View Full Version : The Roach

02-09-2010, 09:03 PM
Just wanted to get some pics of the hex I'm working on up. The leg design I came up with is hideously complicated, which is why there is only one in the picture (and it's only half finished). It's about 14 inches by 10 inches, with a 3.5 inch femur and 7 inch tibia. I'm building it with only had tools, which is making for slow going. It should weigh about 8 or 9 pounds finished. I'm planning on doing some vacuum formed or fiberglass parts to hopefully make it look less like a cockroach, but I think at this point that's all it will ever look like to me. I figure a couple or three weeks till I'm ready to start the software side of the build. Anyway here's the pics (sorry about my crappy flash).

02-09-2010, 09:41 PM
You design looks great! I see a few servos in there. Have you done any preliminary calculations to make sure that servos will be able handle that kind of load?

I built a Hex a few years ago about the same size half the weight with a 5" tibia. The HS-645s collapsed under the force of gravity.

02-09-2010, 10:02 PM
I don't know these "calculations" you're talking about. Seriously, though I really can't bear to do that much trig and such until I figure out the final walking geometry.
I propped up the legless side, and the one leg held up the whole bot with batteries and the rest of the servos on top. Final weight is is just a guess based on the shipping weight of everything I ordered. The servos are 645's though, I don't need help being nervous about the torque ratings....
If it won't stand I'll start cutting out aluminum.;)

02-09-2010, 10:04 PM
Looks like an interesting bot :p

02-10-2010, 07:19 AM
I mean; did you estimate the motor requirements for your project when you came up with the dimensions. It's always a good idea to do some simple math up front; sanity check.

It is about 5" from the center of the robot. The HS-645 has 133 oz-in of holding torque at 6v (much less operational torque).

133/5 = 26.6oz per leg. You have 6 legs; 26.6 * 6 = 159.6 or 9.975lbs. These preliminary calculations are not at all exact, just a rough estimate. Plus I'm assuming the force applied to the servos is 90 degrees.

You're not going to get 133 oz-in from the 645s more like 100 oz-in at run time. The front and back legs are further from the center of the robot so the lifting capabilities will be less than the center legs.

I'm not trying to freak you out. When I saw you bot it brought back memories of my business partner (at the time) designing mechanics that could never work with the motors he was trying to use. Geez that was 10 years ago... Anyway, I think you might be at the limits of what your servos can handle.

Maybe a second opinion is in order. Anyone?

02-10-2010, 09:18 AM
It would probably be a good idea to lighten some of the parts. The square coxa mounts look really heavy. The chassis parts don't have to be solid all over.

02-10-2010, 12:21 PM
Thanks for the insight, I really am just playing with the design at this stage.

Mike, I see where you are going with the torque on the servos, but in the picture the angle on the femur is closer to 45 degrees and the tibia is basically vertical giving about 3 inches of ground clearance.
If the torque from the servo is approx 100 oz in, that gives me 28.5 oz of force at the end of the femur, double that for 45 deg and I get 3 lbs per leg. Even if my memory of physics is completely wrong and I can only support 28 oz per leg a nine pound bot should stand, though not walk.

Adam, the coxa are actually very light, thin walled aluminum. The chassis will definitely get some holes cut in it as soon as I get everything mounted the way I want it.

And as a final note I weighed everything on the bathroom scale and it came out to 180... I mean 5.6 lbs, which means that packaging weighs a lot more than I was accounting for.
Again thanks for the interest I'll keep things updated.

02-23-2010, 10:09 PM
Some progress, should have all the major structure complete this week. Here I have Roach standing on its own power thru RC, because everybody was making me nervous about the weight(probably with good reason, as I'm just winging it on the construction).


I tossed my spare battery on top just to add some load. I'm sure you notice that I've abandoned the cable actuators on the tibias. They were too difficult to machine with any precision on my drill press, and the cables started to devour any nearby aluminum within very few actuations.

I need to repaint the thing as the wind kicked up right after I finished spraying everything.

I went ahead and ditched the solid Al top and started a fiberglass piece to "pretty" things up and save some weight. Heres the form ready to be finish sanded, painted, and waxed.


Lesson for today is that shoe polish is not an appropriate wax for fiberglass work (I know, don't say it).The form came out pretty nice after a liberal dose of acetone and a screwdriver, but the plug didn't make it.Picked up some nice hard carnuba after learning that lazy don't pay.

For a first attempt at fabrication with fiberglass it is surprisingly easy to work with, but with lots of waiting. No problem there though, I have lots of other work to do. I'm putting together a simple power circuit, and trying to dream up a way that I can guarantee full amps to my servos without running all that power through my roboard (which should hopefully be here soon).

I'm thinking that a diode protected way to feed my spare battery into the power leads feeding at least the 12 leg servos would work, but I'm not real sure how to put that together or if it's even necessary. Can the roboard handle 8A+ @ 6V on 18 PWM's? I haven't seen any information about power throughput.

02-24-2010, 12:07 AM
Love your design maconSwift. but i would seriously be concerned about your weight. weight has a funny way of increasing when you least expect it.

03-27-2010, 01:30 AM
ROACH is coming along nicely. Been having some issues with my Roboard, but tyberius is working on it for me.

Got "final" build pictures up, and guess what.....

(sorry they're so small)


Too heavy. The following video is definite servo abuse,(ignore my weird hand gesture, it's late)


but that's OK. Learning C++ is slow going for someone who knows absolutely nothing about programming. Gonna start reworking the leg design while I figure out the code. Probably be months before I have the thing walking. As long as I don't blow up any servos it should fit the bill as something to keep me busy without costing any more money.