Channel Phase Shift Measurements and Phase Equalizing Fi

[stevieboy01]

I am attempting to characterise a phase shift in a channel so that I can compensate it at the receiver end using a phase equalizer. I need a method which produces a phase shift curve with more measurements than a Lissajous method produces. I was thinking possibly a phase shift oscillator may be the way to go with each stage producing a 45 degree shift, but I am unsure how to use this. I want to define the frequency and calculate the resulting shift. Also, I am not sure how to use an equalizing filter. I would probably use an all pass filter but how can I define where the curve goes? <p>Please help. I am mightily perplexed!!!<p>Steve

[Ron H]

To do the analysis, you need a vector network analyzer (VNA). This applies a swept frequency to the input and plots amplitude and phase on the output.
Phase equalization using allpass networks, in my experience anyway, is part science and part art. It would help if you could simulate the equalizer, using the output data from the VNA as the stimulus for the allpass network.
If you can digitize the signal, I think there are equalization alternatives that have no analog counterpart.

[jwax]

Can't help you, but I love that term, "AllPass Filter"! Sounds useful.

[Bernius1]

Color TV uses the color burst to sync the local oscillator in phase, against which the remaining line is referenced. If your transmission is packeted data, maybe you can pulse a reference signal between the packets at a different frequency, to a PLL in the recvr.

[Dean Huster]

I might point out, Stephen, that the traditional RC phase-shift oscillator shifts 180/n for each RC section where a 4-section oscillator will have APPROXIMATELY 45° shift for each section. The actual shift for each section will be strictly dependent upon the R and C values, their tolerances and their tempcos. To me, that eliminates the RC phase-shift oscillator as any serious contender as some kind of standard for a 45° phase shift unless you used very carefully-matched components with a small tempco. Even then, your final frequency may then be an issue if you have to have both the four shifts and the final frequency to be accurate. Quite a balancing act!<p>Dean