I’ve tested several tool finger designs for the fixed tool changer (eventually to be used on the rotary tool changer), and while I don’t have a final design yet I thought I’d share some data.
The goal here is to ensure retention while minimizing required force to grab the tool. These are PLA, but the production version will likely be Nylon. In the case of this fixed tool changer, the retention required is primarily going to be determined by the maximum speed of the table, which is 100 in/min. For the rotary tool changer, there may be higher accelerations, particularly at the end stops.
Using a postal scale and slowly increasing pressure, I measured the force required to insert into the different designs. They are all the same basic geometry with some minor tweaks (except for #2, which is flat):
Design #1: 11 pounds. Tapered in both Y and Z along the X-axis (which you can imagine as a line that splits the finger symmetrically).
Design #2: 12 pounds. 2D, flat sheet. Distorts significantly while trying to insert the tool holder.
Design #3: 15 pounds. Tapered only in the Y direction along the X-axis.
Design #4: 11 pounds. Added two additional bending locations and tapered in the Y direction, but this time the taper goes outwards instead of inwards.
Design #5: 21 pounds. This was my original design.
Now, some of you may be saying, “why didn’t you just simulate this and save yourself a lot of time?”. Let’s take a look at the results of a simulation on Design #4.
This is showing the total deformation with 11 lbs of force applied. The force location might not be quite right, but that’s not critical here. Notice that maximum deformation is 0.005 mm — in other words, 5 microns. Now that amount of deformation is nowhere near the required deformation to seat the tool holder in the finger (approximately 4.25mm).
The first point of error in comparing these is the scale (and my measurements which are probably somewhat subjective). A better method for this would be using a setup where values could be pulled directly off the load cell, preferably with some tunable power source pushing the tool holder (like a pneumatic cylinder).
The other major factor is that these parts are 3D printed, so they deform more easily. Simulating 3D printed objects requires playing some funny games with things and is overall not that fun, so physical testing is easier and more straight forward.