PDA

View Full Version : [Project] My Hexapod (not just yet) Walker **Un-named**



Superlaxstar112
07-21-2008, 03:23 PM
This is my first robot that was not designed to fight other robots. I've been competing in BBIQ battlebots challenges since I was a freshman in High School. I am now going to be a senior in high school and I figured that it was finally time to try something else. I had some cash to burn and decided to build a walking robot.I got some HS-322HD servos and started playing with them. Looked around on the internet for a bit and found this helpful community. Trossen robotics has been a great resource over the last few weeks. I started with some simple ideas that I modeled in Solidworks. I eventually moved towards a symmetrical hexapod walker. Finished my drawings and models and then I hit the machine shop.

Goal: Create a robot that can walk in any direction (except up) and that can be independent of cables and wires leading away from the bot.

Materials:
1 7.2 Volt 4200 mAh NiMH Battery for Servos
1 6.0 Volt 2200 mAh NiMH Battery for Pico-ITX
1 Pico-itx computer
18 Hitec HS-322HD servos
6 Hitec HS-645Servos
Bunch of screws a nuts (4-40 and 2-56)
18 Aluminum servo horns
1 SSC-32 Servo controller
lots of patience

Final Cad:
http://i95.photobucket.com/albums/l144/superlaxstar112/hextest.jpghttp://i95.photobucket.com/albums/l144/superlaxstar112/hextest.jpg

Drawings of some of the parts:

http://i95.photobucket.com/albums/l144/superlaxstar112/bodycircle.jpg

http://i95.photobucket.com/albums/l144/superlaxstar112/bottomofleg.jpg

http://i95.photobucket.com/albums/l144/superlaxstar112/hiptoknee.jpg

I had some friends help me out with the CNC router and I began to make the parts. I used scrap lexan that was lying around so some of the parts had some scratches on them but nothing too serious.

Some of the finished parts with some servos attached (on my kitchen table)

http://i95.photobucket.com/albums/l144/superlaxstar112/IMG_1009.jpg

http://i95.photobucket.com/albums/l144/superlaxstar112/IMG_1010.jpg

http://i95.photobucket.com/albums/l144/superlaxstar112/IMG_1011.jpg

And then here is the **Almost** finished assembly. It is missing one leg because I didn't have a screw to attach the last servo horn. But it still stands on its own. (On my bedroom floor)

http://i95.photobucket.com/albums/l144/superlaxstar112/IMG_1012.jpg

http://i95.photobucket.com/albums/l144/superlaxstar112/IMG_1013.jpg

http://i95.photobucket.com/albums/l144/superlaxstar112/IMG_1014.jpg

I just ordered 6 more servos (Hitec HS-645) so that should solve my missing screw problem. Which should also be a serious upgrade to what I've been using all over the rest of the bot HS-322HD.

I'll update this after they come. My final goal now is to get it to walk and then eventually put a camera on it so it can facetrack and what not.

DresnerRobotics
07-21-2008, 03:37 PM
Great project!

Now I'm not trying to be a downer at all, just want to share some personal experience I have had with walkers and the weight a complete pico-itx system adds...

I would start doing payload testing asap to see how much that bot can hold and still walk well. While the pico-ITX is pretty light by itself, the supporting components start to add up quickly in weight. With the length of the tibia and femur on that hexapod, your servos are going to be working VERY hard to hold it up. I would suggest no less than HS-645MGs on all verically supporting servos.

You might also look into a lower body plate so that your horizontal hip servos aren't supporting the weight solely on their servo shafts.

If your budget permits it, you might also look into using LiPo batteries to help save some payload.

Great going though! Love the design and can't wait to see it walking. This would be the first pico-itx powered hexapod I've seen.

4mem8
07-21-2008, 03:42 PM
Superlaxstar112: Great job, love to see it working. But as Tyberius has stated watch that payload. Nice job.

Superlaxstar112
07-21-2008, 05:05 PM
I was concerned about weight problems as well. Especially when I did the calculations out to see how much each leg could hold and discovered that besides what it had on it at that moment it couldn't hold much else. I have set it up with the 5 legs it has right now and slowly added weight to the point where the servos started to give and it actually held quite a lot. Now this does mean that Ill have to take its weight issues into consideration when I do the programming. I might have to move one leg at a time instead of 3. I'll have to see when the time comes. I was considering using Li-polys but I don't have any experience with them. As far as NiMH go they are pretty simple, but Li-polys are different am I correct?

