Analog PID control for DC motor position

[FvM]

pi dc motor control

The parameters are often given in motor datasheets or they can be extrapolated from performance graphs of general purpose motors.
e.g. as given byhttp://www.johnsonmotor.com

Rotor inertia isn't always given, but here the total inertia including the load must be inserted in calculation. It could be calculated from construction data or measured from a spin-up experiment. A mechanical damping can be neglected in most cases, I think, but could be estimated from spin-down time with open circuit motor.

 

[oudghiri_m]

op amp pid controller

Here u can find en interesting document for DC Motor

 

[dolpre]

dc motor position control potentiometer feedback

Hi FvM,

i am using S9D150-24CH from SPG motor.
The datasheet of the motor is as following website:
**broken link removed**

so, may i know how can i get the parameters which i need from the datasheet?
i need to know the back emf constant of the motor as well. So, how to get it experimentally? Or there is another way to get it from datasheet?

on the other hand, how am i going to estimate the mechanical damping from spin-down time with open circuit motor?
(i had get the ON/OFF by injecting step input to get the step response of the motor)

Thanks...

 

[FvM]

how to convert pid to analog controller

You can estimate R,K and b from the motor data using some simplifications. Assume that the motors no load current would be actually 2 A and is caused by friction losses proportional to rotational speed that are identical to mechanical damping of the system. Then you get a true unloaded speed of 3117 rpm or 326 rad/s and a Ke of 0.074 Vs/rad = Kt 0.074 Nm/A (Kt estimation from T-N characteristics would be slightly higher at about 0.09).

Friction losses at 3000 rpm are 2 • 0.074 = 0.15 Nm or 0.15/314 = 0.47e-3 Nms damping constant. R estimated from the given operational points is 0.47 ohms. You can use an electrical measurement instead.

Motor L could be estimated from current rise time of blocked motor or with an impedance meter.

Inertia could be estimated from dynamical measurement, e. g. rotational speed or current response to voltage step when motor L is already known or rotational speed response to current step without motor L. Also resonance frequency with a torsion spring would allow an accurate measurement.

 

[dolpre]

control position motor use pid

Hi FvM,

Assume that the motors no load current would be actually 2 A and is caused by friction losses proportional to rotational speed that are identical to mechanical damping of the system. Then you get a true unloaded speed of 3117 rpm or 326 rad/s and a Ke of 0.074 Vs/rad = Kt 0.074 Nm/A (Kt estimation from T-N characteristics would be slightly higher at about 0.09).

1. How can i get the true unloaded speed of 3117rpm? As what i get from the datasheet is 3000rpm.
2. What did you mean by Ke= Kt 0.074 Nm/A? Is it means that Ke=Kt=0.074 Nm/A?
3. For the friction losses at 3000rpm, u mentioned that 0.15/314=0.47e-3 Nms damping constant. May i know where the 314 come from?
4. U mentioned that R estimated from the given operational points is 0.47 ohms. Can you explain more about this?

Hope you can help me...
Thank you very very much...

 

[FvM]

dc motor pid control

1. I simply constructed a "true unloaded speed" in "speed versus load current" diagram by extrapolating up to I=0. This is motivated by my assumption about nature of losses respectively idle current.
2. Yes, Ke=Kt=0.074 Nm/A
3. 314 is 3000 rpm expressed as rad/s
4. The resistance is calculated from Ke and and speed drop versus load current.

 

[dolpre]

analog pid with operational amplifier

Hi Fvm,

Really thank you very very much...
And i found one website which explain about this as well.
https://inst.eecs.berkeley.edu/~ee128/fa04/labs/lab7-motor.pdf



Added after 1 hours 7 minutes:

What i found from experimentally value of Kt is 0.076 Nm/A and it is just slightly higher than calculation from datasheet, i.e. 0.066Nm/A.

How if the motor is with a gearhead, will the value of parameters (Kt,Ke) change?

 

[FvM]

analog position control

I think the correspondence of parameter from different calculation methods is exactly enough, to allow e. g. design of a controller.

With a gear, the sizes observed at the reduced shaft have to be scaled with the reduction ratio, either divided (speed) or multiplied (torque). Ke = Kt related to the reduced shaft would also be scaled. But I think it's useful or common, to do this. One would calculate the sizes for the motor shaft and scale the derived values. Another point is, that a gear increases the mechanical damping.

 

[dolpre]

analog pid design

Hi FvM,

i understand that with a gear, the speed will reduce while the torque will increase.
May i know what is the relationship between a gear ratio, N with torque and speed?
In my case, i found that speed_gear = speed/N but the torque_gear != torque*N.

Ke = Kt related to the reduced shaft would also be scaled.

may i know how to scaled it? Is it just multiply or divide it by N?

How if i found the transfer function of the motor without gear, is Kt/(tau*s+1) and want to convert it to the transfer function of motor with gear ratio of N?

Thanks in advance...

 

[FvM]

analog pid controller

For an ideal gear, Ke and Kt would be multiplied with N. If you observe a torque below the theoretical value, it could be due to gear efficiency below 1, I would expect 60 to 90 % depending on gear ratio.

 

[dolpre]

analogue pid circuit

i calculated, the gear efficiency is 73%, so for this non-ideal gear, the Ke and Kt will be multiply by N*0.73 right?

 

[baonguyenpro]

graphs of pid of dc motor

With MCU for Automotive like MPC5554, DC motor controlling is specified by eTPU (Enhanced Time Processing Unit), Freescale developed the library for it. You just use the APIs to build your application. It's very convenient.

 

[Bogdan4all]

pid circuit motor

Hello!
I have a PID controller made of operational amplifier and i want to ask how do i chose the resistor and capacitor values? How they influence the process?
Thanks

 

[ali_aku_suka_telo]

you can check this blog
Im found it has PID circuit (using ISIS) to drive dc motor and it could simulated

---------- Post added at 07:49 ---------- Previous post was at 07:47 ----------

you can check this blog
Im found it has PID circuit (using ISIS) to drive dc motor and it could simulated