Hello!
This is my first post, so while I have been working after hours to make pictures, I can’t post them just quite yet. You all know how it goes 
Has anyone done any bots with either very long legs, or more then three degrees of freedom per leg? If no, why not? (I am new to legged mechanics.)
Thanks!
Caitlyn
Hi Caitlyn,
I have a hexapod, 3DOF legs. I’m playing with another joint. Probably the biggest limitation is weight. Extra joints add more weight. Then there’s the extra cost of the servos.
3DOF legs seem to work quite well. A foot joint (even passive) might be an improvement. But then there’s the extra IK calculations!
Alan KM6VV
Hello,
Look at most bugs and with 6 legs or more. Or even at the basic (not counting all the small joints like the thumb) human. We have two servos for the shoulder, one for the elbow. two for the hip,one for the knee. Same in animals. This is the (after lots of tinkinring by nature) optumum basic form of animal (on earth). Why would it be diferent with robots? We can try to replecate what is already there, trying to create something else is competing with tens of thousands of years of evolution and trying to create the best animal for our planet (kind of terrain too).
Basicly, I didn’t do any 'cause of that. 
Alan,
The weight issue, are current hexapods approaching the limit for the servos already? Because I am overhanging a large load, I was thinking of using the digital 5645 or equivalent. I know cost is a factor too, but I want a small robot to play with to “risk reduce” what I really want to design: A search and rescue bot the size of a small car… (hey, dream big! 
)
Which joint are you considering adding onto?
Sam,
Think more carefully about nature. Look at your shoulder for instance. It is a 3DOF joint, not a 2DOF joint. And also, when looking at your arm, your wrist is another major player, and should not be ignored. It also is a 3DOF joint. In fact, the same is true of your hip joint, and your ankle. So, just looking at humans, I would have to say that nature designed 7DOF appendages.
And actually, there are great reasons to do things different then how nature did it. For instance, take the car. It doesn’t have legs that it pumps really fast (e.g., the Cheetah), it has wheels that turn. Or the airplane. It doesn’t flap its wings (e.g., the bird), it forces a velocity of air over a shaped surface that creates lift. So, yes, there are perfect rationale reasons for designing robotics differently then nature, because mechanics are not 1:1 equivalents.
–
In any case though, my inspiration is natural. I am not really interested in building a “spider”, but a harvestman (a.k.a. Daddy Long Legs)
en.wikipedia.org/wiki/Harvestman
So my leg design is quite a bit longer then Lynxmotion kit standard.
I will post pictures tonight when I can get them uploaded to my webserver.
I said the basic form. Like no wrists and stuff. But I like your example of te car! Nice one 
For a humanoid I would do it 7DOF or maybe 6. As for a multi-legged creature, it can walk fine with 3. making many legs with strong servos (because you need quite strong servos to build that harvestman since of the looooong legs will be very expensive.) I think lynuxguy had built a CAD version of a 3-4 DOF leg for a multi-legged robot. The legs were very long too.
But it would be interesting to biuld it exacly to the same size, because otherwise it would be a different creature, not the same abilities. But they look too small to make it that way.
You may want to read the below thread where building a larger hexapod was attempted.
I don’t know why you say ‘attempted’? It looked like a pretty successful build to me. Some software problems, sure, but those can be worked out.
Certainly not enough to deter my crazy notions! 
Some leg pictures as modeled Thursday after work. I am not quite happy with them, though they do have some interesting quirks.
Servos at neutral positions:
http://www.siennahanon.com/Robotics/Hex_Leg_Img4.jpg
http://www.siennahanon.com/Robotics/Hex_Leg_Img1.jpg
‘Imitating’ a standard 3DOF leg:
http://www.siennahanon.com/Robotics/Hex_Leg_Img2.jpg
3DOF Mode, for saving power:
http://www.siennahanon.com/Robotics/Hex_Leg_Img3.jpg
with that you’re going to need prettypowerfull motors (with 6DOF and the lenght of the leg 
). it would be real cooland would walk nicely I would say.
I think you will find that pictures do not have to obey the laws of physics or actual hardware limitations. I’ll just say that the legs you have drawn have a 99.9% chance of failure to operate as you think they will. Far too much stress and loading for normal servos to handle.
I wouldn’t let the naysayers discourage you. I think it’s a cool project that is worth exploring. Maybe you can pull it off if the servos are strong and everything is as light as possible. For example, you may have to drill lightening holes in the leg tubes.
