Mr. Dave , I'm totally agree with you I did graduate in physics and I´m teaching, how ac drive runs an induction motor for 25 years at YASKAWA , but I' m a mechanical technician too, and all this disciplines I use to teaching costumers, your class it's a master class thx for that , I fell like a young boy now , well done Dave
there is an interesting topic which I had reserched. I considered the switching frequency as an additional time constant within the control plant. The time constant of the motor was more than 10 times bigger than the time delay introduced by the pwm switching frequency. Normally we would neglect that pwm time delay but....I had reserched that influence of switching frequeny on the controller within FOC. I used pole-zero compensation to calculate the controller gains. I used PT1 pkant to represent that pwm tike delay. Result was that the PI controller which does not have a third zero can not compensate for that pwm time delay so there was always an overshoot or a super slow pi controller without much dynamic in the inner current loop, the Iq loop. So I used an inner PID controller et voila, extreamly fast in response to torque changes without any overshoot. I found it extreamly useful in current motor control of big high torque PMSM and at low pwm frequencies at and below of 20kHz.
Excellent video. Very very informational and definitely something I was looking for. One doubt I have over what he says is that the motor goes faster as the current increases. I am not sure about that: yes the torque should increase, but the speed should depend on the frequency in my opinion
I agree, the amplitude would have to do with the amplitude of the torque, and the the frequency has to do with the speed. I believe his statements are incorrect in this regard.
Yea I believe that's a mistake as well since the torque will increase as the Iq current increases. The speed of the rotor isn't proportional to frequency though it's proportional to the voltage amplitude of the motor. This is why the Kv terms of most motors are listed since they are in units of angular speed per volt (e.g., RPM/Volt). Also, I think Mantis might be onto the right idea. The motor is loaded so much the speed gets reduced and as the torque increases to overcome the load it will speed up.
I bought brand new vacuum cleaner on yearly fair, under price. So i went home and figured device turns off after few seconds. Motor is ok, but control is bad, 8-pin blank chip dead ( pin 8 connected to A1 triac and i dont know marking) . Now iam watching this ! I miss my study times... Very nice video !
Good presentation, A question is current loop tuning how we will set the R/L Zero , Kb gain value in digital control , how it depends on the sampling frequency .
Hi, for questions regarding specific design advice, check out this E2E forum where you can search questions or ask the experts: e2e.ti.com/support/applications/motor_drivers/
Good afternoon! Vector control theory says that the resulting stator field vector must be kept perpendicular to the rotor magnet. If this is one magnet with two poles, then it is not difficult to imagine and understand. But in life we have a lot of magnets. For example, there can be 12 of them on a scooter motor wheel, and the rotor is located outside the stator. How to present this corner there? How is the magnetic field located?
Great Video, Thanks Guru. I understand the Ka, Kb parameter identified from R&L but how this is linked in a digital PI . how to get those values for a Zero cancellation digital PI.
I love these videos! Thank you for them. I'm trying to figure out how port my Simulink models into HDL code so that I can implement this on an FPGA device. Does anyone know how to do this?
There are tutorials on mathworks website on how to do it, there's something called simulink hdl coder which does this, but i guess it only works on Spartan fpga devices
@@education7809 . I tried to buy a copy from Mathworks. They have a deal for someone for small companies to buy a license, but it's $4k at the time I asked. Normally it's $12k. But if it saves time and prevents mistakes, then I think it's worth the cost. However, this is not something a DIYer like myself can afford at this time.
Somewhat frustrating. The audience participation can't be heard without a microphone. You might say there is no audio feedback from the listeners perspective. Thumbs up anyway.
Isn't there a trick to locate the differentiator to some other point in the block diagram so that the differentiator does not put out an impulse when set point is changed
1:17:40 Question, are the peak values of the three-phase current always the same size or do these values vary with respect to each other during the control by the foe algorithm?
As far as phase distortions, when I studied DSP we played around with more than just the standard analog filters. Is that an approach that is used in control design? My brain always goes straight to audio whenever filters are brought up. What about high order IIR filters? Is that something that is avoided due to delay and complexity?
I think i figured it out. Correct me if i am wrong but for both controllers the units of the controller actually cancel out the inputs and you end up with output in terms of amps or volts (assuming radian quantity is m/m unitless). This only holds in the case where that damping term delta is not the coefficient of viscous friction but a unitless coefficient. As for the number of poles, he says its number of poles so its number of poles. I guess a lot of the other stuff i was reading used pole pairs so i was a bit confused when i saw poles instead for the first time. The math works out for the 3 phase torque constant so if you have poles pairs just multiply by 2. In my eyes, the strategy is to quantify a physical phenomena or reaction, ie inductance as a slew rate or something, and relate it to the error or integral of the error to get something that isnt just an empirically tuned constant. For example, the proportional terms P is simply L(Vs/A)*w(rad/s system bandwidth)*current error(A) which works out simply to volts*rads or just volts.
