BLDC motor Accelerator problem

C

Thread Starter

Chaitanya B

Hi,

I am designing BLDC drive with a Micro-Controller based controller for E-bike application. Presently the controller is simple Open-loop (without Speed and Current Control)the PWM(0-70%) varies with Accelerator(0 to 5V).

Motor is driven smoothly and works well for most of the condition. This is when Accelerator is moved slowly say 0 to 5 in 1 sec. if the Accelerator is varied from 0 to 5V rapidly. At the Start I am getting very high In-rush and my motor is oscillating and jerking

Is there any way using Software to control this Rapid Start?

Does PID or Closed Loop over Speed or Current will help me to solve this problem?

My Motor is 48V,350W,
6 pole,Hall sensor motor.

-Chaitanya
[email protected]
 
Of course closed speed loop with inner current loop will help you. BTW you may use Halls for speed feedback as well.

Pay attention on value of PWM frequency also because your motor may have low inductance leading to high current ripple.
 
C
Thank You very much :)

I am having 20KHz frequency fixed. Do you think at start I should have some different frequency?
 
C
Inner and outer loops is there any difference?

I mean I was thinking PID over Current and PID over Speed.

Please tell me the significance of inner current and outer speed loop!
 
Cascade structure assumes PI-filter for inner current loop and PI-filter for outer speed loop as a rule. Inner loop bandwidth must be 3..5 times higher than outer speed one.
 
C
Thanks Leon.

I have closed the Speed Loop with Hall Feedback and Motor start is pretty smooth.

Do you have any reference or e-book name for the cascaded control and simple PID basic tuning practical information.I tried to search but many things are pretty complex,so I thought let me ask the real expert.

-Chaitanya
 
P

Peter Nachtwey

> Inner loop bandwidth must be 3..5 times higher than outer speed one. <

Where is this written? Assuming the motor and load has a simple velocity model K/(tau*s+1), which is suitable for PI inner and PI outer loop control, have you worked out the math or seen another source that has proved your statement?
 
You need to use 2nd power model with electromagnetic and electromechanical time constants for synthesis of velocity control with cascaded inner current and outer velocity loops instead such simply model.
 
P

Peter Nachtwey

Then there should be an inner PID loop not a PI loop. Still where is it written that the inner loop must be 3 to 5 times faster than the outer loop? You dodged the question. I bet the ratio can be much lower than that depending on how I implement the inner and outer loop controls. I bet I can get it close to one to one by using a PID on the outer loop. The zeros added by the outer loops Kp and Kd will extend the outer loop's bandwidth significantly. If I put Kp and Kd in the feed back path only the inner loop will have no zeros to extend its bandwidth so the ratio of the inner loop to outer loop bandwidth will approaching one to 1.

The point is that general statements should not be made. Each system must be evaluated on a case by case basis.
 
P

Peter Nachtwey

So you would use information for one process control system and apply it to all BLDCs?

Each system is different and the equations will work out differently. A steam/temperature control system is much different from a BLDC system. John Shaw's stated ratio of inner to outer loop bandwidth of 4 may have applied to a certain type of system he has worked on and it is given without proof or calculations.

Yet people repeat ratios like 4 to1 as if it applied to all systems and they repeat what others have said or posted without questioning the source, the system it applies to, or if there is a proof for the original statement.

The real solution can be found by doing symbolic calculations using a CAS like Mathematica, Maple, Mathcad or wxMaxima. Once you get the procedure down you can modify the plant and the inner and outer loop controllers then see how those changes affect the relative bandwidths very quickly.
 
Peter,

refer in Internet bandwidth ratio is of course first approach only. For velocity-to-current bandwidth for example it may be higher due to load-to-motor resonances. You may review such definition for one of Elmo servo amps - http://www.elmomc.com/products/cello-main.htm
And real solution must be based on real plant behaviors and not on simulated calculations only.
 
Top