Thank you very much for your clear explanation Dr. Sandun..you really help me to understand the hard concept with an excellent explanation..May God Bless You
Thank you very much for sharing this content. Very well explained. P.S. I'm also a big fan of old Halo games, mainly CE and 2. If you didn't know there's a mod for Halo 2 for vista OS that allows you to play it on modern windows OS and has active multiplayer servers. It's called Project Cartographer
Hi Kuruppu, This is One of the excellent video I have come across...Nice Explanation......Thank You....... Can you please make videos on saliency of the motor and sensorless approaches..
In the case of the Internal magnet machine with reluctance, for correct sensor calibration, will the back emf peak still allign with the rotor position sensor 0, the same way? I think there will be some reluctance flux in the Q axis, that may change this approach...
I don't believe so. Because reluctance is not a flux component. It is a torque generating mechanism. This should work for IPMs too. I don't have a machine that I can demonstrate this at the moment. I will post more data once I acquire one.
Thanks sir, very helpful. Do you also have any calibration methods for GMR type sensors having different kinds of errors such as amplitude error,phase error and sensor mounting error? Or Can you recommend me some references
Thank you for a brilliant video. Didn't think of using the Back-EMF for positioning the rotor, just what I needed. But I have a question. Why not just activate phase A i.e [100]. When the rotor is locked then the rotor angle should be 0deg. or?
Yes, this is a common technique. But there are issues with this approach when there is cogging torque. This align and run method is quite popular with sensorless algorithm startup.
@@SandunKuruppu Thank you for replying my questions. I was only talking about the calibration of the external sensor. If we activated the phases (Phase A to VCC and Phase B&C to GND) and the read the external rotor sensor angle. The we could consider that value as 0deg.
Hi, thank you very much for video, at some point in the video you are saying that you've talked in another video about electrical vs mechanical angle for multiple pole pair machine Could point me which which video is it please? thank you in advance
Hi If back emf is deterministic,can we apply a sinusoidal pwm with relative offset from a lookup table for three phases (with a tiny amount of phase advance to account for position loop update delay)to obtain maximum performance with the position feedback from encoder,avoiding the complex foc control? What will be the fallacies in this scheme of control,and is there a any simple way to accommodate the cross coupling between the phases in the above mentioned scheme? *Torque constant and back emf constant for the motor are known in advance
Hello, Back emf is deterministic only when speed is constant. So it wont work beyond this condition. Also, here we are regulating flux through indirect regulation of current. You will need to change the amplitude of the voltage to keep up with changing speed conditions as the back emf changes with speed. The whole idea of rotor reference frame transformation is to avoid considering the complex coupling factor that are seen in stator frame reference frame. Cross-axis decoupling can be done in Rotor Reference Frame, but has very little benefit. Hope this helps!
@@SandunKuruppu thanks for the explanation. yes i need to accommodate for the back emf changes the changing duty cycle of the pwm. My requirement is a simple positioning one with no speed control or changing torque demand. The load will always be constant. So I am looking into a crude model based control with feedback only being the position encoder(high ppr one) value.but i am some what lost in details. any help will be appreciated.
@@dhuwarakesh6608 Yes, this is possible. This is a form of open loop or feedforward control. If you have a D-Q model, then you can achieve this very simply. This has been done in the past.
@@SandunKuruppu i referred some old methods all I could find was hysteresis based control with current feedback other than that it was all some flavour of DTC or FOC.if you have any reference it will be of much use for me.i have some one or two but they are just a speed feed forward control but not based on DQ model.
@@SandunKuruppu with motor model we have to decide upon the direct axis current and quadrature axis current first then perform a inverse park transform to be obtain the three phase current(comes from a look up table based on position) and then taking back emf into consideration apply the duty cycle for the required current...feed forward term I can't get figure out..am i getting it wrong?
@@SandunKuruppu : Hello sir, Could you please make some videos on fault tolerant control for dynamic performance improvement of bldc motor drives and on hall sensor fault diagnosis, identification and reconstruction of hall signals. This ll be a great help. Thanks in advance
@@SandunKuruppu Thanks sir. It will be very helpful if you could please give some explanation on the possible reasons for hall sensor glitches. While testing a hall sensor based bldc motor I could observe glitches in the hall sensor itself. Is it like because of hall sensor pcb mounted on the motor may have some grounding issue, or something else.
