PDA

View Full Version : PC Controlled Jib Arm



davidhere40
02-15-2012, 03:40 PM
I'm making a computer controlled jib arm, the basic idea of a jib arm is this: http://protog.com.au/images/CobraCrane_Fotocrane.png
I'm trying to figure out how to motorize it. I want durability, because it's for an industrial application and will be in constant motion. It's not for photography or video. It's basically a really long robotic arm.
I could really use some help making the right mechanical and motor/controller choices.
Here is what I have so far:

I want to motorize three dimensions:
1) Telescoping in/out movement: I have added a slider to the jib. It is going to be driven by a v-belt. This allows it to extend like a telescoping arm. It will be able to extend 6 feet. The slider will only weigh about 4 pounds max (about 2 pounds for the telescoping square aluminum tubing and less than one pound for the payload). I'm also considering to use a counter weight on the other side of the v-belt that moves in the opposite direction. It would balance the arm at the fulcrum as well as would prevent gravity from creating any net force on the v-belt.
For the v-belt, I was planning to use a sheave and pulley with a radius of about 1 to 1.5 inches. I just guestimated though when I purchased it, because I have no experience here. V-belt was chosen for low cost at long lengths, simple maintenance, mechanical simplicity and durability, but I have no experience with them.
2) Left/Right of the entire jib. The jib has a bearing for smooth lef/right roation at the fulcrum (the point it balances on).
I'm not sure what actuator type to use here. I'm considering either a stepper motor or a linear actuator.
3) Up/Down: Plan to use linear actuator for this. but might consider using the steppers also, if it's more durable.

I've been working for a few days on the in/out (telescoping) actuator. I thought about using a 640 oz-in nema 34 stepper.
I calculate the real torque (68% of what they claim based on torque charts in the specs manual) to be:
17.4 pounds of torque at 1.5 inches (the radius of my v-belt sheave)
But, my load inertia to rotor inertia is 36 times with an 8 pound load and 1.5 inch radius sheave. I could possibly get it down to a multiple of 6 if I make the load 6 pounds and the sheave 1 inch in radius. I'm not sure the v-belt will even do that though.

Speed is important. I want to able to move something like 9 inches a second at top speed.

Is this design totally off?! What would you do to improve it or make the idea work? Any help would be much appreciated.

Thanks,
Dave

Slugman
02-15-2012, 05:11 PM
Couldn't tell you if the design is off, but if you can somehow make the pulleys adjustable in the design it may help later. I am assuming you are going to be able to re-tension the V-belt. I.E. Allow room to fit different size pulleys at one end (Preferably both ends) so you can change sizes if required.

I haven't tried to do all the calculations, so this is just a quick comment after reading your post, but one problem with telescoping is the added 'slop' in the arm due to the gaps between the telescoping sections, which is worse when it's extended. I hope your desired accuracy is able to be achieved with your setup. The usual robot arms such as the ones making cars etc have huge joints just to get the accuracy they need. A bit hard to explain, but larger diameter = smaller arm slop assuming the same tolerances in a bearing.

I just googled for some working examples, & found the "JonyJibs" at http://www.videouniversity.com/shop/jonyjibs-and-accessories/
Could you modify something like that to do what you want? Wheels are notoriously expensive to re-invent. Trust me - I did just that - It took me three versions, & I'm still adding up the total cost. :rolleyes:

Quantum
02-15-2012, 06:04 PM
Sounds like a cool project.

The way your starting out looks good but the one problem is your looking to go 9in per sec. That's 540in per min and that is fast, real fast.
Stepper motors wont work with that kind of speed. You should look into servo motors.
You still need some type of ball screw setup to make this work. The slop from a v-belt will kill you, it might and probably will slip.

I've built my own CNC Mill and converted it to a router. I chop AL at 30in/per min .06-.1 deep and can raise the head about the same speed.
I'm using the same motor to lift my head. 640oz/in

A more reasonable .5- 2(max)in. per sec maybe a chain setup.

Not trying to kill your project but just trying to save you some time/money as I've tried some of these things in the past.

All industrial robots use servo motors in the joints with feedback. There strong, fast, and expensive.

davidhere40
02-15-2012, 08:56 PM
Thanks so much for the ideas!

Accuracy isn't important actually. It is driven and guided by computer vision for the most part. So, really I technically don't need any feedback or even stepper motors. I was originally thinking to just use brushless motors and a USB controller. But, the complete lack of position control made me think I better use steppers. Plus I had a really hard time finding good hardware.

What if the slider was only 2.5 pounds? I can do that if I don't use a counter weight. It is less inertia, but without a weight on the other side of the belt, gravity will add 1.8 pounds of negative force when the arm is at 45 degrees. My original thought was that inertia was better than a constant negative force.

I'm afraid to use certain types of mechanisms like linear actuators. Because I don't trust their durability for a reasonable price. Price is important to make the robot cost effective. But, so is speed. Linear actuators are both slow and very expensive for high quality. I did get one, but have no idea what's in it! Not good :( Plus, because its completely closed, I'm afraid of invisible wear if I go cheap and future break downs.

This is why I was thinking of using belts because at least the wear is visible, they last long, the components are cheap, and repair is very very fast. If a belt breaks, just about anyone can put a new one on.
Since accuracy isn't important, do you think using belts will be ok for me? Slippage is ok. I'm hoping it doesn't require daily tensioning or something, which would be a deal breaker. Maybe a spring for auto tensioning? And maybe monthly adjustment by hand? Do you think this is approach might work? Regarding the telescoping... it's not really telescoping. It's really just a square tube that is hugging a second square tube using simple skate bearings. At the end, the bearing is attached to the main square tube, allowing the moving one to slide, but stay where it should be.

On your existing CNC machine, if you used a belt drive (rather than screw); full bipolar stepping; and slowly ramped up the step frequency, do you think you could accelerate your system to the types of speeds I'm talking about? I mean, maybe I don't need 9 inches per second. I certainly don't need that all the time. The most common movements will probably only be about 1 foot. So, if it could do that in let's say 3 seconds, but be able to move faster over 6 feet, that would be ok. What I don't want is for it to go 1 inch per second and take over a minute to fully extend that arm. That would be a serious problem for the arm to do enough work to make it worth the investment.

Another reason I have been moving away from servos is they have a low duty cycle and I believe they wear out because of the gearing. Plus, most don't have continuous rotation. Maybe I'm wrong, but that's why I've been moving towards steppers or just brushless motors.

Regarding existing jib arms, I thought about it, but they aren't quite right for me because they typically have more filming features than I want that are not useful and are missing the most important feature: sliders. Also, the cost of jib arm can really be accomplished for about $150 or less! It's the simplest part of the system :) The part that is going to drive me crazy is the actuation.