anyone know inductance in servos?

[Admin]

I need to know the specific/typical inductance in a servo (or any el-cheapo motor).

Ideally I'd just crack a servo open and hook up an inductance meter to the motor . . . but I don't have an inductance meter . . .

Anyone can help me out, or know ball-park values?

[paulstreats]

Here is the datasheet for el - cheapo motors (the type found in most cheapo toys around the world)

http://www.farnell.com/datasheets/21702.pdf

Theres no mention of inductance though unfortunately. Though the tables do show just how inefficient they are.

The people who are likely to know the inductance of a servo's motor are at the open servo project. Maybe you could find the discussion forum and ask there

[Kohanbash]

Hey
I just quickly checked the inductance of two servos I had laying around.

For Hitec HS-311 servo the inductance between:
+v and gnd  = 18.5mH
sgnl and gnd  = 45mH
sgnl and +v  = 78.3

For Hitec HSR-1425CR servo the inductance between:
+v and gnd = 15.2mH
sgnl and gnd  =  38.1mH
sgnl and +v  = 90.5mH

Thats an interesting question. What is the application?

[Admin]

Hmmmm I think the motor needs to be removed from the circuitry to get a proper inductance reading.

I checked my multimeter at home and like 10 at work but none of them test for inductance.
I might buy one for ~$150 Monday

Basically what I'm doing is trying to figure out where all the power usage needed for my robot fish goes to. I've done all the mechanical calculations for energy consumption, but have only been able to account for ~2%.

Just measuring wattage going to a servo doesn't tell me what the energy is being used for, just that it's being used.

So my theory is a lot of energy is being lost when I'm reversing motor direction at full speed at several hertz. I just need to know how many joules are being used just in creating/destroying the inductance field, and if I know inductance, I can easily calculate it.

[Kohanbash]

Hi
How much inductance do you expect to find? Also how are you finding the input wattage?

The actual motor is 1.6mH

Also I have a servo thats taken apart and a meter with me right now

[Admin]

Hmmm are you sure its 1.6mH and not 1.6H?

Using this equation:

E = 1/2 * L * I^2

with .5A current (I have current sensors on my servos),

I get .2mW

But my motors drain ~3W . . .

Hmmmmm where else does the power go to in a free-spinning motor, other than friction and inductance?

[Kohanbash]

Hi
yep mH not H. 1.6H also sounds to high for a small servo motor (even thou there could be some weird mutual inductance in the servo that would partially sum the three numbers in my original post).

Where is your equation from? Here is a quick power derivation:
P=I^2 * X   and X=2 PI f L
so Power =I^2 * ( 2*PI*frequency*L)

Also what voltage is being applied to the motors (taking any PWM into account)?

Also have you tried another servo just to insure that nothings busted in your other servo?

[Admin]

Where is your equation from?

http://en.wikipedia.org/wiki/Inductor

Power =I^2 * ( 2*PI*frequency*L)

I haven't a clue on the PWM frequency going to my servos . . .

Also what voltage is being applied to the motors (taking any PWM into account)?

Also have you tried another servo just to insure that nothings busted in your other servo?

Its just a hobby servo, getting about 6V. At that voltage and .5A under full load, thats about 3W - typical of servos. I am using multiple servos that all do about the same.


The next best solution I can think of is command my servos with and without load, then subtract the measured power values to get just the power needed to use the motors. But unfortunately servos act differently under loads so not sure the accuracy on that . . .

[Kohanbash]

Can you try controlling the servos with another brand/type of controller. Maybe the control frequency is slightly off or the inductance is not matched properly between the servo and the controller.

[Admin]

I'm reviving this thread because the problem has come back again for something I'm working on.

I just need to get a ball park value, so lets assume the motor is off, then suddenly I turn it on to 100%, how much energy did it take in terms of inductance as time -> infinity?

How does applied torque and motor rotational velocity affect this? Or is that just momentum, entirely separate from inductive energy?

I'm sure a physics major could solve this!

[Asellith]

I'm a little rusty of motor physics but your coupling that inductor to a magnet that might be where the most energy is lost. The magnetic link could be poor because of cheap manufacturing. Maybe use the meter to read the inductance as the motor spins? The magnetic fields generated in the motor may be where all the power is going.

[Admin]

I ran a test.

I have 5 servos, with nothing attached to them, going through some preprogrammed rotations. This requires an average 10W. This is the energy required for five servos to do zero useful work.

Now when I attach my robot stuff to the servos, they require a sum average of 13W. So only 3W are for real work - thats about 23% efficiency, horrible! And I suspect a good percentage of that 3W are lost in servo gearing inefficiencies, so servo efficiency is surely below 20%.

I'm trying to calculate the exact losses of energy in my robot, and as the majority is wasted in the servos, I need to figure out what to change to fix it. Going to buy some brushless servos to see if I can get a better efficiency. I'm also considering scaling, so I'd need to understand algorithmically the energy stored as inductance in the motor.

[Asellith]

Have you considered maybe a different type of locomotion if your changing things that fast in a repetitive motion can't you design your own linkage and do it with a good DC motor and eliminate the manufactures inefficiencies or maybe a linear actuator. I'm assuming muscle wire will not cycle fast enough and probably is not strong enough for the application.

Does the servos feedback matter or are you just using them as easy and cheap dc gear boxes? Might mean a major redesign but without feedback and a "homing" position system you could use a stepper motor maybe. Just some thoughts.

[Admin]

The position feedback is *very* important. And muscle-wire sucks, horrible in every aspect.

I just processed a bunch more data. Given the new data, I'm dropping the typically servo efficiency to closer to ~15% average (it ranged from 2% to 23%). This is for un-modified servos. I'm sure modified servo efficiency, with continuous motion, has a much higher efficiency. I just haven't had a reason to measure it yet . . .