6 leg robot

[uaefame]

Hello everyone,

I am trying to build a 6 leg robot. I have few questions about motors.

The robot won't be heavy must be less than 10kg.

I wanted to know which motor will best serve the job.

Stepper motor? DC motor? Servo motor?

a- Which motor will consume less time to program?
b- Which motor is best suited for the job?

Question a & b can have different answers

Best regards,
Hesham Ismail

 

[dknguyen]

RC Servo motors are by far your best bet both for less development time and suitability for the job, and most importantly cost. "Less than 10kg" is still very heavy even before you take into account the leg length and robot design. You definately want to go for the strongest servo motors you can find. The Hitec 7955TG (titanium gears RC servo) or 5990TG (robotics version of 7955 with a pivot join on the other side of the output shaft) or 5980G (steel gear version of the 5990TG) are among your best bets, as is the JR 8711. THey are pricey though, but it comes with the territory of legged robots. BUt I would try and keep the whole thing under 3kg. THen it might cost you just $1500 in motors for 2DOF legs. If you want 3 DOF per leg...do the math.

The only other servo-type motors that are stronger are the:
http://www.tribotix.com/Products/Robotis/Robotis.htm
available at
**broken link removed**

THat works out to be $5-6k in motors for 2DOF.

These are very expensive though but any other solution involves building/getting an servo controller, getting a motor from a manufacturer, a position encoder that fits on it, and a gearbox that fits on it. Having so many things attached to one motor almost makes it necessary to get it from a manufacturer so you don't have to go trial and error with the encoder and gearbox. Very very expensive.

WHy is your legged robot so heavy anyways?

For that much money (and less even) you could just get a tracked robot from:
**broken link removed**
**broken link removed**
that would have greater payload, lower power consumption, greather reliablity, and greater climbing performance and speed with less development time. Yes, legs beat out tracks, but only if you have muscles instead of motors.

 

[uaefame]

It is not heavy the competition that i will join require the weighted to be less than kg. I didn't build the actual robot yet i am in design stage right now.

Thanks for the valuable information.

 

[dknguyen]

THen servo motors are still your best bet in development time size, cost programmability, assembly except now with a much lower weight they can be much cheaper.

 

[uaefame]

I need servo motor than can be shipped to my country a week or 2 weeks max. Any online website that can perform this job is acceptable.

 

[dknguyen]

https://www.servocity.com/ Might do it.

**broken link removed**

But of course, you have to call them to tell how soon you need it and shipping.

 

[Nigel Goodwin]

Servos are obviously the only way to go, and there are loads of examples available on the net to steal ideas from.

 

[Roboticinfo]

As everyone has stated servos are going to be your best bet

Here's a couple of links to get some ideas

Norris Labs

**broken link removed**

**broken link removed**

YouTube - Phoenix ripple walking

Now these are mostly just websites to spark your creativity or
to implement your reverse engineering skills, it may not be the best websites but it gives you a general perspective on walking robots that use servos.

**broken link removed**

Has a couple of walking robot kits that you can take a good look at and get some really
cool ideas to start designing your walking robot.

There is also a book called insectronics by Karl Williams,
he has a design on a walking robot in there that is pretty simple.
Although there is so much material online to research about
walking robots I wouldn't even bother.

I might include it has a cool chapter on grippers
although that is irrelevant to the question you are asking.

Good luck

 

[Cotowar]

Have you considered using linear actuators as opposed to servos?

I know servos are a good choice, but for the motion of the leg lifting up and down, you might think about doing a linear actuator with some small gauge wire running through to the tip of the leg. As long as the wire is in a closed off environment, you can pull it to straighten the leg, and push it to bend the leg. This will save you a crap ton of time if you are thinking about putting multiple joints in your legs and making them work. I'm not really sure how you would go about programming something like this, but basically if you supply power to the actuator you can get the leg to open, and if you reverse the direction, you can close it. It will be very accurate, and depending on what actuator you buy, it can be very fast as well.

Just an idea, you should really look into it further though if you are considering it.

 

[Roboticinfo]

Have you considered using linear actuators as opposed to servos?