Doing things that others think is difficult or impossible is how progress happens.
hmm, how about avoiding both the naysaying and the charge-ahead without thought processes and instead do some torque requirement calculations to see if any servos that fit the form factor you have chosen are actually capable of doing the job.
That would be intelegent! Do a couple calculations to figure out what are the servos torque required.
I have been reading this post and i know they have a little bit less limitation movment wise, but you could always make a hex with linear actuators
they are a bit more costly that servos $80+ most of them, but i have found many with pots and feedback for under $100, if not under $90 that have pots in them and 110 lbs of pushing force
also you would be able to make the thing as heavy as you want, more or less, probably be able to carry quite a load too
again, the movment of the hex would be limited, but you wouldn’t have to worry about strength or weight, and if anything you could do some fancy mechanical-ish to give yourself more movement freedom, and if you plan on usig the digital servos like u said, then it would actually be cheaper, cause arn’t those like $120?
Just my 2 cents
The Giant Hexapod worked great. Really cool to see walking. It was adding the tracks underneath that wouldn’t have been possible due to weight.
With that long of leg segments you will need something more than a typical servo. Even an HS-9555 will barely be able to operate the lgs while hanging in the air.
However, if you were to integrate this into your design:
servocity.com/html/spg400_kit.html
They are only $40 and will greatly increase your available torque. The plastic gears will be fine for your application.
An HS-645 with this box will be cheaper than an HS9555 alone and provide much more power.
It would also solve another concern I have. Even BB servos can only handle so much side load. With Metal servo horns, you could increase their strength but the side load on the servos will be too much. The servo gear boxes use much larger bearings and shaft and are more robust.
They will add some weight but you can lighten the gear boxes considerably by drilling lightening holes in the spur gear and the ABS plates. Plus their increase in torque will outweigh the increase in weight considerably.
Just a servo alone, regardless of the type, will not have enough torque to operate legs of that length. And they likely will not take the side loading.
Another option would be to use Firgelli’s mini actuators. They have considerable torque and are quite light. They cost less than an HS9555 and if positioned correctly, could provide enough torque to operate the legs.
Also, what I would do is built the 3DOF part first. Test out it’s abilities using the gearboxes and see what you think. It will give you a good idea if they will be able to handle another segment. If you think so, then continue to built the other 6 segments.
And just so you know, if you start this project, know that it will cost you well over $2500. And that’s using HS645, gearboxes, lynxmotion brackets and tubing. Not including bearings, shafts, hubs, body, electronics, batteries, etc.
Not trying to discourage you, but from experience, I just wanted to give you a heads up.
I would not go with servos at all, but find some nice powerful motors and controllers of linear actuators(big ones).
I know it will cost a lot of money!! Thats why behind the scenes I would be building it in stages (Kind of like Linuxguy and his octal rover thing)
How do you do torque calculations? My background is in software/ai, its just I find this mechanical part of the robots so much more fascinating. Is torque simply a sum of all the weights times the distances of each weight?
Anyone know how much SES equipment weighs?
I figure I have four possible approaches:
Build it with RC servos:
Seems the community consensus is that this is not really feasible.
Build it with the Servo City gearboxes:
Will have to look into this one.
Build it with gear motors:
My initial thought with these is that I can use miter gears to put the motor in line with the leg (and hence perpendicular to the axle). I was looking at LM’s 14rpm motor for this purpose.
Use linear actuators:
I have an idea in my head for how to get 270 degrees of rotary motion from one of these. I need to model it, see your opinions. The major con is that I couldn’t use anything off the shelf. Anyone here do small CNC jobs for other forum members?
Hey Sienna, have you looked at quickster47’s CH3-R? He added LPAs to make “pincers” to two of the six legs. I suppose you can add the LPA to all six of the legs and have 4DOF legs. Use the longer C brackets for longer legs and maybe use the SES-based feets, i.e. hubs and tubes. You can probably use a 6" long tube or maybe even longer if you have some beefy servos (dare I suggest 24*HSR5990TGs? )
No, I haven’t seen his hex yet. I will have to!
The reason I wanted a ‘leg rotate’, was to enable the bot to climb stairs a lot easier.
And 24 servos of any type are already looking to break my budget, so 24 5990s would kill it stone dead
