What a Graceful way to manipulate Maxwell for CEM love it !!

This was one of the first lectures I recorded and it was done live. I soon stopped doing this and recorded all the lectures on their own. Sorry!

@veronicanoordzee6440

5 жыл бұрын

Better quality.

Thank you for posting series of high-quality lectures! At 24:58, the model with complex dielectric constant and conductivity is probably because the imaginary part of the dielectric constant models the dielectric loss, and the author wants to distinguish the dielectric loss and conduction loss. In your discussion, you are assuming dielectric loss is 0 (which is totally fine in most cases depending on what you are doing). For more on dielectric loss, you can refer to Jin's book "Theory and computation of electromagnetic fields" in section 1.7.4.

First of all, thanks for the nice lectures! Regarding Maxwell and Maxwell's equations 2:35 : I don't think the way it is put here does Maxwell justice. Firstly, to invent an equation is exactly that, writing it down. After they had been written down, sure, many great scholars have contributed in deepening the understanding of them and improving upon the formalism. But we shall not underestimate the creative genius of Maxwell to have put all electromagnetic phenomena onto a common theoretical footing. As far as I know it involved him postulating the existence of what we now call "Maxwell's displacement current", a term he added to the equations, such that the equations describing electric and magnetic phenomena known at the time made more sense in the context of the (at that point still undiscovered!) electromagnetic waves. Of course it is easy to speak from a high horse as a modern day scientist or engineer about the people back then who knew so little compared to us. But let's not forget that we are only where we are today BECAUSE OF THEM! Standing on the shoulders of giants. And Maxwell surely was one of them!

@empossible1577

2 жыл бұрын

Great points!

Bless your soul for going over sign conventions and the consequences for refractive index and dielectric constant.

@empossible1577

4 жыл бұрын

Yeah, that is crazy to figure out as a student. The big secret is even the more experienced people struggle with it!

Thank you for sharing this lecture.

Thank you

Thank you for the wonderful video. I have a question. At (40:55), and also in many textbooks it is said that the second set of equations are Fourier transform (time) of the first set. I understand the "Fourier-transformed" set of curl equations can be obtained by assuming complex sinusoidal variation in the fields. However, why is the operation called Fourier-transform? The Fourier-transform operator has an integral also associated with it from -infinity to +infinity.

@ManderSeis

2 жыл бұрын

Both sides of the equation are from -inf to inf, the integrand also has to be equal, so it can be left out on both sides.

wish we had video on the blackboard(i assure the lecturer was drawing somewhere)

Thank you very much. I need the Matlab program for this lecture. Please help me.

@empossible1577

6 ай бұрын

There is no MATLAB code for this lecture. What are you looking for specifically? What do you want to calculate or visualize?

I have a question regarding the wave equation. I understand that the most general form are the ones with the mu and eps inside the rotational operator. However, I have found the optics litterature that they call "inhomogeneous wave equation" the equation "(Laplacian + k²(r)) . E " and "homogeneous wave equation" the equation (Laplacian + k²) . E. Is this valid because in Optics, the term mu is supposed not to be depending on the position? In this set up, we end up on the first equation, and if we also suppose that eps doesn't depend on r, we end up on the latter.

@empossible1577

9 жыл бұрын

***** You are right. At optical frequencies, the mechanisms leading to permeability are too slow to manifest themselves significantly so mu can be ignored in almost all situations. This lets you derive a wave equation in terms of the electric field in the form you described and it is still valid inside inhomogeneous media. You can't however, do the same deriving a wave equation for the magnetic field because the permittivity is still significant and inhomogeneous.

This is a great lecture. I may try to take one of your courses. So if permittivity in the time domain is always real, how do you model loss and dispersion? I can see dispersion being an impulse response. Does the dielectric loss just become a finite conductivity, like you showed in the frequency domain. I use CST transient a lot for antenna design.

@empossible1577

Жыл бұрын

Thank you! This is actually quite an old video now. I recommend using the course website as your main portal to the content because you will always get the latest version of the notes, videos, and other learning resources. empossible.net/emp5337/ For time-domain analysis, loss is incorporated through the conductivity term. In the time-domain, the constitutive relations are actually convolutions. D = eps * E and B = mu * H where * is convolution. This is how it is treated and the match can get a bit ugly as you can imagine. I have some brief notes about doing this for finite-difference time-domain analysis in Lecture 2f at the following link. It is in the section called "Frequency dependent materials." empossible.net/academics/emp5304/ The main complication with dispersion is that dispersion is a frequency-domain concept so it is ugly in the time-domain. Hope this helps!!

