We spotted an object in our gallery called 'part of a robotic clamper'. It was used by designer André Bois to fix a broken robot from 1984. The work caused quite a bit of excitement here at Shapeways (drop the word 'robot' in a group of techno-geeks and they'll be talking about little else all day), so I decided to find out more about this project. André really took his time to answer all of my questions. Thanks, André!
Q: You wrote that you 'found an unused and broken robotic arm' - can
you tell us a bit more about that? (Where did you find it / what was
the arm made for?)
A: This robotic arm, called "Robot Youpi Y101", was part of the
pedagogic course given during first four years of secondary education
in French schools, way back in 1984. Yes, it is that old. The original
French company that built it is longer in business, as well as the
companies that later were able to do support for Youpi.
Somebody teaching there recently told me that it was unused, and he was not sure if it worked or not.
As I like robots, I did my best to get it to work. Electronics part was
ok, just had to replace a fuse. Impressive, after all these years. But
the robotic fingers were broken ! This seem to be some kind of
bug-as-feature : before risking to crush your hand, the robot breaks
The only web documentation I was able to find (French only, of course) :
Originally there was a kind of fork, and when the hand is closing, it becomes a lever, and can break as the motors don't have stop sensors.
So I took measurements, and photographs straight from above the hand, then started Blender on my Ubuntu. This was the first time I modeled something that needed exact dimensions in
Blender... I learned a lot of nice shortcuts, and I was surprised to
discover that Blender is very capable for this kind of CAD modeling. I
mapped the photos on quads within the scene, to have a good vision of
what was physically possible and what was not. Yes, when I did the
modeling, I no longer had access to the robot and its parts, only the
photos and measurements. Too easy otherwise
Q: What were the hard and easy parts of modeling and ordering this?
A: Hard but actually very fun : modeling from scratch something
physical ! this part is not CG fake graphics, nor a re-creation : I
wanted to end up with a much better and stronger design of the part.
Thicker sections, reduced margins, counter balance of traction, but all
this will still have to move along the full course, with other mobile
parts. Blender helped me to model the articulation, as I could not see
the real robot during the modeling phase.
Hard : first time using Shapeways : how will the printing end up, will
this be rough, unprecise ? Will the materials be strong enough ? Which
one to choose ? By the way, your website should REALLY explain better
the different properties of the materials, especially for 3D printing
replacement parts. Lots of questioning, but it turned out ok. Both the
"Cream robust" and "White strong and flexible" parts were nice. [note from the editor: yes, we're working hard on improving that section]
Hard : had to create a Paypal account from a Dutch page (you should fix
this by the way, it has already been mentioned in the forum). Took 2
hours. Very easy : ordering, and getting the parts. Less than 10 days,
perfect. As I said in the forum, between hearing about this broken
robot and actually fixing it, it took only 15 days. I was very lucky to
hear about Shapeways at that time ! [note from the editor: thanks for reporting that PayPal issue - it's been fixed!]
Easy (mostly) : installing the part. Very nice precision. I had to sand
a bit (less than 0.1 mm) the inner side of holes, to be able to fit the
metal axes in, as shown here.
The Cream robust was easy to work, the nylon took some minutes more
Q: Does the replacement part work as well as the original?
A: In fact, way better. As it is stronger. I can force the fingers motor without problems !
Q: How do you feel about using Shapeways for creating replacements parts?
A: Price is a bit high. For the robot, as I don't have access to CNC
machining, I could never have fixed both fingers with less than $70
without 3D printing. And I wanted something strong, so no "hollowing
out" nor "thinning" to make the parts cheaper.
My pedal bin recently had its plastic support broken (bad mechanical
design again), and my first reflex was to think "I can design a better
replacement part, and 3D print it". But it would still cost a bit more
than a complete new pedal bin, as the part is quite large ...
So I think 3D printing is very worthwhile for small parts, and very
specific parts of costly hardware that nobody supports anymore.
Customization is interesting too : have something stronger, or funnier,
or prettier than the original.
Ease of use : perfect. Sure, knowing Blender helps. Tips on sanding can
be useful, I saw you linked an interesting PDF about this. On your
side, as I said, you should explain better the tradeoffs between the
possible materials. And make apparent than even when ordering multiple
parts at once, they can be received separately.