Since I am involved with Openrov and follow the forum, I have seen a lot of speculations about grabbers or manipulators. Also for me this has always been a goal. Now that I have finished my rov and made a couple of modifications I turned my interest to this. I imagined a lot of different designs using servos, but then I need to be able to program these and deal with electrical connections. When I redisigned my endcaps, I already potted some extra wires for this. On the other hand I had a magnetic option in mind, which I did now first. The idea behind it was to use the camera-servo pins to plug in the servo for the magnetic actuator. I thought, I don't use the camera-tilt, when I use the grabber, because I want to see the grabber!
so I made an actuator with a servo inside (to save space) and magnets on one end.
here are the plastic-parts I milled prototypes and made molds to cast a number of copies, so I can play and modify, witout always having to start from the beginning.
My central rod (see my other posts) acts as a hinge. The position of the servo arm was chosen so that towards the end of the movement (grabber closed) the maximum of force is generated and the force is almost parallel to the servo-horn, so it is sort of self-blocking without consuming to much current at the "closed"-position:
The magnets outside of the tube are glued in another plastic part that has little weels that ride on the tube to reduce the friction generated by the manets (without the weels the magnets needs 150g to be moved, with the weels almost nothing!). This plastic part is then connected with the grabber by a pushrod.
Here are the pieces in and outside the tube to demonstrate their position:
Below you can see the grabber with all pieces. It takes less than 10 seconds to mount or dismount it!! It just snaps on the rov's threaded rod and is secured with one screw. Some of the plastic parts are made from self casted syntactic foam(resin with microballoons), to make the whole grabber neutral bouyant, so it doesn't mather, of it's on or not. The claw itself is also secured on the carbon rod with one screw, which allows to rotate it (I have already a modification for the rotation by a second magnetic coupled actuator, but there again I would need to program more servos)
the claw opens and closes nicely when I use the camera buttons.
the closing force at the tip is appr. 200gr, at the inside 300gr. This should be enough, because the rov can not lift a lot.
I am sure, this approach can still be modified a lot.
I think there can be a lot of different designs and improvements. I consider it just as some sort of "proof of concept"
So I am happy to see what others do with these ideas and look forward the your feedback!