Well, it's also correct, tfd.com/z says it's th B.E. pronunciation. Of course it's also useful in speech to distinguish it from the letter C. N.B.: The same holds for 'process' he admonished Richard for. Obviously Richard is from England, and he's just teasing him.
![Ice On My Teeth - ATEEZ エイティーズ 에이티즈 [Music Bank] | KBS WORLD TV 241115](http://i.ytimg.com/vi/c-n92OgbbwA/mqdefault.jpg)
I would have passed control system theory with an A if the lecturer taught like this. This man is a great engineer...and also a presenter!
I thought I will get A till I saw the question paper.
I absolutely love this guy's teaching style. I wonder if he's still with Texas Instruments in 2023.
Mr. Dave , I'm totally agree with you I did graduate in physics and I´m teaching, how ac drive runs an induction motor for 25 years at YASKAWA , but I' m a mechanical technician too, and all this disciplines I use to teaching costumers, your class it's a master class thx for that , I fell like a young boy now , well done Dave
I am from Germany and we actually have two different words for control :D
Steuerung - Control with no feedback
Regelung - Control with feedback
Some people operate with NO feedback. 🤣
13.35 I understand his feeling when I learnt control system undergrad class in 80'. The same puzzle came to mind.
there is an interesting topic which I had reserched. I considered the switching frequency as an additional time constant within the control plant. The time constant of the motor was more than 10 times bigger than the time delay introduced by the pwm switching frequency. Normally we would neglect that pwm time delay but....I had reserched that influence of switching frequeny on the controller within FOC. I used pole-zero compensation to calculate the controller gains. I used PT1 pkant to represent that pwm tike delay. Result was that the PI controller which does not have a third zero can not compensate for that pwm time delay so there was always an overshoot or a super slow pi controller without much dynamic in the inner current loop, the Iq loop. So I used an inner PID controller et voila, extreamly fast in response to torque changes without any overshoot. I found it extreamly useful in current motor control of big high torque PMSM and at low pwm frequencies at and below of 20kHz.
Like a motor control bible. Excellent work 👍
The teacher is freaking awesome
Thank you for this awesome experience. I am foreigner and always wanted to experience how in other countries they explain this things !
finally i understand pid's a little more. thank you
Excellent video. Very very informational and definitely something I was looking for.
One doubt I have over what he says is that the motor goes faster as the current increases. I am not sure about that: yes the torque should increase, but the speed should depend on the frequency in my opinion
Purple Spark ⚡️ Perhaps he meant that a motor under load goes faster
I agree, the amplitude would have to do with the amplitude of the torque, and the the frequency has to do with the speed. I believe his statements are incorrect in this regard.
notice that in the two diagrams, the one on the left speeds up when the rotor on the right speeds up.
Yea I believe that's a mistake as well since the torque will increase as the Iq current increases. The speed of the rotor isn't proportional to frequency though it's proportional to the voltage amplitude of the motor. This is why the Kv terms of most motors are listed since they are in units of angular speed per volt (e.g., RPM/Volt).
Also, I think Mantis might be onto the right idea. The motor is loaded so much the speed gets reduced and as the torque increases to overcome the load it will speed up.
Excellent presentation!
Thank you.
I bought brand new vacuum cleaner on yearly fair, under price. So i went home and figured device turns off after few seconds. Motor is ok, but control is bad, 8-pin blank chip dead ( pin 8 connected to A1 triac and i dont know marking) .
Now iam watching this ! I miss my study times...
Very nice video !
Could you tell me how to get in touch with the teacher?
Thanks for the upload TI!
Motor Control Seminar - Richard Poley & Dave Wilson, Texas Instruments: ruclips.net/p/PLuUNUe8EVqlkaAtIzzdF7KUDvQIJ4KM_0
Good presentation, A question is current loop tuning how we will set the R/L Zero , Kb gain value in digital control , how it depends on the sampling frequency .
Hi, for questions regarding specific design advice, check out this E2E forum where you can search questions or ask the experts: e2e.ti.com/support/applications/motor_drivers/
Thank you , I found it from the TI forum e2e.ti.com/support/microcontrollers/c2000/f/902/t/312133
In 38:42 minutes you are giving your website to ınderstand mathematical background. Howeer I can reach it. Could you please share a link.
Great videos! where i can find materials from TI motor blog now?
great vidéo Mr .
Good afternoon! Vector control theory says that the resulting stator field vector must be kept perpendicular to the rotor magnet. If this is one magnet with two poles, then it is not difficult to imagine and understand. But in life we have a lot of magnets. For example, there can be 12 of them on a scooter motor wheel, and the rotor is located outside the stator. How to present this corner there? How is the magnetic field located?
