how is it possible to implement higher order modes in HFSS?
It is fed by a 4 slotted waveguide.
The dimensions of the cavity are 64cm x 53cm x 30 cm
as a result the resonance frequency is 2.45 GHz.
Unfortunately the dominant mode is TE645 and I can not simulate it in HFSS.
Does anyone know how can I excite the waveguide by TE645?
Regards
Hi,
can you also post the cavity geometry in SAT format. Maybe I can have a look in CST...
F
I wish ACIS SAT 8.0 is compatible with your system.
Thx anyway waiting 4 your comments
regards
Hi
I had a look at your structure and there seems to be some misunderstanding.
The waveguide feed (dimension 8.6*4.3.cm) only supports the TE10 mode at 2.45 GHz. All other higher order modes are already bellow cut-off. You therefore want to excite the waveguide just with the fundamental mode.
The cavity is much larger and will support many modes. I guess this is what you mean with TE645 mode? If the TE645 mode is excite or not by the slots in the waveguide depend on the geometry and NOT on the excitation of the waveguide feed.
All this is valid if you use the "driven" solution.
They are other points which have to be considered if you would like to use the eigenmode solver. The cavity of this size supports already many modes. In your structure you have already altered the geometry from a pur box (which could be solved analytically) by slots in the Waveguide and some material which will be heated. Therefore the "pur analytical" mode TE645 does not exist anymore.
Because of the size of the structure, the eigenfrequencies and modes are very close together as well. Slight changes in geometry/materials will change the mode pattern inside the cavity. The same is true for numerical errors. If your mode calculation is just wrong by 1% you might already "hit" another eigenmodes. (I guess this exactly how a mode stirrer works).
Attached please find some quick results including field pattern and 2.45 GHz and 2.44 GHz as well as the S-Parameter. You see many resonances in the S-Parameter. I obtained the results with CST MWS TD solver in just a few minutes. I used a rather coarse mesh for this just to demonstrate the principles. The cavity is loaded with some lossy material.
F.
hi,
thx a lot for your time to analyze my case.
You have explained the case very simple and deeply.
You are right in WG only TE10 and in Cavity many modes exist.
What I am trying to find is the best mode in cavity.(TEmnp)
m@Length
n@Width
p@height
Since I have 4 slots in longitudal (z) direction, m must be at least 5 (which means 4 peaks for cosine), n should be an odd integer in order to feed the system by one wg, p should be as small as possible.
Depending on p, i will decide where to put the load for max absorption.
This job is like a puzzle and i like to try it.
By trying to find the best combination for mnp, i am trying to transmit max power from source to load.
I have been working with it for two days and about 2 find a solution till weekend.
Do you have any idea about above mentioned optimization?
Is there a way that i can do it more quicker?
ps:just want 2 let you know for a useful site that i encountered:
http://individual.utoronto.ca/yaxun/index.htm
Thx again 4 your help
Hi
nice link!
I have to admit that I'm not a MW oven expert But I'm afraid you are right. This is a very tough puzzle and I'm not sure if this can be solved.
I believe they are just to much unknowns and uncertainties in your model. They re to many modes with different field patterns to close together.
Even if you could solve this model numerical with any field simulator within a accuracy of much less then 1% it would still be doubtful if you can determine all parameters in the real life (material parameters, geometry e.g.) with an equal accuracy. Furthermore, the material parameters and dimensions might change with time during the heating process. I would believe that this easily could lead to mode jumping..
Again, I'm not an expert in this field. Would be nice to know if anyone ever measured the "time constant" of a MW oven. How long does the mode and S-Parameter remain unchanged if the oven is not touched from the outside and they are no actively moving parts inside. Comments appreciated!
F.
Dear RFSimulator,
It is nice to get your ideas.
I agree with you that less than 1% may lead to mode change.
Anyway there is no change for me then believing that HFSS works perfectly.
Let me ask you two questions:
1-)
I want to see the modes on inside the cavity to be sure that they exist.
For example TE432.
ıt means that I sould see 4 mode in the x direction, 3 in y,2 in z right?
As soon as I animate the E filed on the surface of the cavity, there appears peaks with red colors.They show the highest intensity of E.
Can I accept the number of these red colors as modes?
For exaple in your simulation (cavity_03.gif) there seems 4 peaks in z direction.
Is that mean that it is something like TEmn4?
2-)
regarding the link h**p://individual.utoronto.ca/yaxun/calc/rect_res.htm
can u tell me the directions of the waveguide.
It seems that Lenght is in x direction, but I must be sure about it.
This makes me ver doubtful because all the books puts length of the waveguie in -z direction.
This is important for me because of forming E(y) inside the cavity.
I must be sure Is E(y) proportional with six(x)cos(y)siz(z) or cos(x)sin(y)sin(z) or something else?
After being sure about the formula of E(y), I will decide where to put the load for max absorption and be sure if there are 4 peaks in the length direction.
You know number of peaks of sin and cos are different in a same lenth.
Regards
Yavuz
Hi,
the field plots I generated are simuilated with the transinet solver and results of a excitation at 2.44 and 2.45 GHz respectively. Because of the excitation and internal inside the cavity are a linear combination of several modes. Certainly on of the modes might be pretty dominant and if you like you can use your classification.
F.
It's just a simple rectangular cavity. You can calculate the resonances from the theory easily.
Right, you have to see the appropriate field maximums.
No. Not exactly. To see the pure mode you have to use Eigenmode solver.
In a case of waveguide excitation each mode has a loaded quality factor.
So, if the modes frqs are very near you will excite many of them simultaneously.
The picture you see is a sum of few modes depending of their Q-factors and coupling coefficients.
You can't form the particular mode inside the cavity. It's a properties of cavity itself and doesn't matter of connected waveguide.
Again the actual field in the cavity is a sum of all exited modes.
In fact waveguide excites indefinite series of cavity modes because of loading !
Of course you can use approximation and take into account only few of them.
dear navuho,
thank you very much for your comments
regards