Replacing servos with brushed motors. Some questions!

Hi all guys.
I am developing a quadruped robot.
This robot using:
A) n°1 Arduino MEGA 2560 board
B) n°1 Adafruit 16ch servo driver board (where I attach & power up the servos)
C) n°12 Power HD 1201MG Metal Gear Servos - 13.2 kg (183 oz in), .14/sec

I got great result with my robot walking good; here the result:

All ok but servo motors has some limitations.
The fact is that until servo using toothed gears inside so it's very difficulty to make a decent shock absorber system.
And in a quadruped robot, a decent shock absorber system is a really important matter.

So I really would to know if, using my Arduino MEGA 2560, I can replace the servos with brushed motors.
In fact using brushed motors I can make sure a better shock absorber system due to the fact that no toothed gears are inside a brushed motors.

Then some questions:

  1. Can you provide me the name of some brushed motors similar (in torque) to my actual servos?
    So a motor that can handle 13kg of torque.
  2. I know that I can use shields to make my Arduino board handling brushed motors. Can you tell me what is the better shield to handle 12 brushed motors (13kg of torque).
  3. Currently in my project I am using 4 tactile sensors + 1 MPU6050 gyro sensor + 2 SRF05 ultrasonic sensors too.
    So If I will use a brushed motor shield, can I get problems using these sensors?

Thank you in advance for help!

Brushed motors run at high speed and have very little torque. That's why many motors have gears on them to reduce the speed and increase the torque. So you still have the problem of the toothed gears.

Try building a hydraulic robot.

How do you anticipate position control?

In first thank you for answers guys.

Brushed motors run at high speed and have very little torque. That's why many motors have gears on them to reduce the speed and increase the torque. So you still have the problem of the toothed gears.

Try building a hydraulic robot.

Ehh idraulic matter is probably too far from me :slight_smile:
I am pretty sure that exist some brushed motor working (without gears)at 13kg of torque with same servo dimensions.. mmm

How do you anticipate position control?

what you mean?

aldoz:
what you mean?

Well isn't the whole idea of a servo that you tell it to turn to "x" degrees?

Hi

Quote

How do you anticipate position control?

what you mean?

When you use your servos to move your robot, you signal to them what angle or position you want them to go to.
The servo has all the feedback and control built in.

If you use motors you will need feedback so that you can stop them in the desired position.

A 13Kg? ? ? (13 kg/cm) torque servo, uses gearing on its motor.

A stand alone motor, no gearbox, with that amount of torque at its shaft will be very large and heavy compared to a servo.
Also you would be trying to get a motor shaft to turn only a fraction of a turn you would need position feedback and then added programming to use the feedback, that is why they make servo units.

Tom.... :slight_smile:
Are you sure its your project, I'm surprised you know very little about how servo units work.

TomGeorge:
Are you sure its your project,

A literal Doubting Thomas... user names match

LOL guys :slight_smile:
Yeah it's my project, just my english make me bad jokes when I try to understand your words.. :roll_eyes:

My past project, made before the quadruped, was a moving platform (pitch,roll & vertical travel) using 3 winch motors (damn heavy toothed gears!) and sure, I was using 3 potentiometer to "know" where the motor stirrups are positioned.

In fact in my quadruped, If I will use 12 brushed motors,so I should use 12 potentiometers and the dimension of my robot is too little to put so many pots...

Again, from your words, I understanding no way to get a brushed motors (same servo dimension) with 13kg of torque..

So..
probably if I want to use brushed motors instead servos so I need to make my robot bigger.. :\ and this is not the best choice for now.

Hey aldoz,

to keep it very simple: yes, it is the gears that turn your motors' speed into torque. you will not get the same amount of torque from a motor without gears if you keep the same size and same price(!).

investing more money will always get better results, but with the higher price you will get a much better servo also, so again same size & same price: see above... :wink:

... and, as already said: there is the missing position feedback, which again generates additional size/price

A motor with a pot is a servo....