Great Video, Thanks Guru. I understand the Ka, Kb parameter identified from R&L but how this is linked in a digital PI . how to get those values for a Zero cancellation digital PI.
I love these videos! Thank you for them. I'm trying to figure out how port my Simulink models into HDL code so that I can implement this on an FPGA device. Does anyone know how to do this?
There are tutorials on mathworks website on how to do it, there's something called simulink hdl coder which does this, but i guess it only works on Spartan fpga devices
@@education7809 . I tried to buy a copy from Mathworks. They have a deal for someone for small companies to buy a license, but it's $4k at the time I asked. Normally it's $12k. But if it saves time and prevents mistakes, then I think it's worth the cost. However, this is not something a DIYer like myself can afford at this time.
Somewhat frustrating. The audience participation can't be heard without a microphone. You might say there is no audio feedback from the listeners perspective. Thumbs up anyway.
what do you think about the wire attached on her dress?
+Tulak Adik u wot m8???
+CaptainAthiest ?
`
Interesting didactic and really enjoyable speech!
Great Video, Thanks Guru. I understand the Ka, Kb parameter identified from R&L but how this is linked in a digital PI .
Very good explanations, thank you!
Thank you Sir. I am a fan your teaching... sivarajan
Of your teaching..
Isn't there a trick to locate the differentiator to some other point in the block diagram so that the differentiator does not put out an impulse when set point is changed
1:17:40 Question, are the peak values of the three-phase current always the same size or do these values vary with respect to each other during the control by the foe algorithm?
Nearly every electronic problem can be solved with duct tape and a Schmitt trigger...
Tried this on my Air traffic control console.
It works, thanks for the tip.
I can make this motor run.. just give me some duct tape, fishing line... and a brushless motor controller 😎
Can I get the link of the last video of this series which focuses on implementation with registers?
Really nice lecture
Really great thanks so much
As far as phase distortions, when I studied DSP we played around with more than just the standard analog filters. Is that an approach that is used in control design? My brain always goes straight to audio whenever filters are brought up. What about high order IIR filters? Is that something that is avoided due to delay and complexity?
I thought he explained why you did not want high order filters. It introduces more phase lag. Bigger problem for the higher frequencies.
Where can we get the slides?
Is the Richard video (previous presentation) available on youtube? What's the name?
ruclips.net/video/Kr_wmPBUsyY/видео.html this one
hahaha!!! lol
I don't think that this could be stated more elegantly! Manager: uneducated observer!!! :D
Thanks very much!
thank you very much, it's helps me a lot
The best video.
Amazing Class
What happend on day 1 on the seminar
does anybody know if P at 43:32 refers to the number of poles exactly or does he mean pole pairs?
I think i figured it out. Correct me if i am wrong but for both controllers the units of the controller actually cancel out the inputs and you end up with output in terms of amps or volts (assuming radian quantity is m/m unitless). This only holds in the case where that damping term delta is not the coefficient of viscous friction but a unitless coefficient.
As for the number of poles, he says its number of poles so its number of poles. I guess a lot of the other stuff i was reading used pole pairs so i was a bit confused when i saw poles instead for the first time. The math works out for the 3 phase torque constant so if you have poles pairs just multiply by 2.
In my eyes, the strategy is to quantify a physical phenomena or reaction, ie inductance as a slew rate or something, and relate it to the error or integral of the error to get something that isnt just an empirically tuned constant. For example, the proportional terms P is simply L(Vs/A)*w(rad/s system bandwidth)*current error(A) which works out simply to volts*rads or just volts.
I recommend watching it at 1.25 speed
atleast upto the end of first question
anyone found the blog?
Thank you
what kind of component were used to cotrol motor in the early 1900s or late 1800s?
Sorry, but zed-transform sounds much better to me :D
Well, it's also correct, tfd.com/z says it's th B.E. pronunciation.
Of course it's also useful in speech to distinguish it from the letter C.
N.B.: The same holds for 'process' he admonished Richard for. Obviously Richard is from England, and he's just teasing him.
Never heard it even once until my second run at college, and then it seemed like everyone used it. I prefer NATO phonetic alphabet.
Where is the blog now?
Hi Zhitai-- you can read the blog here: e2e.ti.com/blogs_/b/motordrivecontrol/archive/2014/02/21/teaching-old-motors-new-tricks
Great! Really excellent materials for newbies like me!
@@TexasInstrumentsThis link is not available. Please share the updated link
5:25 How self-absorbed do you have to be to believe that you don't speak with an accent?
crumbs
You must forgive his pronunciation, he's not English.
Motor Control
👍👍👍❤️
Which L is he talking about? My design software has four L values: Ld, Lq, L self inductance and L mutual inductance.
Im guessing Ls in an idealized world
An American correcting English speakers on pronunciation is a bad start.
Soup