When i see Brian Douglas video first i hit like button, take my pen and a paper, take a coffee and then i enjoy a 1 year university class equivalent video. PS: Now audio are nice!
No offense, but your university must teach very little... What Brain covered in all of his videos, are about 3 month worth of our contents. Although brain explains the general concepts well, he does not cover any in depth content.
I watched this right after my exam of control theory, because I failed at the state space model and the discrete modeling... I was blown away by this explanation and finally understood what my professor can't explain... Thank you very much!
man u r so great !!!! i 'm from brazil and here we don't have leraning like that !!! the world thanks u so much ! and another thing, thanks for u speak so clearly !!! if wasn't it, at least to me would be impossible understand !
Brian! Thank you so much for your efforts in putting these videos together, they make an otherwise out of reach knowledge set available to anyone interested and that is the coolest thing anyone can do! Your an awesome person for doing this and your time and efforts are deeply appreciated! great video, cant wait for more on discrete control :) !
Amazing videos! Really gave me a much better view of what was taught at control class. Thanks a lot for sharing this amazing videos! Best teaching methods!
I just love your lectures. I did not know there is a way of donation for your work. I am doing it soon. I am communication engineer but I always had interest for Control and I want to be the part of it in my career later. Keep going please :)
thank you for those videos they are really helpful and I like your way of explaining there is a small mistake in 19:46 i think you forgot to multiply the 's' to the time delay
Hi Brian, I am a roboticist. Mostly I deal with problems related to modern control theory. So if you plan for more tutorials, I would like to see Modern Control theory. Thanks for all awesome lectures.
Nooooooooo! I thought I could get through this entire video without making a mistake! Thanks for pointing it out. I'll add it as the first (of probably several) to the errata list in the description. Cheers!
Haha np Brian! I've got another question: Is there no-hold's barred best theoretical way to approach discrete problems, or is the implementation hopelessly tied to the method (ZOH, Pade, Tustin et al)? I know Tustin is more accurate than ZOH etc... But rather, is there an implementation agnostic way to design for discrete? I'm looking at all these transformations and don't see a catch-all approach. You got damn near sorting it all out in your 3rd video!
Reading Franklin, Powell, and Emami: "Discrete design is an exact method... that entails (a) finding a discrete model of the plant G, and (b) designing compensation directly in it's discrete form... requires a sample rate to be chosen in advance. A practical approach is to commence design using discrete equivalents, then tune the result using discrete design." Dang, sounds cumbersome!
Hi Mr. Douglas, I heard someone said "Vacum tube amplifier is introduce even harmonics". If I give signal into the amplifier with only one frequency (pure sinusoidal with only fundamental freq), will I get even harmonics if I perform FFT to my output signal? Or still just one frequency ? Thank U
I can't find the formula for transforming e^-kT into z/(z-e^-T) I know I could find it on a look up table but what is the mathematical formula to do any z-Transformation with Zero Order Hold
If you talk in spanish and are a little bit drunk, a good recomendation would be to play this video on 0.75. It works for me nicely. And its a kinda funny thing tough hehe
@@BrianBDouglas Thanks! I realized you're actually working on a book, so I'll be supporting that. Can you recommend what toolboxes to add to matlab for discrete control system design?
Why do I get 0.6321 / (z - e^-t) if I do c2d(1 / (s + 1), 1) with 1 as sample time instead of z / (z - e ^ -t)? Ah! I asked this too soon before watching the rest of the video. :-)
Hi, I am trying to implement a discrete transfer function in matlab simulink. What I am doing is, 1. Writing s domain transfer function in command window 2. Converting it into discrete using c2d command with required sampling time. Here I am getting desired bode plot of discrete transfer function. 3. Taking that z-domain transfer function coefficients and providing it in a "discrete transfer function " toolbox in simulink.but here I am not getting desired freq response . Where I am doing wrong?
So I guess when your sample rate is sufficiently high enough that the sampled input signal is almost a perfect analog signal, you don't really need to discretize it and you can just stay in de Laplace domain to design your controller... right?
Converting the impulses to steps when sampling a continuous signal with a clocked comparator shall always result in an error(signal out of range) because stepped sampler shall always have a transient delay attached to it by virtue of the component's physical properties. For eg. if the incident signal voltage is x at time 't', and it takes time 't+s' for comparator to trigger the appropriate logic, before it is ready again, then we have already missed the x values for s duration. Then you are saying we fill in the gaps with stepping. It would only add to lower efficiency with no improvement in accuracy. Having a separate isolated discreet controller to manage the 'flowing' continuous signal is better in every way... Better power, efficiency and smoothness.
In my uni, we're taught to discretise: 1. a control with tf([num],[den], ts] with num and den previously calculated by sustituting s=s(z) 2. a plant using the function c2d I've interpreted that: 1. when you discretise a controller, the sampling doesn't rly matter bc you're working in a discrete dimension 2. when you discretise a plant, the input zoh matters -> the method of discretisation matters Hence, Im confused by: How you clustered the elements in the previous video the discretisation of {the plant with the AD converter} and {the control with the DA converter}. This is pretty much blocking me. I don't understand how you decide what substitution to make in the function s(z). I don't understand how this translates to the function c2d and why it isn't used by my uni to discretise the control.
