# linear analog servo motor drive

#### ammar87

Joined Jan 7, 2013
12
hello my friends
i am trying to build a linear analog servo drive for a servo motor and YES it is analog not a PWM for many reasons one of them :is that this drive is a part of control system i am working on and i want to reduce the noise to minimum.
so i dig the web and i found several power op amp: like L165,L2720, LM675,LM12
i am kid of confused because all i want is :
Class-AB linear power Amplifier with no distortion around zero
i am trying to use the power op-amp rather than power transistors to reduce the number of components and the issues related to them .
here is the motor specifications:
http://www.specamotor.com/en/Faulhaber/motors/2338-006S/datasheet_motor.html
i suppose to control the position in bi-direction.
the questions:
1-which power op-amp should i use?
3-is this circuit correct to control the motor position in bi-direction?

any help will be more than appreciated

Last edited:

#### tshuck

Joined Oct 18, 2012
3,534
I'm curious as to your reasoning of using an analog control signal on a servo. What happens when the motor(an inductive load) induces a voltage on the control line? Does it change its commanded position?

Is this for a R/C servo, or the more general servomotor(a motor with positional feedback)? If this is for a R/C servo, you won't get much(any?) benefit from changing it to work with an analog command. The reason of using digital is that it is much more impervious to noise than analog...the reason you gave for going to analog!

#### ammar87

Joined Jan 7, 2013
12
no it is not R/C servo
it is servo motor 6v/1Amp
used in this system:

i bought the system without the linear servo amplifier because they selling it with power supply for $2000 and that's too much for it . What happens when the motor(an inductive load) induces a voltage on the control line? Does it change its commanded position? no the command position is given from DAC it should not change by induced voltage, may be i need to use freewheeling diode,but anyway i need a circuit with good power op-amp Thread Starter #### ammar87 Joined Jan 7, 2013 12 the position feedback comes from encoder ,or pot ,or tacho Thread Starter #### ammar87 Joined Jan 7, 2013 12 #### tshuck Joined Oct 18, 2012 3,534 Is this some sort of inverted pendulum? Or is the rest of it cut off in the image? Correct me if I'm wrong, it seems that the whole thing is controlled by a single motor(Brushed DC?) with two sections being monitored: the angle of the bar and the angle on the rotating table. If that's the case, your design would be better suited to use a digital approach, since your feedback is a digital signal. The amount of disturbance on an inverted pendulum will be nowhere near the quantization error introduced in a decently designed digital control system. The motor could be driven with a H-Bridge if you have a single power supply, or a push-pull for a dual-supply solution. If you are set on using the op amp, TI makes some pretty decent power op amps, one that popped up right away seems to be adequate, the OPA596. The application suggestion even suggests to be used with a servomotor. Thread Starter #### ammar87 Joined Jan 7, 2013 12 Is this some sort of inverted pendulum? Or is the rest of it cut off in the image? yes this is rotary inverted pendulum i will use it in my master thesis Correct me if I'm wrong, it seems that the whole thing is controlled by a single motor(Brushed DC?) yes it is Brushed DC and it is the only actuator in the system with two sections being monitored: the angle of the bar and the angle on the rotating table. yes we have 2 encoders one for the pendulum rod and one for the arm the controller is build in Matlab simulink and the control signal is outpted using DAC of Dspace data acquisition card. i don't want to use PWM because : PWM drives are very efficient, but are electrically noisy as they operate by pulsing the motor at full supply voltage at typical frequencies of 4 kHz to 30 kHz. This pulsing tends to saturate everything electrically in the surroundings, often including the intended operation. A second side effect of using PWM drives shows up in ultra-high precision systems requiring nanometer precision. Due to the pulsing nature of the PWM drive, the motor will tend to dither causing position error that cannot be tuned out. i read this from a datasheet of a linear servo amplifier designed for another motor not the one i am looking for #### tshuck Joined Oct 18, 2012 3,534 PWM drives are very efficient, but are electrically noisy as they operate by pulsing the motor at full supply voltage at typical frequencies of 4 kHz to 30 kHz. This pulsing tends to saturate everything electrically in the surroundings, often including the intended operation. A second side effect of using PWM drives shows up in ultra-high precision systems requiring nanometer precision. Due to the pulsing nature of the PWM drive, the motor will tend to dither causing position error that cannot be tuned out. The environment will see to it that your system will never reach nanometer-scale accuracy, and as such, isn't a concern here. As far as it electrically saturating its surrounding, I've never had problems at such low currents. At 20A+, I would be a bit more wary of the interference(in the system)... #### BillB3857 Joined Feb 28, 2009 2,565 Dithering isn't always bad. If at a proper level and frequency, it can impart a stiffness at zero speed and improve response from zero speed. Thread Starter #### ammar87 Joined Jan 7, 2013 12 tshuck of course it is not a nano-scale , but i am curious way the company not use PWM to drive it ? #### tshuck Joined Oct 18, 2012 3,534 tshuck of course it is not a nano-scale , but i am curious way the company not use PWM to drive it ? Any number of reasons. The company(which company is it, by the way?) could have devised this to show off their op-amps, or to show a different way to control the motor than what everyone else does, or to show that this controller could be implemented without a digital processor. The pendulum is underactuated and extremely non-linear due to the gravitational forces and the coupling arising from the Coriolis and centripetal forces http://en.wikipedia.org/wiki/Furuta_pendulum Thread Starter #### ammar87 Joined Jan 7, 2013 12 Any number of reasons. The company(which company is it, by the way?) it is Quanser The pendulum is underactuated and extremely non-linear due to the gravitational forces and the coupling arising from the Coriolis and centripetal forces that's totally true ,it is not easy to control #### tshuck Joined Oct 18, 2012 3,534 It seems like it was the first guess, they are trying to show off their products, namely, this one. They call it their "Hardware-In-The-Loop" power-module. It interfaces with their software in order to control the system. As we've been discussing, this is not the only method to make this control system. A company will rarely design a kit that doesn't use/demo that company's parts... Thread Starter #### ammar87 Joined Jan 7, 2013 12 It seems like it was the first guess, they are trying to show off their products, namely, this one. They call it their "Hardware-In-The-Loop" power-module. It interfaces with their software in order to control the system. exactly they trying to show off their products not only that , but also sell it for$2000 and refuse to give warranty if i refuse to buy this power module!
so \$2000 is too much for power supply and servo drive , so i decided to buy the system without it ,means no warranty. anyway Texas instrument have this circuit :

i think this fit my application , what do you think?

#### BillB3857

Joined Feb 28, 2009
2,565
My guess is that there is a LOT more to their drive than what you have shown. The feedback system, compensation network, PID control, etc all come into play to let the vertical pendulum do its thing. Or are those missing functions what you are wanting to build/duplicate?

#### ammar87

Joined Jan 7, 2013
12
My guess is that there is a LOT more to their drive than what you have shown. The feedback system, compensation network, PID control, etc all come into play to let the vertical pendulum do its thing. Or are those missing functions what you are wanting to build/duplicate?
the control algorithm (PID , LQR etc) is built using Matlab , so the controller is digital inside the computer (PC) ,the output of the controller is the control signal which represent the DAC voltage , now this DAC voltage can not drive the motor directly so we need some sort of driver basically capable of provide enough current to the motor .what i want to build/duplicate is the motor driver .

#### THE_RB

Joined Feb 11, 2008
5,438
Have a look at the Pololu.com motor drivers, they are small hobby robot type motor drivers and will easily drive a 6v 1A motor.

If you prefer you could try Sparkfun or RobotMarketplace, they all have small cheap DC motor drivers.