@@SandunKuruppu At 0:43 the rotor is resting at electrical zero angle position and the vertical bar are physical conductor of Phase A, if you are considering angle between d-axis of rotor and phase- A axis. Accordingly the initial rotor position at 12:40 should be 90 degree not 0 degree. Flux linkage and voltage output analysis are correct. So clarification is required, whether the d or q-axis of the rotor is used to calculate the electrical rotor angle.
@@manasjena949 The starting position can be defined with respect to any stator reference point. In the first case it was defined horizontally and in the second case it is defined vertically. This is only the starting position and it can be a constant offset correction.
@@SandunKuruppu yes sir like while going from Iq to Vq we basically uses pi controller to do that in simulation I am talking about that how to design but I got research paper how to do that. But now I am getting stuck on how to get Iq reference current from speed and Id reference current (considering filed weakening otherwise we can put it zero ) .
@@SandunKuruppu and sorry for any inconveniences from my side I just want to learn how to design complete FOC step by step. As it will be helpful for mtech
Thank you so much, I really found this channel really really helpful.
Happy to hear that!
Thank you very much for your clear explanation Dr. Sandun..you really help me to understand the hard concept with an excellent explanation..May God Bless You
Glad it was helpful!
The explanation is perfect, Thanks for it
Glad it was helpful!
Thank you very much for sharing this content. Very well explained.
P.S. I'm also a big fan of old Halo games, mainly CE and 2. If you didn't know there's a mod for Halo 2 for vista OS that allows you to play it on modern windows OS and has active multiplayer servers. It's called Project Cartographer
Thank you Derp! I will check it out.
@@SandunKuruppu No problem! Halo 2 Vista/Project Cartographer plays much more like the original xbox version than the MCC version.
Hi Kuruppu, This is One of the excellent video I have come across...Nice Explanation......Thank You.......
Can you please make videos on saliency of the motor and sensorless approaches..
Thank you! I will add these topics to my future video list. I do have saliency covered in one of the videos, not sensorless control through.
In the case of the Internal magnet machine with reluctance, for correct sensor calibration, will the back emf peak still allign with the rotor position sensor 0, the same way? I think there will be some reluctance flux in the Q axis, that may change this approach...
I don't believe so. Because reluctance is not a flux component. It is a torque generating mechanism. This should work for IPMs too. I don't have a machine that I can demonstrate this at the moment. I will post more data once I acquire one.
Thanks sir, very helpful.
Do you also have any calibration methods for GMR type sensors having different kinds of errors such as amplitude error,phase error and sensor mounting error? Or Can you recommend me some references
You would be better off speaking with the sensor manufacturer. They will be able to help you with this.
Thank you for a brilliant video.
Didn't think of using the Back-EMF for positioning the rotor, just what I needed.
But I have a question. Why not just activate phase A i.e [100]. When the rotor is locked then the rotor angle should be 0deg. or?
Phase A and Neutral? You need a return current path and based on the return path the flux alignment can change.
@@SandunKuruppu I was thinking Phase A to VCC and Phase B&C to GND. Then the rotor will stay in a very specific position.
Yes, this is a common technique. But there are issues with this approach when there is cogging torque. This align and run method is quite popular with sensorless algorithm startup.
@@SandunKuruppu Thank you for replying my questions. I was only talking about the calibration of the external sensor. If we activated the phases (Phase A to VCC and Phase B&C to GND) and the read the external rotor sensor angle. The we could consider that value as 0deg.
Depends on the motor. If there is lot of cogging, it wont be the actual zero.
Thank you sir.
Most welcome!
Hi,
thank you very much for video, at some point in the video you are saying that you've talked in another video about electrical vs mechanical angle for multiple pole pair machine
Could point me which which video is it please?
thank you in advance
ua-cam.com/video/VSPr7_d4rjA/v-deo.html
thank you sir
Most welcome!
Nice video. Can you tell which book you refer for PMSM control?
There is no specific textbook for this video. This is based on my experience.
Hi
If back emf is deterministic,can we apply a sinusoidal pwm with relative offset from a lookup table for three phases (with a tiny amount of phase advance to account for position loop update delay)to obtain maximum performance with the position feedback from encoder,avoiding the complex foc control?
