You could bolt some of these cup magnets into the holes in the high-strength chain. The magnets are very strong and could easily support a small robot.
The magnets are larger in diameter than one link of chain, so you’d probably have to stagger them to get it all to fit.
I would be concerned about generating enough force to lift the magnets off at the end of the track, but hopefully the fact that it “tips” off would allow it to work efficiently.
Now I’m interested in seeing if this would work, so I’ll have to round up some parts and try this out (the magnets are all over the house holding stuff up).
rare earth magnets are VERY VERY strong!!!
trust me!
i have experience with them
if you attach the magnets on so their surface area is fully on the wall, your robot will NOT have enough power to pull them back out again
especially if you need enough magnetic force to lift a whole robot…
it would have a hard time trying to pull even a 1cm radius rare earth magnet
I decided to prototype one as an experiment. I used 12 of the cup magnets I linked above attached on a stagger pattern to a 24 segment high-strength chain.
The magnets I used are rated for around 20 lbs of pull force, but since they are lifted off the surface by a tipping action, the tracks don’t have to apply 20 lbs of force to retract the trailing magnet.
It does take more force than I think would be desirable, but it can be done. I was able to make the contraption advance up the wall by applying a rotating force to one of the axles by hand (using a gear as a handle). Nothing broke or twisted, though it was a very jerky ride. Since we’ve seen that Vex can be geared so low that it can twist an axle, I’m convinced it could be made to be driven from a vex motor.
The trick here is that the track only ever lifts one magnet at any time, but many (4 or 5 in my case) are in contact with the metal wall. I think it would be better with twice as many magnets that were each only half as powerful - that would provide the same draw to the wall but with only half the force needed to advance each track. Perhaps adding felt pads to the faces of my magnets would have the same effect.
Oh, and with 4 or 5 of these magnets holding the track on the steel, it could easily have supported a battery, motor, gearing, microcontroller, etc. The only way I was able to get this off the steel was to walk it to the edge - I couldn’t pull it straight off.
The other thing to keep in mind is steering. A pair of these tracks would have 8 to 10 magnets in contact with the surface, and I just don’t see skid steering working very well. It might work but with a large turning radius (no rotating on the spot or stopping one track to make a tight U turn).
Nicley built, Quazar. The only problem I see now with that is that the magnets are now moving, possibly creating electrical current in the wiring. Be careful how you set it up.
Just looking around, I found these magnets that might be a better choice.
They would need to be mounted using a #6 flat-head screw (not the usual #8 that Vex uses). These are about 1/6 as strong as the cup magnets I used above, so I think I’d install twice as many of them. That should provide sufficient holding power (particularly if 2 tracks were used) but would not be absurdly difficult to lift off the trailing edge.
As for bot geometry, I thing you would be best off going with a wide wheel-base (track-base?) and keeping the tracks fairly short (24-32 links seems about right). It might look a bit odd, but I think you could steer better that way.
I had a few thoughts for derivative designs. As I don’t have either the parts on hand or the time to try building them myself, I’m posting them for the use or criticism of anyone who’s interested:
It appears that you could put one of the smaller magnets in each of the holes of an attachment tread. That would get you closer to the “half as powerful” you mentioned a few posts ago.
As the smaller magnets have through holes, you could stack two of them on a bolt.
I’ve never handled the Vex chains and I failed, in my brief attempt, to find the info on the Vex site, so I don’t know how one couples and uncouples the links. It may, however, be possible to make a chain that is just large enough to go around one of the drive sprockets. If so, one could make a magnetic wheel by placing the chain around the sprocket and studding it with magnets. The obvious downside is that there would be only one link at a time in contact with the surface, reducing the available holding force. OTOH, one might well get improvements in the ability to turn from having shorter contact areas.
Speaking of holding force, perhaps this is a good time at which to point out that, when one is trying to stay on or climb a vertical surface, the force resisting sliding downward is not the force of attraction between the magnet and the surface, it is the friction between the contact area of the magnet and the surface. That friction is the product of the attractive force (between the magnet and the surface) and the coefficient of friction between the two surfaces. This presents an opportunity for some experimentation. Putting a layer between the magnet and the wall decreases the attractive force (greater distance and, perhaps, interference in the field by the material), but may well increase the coefficient of friction.
