Hi Brooke, acc. my understanding, the characteristic impedance of a transmission line (ideally losseless) is constant and waveform independant, as given by the relation of inductive and capacitive values (muh and epsilon) equally distributed over the line.
As long as the physical parameters of the propagation path do not vary with the frequency (due to special material behaviour or too big cable size), there cannot be a dependency on frequency of the impedance Z. Therefore the impedance Z of a transmission line can as well not be dependent on the waveform, as all not sinosoidal waveforms are just a sum of a spectrum of sinosoidal waves (spectrum of frequencies). An even very short pulse is as well just a sum of sine waves with few periods over a wide band (the reason why I think there are no real digital rf signals, just the modulation information may be called digital). The safe way to transfer a proper waveform is to apply properly matched transmission lines conforming its characteristic line impedance Z. That I cannot measure the correct line impedance under some circumstances, is another issue. If, in relation to the wavelength, the length of line is going to be short (as on low frequencies), then I am running into problems, similar as it is the case in freespace measuring fields under near field conditions. Already on lines length below a quarter of wavelength it is hard to get acceptable measurement values, and at a 100th of a wavelength or less I can practically ignore the effects of reflections, output = input. Short lines behave capacitive and at short distances the voltage is dominant in measurements! At audio frequencies I cannot talk anymore of reflections. Therefore, at very low frequencies it is not anymore of importance to use matched transmission lines for dist. of a few meters, but the lines still do have their characteristic impedance! Summarizing my opinion, a 50 Ohm transmission-line does have a Z of 50 Ohms on a l l frequencies, the charact. impedance remain c o n s t a n t (unless the design or the material fails physically, which occur mainly in the microwave region). Even if it is not of importance to use transmission lines of a specific value, it does not harm either but it may be still important to use a matched load. For signals of low frequencies as the pps with short risetimes, perhaps it would be better in effect to apply lines of hight impedance (eg. 75 --> 95 or 100 Ohm) , because the lower capacitive effects and dielectric losses, but match the end with the adequate load. There are of course other solutions, eg. when working with cmos: 'open' end at rx for a double signal level, then eliminating the reflections at the matched tx output, but that is another point and I think already discussed on this place. kind regards, Arnold On Sun, 28 Jan 2007 13:02:39 -0800, Brooke Clarke wrote: >HI Didier: >It's my understanding that the term impedance can only be applied when >sine wave signals are being used. So for pulse work you might look at >the harmonic content and try to match all those frequencies. >Long ago Bob Grove promoted the idea of using 75 Ohm TV coax for ham >antennas at 2 meters and higher frequencies because it had lower loss >than 50 Ohm coax and was much lower in cost. For ham applications the >VSWR due to the coax impedance was much smaller than the inherent match >(mismatch) of the things on either end. >It's only been in the last few years that I understood that the >impedance of a transmission line is only a constant value above some >frequency and below that is no longer a constant. So, for example, >audio signals can not be transmitted using "transmission lines" of >constant impedance. For more see: >http://www.pacificsites.com/~brooke/Zo.shtml >Have Fun, >Brooke Clarke _______________________________________________ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