@rfengr00

Жыл бұрын

@@empossible1577 yes that helps. I believe the software I use, uses the Debye (sp) dispersion model. I’ll have to read up on that. Can any of the time domain methods be used to solve inside the conductor? I’m interested in the low frequency (kHz) region where inductance transitions from internal to mainly external , and the RF currents flow deep, as this has interesting dispersion effects on Z0 for TEM lines, as the R and L are frequency dependent.

@empossible1577

Жыл бұрын

​@@rfengr00 In general, any method can be made to do anything. Also, since we live in a time-domain world, there is nothing a time-domain simulation cannot do. I don't know all the details about what you want to simulation, but I think both time-domain and frequency-domain will work for you.

What is the best CEM method for RF cavity resonators?

@empossible1577

Жыл бұрын

First, don't be too paranoid about finding the "best" method. Many different methods will get you an accurate answer. The "best" method would depend on the geometry, the material properties, and what it is you want to learn about the resonator. For example, if it is just air and metal, the method of moments may be the "best," but depending on geometry a Fourier modal method could also work. Without knowing anything, I would probably have to say that the finite-element method has the best chance of being your best method. However, FEM will be much more difficult to learn, formulate and implement than something like finite-difference time-domain. Having given a lot of thought to the "best" method, I have come up with what I think is the best first method to learn to get started in computational electromagnetics. I think that method is finite-difference frequency-domain, with a close second for finite-difference time-domain. Here are two learning resources for the complete beginner to get started with these methods: Finite-Difference Frequency-Domain empossible.net/fdfdbook/ Finite-Difference Time-Domain empossible.thinkific.com/collections/FDTD-in-MATLAB Here is a short video on the FDTD video course: kzread.info/dash/bejne/p3ad0tSCfpvNlrw.html Hope this helps!!

Dear Prof. Raymond first of all I would like thank you for the wonderful work that you have done, I find what I realy need for. So, in my project, I have to analyse a Substrate Integrated Waveguide (SIW) using FDFD, it's a periodic structure formed by a substrated metalised in the both sides, two rows of vias are add in direction of propagation, I have used formulas from you lectures (lecture 10, 12, 13, 14 and 15 ) to do the job unfortunetly results are very bad, matlab codes work for the rectangular waveguide but not for the SIW, for this reason I have some questions : 1)- should I use 3D-FDFD or 2D-FDFD eigen-value to analyse my structure ? ( I know it's a stupide question but I need your advice 😊😊😊) 2)- in the slide 15 , FDFD Extrat, in case of 3D-fdfd how can I calculate propagation constante? you had mentioned that k0 is unknowen and is the eigen value but is given at first, so what I have to do ? 3)- until now building a geometry in matlab is a nightmare for me 😄😄😄, I'm begineer abviosly, can you give me a good refrence where I can prove my self with geometries ? 4)- do you mind if I send you my matlab codes just to make a look up ? thank you very mutch

@empossible1577

6 жыл бұрын

I think it is always good practice to reduce the dimensionality of your problem if at all possible. The simulations run much faster and you can iterate through more design options. Once you have done this, you perform a slow 3D simulation based on your 2D results to confirm your design or perhaps make some final tweaks. I think you can reduce your problem using the effective index method. To do this, analyze the metal-eps-metal stack as a slab waveguide and calculate the effective refractive index. Given this, you can make a 2D simulation like it was a topview of your problem. Fill the grid with the effective index you just calculated for the dielectric regions and fill in small metal circles for the vias. You should be able to form your waveguides and circuits and produce your designs. As a final confirmation, consider running a 3D simulation in a commercial FDTD software package like Lumerical or CST Microwave Studio. I think you will be pleasantly surprised how close your 2D simulations come to your full 3D simulations with orders of magnitude more speed. As for practice building structures, see Topic 2 here: emlab.utep.edu/ee4386_5301_CompMethEE.htm There is an entire lecture dedicated to building geometries into data arrays, which is one of the problems you are facing. Hope this helps!!

What prerequisits are necessary for this course? ...and do you know of any online course which teaches them ? Thank you.

@empossible1577

7 жыл бұрын

I do not know of any online courses for all of these methods. Sorry! For prereqs, you must have a background in electromagnetics, programming, and linear algebra.

@navjotsingh2251

4 жыл бұрын

CEM Lectures what if you know programming, linear algebra and vector calculus but you have never done electromagnetism before? Could I still follow along with little problems?