DresnerRobotics
07-21-2008, 05:19 PM
I was concerned about weight problems as well. Especially when I did the calculations out to see how much each leg could hold and discovered that besides what it had on it at that moment it couldn't hold much else. I have set it up with the 5 legs it has right now and slowly added weight to the point where the servos started to give and it actually held quite a lot. Now this does mean that Ill have to take its weight issues into consideration when I do the programming. I might have to move one leg at a time instead of 3. I'll have to see when the time comes. I was considering using Li-polys but I don't have any experience with them. As far as NiMH go they are pretty simple, but Li-polys are different am I correct?


Holding torque and moving torque aren't always the same though. I would create a walking gait first, then have it run a simple walk forward sequence for say.. 10 seconds. Add 100 grams of weight and try it again, and so on. See how much you can really add before your gait performance starts to suffer. Also keep in mind the servos you're currently using are nylon geared, and pushing them to their limit will indeed wear them down sooner.

Alternating tripod gaits will require even more torque, as you're supporting the same amount of weight on less legs each time you take a step. Might be wise to start with a static gait.

Ideally I would invest in 12 HS-645MG servos for the vertical lift joints and go from there.

I would suggest reading up on LiPos, do your research before committing to them. They are dangerous if improperly used. The most basic things to watch out for are over charging and over discharging. There are regulator circuits you can add on to monitor and control this, but its still something I highly suggest you research before trying.

Superlaxstar112
07-21-2008, 05:22 PM
I've never used LiPos before cause I've heard they explode and well I've never really wanted to risk it. I will definitely look into them though because they will definitely save weight. Another thing is, this is my first walking robot. I am a total newb when it comes to any of the vocab and lingo. I tried to design this on my own and see what works and what doesn't without taking too many ideas and what not from others. With that in mind... what is a gait?

DresnerRobotics
07-21-2008, 05:32 PM
LiPos are safe as long as the user knows what they are doing and doesn't overcharge/discharge them. Like I said, you'll want to do your research on them before considering them.

Weight will be your biggest enemy with a project like this, and I honestly am not sure the servos you have will be sufficient.

A Gait is the type of walking sequence any walking animal or machine uses. Alternating Tripod is an example of a walking gait for a hexapod. I've heard the 'one foot up at a time' type of gait described as a static ripple gait.

ooops
07-21-2008, 06:13 PM
Just a thought guys, and I am letting the cat out of the bag here, but what if you mounted one battery on each of the legs below the servo. It would take a lot more wire to series them, and I suppose there would be a voltage drop, although minimual, but basicly you would remove the weight of the batteries on the legs that are touching the ground during the gate. The trade off is that you add mass to the leg or legs that are in motion, but seems to me that would be a small price to pay to increase the payload by the wieght of the batteries that are touching the ground.
Please share any thoughts you have on this.

DresnerRobotics
07-21-2008, 06:24 PM
Just a thought guys, and I am letting the cat out of the bag here, but what if you mounted one battery on each of the legs below the servo. It would take a lot more wire to series them, and I suppose there would be a voltage drop, although minimual, but basicly you would remove the weight of the batteries on the legs that are touching the ground during the gate. The trade off is that you add mass to the leg or legs that are in motion, but seems to me that would be a small price to pay to increase the payload by the wieght of the batteries that are touching the ground.
Please share any thoughts you have on this.

You'd kill those leg servos. You would still have to support the weight of one of the batteries at a time on the 5 other legs as well, not to mention the additional stress on each leg holding the battery. This would also probably wreak havoc on your walking gait accuracy, having to basically throw all that weight around.

I guess its possible, but certainly not the approach I would take.

Zenta
07-22-2008, 05:30 AM
Hi!

Interesting project you got there! Very nice built.
I'm sorry but I'm a bit worried about your coxa design. As pointed out by Tyberius, the coxa servo (hip vertical) are probably eventually going to break caused by the stress/load on the servo axis. Especially when you've added more weight and also trying to make it walk.
Another alternative is to reduce the femur length to the half of it, that would also give your hexapod much more load capacity.