I know servos are a good choice, but for the motion of the leg lifting up and down, you might think about doing a linear actuator with some small gauge wire running through to the tip of the leg. As long as the wire is in a closed off environment, you can pull it to straighten the leg, and push it to bend the leg. This will save you a crap ton of time if you are thinking about putting multiple joints in your legs and making them work. I'm not really sure how you would go about programming something like this, but basically if you supply power to the actuator you can get the leg to open, and if you reverse the direction, you can close it. It will be very accurate, and depending on what actuator you buy, it can be very fast as well.

Just an idea, you should really look into it further though if you are considering it.

Well that wouldn't be a good idea
for the most part linear actuators are extremely slow, sure you might be able to locate
one that can travel maybe 2" inches a second but that's still rather slow for a walking robot. Not to mention linear actuators are extremely pricey, you could buy almost two heavy duty servos for the price of one linear actuator. The only way to achieve a walking robot with a cylinder like piston would be to use an air cylinder. The negative part of going that route is it's extremley hard to get the timing right with the solenoid valve if you are a beginner programmer not to mention you would have to have an air compressor on board or
long hoses that would connect to the solenoid on board the robot.

Servos will be your best bet and cheapest
not to mention easiest when it comes to programming.

 

[uaefame]

Thanks for the feedback, I will be using servo motor. I am still looking for a servo motor for my project. I am not sure which size to select? Is there a guide on selecting the size of servo motor.

 

[blueroomelectronics]

You'll have to calculate the torque you'll need for your robot. The larger ones are very expensive. A small hobby servo should suffice for a small six legged robot.

 

[Karbos]

Another problem you will have to look out for is stripping of servo gears. The "Karbonite" gears are just fancy resin gears. I would (if your budget allows) get some all metal gear servos. Might as well get a fancy pair, or a lot of cheap ones.

 

[uaefame]

Thanks again for the replies,

Budget is not a problem right now, the only main problem is time.

I am looking for servo motor that can handle 30kgcm and 7.5kgcm. Any good website around that ship to UAE. ASAP.

Btw, what you mean metal gear servos can you define an example and its cost?

 

[blueroomelectronics]

Try either RobotShop or Pollu electronics. They'll have specs there too.

 

[uaefame]

I decided to got with ds8911 JR servo motor its 25kg.cm.

I will be using 3Servo motor per leg.

I have 6legs. Therefore, total number of Servo Motors will be 18. What is the recommended pic to control 18 servo motor. Do you think it can be controlled using pic16F818.

If not, recommend to me serveral pic that can do the job?

Thanks in advance

 

[uaefame]

Any ideas which PIC to use to control serval servo togather.

Does AN0-AN4 mean I can control 5servo motors?

I am totatlly confused about selecting PIC that turn 18 servo motor?

Any help will be appericated!

Thanks in advance

 

[blueroomelectronics]

When did your design start using 18 servos ?

 

[Cotowar]

When did your design start using 18 servos ?

hes got 6 legs, and probably has two joints per leg, as well as the joint where the leg connects to the body, for a toal of 18 joints. 1 servo at each joint, and you have 18 servos.

now, my question is this. In general computing, its easier to layer things, as you can make each task less complex, and can more efficiently route traffic. This can be seen when making a computer cluster. its easier to fill a switch with nodes, and then connect taht switch to a second switch, that will in turn connect another node filled switch. its slightly slower, but its easier to understand what's happening.

Is the same idea possible from a µC point of view? for example, could he have three servos per µC, and the 6 µCs connected to a master µC? The master would not directly control the servos, but would instead tell the slave µCs what to do, and they would in turn control the motors. This sounds kind of redundant seeing as they make TQFP µCs with 100+ pins, but still, imagine this on a large scale. even at a small scale, you could take a Java-like approach and re-use the code in each of the slave controllers. Then you have the master designate which signal to send to which slave. It cuts down on overall coding, as you can program multiple devices with the same information, and cuts down on troubleshooting as there is less going on so there is less to check through. Can someone tell me if this can be done? I'm really good at thinking like this, but I don't know enough about µCs to know whether or not this mode of thinking is practical.

Oh, and I'm signing up for a digital electronics class, as well as a semiconductor applications class next semester. Hopefully they won't diappoint.

 

[uaefame]

Hmm, I am using 18servo because I have 6 legs and each leg I am will have 3DOF (Degree of freedom).

Just wanted the name of the chip so I will be working on it instead of mastering a wrong chip I am working on PIC16F818 right now you think it can handle 18servo or I should be using 2or 3 PIC 16F818?

If you know a better PIC please recommend the PIC name

Thanks in advance