If one can’t fit a chain snugly to a single sprocket, that design could be approximated by having two, coplanar sprockets with the line through their axes perpendicular to the surface to be traversed. The sprocket that is farther from the surface could be used to transfer drive power to the chain and to provide appropriate tension. This design strikes me as being well-suited to building a vehicle with high “ground” clearance.
A low-profile variant of the previous design is to have the line through the axes parallel to the surface to be traversed, using a smaller sprocket to provide tension and, perhaps, to transfer power to the chain.
Lastly, I point out that, as the force keeping the vehicle from sliding down the metal wall is the friction, there’s no fundamental need for the magnets to be the driving contact element. In theory, one could use magnets with very low friction in contact with the surface to provide attractive load and conventional wheels to move the vehicle. Alternatively, one could have magnets arranged, around the wheels, in a plane parallel to the surface to be climbed, but not in contact with the surface. One could also use a track passing between magnets that are fixed relative to the vehicle and the surface. The foreseen advantages of these designs are that they do not require “sticking” and “unsticking” and that the number of magnets is reduces, as they are spending all (rather than about 1/3 for a magnet-studded tread) of their time adhering the vehicle to the surface.
I think that would work very well. One reason the ride was rough was the staggered pattern - the track would lean one way and then the other as it struggled to pull off each magnet. Having it balanced would probably work better and put less stress on the plastic links.
I don’t think that would add much pull. The magnetic strength falls off very quickly with distance, so the rear magnet wouldn’t contribute as much to the holding force. The small magnets have a bigger brother that might be a good option instead of stacking (twice the pull force, and a #8 diameter mounting hole).
I thought about that, and yes you can make a chain that is tight around a sprocket. I think you need to keep a few magnets in contact at all times. The magnets really only hold strongly if they are completely in contact with the surface - as soon as you tip them a bit, their holding power falls off considerably (the track solution I made counts on that). A tight track or wheel with magnets wouldn’t provide enough pull to hold the robot in place while you were “in between” magnets.
Yep. The magnets I used were so strong that I had trouble sliding it around my hand - I think a pair of these tracks could easily hold a full vex robot up without sliding. A thin silicone rubber pad would reduce the pull force and increase the friction, which might be a better option than going with more, weaker magnets.
It also occurred to me to replace the rollers on an omni wheel with small tube magnets (I didn’t find any exactly the right size). I think this would have overcome the steering problem, but I don’t think it would provide enough holding force because only 1 magnet from each wheel would be contributing to the holding force, and only the magnets aligned perpendicular to gravity would be resisting sliding.
Agreed, though I think such designs are more sensitive to wall uniformity, but would certainly be a smoother ride and probably simpler to build. I’ve seen wind-up fridge toys that use this approach. If I find some time, I’ll try to mod my design to use this approach and see how it goes. The trick there is going to be to keep the magnet a very precise distance off the wall surface. Too far away and it’ll fall off; too close and it’ll grab. The sweet spot is a very narrow range.
I think you are asking whether to get Vex Motors or the High Strength Sprocket and Chain Kit. You can’t use the Chain Kit without motors, so I would go with the motors.
The Vex motor has a built-in controller, so you wouldn’t need an L293D to drive it. You simply provide it with a hobby-servo style PWM signal and it will run.
If you want a motor that you can drive using a L293D, you should look at the Vex VB-1 motor (sometimes called the VEXplorer motor). It is a normal DC motor without a built-in controller, and you get two of them for US$20.
Pictures of the controller board are located here and here.
The driver chip is an Infineon BTS-7700-G, and you can download the datasheet here.
There is also a microprocessor (one of the small chips on the back of the board) that takes the servo PWM signal and converts it to a pair of PWM signals suitable for driving the H-bridge in the Infineon part.
The signal that you send to the controller board is a simple logic-level PWM signal that is commonly used by hobby servos. About 60 times a second, you send a pulse to the motor. The width of the pulse specifies how fast the motor will spin. A 1.0ms pulse will spin the motor full reverse, a 1.5ms pulse will stop the motor, and a 2.0ms pulse will spin the motor full forwards. Pulse widths between these values will spin the motor at a proportional speed.
