does TEM mode mean uniform field distribution?

[Decesicum]

I read in some articles that TEM mode means uniform field distribution. Is this true? Because I have simulated a TEM cell (for which higher modes are theoretically excited above (about) 300 MHz) and it has only TEM mode up to about 1.8 GHz! I mean I don't see any longitudinal E-field components or they are very weak, much weaker than the transversal ones. At 2.2 GHz I don't have TEM mode anymore, only TE (I suppose that there are few higher TE modes excited and summed) But when I have only TEM the field is not uniform (its amplitude is not constant over space) Why do they say that a TEM mode aproximates a uniform plane wave?
Thanks!

 

[jiripolivka]

TEM propagation mode is used in coaxial , stripline and similar wave-guiding structures. On a homogeneous line, the field distribution is uniform. Any discontinuity changes it; certain such discontinuities are used intentionally to match device and line impedance, etc.
When the wavelength on a line becomes comparable with wave-guiding structure cross-section, other propagation modes appear and the field is no more uniform. Loss is due to that we cannot reasonably collect the energy back in all occurring modes. Often such structures are used to radiate the energy into space, to make filters, etc.

 

[Decesicum]

I understand. What I say is that the field is never uniform, nor when TE/TM dont exist. I am analyzing a TEM cell for which the field should propagate -at least theoretically- only in TEM mode at 100 MHz. Indeed I do not see any longitudinal components (in simulation) but the field is uniform only in certain regions. I attach a print screen to show what I want to say [ https://obrazki.elektroda.pl/88_1337417290.jpg] Or maybe I am not clarified with what "uniform field" means. When the frequency is increased above the cutoff frequency of TE01, the field is still TEM (up to quite a high frequency) I only suppose that the higher modes are too weakly excited to have a visible effect, although theoretically I should find also longitudinal components. This is probably also because I didn't yet introduce anything in the cell to distrupt more the field distribution (more then the tapers do, probably). Please correct me if I am wrong.

 

[WimRFP]

Hello,

TEM does not automatically mean uniform. A coaxial cable has TEM conditions also, but the field isn't uniform.

A spherically propagating wave can have TEM condition (as E and H are perpendicular to the propagation direction). However with increasing distance, the power/m^2 drops, hence E en H. From a practical standpoint, a spherical wave can be considered "plane" in a certain (small) volume (think of a far field antenna test range).

If you have a TEM wave with perfect plane wave front, then E and H are exactly uniform in the complete volume. Some curvature in the wave front automatically means that you have (some) longitudinal component and then then E and H will no longer be uniform.

 

[jiripolivka]

I understand. What I say is that the field is never uniform, nor when TE/TM dont exist. I am analyzing a TEM cell for which the field should propagate -at least theoretically- only in TEM mode at 100 MHz. Indeed I do not see any longitudinal components (in simulation) but the field is uniform only in certain regions. I attach a print screen to show what I want to say [ https://obrazki.elektroda.pl/88_1337417290.jpg] Or maybe I am not clarified with what "uniform field" means. When the frequency is increased above the cutoff frequency of TE01, the field is still TEM (up to quite a high frequency) I only suppose that the higher modes are too weakly excited to have a visible effect, although theoretically I should find also longitudinal components. This is probably also because I didn't yet introduce anything in the cell to distrupt more the field distribution (more then the tapers do, probably). Please correct me if I am wrong.

The figure you show clearly demonstrates the problem with inhomogeneities what the tapers are.
Within the conical taper the field does not conform to parallel lines; where the other taper is, it causes a reflection back and the effect are the yellow arrows disrupting the field.
A perfect TEM line transports a completely uniform field just as the math shows. Any change in geometry or impedance disrupts it. Where the wavelength approaches the cross-section dimension, other modes occur, disrupting the field uniformity.

 

[Decesicum]

Thanks for all the answers! So it is quasi-TEM mode. But anyway, in the rectangular section (central guide) I still can assume the field is uniform, at least in a certain volume, can't I? I think about exporting the 3D computed values and to extract the volume in space where E-field varies with no more than few percent. Do you know how small the field variations are allowed to be (in % or dB) in order to still consider the field uniform?
And one more thing: higher modes are excited by changes in geometry (tapers in this case) just because the change in geometry means changing in impedance? And impedance changes causes reflection, which in turn disrupt the field distribution. Is is so?

 

[jiripolivka]

Thanks for all the answers! So it is quasi-TEM mode. But anyway, in the rectangular section (central guide) I still can assume the field is uniform, at least in a certain volume, can't I? I think about exporting the 3D computed values and to extract the volume in space where E-field varies with no more than few percent. Do you know how small the field variations are allowed to be (in % or dB) in order to still consider the field uniform?
And one more thing: higher modes are excited by changes in geometry (tapers in this case) just because the change in geometry means changing in impedance? And impedance changes causes reflection, which in turn disrupt the field distribution. Is is so?

I have not done any deep research in this problem. The start or end of the taper is apparently not smooth enough; it may not change the impedance but as I see it, the change in inner/outer diameter, while their ratio is not changed, tilts the "reference plane". Real conical transitions always are optimized for VSWR limit, and it is never exactly 1:1.

As the TEM device you study will serve as a TEM launcher or a test chamber, there will be objects in it that will disturb the "clean" field. The basic VSWR without any test object will also depend upon the termination connected to the "output" taper which is also not perfect. Having the VSWR optimized determines the best achievable field "cleanliness".