Amazing platform with a lot of potential (obviously invested since I have one sitting in my office ). Some thoughts that have come to mind, as inspired by the workshop and since I’ve had a Jubilee sitting next to me with a bunch of other automated equipment sitting beside it.
A force sensor on the tool carriage (?) would be really good. There’s such a high risk of crashing a custom tool into the build plate, and this would be a fairly robust solution (i.e., if an unexpected force is applied, stop all activities), very similar to what is possible with the e-series of UR robotic arms (e.g., UR5e). Highest priority would be “vertical” crashing, but lateral crashing would also be important. Thoughts on how this could be implemented? (OK if detailed discussion goes to the discord, and then we link that here)
Extend the build platform past it’s “normal” area. This would extend the usable area and make it a lot easier for external devices to interact with Jubilee.
Adding a small downward facing camera to the tool carriage. This would make it a lot easier to do things like automatic recognition of AprilTags, which would then also reduce a lot of the overhead and burden of trying to manually change offsets.
There was some discussion on the open lab automation discord about using the stallGuard capability that the stepper drivers on the Duet theoretically have. (Stall detection and sensorless homing | Duet3D Documentation). This seems like it would be capable of detecting crashes. I haven’t looked into if it is something that can be turned on and off on the fly. I would imagine that the force threshold to detect a crash might be lower than the force needed to pickup a pipette tip so it might need to be something that is turned off/on for individual actions. Are there more direct sensing methods like some sort of piezoelectic plate that gets sandwiched into the tool changer carriage mount somewhere that could work?
what are you imagining here? adding an extra inch of plate to the sides so you can mount more things here? I know machine agency/Blair has mounted a sharps container off to the side and we have developing a deck-side pipette tip trash bin on the todo list. mounting heavier things outside of the footprint of the kinematic balls might exacerbate existing bed stability issues that we need to find a solution for.
An integrated camera would be great. There is probably space to fit one in there without redesigning the whole thing? Do you know of any good camera options with a smaller footprint than the raspberry pi camera?
The stall guard and a piezoelectric plate both seem like good possibilities! The OT2 pipette stop mechanism is a sort of vertical crash prevention I’m now realizing. Something like a piezoelectric plate would probably do the trick, too. Might look into these more.
Wouldn’t have to be on all sides, but something to extend outside of the frame might simplify some external automation integrations. Like you said, the tip disposal attachment makes a lot of sense, and helps address some space limitations. Does bed stability issues mean unequal forces on the rods due to distribution of the bed weight and therefore a higher likelihood of an unlevel bed during operation?
I’m not as familiar with the interior of the tool carriage design, but with an appropriate inset, do you think the Raspberry Pi Camera 3 could be accommodated? Tbh Pi camera 3 is probably the best option.
Another option is a “displacement probe” to each tool, in the style of Jubilee’s OT2 pipette tool. It could work too, the design is already implemented, and perhaps will be better/simpler in the end.
This may be why, to my knowledge, large CNC machines don’t sense force. It may not be as reliable as expected.
Placing load cells on the base plate instead could be a better solution for our bot.
+1
Do you have experience with OpenCV or similar software?
I’d just make a larger robot instead. Jubilee is a 3D printer at its core, and it’s ok. It’s motion system is meant for that exactly, and so introduces a few tradeoffs.
This includes the possibility of crashing the X crossbar against labware taller than the tool.
Amazing. Thanks for this thorough assessment! This offers a really good perspective.
More and more, I’m having the itching suspicion this might be the case… the thought was maybe instead of repeating this design over and over, something could be built-in to the tool carriage, but if the repetition of the displacement probe design becomes trivial, very well might be worth it like you mentioned.
I’m planning to do a simple tutorial soon. This would be some good motivation for me to get to it sooner. If I implement this, is this something you’d be willing to try out?
Good point. I’m still curious to hear about the stability issues, since I’d still have some interest in hanging a few pieces “off the side”.
PS: Load cells / piezos /etc. still have many uses though. For example for probe-less bed leveling, moving at very low speed (and thus low vibrations) or so.
). Some thoughts that have come to mind, as inspired by the workshop and since I’ve had a Jubilee sitting next to me with a bunch of other automated equipment sitting beside it.