IMPUSLE=Hemrholtz PV=NRT INT(Pdv+Vdp) Billinear is simpson's rule of integration simpson INT(INVERSE) signal Matched is comparison accurate with a phiso comparator INFCOS? Linear Interpolation is front order hold is sorting Amoritized Zero order is hashing no timestamp This is what i think it is
In Spain, in my college, professors are recommending your videos to their students. Nice work
TU/e Eindhoven the same
same here at philippines
In Chile is popular too, beautiful content!
Same here in Argentina!
I'm also in Spain and thank god I found his videos. lifesaver
On behalf of all students who were forced to go online this semester, thank you for these videos!!
When i see Brian Douglas video first i hit like button, take my pen and a paper, take a coffee and then i enjoy a 1 year university class equivalent video. PS: Now audio are nice!
Much appreciated!
Jéfferson Pimenta true story
No offense, but your university must teach very little... What Brain covered in all of his videos, are about 3 month worth of our contents. Although brain explains the general concepts well, he does not cover any in depth content.
I am addicted to control system because of him
I watched this right after my exam of control theory, because I failed at the state space model and the discrete modeling... I was blown away by this explanation and finally understood what my professor can't explain... Thank you very much!
Even years after uni my understanding is better having watch your channel
12:43 THANK GOD BRIAN!!! I have being looking for this answer for the last few hours!!
Just realized you're back at making videos! Thank you very much Brian!!
It is incredible how complete, perfect and the way you explain. Greetings from Argentina. There are few teachers like this here.
Well, you're starting to go from covering one of my college courses (control theory) to two (digital control). Seriously awesome job!
Thank you!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! you are benefiting all people who don't know what discrete system is
man u r so great !!!! i 'm from brazil and here we don't have leraning like that !!! the world thanks u so much ! and another thing, thanks for u speak so clearly !!! if wasn't it, at least to me would be impossible understand !
That's probably the best video class I've ever watched
This is actually one of your best videos . You deserve way more subscribers!!! From an ECE student :D
This is literally SO helpful and I really appreciate your attitude towards spreading the knowledge!! Thank you!!
Brian! Thank you so much for your efforts in putting these videos together, they make an otherwise out of reach knowledge set available to anyone interested and that is the coolest thing anyone can do! Your an awesome person for doing this and your time and efforts are deeply appreciated! great video, cant wait for more on discrete control :) !
excellent, this video has greatly improved my understanding of discrete systems.
Amazing videos! Really gave me a much better view of what was taught at control class. Thanks a lot for sharing this amazing videos! Best teaching methods!
When you work in black screen, you must not turn to white screen suddenly. Beyond that ... GREAT WORK !!
The best explanation about this topic ever!!!!!! Very, very nice work!!!!
Yeaaahhhhh your videos are back. I am so exited to see the new series. Thank you very much for your efforts. Keep on coming.
Brian, you forgot the "s" in the expression L(V(t)) = V(s) = sum Vk[(1/s)*e^(-kT*s)-(1/s)*e^(-(k+1)T*s)] (minute 19:44) ;)
I just love your lectures. I did not know there is a way of donation for your work. I am doing it soon. I am communication engineer but I always had interest for Control and I want to be the part of it in my career later. Keep going please :)
thank you for those videos they are really helpful and I like your way of explaining
there is a small mistake in 19:46 i think you forgot to multiply the 's' to the time delay
This man right here is the G.O.A.T
very good, this classes are helping me a lot ! a hello from Brazil to you.
this is amazing,
i just found your channel and it seems you answered all my questions
thanks for this fabulous job.
GOODLUCK
Great video as usual, code breaker class video with deep insights.
Your videos are awesome! Thank you
Thank you very much Brian!
Hi Brian, I am a roboticist. Mostly I deal with problems related to modern control theory. So if you plan for more tutorials, I would like to see Modern Control theory. Thanks for all awesome lectures.
thank you sir for this kind of clear and excellent explanations, really helped it.
You are the real MVP
Thank you
Thanks for sharing those videos, las semester I got an A on both my controls class thanks for making those concepts understandable..
Hi Brian, at 19:43, should your exponents have an s?
Nooooooooo! I thought I could get through this entire video without making a mistake! Thanks for pointing it out. I'll add it as the first (of probably several) to the errata list in the description. Cheers!
Haha np Brian! I've got another question:
Is there no-hold's barred best theoretical way to approach discrete problems, or is the implementation hopelessly tied to the method (ZOH, Pade, Tustin et al)? I know Tustin is more accurate than ZOH etc... But rather, is there an implementation agnostic way to design for discrete?
I'm looking at all these transformations and don't see a catch-all approach. You got damn near sorting it all out in your 3rd video!
Reading Franklin, Powell, and Emami:
"Discrete design is an exact method... that entails (a) finding a discrete model of the plant G, and (b) designing compensation directly in it's discrete form... requires a sample rate to be chosen in advance. A practical approach is to commence design using discrete equivalents, then tune the result using discrete design."
