Hussein, this is a common question I get from my students. Don't treat it as a units conversion, rather a closed loop system. Think of it from a control system perspective where the voltages are applied based on the current error, because voltage is the variable we can control and current is the variable we want to control. It is similar to controlling the gas peddle we when driving a car. Hope this helps!
@@SandunKuruppu Hi Sandun, can you please provide some references to understand hall offset in càse of a ipmsm motor with hall sensor and how to findout. Is it like the angle difference between theta when changes from 360 deg to 0 deg and U- Phase back emf zero crossing falling edge. Can u give some insights on sensor calibration how usually done . Thanks in advance.
@@jubul1988 Let me first try to understand your question here. Are you using discrete hall sensors for trapezoidal control or a linear hall sensor for field oriented control? So if it is a linear hall sensor, the sensor offset calibration is same as non-salient PMSMs. An IPM is a salient PMSM where, Ld is not equal to Lq. Therefore, you can get reluctance torque by injection Id. This is a concept called maximum torque per ampere. It has nothing to do with sensor calibration. It places the stator current vector and an angle away from 90 degrees to excite the reluctance effect for torque.
It was a good explanation, I spent hours try to understand and read So you didn't use a clarke transform with your technic, right? But if i use both transforms, how can I measure the current?, that part is still unclear for me , I think if i measure directly my microntroler will burn. Should I pick up measure directly to my microcontroller? thanks!
Azael, I'm assuming you are asking about FOC. Uou do have to use Clarke and Park transforms, but in this case they are combined to one transform called the reference frame transformation. So in a way, I did use them. You have to measure phased currents using current sensors (shunt or hall effect based current sensors) and read them into a microprocessor. Then you have to apply reference frame transformation.
That I believe would be a speed control problem. The speed control loop will establish the desired Id and Iq. There may be other strategies to setup Id and Iq, when the load is varying. When load is varying, typically, the motor does not regulate torque (if there is a rigid coupling between the load and the motor/PMSM).
how foc method works for sensored motors? Bcz foc uses Clark and parks transform ..For which it takes current..So how with hall sensors Foc method works?
Hello sir, can you please explain why we used PI controller not PD controller to get vd,vq out of id,iq Although the equation of vd,vq contain derivative of these currents not integrals of them
The current regulated plant has a pole and therefore a PI is sufficient to move the pole and achieve the dynamics. Not having an integrator (like in a PD) results in zero steady state error.
Hello Ankit, that is a good question. There are number aspects that you have to consider in calculating reference such as voltage available, motor specifications and operating point. This depends on the application. So there is no straight forward answer.
The phase angle theta is varied by varieing frequency coontinuesly this way we get a 90degree phase shift between rotor vector and stator vector at any condition of rotor position. Is this statement is correct.pls help me.
Naseeb, what do you mean by phase angle is varied by varying frequency? Specify the time stamp of the video. PMSMs are synchronous motors. You can only control the relative placement between the stator and the rotor field. That is done through the inverter voltages applied. Frequency depends on the motor speed and number of poles.
@@SandunKuruppu that statement not include in your video its my mistake what is the function of inverter in pmsm drive. how the motor speed depends on supply frequency.
Hi Sandun..May I know about any simple method to execute in lab for the measurement of offset angle for Hall sensors in PMSM application.I referred Mathworks approach but the measured offset is not correct for the motor that I have. So any suggestions from your side would be much helpful.
Hi Ravi, there are different methods you can use the depends on the motor you have (IPM or SPM), what type of signals you have access to and if the motor phases can be removed from the inverter. The well known approach is to observe motor back emf (no inverter attached) and position sensor signal while the motor is spun at ~100RPM to find the alignment between the position sensor signal and back emf of phase A. Let me know if this is not doable and there other methods to calibrate on the dyno with the inverter attached.
@@SandunKuruppu I used a method where the motor is run on open loop with only Vd applied..The position signal is provided and I logged the measured theta from hall sensors ..The value of theta corresponding to zero fall of open loop position input signal was taken as the offset..But this method gives me a higer offset value in the range of 280 and above...Is it fine or any mistakes in this approach?
@@ravikodoth7284 When you say open loop, do you mean using an internal ramp? I guessing the motor is driven open loop by the inverter? I'm not clear what you mean by "Value of theta corresponding to zero fall of open loop position input signal.." Could you clarify? You can send me an email to the email shown on my channel. A picture of your calibration setup will be also helpful.
Hello sir, In Matlab simulation, Obtaining speed from rotor position by utilizing derivative block causing disturbance and fluctuations of speed. How could we obtain accurate speed from rotor position...while FOC of PMSM..
Pavan, In real applications, the speed is calculated based on rotor position. i.e. by taking the derivative. How much noise are you seeing? What kind of noise? There shouldn't be any noise in your simulation. Alternatively, if you are simulating a dynamometer (dyno), the speed should be held constant after a transient while the PMSM regulates torque. So, some information is missing in your question for me to help you.