... kinda.

Torque is not measured in units of mass or force or force/distance or mass/distance.

Torque is force times radial distance, units newton-metres, kgf-cm if you must (but
don't, working in Nm is much simpler for calculations:

power (watts) = torque (Nm) x angular velocity (radians/s)
force (N) = torque (Nm) / radius (m)

You can approximate the acceleration due to gravity as 10 (its actually about 9.8m/s/s),
and the conversion from rpm to rad/s is also about a factor of 1/10. (actually 1 / 9.55)

Another way to think about torque is as energy per radian, which nicely parallels
force as energy per linear distance.

The Nm is a really convenient unit to get a feel for too:

0.1Nm - finger and thumb can twist about this hard
1Nm - whole hand on a screwdriver can provide this amount of twist readily
10Nm - both hands on the steering wheel can provide this sort of amount
100+Nm - arm wrestling territory!

Hey aldoz,

to keep it very simple: yes, it is the gears that turn your motors' speed into torque. you will not get the same amount of torque from a motor without gears if you keep the same size and same price(!).

investing more money will always get better results, but with the higher price you will get a much better servo also, so again same size & same price: see above... :wink:

... and, as already said: there is the missing position feedback, which again generates additional size/price

The servos has 2 great advantage:
A) you can know easly the position
B) you can get a lot of power (for such little dimensions!)

But has a bad attitude :
they broken when a fast external force applied to they (these little tooths are really delicates).
Classical example is the quadruped turned on, with the 4 legs in position before to start to walk;
If I lift the robot and I make it fall (from 15 or 20cm) sure my servos will get some damage..

Sure you can spent a bit more to get alluminium tooths but the problem remains.
Again, if you reach the end of rotation you can broke the servo too!

A motor with a pot is a servo....

... kinda.

In fact yes :slight_smile:

Torque is not measured in units of mass or force or force/distance or mass/distance.

Torque is force times radial distance, units newton-metres, kgf-cm if you must (but
don't, working in Nm is much simpler for calculations:

power (watts) = torque (Nm) x angular velocity (radians/s)
force (N) = torque (Nm) / radius (m)

You can approximate the acceleration due to gravity as 10 (its actually about 9.8m/s/s),
and the conversion from rpm to rad/s is also about a factor of 1/10. (actually 1 / 9.55)

Another way to think about torque is as energy per radian, which nicely parallels
force as energy per linear distance.

The Nm is a really convenient unit to get a feel for too:

0.1Nm - finger and thumb can twist about this hard
1Nm - whole hand on a screwdriver can provide this amount of twist readily
10Nm - both hands on the steering wheel can provide this sort of amount
100+Nm - arm wrestling territory!

Thx for info about torque!!

i am still a beginner, experimenting a lot for my own robot project and just start to understand things.
still not sure here which motors to take for what, especially as different tasks have different needs.

parallel to servos, which just solve a lot of tasks of robotics out of the box, i am looking into using stepper motors. they are also easy to control, as you can directly tell them how far to turn, they have 360° and no gears, so nothing to break in there.

also, they are available in a lot of different sizes, while more torque mostly results in bigger size. i have 3 nema 23 size ones (from a cnc kit) with 1.8 Nm (~35€ a piece) and for robotics testing one nema 17 with 0.5 Nm (~17€).

it might be an idea to look up a few of them to see if their parameters of torque, size, speed and price are interesting for you. they even give you the option to work very easy with timing belts easy setups of few but big and sturdy gears.

for a robotic arm their sheer size and massive shafts of at least 5mm makes them sturdy enough not to break when pressure is given to them.