What will be the fallacies in this scheme of control,and is there a any simple way to accommodate the cross coupling between the phases in the above mentioned scheme?
*Torque constant and back emf constant for the motor are known in advance
Hello,
Back emf is deterministic only when speed is constant. So it wont work beyond this condition.
Also, here we are regulating flux through indirect regulation of current. You will need to change the amplitude of the voltage to keep up with changing speed conditions as the back emf changes with speed.
The whole idea of rotor reference frame transformation is to avoid considering the complex coupling factor that are seen in stator frame reference frame. Cross-axis decoupling can be done in Rotor Reference Frame, but has very little benefit. Hope this helps!
@@SandunKuruppu thanks for the explanation. yes i need to accommodate for the back emf changes the changing duty cycle of the pwm.
My requirement is a simple positioning one with no speed control or changing torque demand.
The load will always be constant.
So I am looking into a crude model based control with feedback only being the position encoder(high ppr one) value.but i am some what lost in details. any help will be appreciated.
@@dhuwarakesh6608 Yes, this is possible. This is a form of open loop or feedforward control. If you have a D-Q model, then you can achieve this very simply. This has been done in the past.
@@SandunKuruppu i referred some old methods all I could find was hysteresis based control with current feedback other than that it was all some flavour of DTC or FOC.if you have any reference it will be of much use for me.i have some one or two but they are just a speed feed forward control but not based on DQ model.
@@SandunKuruppu with motor model we have to decide upon the direct axis current and quadrature axis current first then perform a inverse park transform to be obtain the three phase current(comes from a look up table based on position) and then taking back emf into consideration apply the duty cycle for the required current...feed forward term I can't get figure out..am i getting it wrong?
Thanks sir , very useful
Always welcome!
@@SandunKuruppu : Hello sir, Could you please make some videos on fault tolerant control for dynamic performance improvement of bldc motor drives and on hall sensor fault diagnosis, identification and reconstruction of hall signals. This ll be a great help.
Thanks in advance
@@jubul1988 I will add these to my list of topics. Thanks!
@@SandunKuruppu Thanks sir. It will be very helpful if you could please give some explanation on the possible reasons for hall sensor glitches. While testing a hall sensor based bldc motor I could observe glitches in the hall sensor itself. Is it like because of hall sensor pcb mounted on the motor may have some grounding issue, or something else.
@@jubul1988 I'm not sure what you mean. Can you share some scope captures of this issue?
Dear Sir
How Can we use Hall Sensor output data for finding theta more accurately at starting ?
Do you mean discrete hall sensors or linear hall sensors?
@@SandunKuruppu The output is in the form of zero and one so I think it is discrete!
@@anmolmeshram7057 So then this method is not applicable.
Sorry to say but I think Electrical theta angle zero will coincide with phase a- neutral zero crossing. In your video it is different. Why?
Can you tell me the time stamp of the video you are referring to?
@@SandunKuruppu At 0:43 the rotor is resting at electrical zero angle position and the vertical bar are physical conductor of Phase A, if you are considering angle between d-axis of rotor and phase- A axis. Accordingly the initial rotor position at 12:40 should be 90 degree not 0 degree. Flux linkage and voltage output analysis are correct. So clarification is required, whether the d or q-axis of the rotor is used to calculate the electrical rotor angle.
@@manasjena949 The starting position can be defined with respect to any stator reference point. In the first case it was defined horizontally and in the second case it is defined vertically. This is only the starting position and it can be a constant offset correction.
Sir sorry to say this not the right video can you tell me how to calculate pi for FOC for iqs and Ids to vqs and vds
I'm not sure what you mean by pi for FOC? Can you calrify?
@@SandunKuruppu yes sir like while going from Iq to Vq we basically uses pi controller to do that in simulation I am talking about that how to design but I got research paper how to do that. But now I am getting stuck on how to get Iq reference current from speed and Id reference current (considering filed weakening otherwise we can put it zero ) .
@@SandunKuruppu and sorry for any inconveniences from my side I just want to learn how to design complete FOC step by step. As it will be helpful for mtech
@@SandunKuruppuOr could you suggest me some source where I can learn and grasp the knowledge how to design everything.
@@SandunKuruppu Field Oriented Control
Thanks sir
Welcome!