Dang, sounds cumbersome!
Wow, this was exactly what a nedeed to move on in the problem of my homework
Can you please do a video on lypanov functions and lypanov stability !
nice video！better than my college professor！
22:15: Could someone explain how the yellow box is achieved?
you are a freaking monster! just supported you on Patreon! wish i had more money haha
Shouldn't the equation at 20:00 include a 1/Ts due to the sampler as shown in the block diagram???
Really good explanation
great explanation
Man, you're awesome. Thank you so much
Amigo muchas gracias me has ayudado mucho para mí proyecto del un microcontrolador PI :)
Thanks. Very good video
Hi Mr. Douglas, I heard someone said "Vacum tube amplifier is introduce even harmonics". If I give signal into the amplifier with only one frequency (pure sinusoidal with only fundamental freq), will I get even harmonics if I perform FFT to my output signal? Or still just one frequency ? Thank U
Great video! Thank you!
What software did you use to create that whiteboard effect in your videos?
I can't find the formula for transforming e^-kT into z/(z-e^-T)
I know I could find it on a look up table but what is the mathematical formula to do any z-Transformation with Zero Order Hold
Great Video!
Thank you so much!
If you talk in spanish and are a little bit drunk, a good recomendation would be to play this video on 0.75.
It works for me nicely. And its a kinda funny thing tough hehe
Thank you!! You are a very good teacher * ---- *
Do you recommend a textbook on this discrete control theory?
This is on DSP, but I really like this book: www.analog.com/en/content/scientist_engineers_guide/fca.html
@@BrianBDouglas Thanks! I realized you're actually working on a book, so I'll be supporting that.
Can you recommend what toolboxes to add to matlab for discrete control system design?
Why do I get 0.6321 / (z - e^-t) if I do c2d(1 / (s + 1), 1) with 1 as sample time instead of z / (z - e ^ -t)?
Ah! I asked this too soon before watching the rest of the video. :-)
Hi, I am trying to implement a discrete transfer function in matlab simulink.
What I am doing is,
1. Writing s domain transfer function in command window
2. Converting it into discrete using c2d command with required sampling time. Here I am getting desired bode plot of discrete transfer function.
3. Taking that z-domain transfer function coefficients and providing it in a "discrete transfer function " toolbox in simulink.but here I am not getting desired freq response . Where I am doing wrong?
Thanks for the video!
oh, step function , why are you so magical
Nice videos. By the way, having learn the theory behind the control aspects, transfer function etc, any videos on how can we realise the controller ?
Its done in a recent video
You rock man!
Kronecker delta is another thing. It is very different to what you showed in the video.
hey, you are awesome!
So I guess when your sample rate is sufficiently high enough that the sampled input signal is almost a perfect analog signal, you don't really need to discretize it and you can just stay in de Laplace domain to design your controller... right?
amazing!
Converting the impulses to steps when sampling a continuous signal with a clocked comparator shall always result in an error(signal out of range) because stepped sampler shall always have a transient delay attached to it by virtue of the component's physical properties. For eg. if the incident signal voltage is x at time 't', and it takes time 't+s' for comparator to trigger the appropriate logic, before it is ready again, then we have already missed the x values for s duration.
Then you are saying we fill in the gaps with stepping.
It would only add to lower efficiency with no improvement in accuracy.
Having a separate isolated discreet controller to manage the 'flowing' continuous signal is better in every way...
Better power, efficiency and smoothness.
In my uni, we're taught to discretise:
1. a control with tf([num],[den], ts] with num and den previously calculated by sustituting s=s(z)
2. a plant using the function c2d
I've interpreted that:
1. when you discretise a controller, the sampling doesn't rly matter bc you're working in a discrete dimension
2. when you discretise a plant, the input zoh matters -> the method of discretisation matters
Hence, Im confused by:
How you clustered the elements in the previous video the discretisation of {the plant with the AD converter} and {the control with the DA converter}.
This is pretty much blocking me.
I don't understand how you decide what substitution to make in the function s(z).
I don't understand how this translates to the function c2d and why it isn't used by my uni to discretise the control.
Shouldn’t the negative step function be -U and not U?. Otherwise the summation of the negative and positive step functions would not produce a pulse
Great video i pledge support on patreon
Very nice!!!!!!
Watching this in 2021
05:33 and 22.. yea 2022
can you please do tustin as the next one !
2019... Cheers
IMPUSLE=Hemrholtz PV=NRT INT(Pdv+Vdp)
Billinear is simpson's rule of integration simpson INT(INVERSE)
signal Matched is comparison accurate with a phiso comparator INFCOS?
Linear Interpolation is front order hold is sorting Amoritized
Zero order is hashing no timestamp
This is what i think it is
but we are really doing it by hand now QAQ
life safer
this is 2019 hhh
but this is aaaaa 2022
Oh come on, do something about the other subjects as well, control systems isn't the only thing in the world.
But it is a subject which lacks a lot of teaching in an intuitive manner(from text to lectures), Brian fills this gap so I would'nt complain.