Disengaging a motor on a spring powered hurdler

We are working on a hurdler design that will use a large spring to hopefully accelerate the trackball enough to clear the overpass. Since safety will obviously be a concern, we will mount the spring in a protective cylinder, have a cable running through the center of the spring, use at least two strong latches to hold it in place when loaded, have a way to release tension in the spring safely without firing the mechanism, and use a two step process to fire the hurdler on the operator panel to help prevent accidental firing. (Comments on the safety of such a device or additional precautions would be greatly appreciated.)

To retract the spring, we will use a winching mechanism. However, the motor used will have to disconnect from the winch so the mechanism can fire. Does anyone have any suggestions on how to accomplish this? One idea was to use a shifting gearbox with a speed missing, but as we have used all of the CIM motors in the drivetrain, there are none available for use in, say, an AndyMark shifter. We do not have the capability to make our own gearbox. Does anyone know a way to connect a different motor to a shifting gearbox, any other suggestions on how to disconnect a motor, or a suggestion for a winch design where this wouldn’t be required?


Don’t they make pneumatic cylinders only use air in one direction. Single action acting cylinders I believe… Anyways you could use that to pull the spring back, let out the air and the spring expands.

as I ranted in one article, be careful. Springs are dangerous, and you need to pull them back. Even the 2" bores can only produce a maximum of 189 pounds(foot pounds?) of force/ energy. This is going outwards not inwards which is what you need. so at @ 180 b/c you loose energy to heat, a bullet produces at a minimum 1000J. in reality, do you want the energy of a .22 firing your ball upwards? if you ever have fired a gun, .22 is an oversized BB gun essentially.

Have you considered cam designs, as a potential replacement for your winch? Basically you have a spiral, starting with a small radius, and increasing over one revolution to a large radius, then snapping back to the small radius. As the motor turns this, the spiral pushes back on your spring. When it passes the drop from the large radius to the small radius, it releases the spring.

i don’t think you are allowed to store energy (like in a spring)

You are allowed to use stored energy under the conditions stated in the rules.

if you are planning on using air to compress the spring…why not eliminate the spring and use the air itself as a spring?

So to answer the actual question posed, 57 had a mechanism like this last year. You’ll find that the vast majority of devices out there simply won’t release under load. If you’re winding back and then using the latches to control the final release then it’s not as much a problem and you could use an andymark shifter and an andymark planetary adapt an FP to the tranny. That would work fine as long as you back the motor up so the springs take the load before you try disengaging.

That said, our solution was a little different. We used a worm-worm gear combo to drive the retracting pulley. An FP through a BB tranny drove the worm, and that whole assembly was mounted on a pivot and gravity biased to the worm wanted to pull away from the gear. So to retract, we’d use a cam on a servo to push the worm into the gear and retract. To fire, we’d move the cam the other way and then we’d have to reverse the motor slightly. The worm was arranged such that while retracting, it was actually pushed into the gear thanks to the location of the pivot point. Reversing the motor broke changed thing s enough that the worm broke free of the gear and out went the spring.

That sounds interesting Kevin, do you know if there are any pictures of it on the web somewhere? where they might be?

I don’t think there are any. At least, I didn’t take any, and it’s buried inside the robot, so it’s unlikely to be visible in beauty shots. As a caveat, it took build season and most of two regionals to get the working combination of software and hardware. And then an unkind robot picked out the one weak point and broke a delrin flange that attached the pushing plate to the pushing rod. Much of our time during build trying to get it to work was fruitlessly attempting to adapt various pre-made release devices to the system, like a heavy duty fishing reel, etc. That would be how we discovered that many such mechanisms lock under load. I imagine a dog shifting transmission isn’t likely to be much better.

The beauty of the worm gear design was that an appropriate worm meant the system wouldn’t backdrive, and yet the arrangement of it all meant it would get out of the way when firing time came. Were I to apply it to retracting a heavier spring like you would need for this task, I think I’d look very carefully at a small piston to engage and disengage the worm instead of depending on gravity to release it and an admittedly not entirely sturdy cam on a servo to keep it engaged.

Actually, come to think of it, the biggest single advantage of this system is the it’s much more likely to be safe. Using the latches to hold the punch while under load means that if your latches fail, then the motor either has to hold the load itself or slowly release it, and either one takes a lot of amps. You’d have a similar problem if low volts or mechanical problems prevented the punch from fully retracting. With the worm system, the motor can latch or release the punch at any point, so the only dangerous point of failure would be the mechanism that holds the worm on the gear. And if that can only engage the worm and you depend on a spring or gravity for release, then that’s pretty good too…

At any rate, as I said above, I don’t have a picture available. I’ll try to snap one tomorrow or this weekend, but I’ll also try to cad up a schematic of the arrangement and how the forces end up making it work.