Visit Oricom (http://www.oricomtech.com/projects/leg-time.htm) for more information about walking gaits. the wave gait is probably the safest way to walk for your robot, it is slow but strong. BTW, ripple gait is my favorite.;)

Keep up your good work!



Just a thought guys, and I am letting the cat out of the bag here, but what if you mounted one battery on each of the legs below the servo. It would take a lot more wire to series them, and I suppose there would be a voltage drop, although minimual, but basicly you would remove the weight of the batteries on the legs that are touching the ground during the gate. The trade off is that you add mass to the leg or legs that are in motion, but seems to me that would be a small price to pay to increase the payload by the wieght of the batteries that are touching the ground.
Please share any thoughts you have on this.

LOL, yup you just let it out! Actually it is a very interesting idea (I've been thinking about the same too;))! But you'll need a very very different leg and body design. And it would give your hexapod very many limits about walking. Each time you are lifting one leg the COG will move VERY MUCH towards the lifted leg (maybe too much?). Also take a look at the nature; spider, insects, dog, cat and so on..

ScuD
07-22-2008, 06:12 AM
I would also recommend making a double platform, adding a platform beneath the servo's and adding shafts to these, so those servo's have double axes (man I'm having a hard time explaining myself these days)

The most stress in the entire design is on the axes of those servo's, and you'll strip the gears like butter.

Here's (http://users.pandora.be/svendecock/robot/robcomplete.jpg) an example of how I solved this issue, though in my design the servo's themselves are embedded in the platform and not the servo axes, so i needed to adapt the legs themselves.

I still had problems with the servo's though, they only had about 3kg/cm of holding torque, with each step it took it sunk a little deeper untill it was crawling around on it's belly.

One more thing; please do not take any of these comments as criticism, designing your own bot is a thing that is VERY much appreciated around here, we're just sharing our own experiences and trying to avoid you from having the same problems we did.

Oh, and welcome! :wink:

Superlaxstar112
07-22-2008, 08:08 AM
What is a coxa and femur? :-P or which part of the leg? And I do see what you guys mean about stress on the axles. Any Ideas of how I could mod my current design to accommodate the issue without adding too much weight? And yeah if anything you guys are trying to help me not destroy my own robot. I love all this help

ScuD
07-22-2008, 08:34 AM
http://weta.boarsnest.net/LegFig2.jpg

That's probably the first question I asked when i got into hexapod's too :o

For strengthening your bot, my first suggestion is a double-decker platform.
Just add another platform below the one you have already, and make some axles on the servo's to fit through it.

I'd draw a little paint-mumbojumbo schematic but i'm at work right now, I'll see if I can explain better tonight.

Zenta
07-22-2008, 08:45 AM
Hi,

A simple and easy way to add an extra axis/joint on to the bottom of the coxa servo is to use a Injected molded servo hinge (http://www.trossenrobotics.com/store/p/5327-Injection-Molded-Servo-Hinge-Two-Pack-Standard-HD-.aspx).

-Zenta

Adrenalynn
07-22-2008, 10:54 AM
Further with ooops idea is the rotational mass and acceleration. I suspect it wouldn't take long for that to destroy servos and rip out joints.

ooops
07-22-2008, 12:29 PM
Further with ooops idea is the rotational mass and acceleration. I suspect it wouldn't take long for that to destroy servos and rip out joints.

Not defending the idea, but rather attempting to go from white board to trash can if that is appropraite. Quick scale of a standard AA cell (not rechargable - but what was quickly available) the weight is .9 ozs. If one were to cut out a slot in the lexan leg just below the servo it seems without having any lexan to weigh that the net increase in weight would be negligable.
By adding just one battery per leg the overall weight isn't the issue rather the starting and stopping of the additional weight (rotational mass) of the one leg is.
Soooo, how much weight is too much?

Adrenalynn
07-22-2008, 02:44 PM
Soooo, how much weight is too much?

When your servo explodes in a molten mess of plastic, that's generally considered too much. :p

Beyond that - you're on your own for experimentation...

ooops
07-22-2008, 03:45 PM
Beyond that - you're on your own for experimentation...

That is scary and dangerous;)

Superlaxstar112
07-22-2008, 04:07 PM
So I ordered some of those injection molded holder things. Does anyone know the diameter of that little nub sticking out of the bottom? does that rotate freely or should I seat it in a bearing?