Laser cut from a Thingiverse model?

[Dan Linder]

I came across this blog about an inexpensive robotic arm:
    http://geekswithblogs.net/kobush/archive/2012/04/09/149258.aspx

...that uses this Thingiverse model:
    http://www.thingiverse.com/thing:2433

The arm was cut from 4mm plywood, but I don't see how to tell how large the largest piece is, nor if all the pieces will lay on a single sheet that will fit our laser cutter.

I also remember discussions that the laser cutter has problems with 'thick' wood - does 4mm constitute thick?

Can anyone with more expertise in our setup look at the model and give me their $0.02 worth?  I'd like an excuse to come down and spend an evening/afternoon cutting this out but I'd definitely need to be co-pilot on this.

Thanks,
Dan

--
***************** ************* *********** ******* ***** *** **
"Quis custodiet ipsos custodes?"
    (Who can watch the watchmen?)
    -- from the Satires of Juvenal
"I do not fear computers, I fear the lack of them."
    -- Isaac Asimov (Author)
** *** ***** ******* *********** ************* *****************

[Ben Hutcheson]

Kevin and I cut an arm this winter based on that exact model.

4mm is fairly ambitious for our laser, especially until we get the lens replaced (I'm working on it, I promise).  So I re-drew all the connecting slots for 3mm stock; there aren't too many, so this went pretty quickly.  We cut one in 3mm birch plywood, and I think it took two sheets the maximum size of our laser bed to cut all the parts.  However, the design was actually optimized for a CNC mill rather than a laser, and certain details weren't quite precise enough thanks to the typically-negligible kerf of our laser.  Additionally, the weird blobby gear teeth on the gripper claw are terrible.  I don't understand how they ever meshed well for anyone, regardless to the machine used to cut the parts.

To improve on these and a few other perceived flaws in the design, I kept the same basic layout and dimensions and drew a completely new arm, better-suited to lasering and 3mm stock.  I haven't yet designed a gripper I like, but everything else is pretty tidy.  We cut and assembled one in acrylic, and took it to the Robotics Expo at the SAC Museum.  Small children managed to burn out three of the seven servos in a matter of minutes.  This leaves me with some doubts about servos being a good choice for this application, and they were never all that smooth to begin with.  So I'm delaying repairs and replacement while I consider an entirely different approach.

That said, our gripper-less arm is still at the Makery, and I still have all the files.  If you would like to visit the space and play with the semi-operable arm, you can decide if it's worth pursuing for yourself.  You are welcome to cut an arm of your own from either the original or my revised design, if you'd like.  You would just need to use the original gripper design or create your own.

All that aside, why should you need an excuse to come hang out at the space?  We'll be around most of the day Saturday working on the Power Wheels Corvette, driving RC cars, building workbenches, flying quadcopters, and making fun of each other.  If weekends aren't good for you, we'll be doing most of those same things again this coming Tuesday evening, plus we'll have a guest speaker!

--
Support Omaha Maker Group with purchases you make anyway. Shop Amazon using our Affiliates link, and OMG receives a portion of the proceeds. http://amzn.to/1f3i3ve
 
 
Leave lurking behind — come visit us at the Makery at 8410 K Street, #5, Omaha (just off 84th & L). We’re nice, we promise. http://bit.ly/1dKnTmC
---
You received this message because you are subscribed to the Google Groups "Omaha Maker Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to omaha-maker-gr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Dan Linder]

On Thu, Apr 10, 2014 at 1:31 PM, Ben Hutcheson <hutche...@gmail.com> wrote:

Kevin and I cut an arm this winter based on that exact model.

I kinda wondered - the corner of one of my photos of the booth from the Robotics Expo looked like it might have been...

 

4mm is fairly ambitious for our laser, especially until we get the lens replaced (I'm working on it, I promise).  So I re-drew all the connecting slots for 3mm stock; there aren't too many, so this went pretty quickly.  We cut one in 3mm birch plywood, and I think it took two sheets the maximum size of our laser bed to cut all the parts.  However, the design was actually optimized for a CNC mill rather than a laser, and certain details weren't quite precise enough thanks to the typically-negligible kerf of our laser.  Additionally, the weird blobby gear teeth on the gripper claw are terrible.  I don't understand how they ever meshed well for anyone, regardless to the machine used to cut the parts.

If 4mm wood is too ambitious, could we cut 'most' of the way through then carefully using the cut lines as a guide carefully cut with a small jig-saw with a wire blade to go the rest of the way?