Why not use a winch to pull it back to a latch and then unwind the witch letting the cable go slack or mabey using a really weak spring that pulled the cable in some other direction just to keep it from floping around. This way you wouldn’t need anything more than your winch and the latch on the spring.

Or you just buy an automotive winch that free wheels and adapt a kit motor to the planetary it comes with. Just fire a pnuematic to lock the winch in and release the free wheel to retract or release.

OK JUST as an example… (I lived in Europe… I do dangerous things…) NOT FOR ACTUAL USE but for illustration purposes… if you has two identical Metal hole saws…

and Meshed them teeth to teeth, on a shaft, they would lock in one direction. One attached to the winding spool (fixed in place) , one attached to the motor such that it can retract away from the … winding spool hole saw… , You can wind the spring back… Retract the Motored saw… and Yell "SCOTLAND Es FRE’!!! Or something equally appropriate when looking for deadly outcome with a projectile…
The concept is having two plates that lock and then releasing them, by separation… re engaging them to start the cycle over…

Um… In case you missed it… DON’T ACTUALLY USE the Metal Hole Saws!

In fact… forget I said anything!


See, we too were thinking along these lines last year. The problem is that any of these mechanisms that we found would lock under load whether by design or lack thereof. To the former, I’m absolutely certain that the freewheel on an auto winch is designed to lock under load. Or, at the very least, be very difficult to release under load. Imagine the consequences if you could easily bump the freewheel while using the winch to lift/tow/pull something.

Similarly for the locking plates. Going down the spectrum from profiles that don’t lock under load to those that do you have:

  1. If you trace an imaginary pencil from right to left along the “hole saw” teeth, it would never have to move to the left to follow the teeth. This triangle-wave like profile gets harder and harder to hold together the more torque you apply. Easy to release, but hard to load.
  2. Your imaginary pencil goes straight vertical tracing a tooth. This sawtooth profile is probably the best bet. Dog shifters like the Andymark shifter rely on flat surfaces like this. Holding thing together isn’t particularly difficult, but to release them, you do have to overcome the static friction of the two surfaces being pressed together.
  3. If your imaginary pencil has to move to the left at some point while you’re trying to trace to the right, then you’re really in trouble. This profile would lock together very well under any kind of pressure, but would be very difficult to pull apart, as to do so, you’d have to turn either the input or output side by pulling apart hard enough to backdrive it.

So, I think dog shifters and veeery mild profiles in category 3 would work, but they get harder and harder to release as you put more and more pressure on them.

This can be fixed by “re-working” the mechanism. We made a compact shoulder out of a modified winch in 2005 and never had a problem once we worked it over. It cost us $70 and we had to machine 2 parts (motor mount and input shaft) to get it to work. It’s a cheap solution that you can make work with minimal effort if look around.

We were thinking of using a gear with teeth missing to pull it back, although that would require a non-winchy device.

Why not mount a winch so it can spin freely, then use dog gears to engage/disengage it?

I like your idea about the shifters. Andymark makes planetary gearboxes that changes the output of a Fisher-Price motor into that of a CIM. Two FP’s Into that, into a Gen2 with one of the speeds taken off should be plenty of force to compress a decent spring, assuming a post gearbox reduction. It all comes pre-made, takes little modification, and is actually quite light.

You are allowed, as long as it’s safe. See <R01> and <S01>. (That’s a subjective call, and the officials will make their ruling based on design and operation.)

Those are prohibited by <R89>, because they do not share part numbers with the items on the Free Pneumatics Order Form.

We are planning a pneumatic actuated catapult, and will also have to figure out how to unlatch it, because it will start with the cylinders full of pressure, and halfway extended. The latch will have to release the catapult against full force of what will probably be two 2" bore cylinders. However, the cylinders will probably be located relatively close to the pivot, and the latch can be located at the opposite end, so this will give the latch mechanism quite a bit of leverage. We can also add more leverage by making the latch a lever, actuated by a long stroke, small bore cylinder. I think it will work out ok, but we’ll see today if we can draw up a design and run some numbers.

also one could use a steel pin that engages two holes in the parts, and the part with these holes could be made of a plastic material, with a low coefficient of friction. This might make it easier release the mechanism. It could be used on a motor design, by having the pin go sideways into a rotating plate. You would want it to be rather large in diameter, to get the force down to a reasonable level.