Well, we’ve figured out that you aren’t going to build a 3-legged “swinging” robot, but here’s an interesting animation study by a friend of mine.
drzeus.best.vwh.net/wotw/other/gait/
Alan KM6VV
I have been doing some more poking around with the torque needed for each leg of my long leg quad robot. At this stage im not sure if anything ive done with the math is acurate at all. Most of the end results I got in the tutorial and the calculator seem too low(meaning very little torque required).
So far I have these as my variables for this tutorial
robotshop.ca/robot-leg-tutorial.html
W4 (my payload …or as far as i can tell thats what that means) 1.2kg
L3 L2 L1
100mm 762mm 762mm
3.93700787 inches 30 inches 30 inches
W3 W2 W1
190g 190g 190g
6.7oz 6.7oz 6.7oz
Additional information on the part specs can be found below
Please any help on working out max torque requred would be super cool.
Carbon Fiber Tube
OD inches WEIGHT lbs / ft.
.825" .062
Total weight Each leg segment
Segment length = 0.762 meters
0.070306817 kg
SERVO
Weight:
Grams OZ
0.19 kg 6.70
4.8V 6V
Speed(sec/60°) Torque Speed(sec/60°) Torque
Kg-cm Oz-in Kg-cm Oz-in
0.15 35.00 486 0.12 42.00 583
Ok thought I would post a quick image of the concept thus far, I have not added the payload yet…but I have used mostley standard parts for the servos and joints…
http://i873.photobucket.com/albums/ab294/navanod/tripod_01.jpg
another image this time with the first test of a carbon fiber material
http://i873.photobucket.com/albums/ab294/navanod/tripod_01_02.jpg
Nice images but It looks like the servo shaft is going to be under a lot of stress bolted to the triangular body piece. There seems to me a lot of stress at just about every single servo joint. A better design would to remove the servos and use cables and pulleys at the joints and use geared motors to pull the cables. Then I think you will have a design you can work with.
hmm not really sure how to go about cable design…any robots you know of can you post some info? I know the joints willbe at alot of stress but i think in the end I will make carbon fiber joints. To me at the moment whats most important is ti see how much torque I will need…to see if i can find a servo that can move it…,then i alter the design from there
Your design is definitely the direct way to do this vs. a cable pulley system, and perhaps using ultra light materials you can make it work. Eddie mentioned balsa wood which is about how light you want to make the poles. I’m not familiar with carbon fiber material since I have never worked with it, but if it is a light as balsa wood or lighter, then it would be a great material. You can also save a lot of weight if you power the design using tethered wires, but I’m not sure after adding the electronics, screws, and glue, what the end weight will be. The only way to know for sure how things are going to work is to build a proto type, then you can see what needs to change.
Not sure what the problem is. The torque on the servos is minimal when the legs are vertical. The torque will go up dramatically the farther you move the servos, but you shouldn’t need to move them much at all.
I assume you are just going to push the COG around to lightly drag the lightest leg in a direction. This will never work like the animation from above. The servos only have 180° of throw to work with.
The only thing you need to do to strengthen the horizontal throw servos is add a lower plate to the chassis.
the carbon fiber legs(which are a pre made tube i can buy have a given weight per inch so i can work out what they way fairly evenly). The specs for the servos are above and they are most heavy part 6.7oz or 0.19kg. I will be powering and controlling the tripod from a back pack including all the given control parts(batt, servo control, power amps). The only weight is the payload which is about 1kg or just a bit less(payload is on top).
All the specs so far are in a few posts up the page. Further more it looks like I might have access to one of the local Uni’s here to work with some industrial design guys as part of a class project…so I should have the means to CNC the parts, or the carbon fiber moulds:) Im good at designing and putting parts together but not so good at the math…which is why I allways ask for help…or im going to keep re inventing the wheel so to speak.
So im still trying to figure out how to take what I have so far and work out the torque required…Comming soon I will be posting some 3dsmax files(in scale) along with some rigging and basic IK. But any help on the math would be awsome:)
cheers
Nav
ps sorry if there are lots of typos…I am writting this on a mobile half way round the world in the middle of the night with thumbs that are too big for the keypad
ok so…
if I design it as a tripod(what im trying to figure out now) the tripod will have a similair gait as the following video
youtube.com/watch?v=4FABBKPo … re=channel
If its to crazy to get the torque required because of the cost for each servo then I will consider a quad. maybe lets jsut start with getting it to raise the center pivot as much as possible up and down…animation comming soonish
Yes that is a bot built with SES / Phoenix parts by Sienna. That’s exactly what I meant by shifting the COG to move certain legs. So yeah, it will work.
Great looking concept drawings!
I’d have to agree with the previous posts, way too much stress and torque required just to move the legs a little!
Brackets (universal) that supply a ball-bearing joint opposite to the servo shaft might help, or find servos with this joint built-in.
If the leg material is cheap, and you already have a couple of servos, you might consider just building up one leg, and performing experiments with it. You can make the other two legs and joints rigid (fake the servos), and see what force the servos can provide on that one leg.
1 Kg payload? Why so much? Think of a camera tripod. Find one with similar leg lengths, and see what “payload” it is advertised to support. Might give you some ideas/reference.
Alan KM6VV
the thing is im not sure how much is too much torque…lets just assume that the joint is 100% efficent, how much torque will the servos need for that leg length and weight?
I have to find a good starting for which servo to buy, I am sure there is a servo out there that can do it…but I just dont know what it needs todo yet.
Torque is length * weight, and you probably only intend to “lift” a leg a limited angle, so torque could be estimated:
Sin 30 degrees (leg lifted) would be about .5 * length * weight. 90 degrees would require the full torque (1 * length * weight).
I don’t believe you can have too much torque! Excess torque just doesn’t get “used”.
Alan KM6VV
Ok I found a little cheat calculator here
frictionhinge.com/calc.html
So I put in my greatest input values for the joint that will carry the most stress
Weight .6kg
Length 60cm
Distance to CG 30
Angle (0)
then I get a result of 18 Kgf-cm or 15.63 Lb-in
So yeah thats an important result…because servos in that torque range are not to expensive. My only worry with that calculation is that the center of mass for the leg is probably furtheraway from the pivot seeing as it is carrying another leg and servo too.
If the center of mass is offset, then I suspect it would increase the apparent angle. More torque needed!
Yeah, that could cost ya! 256 oz/in servos can’t be cheap!
Alan KM6VV
I want to add real quick that servo ratings are stall torque ratings meaning the servos are not able to move once the max torque ratting is achieved. So you will need a servo stronger than 15in. lbs. based on your calculations previously. Just something to keep in mind.
This servo would work: (417 oz in.)
Ouch! > $100 !
Yes, that servo looks like it would work. Gonna take one BIG battery!
Alan KM6VV
yeah its on a cable power supply to start with:)…so yeah its alot to drive but can work…there are other servos(not from here) that can deliver up and over of 100kg…then after that torque we head into commercial robot servos $$$. So I guess my main task right now is to keep within the so called hobby range of servos. I mean if there is a servo that you know of that can do more torque then please post. The more we share this knowledge the better our projects and products can become.
thanks for all your help so far guys:)
nav
Something else you might want to consider is the use of helper springs. I can’t remember if this was already mentioned or not, but if you attach springs over each joint, you can really cut the load down. It would be easy to do and help out the servos a lot. Smallparts.com has many types of springs.