pic: OH GOD! WHAT HAVE I DONE???

[roboticWanderor]

loading photo ...

[JesseK]

It really depends on where you have to pick up the game piece from, but I have an idea. For easier driver control, you could make the first arm coming off the large bottom sprocket a 4-bar linkage. Adds a bit of weight and slight complexity for mounting the next joint. Yet you essentially add another option in control the driver without having to add sensors, button, or more programming. This wouldn't work if you had to pick something up off the ground, unless you telescoped the arm out (tradeoff in complexity I guess).

At 70 inches long, we really couldn't make a suggest for the reduction until we see a game piece. For instance, this would be perfect for a 2007 tube using dual-FP's on the base, but for the 2008 trackball you'd want a high-torque CIM with some sort of braking mechanism.

[Aren_Hill]

you've made quite a nifty arrangement of manipulation.

As a driver id be a little intimidated to operate that with any dexterity unless i just had a miniature of it on the OI then it would work fairly well.

idk if you'll be able to pull off 1 fp down at the bottom to move that thing it looks like its getting fairly hefty up top. Im probably wrong though FP's are beasts (our shooter this past year is proof).

this thing would be very impressive to see on a bot

[Andrew Schreiber]

Originally Posted by Aren_Hill

you've made quite a nifty arrangement of manipulation.

As a driver id be a little intimidated to operate that with any dexterity unless i just had a miniature of it on the OI then it would work fairly well.

idk if you'll be able to pull off 1 fp down at the bottom to move that thing it looks like its getting fairly hefty up top. Im probably wrong though FP's are beasts (our shooter this past year is proof).

this thing would be very impressive to see on a bot

I second the feeling that it is nifty.

As it stands right now I would write a series of equations that will give you the position in a 2d plane (aligned along the face of the large sprocket) and then work them backwards. These equations would be pretty scary to figure out but once you do them and solve them for x and y you can just use a series of buttons to control it. If you added on a turret you just have to know that that plane will rotate. In all honesty it actually isn't that hard to control. If your driver wanted to be able to control it he could just have a stick to tell it to increase the distance in the y direction or the x direction.

On an unrelated side note, roboticWanderor, your cads are impressive and an inspiration.

[Aren_Hill]

I'm just usually a stickler for the controls being very very intuitive
hence why in 2007 i built this
http://www.chiefdelphi.com/media/photos/29147
and i wasn't even the arm operator

I usually wanna know exactly what the robot is going to do when i give it an input, as any delays or mistakes can easily lose you a match

plus im a mechanical guy and a mechanical thinker so simple links between control movement and robot movement are a must for me.

[Andrew Schreiber]

Yeah, I am a programmer and as such have a genetic predisposition to distrust mechanical systems.

A physical model would be simpler to do. Perhaps a compromise would work best, a mechanical model which could be used to train some buttons, yet would allow you to fine tune on the fly. Might be an interesting way of controlling a system.

[spc295]

how are you driving the second joint? what is the gearbox on the end of the arm for? is it to drive the second joint or to extend the telescope? if not what drives the telescope?

[Kevin Sevcik]

Loading depends on the center of mass of your arm, and how heavy an object you're trying to pick up. Gearing depends on the motor you're using, how fast you want it to go, and how often it will be moving.

If all 24 lbs of your arm is centered 40in off that pivot, you're talking about 960 in-lbs of torque. One CIM puts out about 25 in-lbs of torque at 40A. So you'd need at least 1:150 or so gear ratio, probably more, given the efficiencies you'd be getting at that ratio. Just for lifting the arm by itself. Four AndyMark StackerBoxes would get you to 1:160 or so, and you could reduce from there as necessary.

Also, this is for lifting the arm from dead flat, if it won't ever get there or you counter balance it, you could improve things.

[FoleyEngineer]

Originally Posted by Kevin Sevcik

One CIM puts out about 25 in-lbs of torque at 40A.

Hi Kevin,

Where do you get this figure from? I'm looking at the performance curve for the CIM and it looks like at 40A it's generating about .8Nm of torque which works out to .59 ft-lbs, or about 7 in-lbs - not 25 in-lbs as you suggest.

Maybe I'm mistaken. Wouldn't be the first time! Anybody?

Thanks!

[Jonathan Norris]

Originally Posted by FoleyEngineer

Hi Kevin,

Where do you get this figure from? I'm looking at the performance curve for the CIM and it looks like at 40A it's generating about .8Nm of torque which works out to .59 ft-lbs, or about 7 in-lbs - not 25 in-lbs as you suggest.

Maybe I'm mistaken. Wouldn't be the first time! Anybody?

Thanks!

Looks like he is taking about stall torque, which is a bit less then 25 in-lbs. Yes we run 40A breakers, but the amperage will spike a lot higher then 40A before the breaker trips. From my rough understanding with 40A breakers on a CIM we can get all the torque out of the motors. The only time I can remember hearing the breakers trip is stalling our drive against a wall...

[JesseK]

What is the bearing/hole plate on the end of the telescope for? It looks like a great place for a sensor, or for the wiring/tubing to come out of, but I could be wrong.

[JVN]

When the game is unveiled... ask yourself how many degrees of freedom you actually NEED. Beachbots, FTW.

-John

[EricH]

Originally Posted by JVN

Beachbots, FTW.

-John

Are you sure about that one, John? (OK, OK, that one has three. One to go up, one to go over, and one to lengthen.)

Since then, one degree of freedom plus any extras on the base.

[JVN]

Originally Posted by EricH

Are you sure about that one, John? (OK, OK, that one has three. One to go up, one to go over, and one to lengthen.)

Since then, one degree of freedom plus any extras on the base.

Eric,
As noted by the "FTW" I was referring to their WINNING robot, not it's ancestors.

I've been a huge fan of the super-elegant 2005, 2007, 2008 Beachbot robots. In 2005 & 2007 in particular I loved the 4-bar designs my teams utilized, but then saw that 330 could score as well or better with just a single joint. Simplicity yields elegance.

-John

[roboticWanderor]

Thanks a lot Kevin Sevcik, that is relly helpfull. what is funy, is that we just learned this stuff in phsiscs class!
Hopefully, an arm like this could be controlled with a phisical model, similar to the controller I built my sophmore year (it won the innovation in control award by the way).
The gearbox that is on there right now is for the extension, as that math was easy enough to figure out.
the upper arm's articulation technically will operate like a 4-bar linkage, but still be moveable (similar to team 60's arm design)