Hey, nice course! Just a question : why do you say Maxwell's equations don't describe how waves interact with materials? In Gauss's Law, there is the charge term, which is related to the material, am I wrong? And in Ampere's Law, there is J which is also a function of the material.

@empossible1577

9 жыл бұрын

Great question. First, I would argue that rho and J are not material properties, but instead just tell us about charges and movement of charges in materials. Second, in high frequency simulations these terms are usually dropped from Maxwell's equations anyway. rho mostly characterizes static charges that don't ordinarily interact with waves. J is incorporated into epsilon to give us a complex permittivity. The parameters characterizing the materials, eps and mu do not appear in the "pure" form of Maxwell's equations. Instead, they appear in the constitutive relations. In this sense, Maxwell's equations (curl equations and divergence equations) do not directly tell us how the fields interact with materials because mu and eps do not appear in them. Instead, that information comes from the constitutive relations. It really is the combination of all the equations that tells us the whole story. Sorry if my words are confusing. I think I just like saying that Maxwell's equations do not tell us how fields interact with materials because it is shocking and forces us to look at the equations differently.

@TheToadAlly

9 жыл бұрын

CEM Lectures Ok. Also I have seen in the litterature that these equations are sometimes called material equations, which is pretty indicative.

@empossible1577

9 жыл бұрын

***** I have never heard them called that, but I like them described that way. BTW, if you want the longer story behind these equations listen to Lecture 2 in the EM21 course.

@ 5:35 How can you have E-field lines if there's no charge around? What's the source of the field then?

@empossible1577

3 жыл бұрын

First, there are no such thing as electric or magnetic field lines. The fields are a smooth and continuous phenomenon sort of like a cloud or fog. There is also a direction associated with the field. Mathematically, you can trace your finger following this direction and draw lines, but the lines are not a physical thing. You can have an electric field without charge in an electromagnetic wave. These propagate through the empty vacuum of space so there exists both magnetic and electric fields without charge. However, it always takes an accelerating charge to generate the wave at the beginning.

@jacobvandijk6525

3 жыл бұрын

@@empossible1577 @ 5:35 you're not talking about EM-fields, but E-fields. They need a source! And EM-fields need a source too. The source is an electrically charged particle that is being accelerated (like in an antenna). P.S. Math is useless if you don't understand the physics ;-)

Hi, Is there any course of Matlab for CEM ? Or which tool you are going to use for computation ? any suggestion Please

@empossible1577

7 жыл бұрын

Every course on this channel uses MATLAB. If you are interested in MATLAB for CEM, you have come to the right place!

@abrarahmed2621

7 жыл бұрын

Yet I am not expert in Matlab wants to learn matlab first and implement Maxwell equations.Is there any course on "Fundamentals of Electromagnetism with MATLAB" ?

@empossible1577

7 жыл бұрын

Not that I know of. There are a lot of good tutorials on MATLAB. Once you have worked through some of those, work through Lecture 2 here: emlab.utep.edu/ee5390fdtd.htm This lecture is not intended to teach MATLAB. Instead, it summarizes most of the skills you will need for CEM. Some other skills are in Lecture 3. When designing the course series, I actually intended the FDTD to be the first, followed by CEM, and then followed by EM21. You may want to start with the FDTD course and work through at least Lecture 12 and the six MATLAB sessions. Then move on to the CEM course. I also teach a basic Computational Methods course, that may have some useful things. Here is the website for that: emlab.utep.edu/ee4386_5301_CompMethEE.htm Hope this helps!

@abrarahmed2621

7 жыл бұрын

Jazak Allah khair sir :)

a little hard...

@empossible1577

7 жыл бұрын

Sorry. This lecture is not intended to teach Maxwell's equations from scratch. Instead, it is meant as a review and to put them in the context for the course. I am actually working on a lecture series to discuss Maxwell's equations for the novice. I am not sure when I will finish that.

Please go back and read the history a bit and you will then give Maxwell his proper credit. Maybe start with Classical Electrodynamics, 3rd edition, J.D. Jackson, Page 238.

@empossible1577

5 жыл бұрын

This lecture is not intended to teach Maxwell's equations from the beginning. It is only intended to review them and illustrate them to better understand some concepts that arise later in the numerical methods. If you want to see a better treatment of the governing equations of classical electromagnetics, watch the videos for Topic 3 here: emlab.utep.edu/ee3321emf.htm