Build a better Servo...
Build a better Servo...
by altwolf
Thu Sep 10, 2009 5:47 pm
Hi all. I was just laying around thinking about hobby servos last night while I was trying to fall asleep, and I wondered why hobby servos are shaped like they are. For the Biped building I am trying to do, the fact that the servo horn is at one end of the servo makes things a little difficult.
I am sure that servo's, with their horns at one end, are the result of the simplest way to set up all the electronics and gears inside, I don't begrudge anyone that and for most things, that is fine. But when you are trying to balance a bot, having that offset from the center of the weight of the servo makes joints and things more complicated to build.
Would'nt it be useful to have a servo with a centered horn? I think it would, personally. Would anyone else agree? What are the advantages and disadvantages? I would like other people's input on any disadvantages they can think of.
Also, why don't servos have a horn on each side? It always seems weird to me that you have to use a dummy socket on the other side to connect a common "c" bracket. Isn't it less than optimal when turning a common "C" style bracket to have all the force on one side? Plus, with a horn on each side, you could more conveniently make one servo turn more than one thing, rather than building complicated bracketing to get the same effect.
Just some thoughts.
How would you build a "better" servo?
I am sure that servo's, with their horns at one end, are the result of the simplest way to set up all the electronics and gears inside, I don't begrudge anyone that and for most things, that is fine. But when you are trying to balance a bot, having that offset from the center of the weight of the servo makes joints and things more complicated to build.
Would'nt it be useful to have a servo with a centered horn? I think it would, personally. Would anyone else agree? What are the advantages and disadvantages? I would like other people's input on any disadvantages they can think of.
Also, why don't servos have a horn on each side? It always seems weird to me that you have to use a dummy socket on the other side to connect a common "c" bracket. Isn't it less than optimal when turning a common "C" style bracket to have all the force on one side? Plus, with a horn on each side, you could more conveniently make one servo turn more than one thing, rather than building complicated bracketing to get the same effect.
Just some thoughts.
How would you build a "better" servo?
by sap1975
Thu Sep 10, 2009 6:45 pm
Hi there
Well… my guess is that the standard layout is the one causing the smallest footprint for a motor, controller and a spur gearbox in a single package while keeping costs controlled.
Sure you could do specialized runs for hobby robotics servos like robotis does but if you would want to have the “horn” axis pull on both sides you introduce a new set of problems. If you want to keep the same footprint that is.
A. You have to offset the gearbox vertically instead of horizontally which has an impact on the loads of the axels which in turn requires that you use stronger/bigger axels.
B. You can no longer read the “position” of the actuator directly and have to introduce either another gear set or a pulley belt rig of some sorts.
C. You have to re-position the “electronics” you would most likely have to modify the layout to make it fit.
There’s nothing impossible in this stuff but it does drive the cost of manufacturing up.
A note on having the horn centered. You should be able to accomplish this relatively simple by bolting a spur gear onto the servo horn and another to a bracket containing a hinge.
Yep I know it would be a lot easier to just buy ready made ones but where’s the fun in that
Anyways. That was my two cents worth.
Cheers.
/Stig.
Well… my guess is that the standard layout is the one causing the smallest footprint for a motor, controller and a spur gearbox in a single package while keeping costs controlled.
Sure you could do specialized runs for hobby robotics servos like robotis does but if you would want to have the “horn” axis pull on both sides you introduce a new set of problems. If you want to keep the same footprint that is.
A. You have to offset the gearbox vertically instead of horizontally which has an impact on the loads of the axels which in turn requires that you use stronger/bigger axels.
B. You can no longer read the “position” of the actuator directly and have to introduce either another gear set or a pulley belt rig of some sorts.
C. You have to re-position the “electronics” you would most likely have to modify the layout to make it fit.
There’s nothing impossible in this stuff but it does drive the cost of manufacturing up.
A note on having the horn centered. You should be able to accomplish this relatively simple by bolting a spur gear onto the servo horn and another to a bracket containing a hinge.
Yep I know it would be a lot easier to just buy ready made ones but where’s the fun in that
Anyways. That was my two cents worth.
Cheers.