Torque ripple does exist to a certain degree in many machines and sometimes negligible. There are many reasons and it is a larger topic of discussion. Sometimes it is the machine design, sometimes the drive and sometimes it is the controller. So it depends. I will post a video on this topic soon!
I will share a Matlab Simulation example for FOC. Are you looking to see what the signals look like with FOC? I can include both simulation and experimental results.
@@SandunKuruppu matlab simulation of FOC on PMSM...this reference current generation is done with other method we cannot take it as constant I think.....as FOC applied in induction motor there id iq is generated with different way
@@beatthetech9463 I dont follow your question. Reference currents are generated based on the torque required. I dont see why it cant be constant. If the torque needed is constant, then the current references are constant. Does this make sense? Same general concept with a difference in implementation is followed in induction motor FOC.
Pavan, you can obtain rotor position using a position sensor or you can estimate with an estimator or for certain types of machines, you can inject high frequency to measure position.
@@pavangongalla9292 Pavan, see my other videos. I show how to design PI controller with bandwidth and how to simulate PMSM as a torque controller using constant speed integrated to create the position signal.
![[TH]FFWS SEA Spring 2024 Knockout Stage - Day 14](http://i.ytimg.com/vi/CU9MV8EEIjE/mqdefault.jpg)
This is sensational 💪🏻 Thank you👌
You're so welcome!
It is a great explanation as always, grateful for the help we got during the project.
Great to hear from you Gayan! Glad it helped. Are you working on anything related to what we did?
It's really a good lecture. Though it's a complex algorithm you made it very simple. Thanks.
Glad it was helpful Mohanraj!
Great work
Thank you!
Thanks sir for your amazing gift in making complicate things into a simple analogy, things that only genius people do.
Thank you!
Best teacher ever seen
Thank you!
Very nice explanation sir 👍👍
Thanks and welcome
Hi Sandun, do you think you may ever go over resonant control in the abc frame?
Thank-you for your videos they are very helpful!
Doron, Can you elaborate on what you mean by resonant control? Not sure if it is a specific term we are not familiar with. Do you mean with a PLL?
awsome explanation.
at the time of 39:40, could you please how the PI controller converts error from the current domain into the volage domain?
Hussein, this is a common question I get from my students. Don't treat it as a units conversion, rather a closed loop system. Think of it from a control system perspective where the voltages are applied based on the current error, because voltage is the variable we can control and current is the variable we want to control. It is similar to controlling the gas peddle we when driving a car. Hope this helps!
Can you make a video about the clark park transformation?
Its already there!
Great lectures sir,
Could you please explain to me how the necessary parameters such as the rotor position may be derived for a sensorless FOC.
There are different approaches used. Estimation algorithms rely on measurements and a motor model to track position.
@@SandunKuruppu Hi Sandun, can you please provide some references to understand hall offset in càse of a ipmsm motor with hall sensor and how to findout. Is it like the angle difference between theta when changes from 360 deg to 0 deg and U- Phase back emf zero crossing falling edge. Can u give some insights on sensor calibration how usually done . Thanks in advance.
@@jubul1988 Let me first try to understand your question here. Are you using discrete hall sensors for trapezoidal control or a linear hall sensor for field oriented control? So if it is a linear hall sensor, the sensor offset calibration is same as non-salient PMSMs. An IPM is a salient PMSM where, Ld is not equal to Lq. Therefore, you can get reluctance torque by injection Id. This is a concept called maximum torque per ampere. It has nothing to do with sensor calibration. It places the stator current vector and an angle away from 90 degrees to excite the reluctance effect for torque.
It was a good explanation, I spent hours try to understand and read
So you didn't use a clarke transform with your technic, right?
But if i use both transforms, how can I measure the current?, that part is still unclear for me , I think if i measure directly my microntroler will burn.
Should I pick up measure directly to my microcontroller?
thanks!
Azael, I'm assuming you are asking about FOC. Uou do have to use Clarke and Park transforms, but in this case they are combined to one transform called the reference frame transformation. So in a way, I did use them. You have to measure phased currents using current sensors (shunt or hall effect based current sensors) and read them into a microprocessor. Then you have to apply reference frame transformation.
How do we get I_qs_refrence and I_ds_refrence as a certain value if the mechanical torque is varying for example with wind in a wind turbine ?
That I believe would be a speed control problem. The speed control loop will establish the desired Id and Iq. There may be other strategies to setup Id and Iq, when the load is varying. When load is varying, typically, the motor does not regulate torque (if there is a rigid coupling between the load and the motor/PMSM).
how foc method works for sensored motors? Bcz foc uses Clark and parks transform ..For which it takes current..So how with hall sensors Foc method works?
The question is not clear here! Current and rotor position are needed to obtain rotor reference frame currents.
Hello sir,
can you please explain why we used PI controller not PD controller to get vd,vq out of id,iq Although the equation of vd,vq contain derivative of these currents not integrals of them
The current regulated plant has a pole and therefore a PI is sufficient to move the pole and achieve the dynamics. Not having an integrator (like in a PD) results in zero steady state error.
Sir in video at 36:21 you have mention about Reference torque.. Please explain how From references torque ID and IQ are calculated.
