Legacy Forum: Preserving Nearly 20 Years of Community History - A Time Capsule of Discussions, Memories, and Shared Experiences.

Not enough torque for Walking motion

Bioloid robot kit from Korean company Robotis; CM5 controller block, AX12 servos..
10 postsPage 1 of 1
10 postsPage 1 of 1

Not enough torque for Walking motion

Post by radiocar » Wed Nov 28, 2007 3:31 am

Post by radiocar
Wed Nov 28, 2007 3:31 am

I played the motion which is provided by Robotis and the robot need to bend its knee to lower the moment for legs motor(ID 15, 16).

I measured the weight of the whole robot and its around 2kg, and I tried to program it not to bend it's knee so much but the motor can't afford and LED is lighted up due to overloading.

If motors remain the same (AX-12), is it possible not to bend it's knee so much to perform walking motion ?

Design in a more dynamic way? Or any other suggestions?

(p.s. Human don't bend knee to walk!)
I played the motion which is provided by Robotis and the robot need to bend its knee to lower the moment for legs motor(ID 15, 16).

I measured the weight of the whole robot and its around 2kg, and I tried to program it not to bend it's knee so much but the motor can't afford and LED is lighted up due to overloading.

If motors remain the same (AX-12), is it possible not to bend it's knee so much to perform walking motion ?

Design in a more dynamic way? Or any other suggestions?

(p.s. Human don't bend knee to walk!)
radiocar
Robot Builder
Robot Builder
Posts: 9
Joined: Sat Nov 17, 2007 8:44 am
Location: Hong Kong, China

Post by JonHylands » Wed Nov 28, 2007 4:31 am

Post by JonHylands
Wed Nov 28, 2007 4:31 am

Check your battery charge state - if the battery is not fully charged, the AX-12's don't work as well...

- Jon
Check your battery charge state - if the battery is not fully charged, the AX-12's don't work as well...

- Jon
JonHylands
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 512
Joined: Thu Nov 09, 2006 1:00 am
Location: Ontario, Canada

Post by radiocar » Wed Nov 28, 2007 4:53 am

Post by radiocar
Wed Nov 28, 2007 4:53 am

Jon,

I plugged in AC source as well, and other motors work in good condition

so the problem is not on battery side, is on the torque of leg motors
Jon,

I plugged in AC source as well, and other motors work in good condition

so the problem is not on battery side, is on the torque of leg motors
radiocar
Robot Builder
Robot Builder
Posts: 9
Joined: Sat Nov 17, 2007 8:44 am
Location: Hong Kong, China

Post by BillB » Wed Nov 28, 2007 8:22 am

Post by BillB
Wed Nov 28, 2007 8:22 am

Are the knee joints simply unable to perform the move - or do they heat up quickly?

My Bioloid only seems to have a problem with the ankle joints. Never had a problem with the knee joints.

I use the standard Biolid walk - but I have a modified standing pose with straighter knees. You know when you have a good standing pose when it keeps standing upright even when you turn the power off.

So I guess in my Biolid the leg servos are not working hard when standing - saving its energy for the walk.
Are the knee joints simply unable to perform the move - or do they heat up quickly?

My Bioloid only seems to have a problem with the ankle joints. Never had a problem with the knee joints.

I use the standard Biolid walk - but I have a modified standing pose with straighter knees. You know when you have a good standing pose when it keeps standing upright even when you turn the power off.

So I guess in my Biolid the leg servos are not working hard when standing - saving its energy for the walk.
BillB
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 232
Joined: Sun Aug 06, 2006 1:00 am
Location: Hampshire, UK

Post by radiocar » Wed Nov 28, 2007 3:25 pm

Post by radiocar
Wed Nov 28, 2007 3:25 pm

BillB wrote:
My Bioloid only seems to have a problem with the ankle joints. Never had a problem with the knee joints.



BillB,

Yep, the ankle joint! That joint suffer a large torque due to heavy weight of Bioloid. That's why Robotics standard walk need to bend the knee joints.

My Bioloid knee joints don't have problem, but i just want to have a straighter knees so that it walk more like a human.

I woke up next day, and the ankle joint still get its led on when i played the walking motion i designed, so should not be the temperature problem :(
BillB wrote:
My Bioloid only seems to have a problem with the ankle joints. Never had a problem with the knee joints.



BillB,

Yep, the ankle joint! That joint suffer a large torque due to heavy weight of Bioloid. That's why Robotics standard walk need to bend the knee joints.

My Bioloid knee joints don't have problem, but i just want to have a straighter knees so that it walk more like a human.