/Stig.
by limor
Sat Sep 12, 2009 9:34 pm
Have you seen the RoboBuilder servos ?
the reason why these actuators come in those black rectangular boxes and are controlled through an analog PWM protocol from the 70's is that there are many millions of RC servos manufactured every year and robotics in relation to RC is a very small market
The ideal generic robotics actuator would be tube shaped, have both rotation and linear motion a versatile ball joint at its end, include a configurable clutch and spring, position, speed and current measurement, 2mbps control bus. Only in Korea such a thing could be developed because of their government funding for such projects. Anywhere else it would be difficult to fund because it would be difficult to justify ROI.
the reason why these actuators come in those black rectangular boxes and are controlled through an analog PWM protocol from the 70's is that there are many millions of RC servos manufactured every year and robotics in relation to RC is a very small market
The ideal generic robotics actuator would be tube shaped, have both rotation and linear motion a versatile ball joint at its end, include a configurable clutch and spring, position, speed and current measurement, 2mbps control bus. Only in Korea such a thing could be developed because of their government funding for such projects. Anywhere else it would be difficult to fund because it would be difficult to justify ROI.
by Bullit
Sun Sep 13, 2009 4:51 am
Actually they are available if you can afford them and not from South Korea but from Germany.
Now if only someone could build servos with this capability at a price hobiest could afford. Last I checked these are quite expensive, but awesome. They have multi, ball and linear axis too. Some models are tube shaped and many allow wires and pneumatics to feed through them. They have brushless motors and harmonic drives.
Now if only someone could build servos with this capability at a price hobiest could afford. Last I checked these are quite expensive, but awesome. They have multi, ball and linear axis too. Some models are tube shaped and many allow wires and pneumatics to feed through them. They have brushless motors and harmonic drives.
by sap1975
Sun Sep 13, 2009 8:13 am
Ahh… if we’re ”allowed” to bring in the Pro stuff i think HarmonicDrives the company not the gearbox that is takes the price. There’s only really two problems one being as you mentioned “if you can afford them” the other being size. Then next thing is obviously that these systems don’t come with pre cut brackets and ready to go motion capture software.
Just to give en idea of the price range, size and power were typically talking from about a 1000£ and up per axis including driver which will give you anything from 20W to 1000W in a 3” x 10” package.
The PR70 you mention is 140x70x70 cm and nominal 96W.
As always in life there’s so much to choose from if you have the money.
I do however believe that Limor makes a good point. The hobby robotics servo is a niche product and designing a proper actuator in this size is not an easy job. Couple that with the fact that people in general won’t even spend a couple of hundred pounds on decent torque servo’s as it is the math just doesn’t add up.
Ohh. and by the way. Altwolf. where are you? You started this...
Cheers
/Stig.
Just to give en idea of the price range, size and power were typically talking from about a 1000£ and up per axis including driver which will give you anything from 20W to 1000W in a 3” x 10” package.
The PR70 you mention is 140x70x70 cm and nominal 96W.
As always in life there’s so much to choose from if you have the money.
I do however believe that Limor makes a good point. The hobby robotics servo is a niche product and designing a proper actuator in this size is not an easy job. Couple that with the fact that people in general won’t even spend a couple of hundred pounds on decent torque servo’s as it is the math just doesn’t add up.
Ohh. and by the way. Altwolf. where are you? You started this...
Cheers
/Stig.
by altwolf
Sun Sep 13, 2009 6:23 pm
Heh, I'm here. Just digesting all this information. It is very enlightening. I need to google and wiki "harmonic drive". It makes me think of this documentary I watched on Disney "Imagineering" and the work that went into developing those animatronic people they use in the theme parks.
The show talked about how the early systems used to have really bad feedback problems caused by sudden stopping movements and the weight of the exoskeletons. For example, when someone would raise their arm and stop, the sudden stopping motion would cause the whole frame to shake.
So they invented a special circuit - this was in the 50's or 60's - that could compensate for the secondary motions caused by sudden stops in the limbs. Of course, the show mentioned that super-interesting tidbit and then moved on to something else with no explanations. Grrrr.
The show talked about how the early systems used to have really bad feedback problems caused by sudden stopping movements and the weight of the exoskeletons. For example, when someone would raise their arm and stop, the sudden stopping motion would cause the whole frame to shake.
So they invented a special circuit - this was in the 50's or 60's - that could compensate for the secondary motions caused by sudden stops in the limbs. Of course, the show mentioned that super-interesting tidbit and then moved on to something else with no explanations. Grrrr.