Thanks
Hello Ankit, that is a good question. There are number aspects that you have to consider in calculating reference such as voltage available, motor specifications and operating point. This depends on the application. So there is no straight forward answer.
Sir can you explain how to have PI controller gains.
It is already there in one of the earlier videos. Start from the beginning of the playlist.
Hello sir,
Is the position is same as rotor angle (rad) in MATLAB simulink block ??
or
could we measure it separately by any equation ..
Pavan,
The position is typically the rotor angle. You could measure it externally too. It depends how you setup the model. i.e. integrate speed.
@@SandunKuruppu ok sir, thank you
The phase angle theta is varied by varieing frequency coontinuesly this way we get a 90degree phase shift between rotor vector and stator vector at any condition of rotor position.
Is this statement is correct.pls help me.
Naseeb, what do you mean by phase angle is varied by varying frequency? Specify the time stamp of the video. PMSMs are synchronous motors. You can only control the relative placement between the stator and the rotor field. That is done through the inverter voltages applied. Frequency depends on the motor speed and number of poles.
@@SandunKuruppu that statement not include in your video its my mistake
what is the function of inverter in pmsm drive.
how the motor speed depends on supply frequency.
Hi Sandun..May I know about any simple method to execute in lab for the measurement of offset angle for Hall sensors in PMSM application.I referred Mathworks approach but the measured offset is not correct for the motor that I have. So any suggestions from your side would be much helpful.
Hi Ravi, there are different methods you can use the depends on the motor you have (IPM or SPM), what type of signals you have access to and if the motor phases can be removed from the inverter. The well known approach is to observe motor back emf (no inverter attached) and position sensor signal while the motor is spun at ~100RPM to find the alignment between the position sensor signal and back emf of phase A. Let me know if this is not doable and there other methods to calibrate on the dyno with the inverter attached.
@@SandunKuruppu I used a method where the motor is run on open loop with only Vd applied..The position signal is provided and I logged the measured theta from hall sensors ..The value of theta corresponding to zero fall of open loop position input signal was taken as the offset..But this method gives me a higer offset value in the range of 280 and above...Is it fine or any mistakes in this approach?
@@ravikodoth7284 When you say open loop, do you mean using an internal ramp? I guessing the motor is driven open loop by the inverter? I'm not clear what you mean by "Value of theta corresponding to zero fall of open loop position input signal.." Could you clarify? You can send me an email to the email shown on my channel. A picture of your calibration setup will be also helpful.
nice
Thanks!
Hello sir,
In Matlab simulation, Obtaining speed from rotor position by utilizing derivative block causing disturbance and fluctuations of speed.
How could we obtain accurate speed from rotor position...while FOC of PMSM..
Pavan,
In real applications, the speed is calculated based on rotor position. i.e. by taking the derivative. How much noise are you seeing? What kind of noise? There shouldn't be any noise in your simulation.
Alternatively, if you are simulating a dynamometer (dyno), the speed should be held constant after a transient while the PMSM regulates torque. So, some information is missing in your question for me to help you.
Iam implementing the FOC of PMSM but the line voltage and currents of the motor are not similar as calculated values...what could be the reason ??
Need more information Pavan. A picture/plot would help!
sir is it a masters class you r teaching for or degree class
For BSEE degree Siddhesh.
Hello sir,
There are torque ripples in the PMSM motor,
What is the reason ?
What is the solution ?
Torque ripple does exist to a certain degree in many machines and sometimes negligible. There are many reasons and it is a larger topic of discussion. Sometimes it is the machine design, sometimes the drive and sometimes it is the controller. So it depends. I will post a video on this topic soon!
can u please share some matlab simulation video of implementing FOC on PMSM it will be really helpful please
I will share a Matlab Simulation example for FOC. Are you looking to see what the signals look like with FOC? I can include both simulation and experimental results.
@@SandunKuruppu matlab simulation of FOC on PMSM...this reference current generation is done with other method we cannot take it as constant I think.....as FOC applied in induction motor there id iq is generated with different way
@@beatthetech9463 I dont follow your question. Reference currents are generated based on the torque required. I dont see why it cant be constant. If the torque needed is constant, then the current references are constant. Does this make sense?
Same general concept with a difference in implementation is followed in induction motor FOC.
@@SandunKuruppu do upload simulation sir... please of FOC on PMSM that's will be really helpful
Dear Sir,
Can you provide videos on multiphase pmsm
Do you mean more than three phase?
@@SandunKuruppu
Yes sir for 6 phase ipmsm foc control
how could we calculate the rotor position of PMSM ?
Pavan, you can obtain rotor position using a position sensor or you can estimate with an estimator or for certain types of machines, you can inject high frequency to measure position.
@@SandunKuruppu in simulation how could I do it
While performing foc of pmsm in matlab simulink, what could be the theta in parks and inverse parks transformation
How to tune the PI controllers of PMSM in FOC control
@@pavangongalla9292 Pavan, see my other videos. I show how to design PI controller with bandwidth and how to simulate PMSM as a torque controller using constant speed integrated to create the position signal.