a layer thickness of 0.4 mm. Whatever layer & gear thickness you use, for best gear results set the shovel segments to the thickness of the gear over your layer thickness, to get one shovel segment per layer.

    To get around the AOI problems with boolean shapes, you can import the gear into another modeling program. Or you can do the modeling in AOI without ever converting to a triangle mesh, and export the file as xml, which skeinforge will read and do the boolean operations correctly. Skeinforge can do them correctly because it slices first, then does the boolean operations; which is much easier than doing boolean operations on triangle meshes.

  4. Wow 80x20mm is a pretty big gear but I suppose it needs to be. That will reduce the backlash and also reduce the force on the teeth.

    I have no idea what shovel segments are or why I need one per layer?

    As you said, it wont make a valid mesh. I don't understand how anybody uses AOI as everything seems to end up non-manifold and un-fixable. I posted a bug report for my simple test case of two cubes. It got closed as fixed, but it isn't, so I posted it again recently.

    I would like to import it into CoCreate, but how can I export it from AOI without it converting to a mesh? I think OBJ and VRML are just triangles, that leaves .pov and .dae. It it possible to convert those to STEP or IGES?

  5. The herringbone gear, from wikipedia:

    is a double helical gear. From the top the helical grooves of this gear looks like the letter V.

    Gearweaver makes a shovel gear, which is similar to a double helical gear. From the top the grooves of this gear looks like the letter U. I could not find a name for this type of gear; I called it a shovel gear because that was the closest shape I could think of. If someone knows the correct or a better name, please post it.

    It would be best to have one shovel segment per layer so that there would be a smooth curve without any discontinuous steps, which would otherwise happen when one layer spans zero or more than one segment.

    The gear produced by the Gearweaver script is, or at least it should be, a correct triangle mesh. It is composed of many triangles, but the shape has no holes so it is straightforward to make a valid triangle mesh. You could then import that into CoCreate as an OBJ file. Then you could manipulate as need be to fit it onto the shaft or coupling.

    What I meant by not ever converting to a triangle mesh in AOI, but didn't state clearly; is that you could do boolean operations on the gear as need be to fit, and you would export the final boolean shape as an xml file. Rather than converting the final boolean to a triangle mesh, which because it would almost certainly have at least one incorrect boolean operation, would be an invalid triangle mesh.

  6. If I make a gear and then convert to a mesh then AOI says it is invalid if I check with the solid editor. As I say virtually everything I make in AOI is invalid, which is why I stopped using it, but somehow the metalab guys manage to do amazing things with it.

    My simple subtract two cubes test fails even on the latest version and even on a new computer. Vik also tried it for me so it seems to be a definite bug, but it got closed as fixed. If it can't subtract two cubes (with no coincedent faces) I can't see how it can do anything.

    CoCreate only accepts STEP and IGES, so I would have to do a two stage conversion with a third app, but I want to avoid a format that converts it to a mesh as CoCreate is not a mesh editor.

  7. After you posted that you made a gear and converted to a triangle mesh, I tried gearweaver again. I then realized that it made the curves First Gear Profile & Second Gear Profile, then a few minutes later it converted those objects to the triangle meshes First Gear & Second Gear.

    I never realized that the curves hung around around for so long, because either I test it with a few shovel segments and the triangle meshes are generated right after the curves, or I test it with many segments, then browse the web for several minutes, and come back to see the final triangle mesh gears. When using gearweaver with many shovel segments, it's best to run it, then do something else for several minutes, and come back.

  8. Would love to see you get this design extruding!

    As to the backlash maybe a rapid reversal when shutting off will unload the drive train enough for the pressure to drop.

    This also might be addressable in the firmware by a look ahead to know when a stop extrude event is up coming.

  9. Enrique,
    Yes I get a mesh after a while, but it is non-manifold.

  10. Freds,
    Yes I suppose the back pressure on the filament will keep the drive chain backlash free until the pressure is zero, at which point it will have stopped oozing.

    I will give it a try when I have made a new extruder controller.

  11. How do you determine that it is non manifold?

  12. Oops, after I posted that comment, I found that when exporting the shape as an stl file, it warns about being a non manifold surface. I'll try to fix the bug, after Monday.

  13. Yes, also the solid editor shows 845 irregular edges and 26 faces.

  14. The "fix all" solid editor option always gives me an array out of bounds error so has never fixed anything.

  15. Thanks for pointing out the solid editor edge & face information dialog. I uploaded the fixed version of Gearweaver to the zip file at:

    If you want to add your own shaft, to eliminate the shaft hole in Gearweaver, set the "Shaft Radius (mm)" to zero.

  16. Enrique,
    Thanks for the fix. I printed the large gear but it did not come out very well. Being so large it is too warped I think. Whereas it would not matter much with a plain gear, with curved teeth they need to exactly match the other gear.

    I will have another go later when I have improved my build quality.

    In the mean time I will try a Meccano worm gear pair for 40:1 reduction on a NEMA17.

  17. Thanks for trying out the big gear. Curved teeth should be quieter, but since warping precludes them, after trying the Meccano gear you could try setting the "Shovel Segments (surfaces)" to one so that it outputs straight teeth.

    Otherwise, I look forward to heated beds, so that people could make gears with curved teeth.

  18. hello, when are you going to post something new on this blog? (im a newcommer and i want pictures!)