Hi

For a plane wave propagating in a source free, lossless dielectric medium, the wave impedance formula sqr[mu/eps] describes the ratio of E to H fields.

In an ideal transmission line (TEM mode assumed here), the E and H fields propagate perpendicular to each other, and so the WAVE IMPEDANCE (The ratio E/H at any given point in space within the TL), is also defined by sqr[mu/eps].

In TL theory we work with "voltage and current", defining their ratio as the "Characteristic impedance of the transmission line": Zc=V/I.

But "V" and "I" are DERIVED QUANTITIES, derived from the fundamental E and H fields by means of line integrals of those fields in the transversal section of the TL, plus boundary conditions, like Vo and Io as applied currents or voltages. (Check the excellent paper by Dr. Williams or other references like Ramo, Pozar, Harrington etc.)

Thus "V" and "I" depend on the geometry and materials conforming the transmission line, and their ratio may or may not equal the quotient E/H. Thus the TEM wave impedance is equivalente to the plane wave impedance for a similar medium, but wave impedance and TL's characteristic impedance can differ.

And this is only for the TEM mode... If you start to consider TE, TM or hybrid modes, more and more "impedances" show up. And there are many situations (e.g.:In a hollow rectangular waveguide) where the derivation of V and I themselves as integrals of E and H fields is not straitforward, or even unique.

Regarding the microstrip, the fundamental mode is not a pure TEM but a quasy-TEM mode: Permitivities differ in substrate and air, and thus there are longitudinal E and H fields to maintain continuity. Nonetheless, for lower frequencies this longitudinal components can be neglected and the static solution is taken as the spatial distribution of E and H. You can compute from this field distribution the C and L values of your transmission line.(e.g.: Pozar, 3.8)

Hope this helps.

Iban

Hi Nguyen

Sorry for my late reply, I've been doing some work on this issue and I forgot about MW101 for a while...

The problem is the resonance between the wirebond's L and the parasitic C of the PCB area where it lands. This resonance lies in the 7-15GHz range, as you know.

I've tried to take it out of my BW (1-40GHz) by using very short wirebonds and thin, high permittivity substrates with wide PCB areas and SMD capacitors (fres>40GHz), but no way.

In fact I have kind of a "solution" (which I don't like much but it is my best bet up to now...): I'm using a single 4mm - 5mm long wirebond (1mil diameter, gold) on either side of my DA's gain stage drain line.
What I've managed to do is to make the resonance go below 1GHz by using theses long wirebonds and Dicap Capacitors (100pF, 12mil wide).

With 4mm, the gain is still affected by some 2dBs at 1GHz (6mm would be nice but impractical). My main concern with this solution is the maximum current my wirebond can withstand... I need about 350mA for the gain stage drain line. I've been reading some papers and documents on the subject and I think that 300mA (200mA more realistic) is the highest current I can achieve for this wirebond legth and diameter... And more wirebonds reduce inductance, increasing fres... Anybody, any help or comments on this?

I've analyzed Coilcraft's conical inductors also, but the PCB pattern they require for soldering makes the resonance show up in my simulations (even halving the area of the smallest one)...

At present I'm also considering Vishay's RFLW 3 and 5 series of wirebondable spiral inductors. Thier SRF falls below 6GHz but if I manage to concatenate a couple of them with shorter wirebonds this could make my day... I'll try to get samples or buy a few (but there seems to be no stock for my country, Spain). The problem with this is delivery time, as I need to carry out measurements right now...

This DC supply issue in wideband amplifiers is a topic I haven't seen covered much in papers out there. We'll publish a simple paper at our URSI meeting this september, but only scratching the surface of the topic... I'd thank if somebody could forward me titles of publications concerning this issue.

All help is welcome. I'll post my advances and measurement results ASAP.

Thanks for reading.

Iban

Hi Steve

Many thanks!!

iban

Hi
Agreeing with the previous answer and from what I've learned, I would say wavelength in a waveguide is modified with regards to the wavelength in an unbounded medium by the m and n indexes of the mode that you consider, as the propagating wave is not a perfect plane wave unless you go very high in frequency...

You can also check your Pozar(MWEng, 4ed. p112), Balanis(Adv.Eng.Electromags,1ed. p354) or Ramo (Fields&Waves in CommElectr, 3ed. p418) or go to the classics: Marcuvitz (WaveguideHandbook, Ch1)

Hope this helps.

Regards

Iban