I woke up next day, and the ankle joint still get its led on when i played the walking motion i designed, so should not be the temperature problem :(
radiocar
Robot Builder
Robot Builder
Posts: 9
Joined: Sat Nov 17, 2007 8:44 am
Location: Hong Kong, China

Post by BillB » Wed Nov 28, 2007 4:03 pm

Post by BillB
Wed Nov 28, 2007 4:03 pm

Starting from a cold servo - how long does it take for your walking motion to torque out?
Starting from a cold servo - how long does it take for your walking motion to torque out?
BillB
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 232
Joined: Sun Aug 06, 2006 1:00 am
Location: Hampshire, UK

Post by radiocar » Fri Nov 30, 2007 2:50 am

Post by radiocar
Fri Nov 30, 2007 2:50 am

Image

Right foot is talking first step

Image

Second Step

Image

You can see that, left foot is lifting up and right foot ankle suffer from heavy weight, the ankle motor may hold this position, but it is not able to move forward (arrow direction)

Absolutely, bend the knee like demo walking can solve the problem, but that's not similar to our human walking

Also, demo walking looks like sliding, two legs never leave the ground!
Image

Right foot is talking first step

Image

Second Step

Image

You can see that, left foot is lifting up and right foot ankle suffer from heavy weight, the ankle motor may hold this position, but it is not able to move forward (arrow direction)

Absolutely, bend the knee like demo walking can solve the problem, but that's not similar to our human walking

Also, demo walking looks like sliding, two legs never leave the ground!
radiocar
Robot Builder
Robot Builder
Posts: 9
Joined: Sat Nov 17, 2007 8:44 am
Location: Hong Kong, China

Post by JonHylands » Fri Nov 30, 2007 5:25 am

Post by JonHylands
Fri Nov 30, 2007 5:25 am

Watch these videos:

phpBB [media]


Also, this one, of a humanoid walking:

http://www.bioloid.info/BrainBot-Walk-01.wmv

Neither robot had overheating problems with the ankles...

- Jon
Watch these videos:

phpBB [media]


Also, this one, of a humanoid walking:

http://www.bioloid.info/BrainBot-Walk-01.wmv

Neither robot had overheating problems with the ankles...

- Jon
JonHylands
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 512
Joined: Thu Nov 09, 2006 1:00 am
Location: Ontario, Canada

Post by BillB » Fri Nov 30, 2007 9:51 am

Post by BillB
Fri Nov 30, 2007 9:51 am

Radiocar,

I normally get problems with the ankle after about 15-20 minutes of intensive use, which makes me think that it may be of the following problems (each of which I have experienced before):

1 - Something is jamming the servo. If you attach the bracket using anything but the shortest of screws then the servo horn gets physically attached to teh servo body. So paerhaps try and check that the screws are of the smallest variety.

2 - The servo torque limit has been set very low. Try changing the torque limit by resetting to factory defaults (by using the dynamixel configurator if you have a usb2dynamixel adapter, or adding a BCP command)

3 - Faulty wiring. Try replacing/swapping the wiring to the ankle.

4 - Finally, there may be a fault in the servo. You can check this by swapping the suspect servo with one you know works. Swap over the servo ID's (using the CM5 terminal) and see if it works.

Hope this helps.
Radiocar,

I normally get problems with the ankle after about 15-20 minutes of intensive use, which makes me think that it may be of the following problems (each of which I have experienced before):

1 - Something is jamming the servo. If you attach the bracket using anything but the shortest of screws then the servo horn gets physically attached to teh servo body. So paerhaps try and check that the screws are of the smallest variety.

2 - The servo torque limit has been set very low. Try changing the torque limit by resetting to factory defaults (by using the dynamixel configurator if you have a usb2dynamixel adapter, or adding a BCP command)

3 - Faulty wiring. Try replacing/swapping the wiring to the ankle.

4 - Finally, there may be a fault in the servo. You can check this by swapping the suspect servo with one you know works. Swap over the servo ID's (using the CM5 terminal) and see if it works.

Hope this helps.
BillB
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 232
Joined: Sun Aug 06, 2006 1:00 am
Location: Hampshire, UK

Post by radiocar » Sun Dec 02, 2007 8:09 am

Post by radiocar
Sun Dec 02, 2007 8:09 am

Thx

I will try it out.
Just want the robot looks like a human :)
Thx

I will try it out.
Just want the robot looks like a human :)
radiocar
Robot Builder
Robot Builder
Posts: 9
Joined: Sat Nov 17, 2007 8:44 am
Location: Hong Kong, China


10 postsPage 1 of 1
10 postsPage 1 of 1