I wonder if the gear meshing problem might have been due to the thinner material?  I'm all up for either doing two runs using 2mm plywood, or a single run with 4mm wood and then lightly sanding the edges to make them flatter.

And where would I purchase 4mm plywood?  When I've looked at HobbyLobby and Menards in the past, they were all in US/Imperial measurements, not metric. Converting 4mm is just under 5/32", and 3mm is a bit over 1/8" - I assume I'd be better off with slightly thicker and more clean-up after lasering.

 

To improve on these and a few other perceived flaws in the design, I kept the same basic layout and dimensions and drew a completely new arm, better-suited to lasering and 3mm stock.  I haven't yet designed a gripper I like, but everything else is pretty tidy.  We cut and assembled one in acrylic, and took it to the Robotics Expo at the SAC Museum.  Small children managed to burn out three of the seven servos in a matter of minutes.  This leaves me with some doubts about servos being a good choice for this application, and they were never all that smooth to begin with.  So I'm delaying repairs and replacement while I consider an entirely different approach.

What did you use for spacers/washers between the arms when it rotated?  For the base I know Menards has a small 4" lazy susan that should help take all the weight off the servo, but for the arms I'd think a teflon washer between two metal ones (i.e. arm-metal-teflon-metal-arm) would give a pretty good "slippery" joint.  Or do you think the overall mass of the arms was beyond the torque of the servos?

I was planning to purchase slightly higher-torque servos such as this one - http://www.lynxmotion.com/p-719-hs-485hb-83-oz-in-standard-servo.aspx - for $16 that provides 83 oz-in.  I need to do the math to make sure, but in the LynxMotion arms that appear to be much heavier use the 133 oz-in (HS-645MG, $30 each) so I'd hope this one would be sufficient.

That said, our gripper-less arm is still at the Makery, and I still have all the files.  If you would like to visit the space and play with the semi-operable arm, you can decide if it's worth pursuing for yourself.  You are welcome to cut an arm of your own from either the original or my revised design, if you'd like.  You would just need to use the original gripper design or create your own.

Thanks - I won't be able to make it down this weekend (life is getting in the way of playing...), but I'll continue to research the options.

 

All that aside, why should you need an excuse to come hang out at the space?  We'll be around most of the day Saturday working on the Power Wheels Corvette, driving RC cars, building workbenches, flying quadcopters, and making fun of each other.  If weekends aren't good for you, we'll be doing most of those same things again this coming Tuesday evening, plus we'll have a guest speaker!

Definitely want to get down - I'll watch the "Makery is open" announcements and swing by when I can.

[Ben Hutcheson]

You could definitely use the laser to mark lines out, and follow them with the scroll saw.  I think you would probably lose a lot of the laser's precision, but could cut thicker material.  I doubt it would improve the gear mesh at all, though - that issue was entirely two-dimensional.  I think the teeth were drawn that way because an end mill can't cut the inside corners of involute gears without some modification.

Finding appropriate plywood for the laser is always a challenge.  Partly because hardware stores rarely carry thin enough material, and partly because they lie about the dimensions.  The only place locally with thin plywood is Midwest Woodworkers, who carry 5'x5' sheets of what they call 1/8" birch, but which is actually 3mm (this is what we have a stack of at the Makery).  Shopping online, look at boat-building suppliers and hobby shops, and once again be prepared for approximate dimensions and possibly also sticker shock.

Our spacers on the arms are turned delrin (or HDPE or something similar) bushings with shims.  I'd suggest 3D printing a stepped bushing instead.  On the base, we used a thin layer of styrene between the rotating plates.  A more advanced system would be neat, but I'm not sure it would actually help things.  The weight of the arms and the friction in the mechanism didn't bother the servos in the slightest.  The problem is that servos are not designed to continuously hold loads at high torque like stepper motors can.  Ours burned up when someone directed the manipulator end of the arm to press on the table top and walked away.  I imagine a similar issue could arise if you tried to hold a significant weight in the gripper.

Using physically larger servos would probably help resolve this issue, as they would have more thermal mass.  But, they would also be heavier.  Higher-torque servos might even exacerbate the issue by permitting higher power draw.  In any case, these are the servos we used, at 70 oz-in of torque.  Despite the Hobbyking origin, Corona servos are fairly reputable, so I don't think the issue is servo quality.  Physically-larger servos with a much lower speed (hopefully signifying a higher gear ratio) would be the best bet.  Analog rather than digital might also help, as digital servos tend to draw more power.  So, basically, yes.  The servos you've selected look like they might be a slightly, but maybe not significantly, better choice than the ones we used.

--