TOM-d:
i am still a beginner, experimenting a lot for my own robot project and just start to understand things.
still not sure here which motors to take for what, especially as different tasks have different needs.

parallel to servos, which just solve a lot of tasks of robotics out of the box, i am looking into using stepper motors. they are also easy to control, as you can directly tell them how far to turn, they have 360° and no gears, so nothing to break in there.

also, they are available in a lot of different sizes, while more torque mostly results in bigger size. i have 3 nema 23 size ones (from a cnc kit) with 1.8 Nm (~35€ a piece) and for robotics testing one nema 17 with 0.5 Nm (~17€).

it might be an idea to look up a few of them to see if their parameters of torque, size, speed and price are interesting for you. they even give you the option to work very easy with timing belts easy setups of few but big and sturdy gears.

for a robotic arm their sheer size and massive shafts of at least 5mm makes them sturdy enough not to break when pressure is given to them.

Very interesting,
I always think about stepper motors but I never start to study about.
sincerely from what I know a stepper motor is a 360° servo?
Sure the precision of movement is high in stepper and servo too.
You say that stepper is 360° so no gear; mmmm I was pretty sure stepper is full of toothed gears inside! maybe just to reach more torque.. or not? (my damn english :frowning: )

Steppers have no gears inside. They also have no feedback, so you don't know where the motor shaft is even though you know it is not going to move unless you command a step. (No suspension ability.)

There must be geared brushed motors that can perform the torque required. It's just a matter of scaling up from the plastic hobby servos to industrial equipment. Unfortunately the cost scales up even faster.

Again, from your words, I understanding no way to get a brushed motors (same servo dimension) with 13kg of torque..

You can get a smaller brushed motor and do some testing to see if the result meets your needs. Metal gear servos are not very expensive, check HobbyKing. There are servos that are made for robots that might be an option.

@MorganS:
Ok, stepper motors has no feedback position but, I know how many the steps I order to make! So I can know the position! right?

@zoomkat:
Yeah I remember I saw some little brushed motor with high power somewhere.. mmm I need to check.

aldoz:
Ok, stepper motors has no feedback position but, I know how many the steps I order to make! So I can know the position! right?

#Position: Yes. A stepper motor has a fixed degree he takes per step. So if he takes 1° per step and you order him to take 360 steps it will make a full turn. If it takes 1.8° per turn (standard value often used) you have to order him to take 200 steps to make the 360° for a full circle. These values vary but are always stated with the motor.

So you can tell the position the motor will end up by the steps you order him to take. Problem is when the motor gets too much resistance, can't move and drops some steps. Then you have no way to tell how much steps he really took and what its actual position really is.
You will have to append an extra rotary encoder for this, which measures the real position. then it is called a "closed loop stepping motor" setup.

I have this thoughts with my desktop cnc mill. Normally, when usage is kept within the parameters of the motors capabilities there should be no/not much steps dropped, which turns out to be not a problem. Upgrading it to closed loops will add another level of complexity, hardware (weight/size) and costs(!), but could sqeeze those few 1/100's mm for even more precise results.

#Gears: No gears in there, it's like a brushless motor, just individual seperate coil pairs which can be controlled seperately to make the steps possible. Definitely different control logic and different controler boards.

In addition to what @TOM-d said about stepper motors you also need some means to identify their position when the Arduino is started. The usual way to do that is to run the stepper until it triggers a switch. That is then assumed to be position zero, or the home position.

This means that a servo is very much easier to use if it is suitable for the task.

You may find some useful stuff in stepper motor basics

...R

thank you Tom and Robin and all you guys, your advices let me to know I need to continue to use servo motors.
Until now they are working fine and to change in stepper or brushed motors it would be too complex for now.

Please take a look to this vid:

ehhh you can see how this mechanical leg can suffer really hard hits without problematics and off course it's a hydraulic leg! Simply no servos can suffer something like that;

I was just thinking about a brushed motor can do something similar if no toothed gears applied on it.
was for this that I was looking for brushed motors.

Thank you so mutch to you all guys; for now I will continue with servos.
Rather.. somebody know some basic Arduino + hydraulic tutorial? just for curiosity :slight_smile: