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?
Since I am working on a project already in place, I'm afraid the recievers have been defined already. They are Radiometrix BIM 418-40 transceivers.
I will elaborate more........
At present we are using a 4 level ASK digital transmission being fed into a DAC for RF transmission. At the receiver end, it is being converted back, naturally using an ADC.
I am firstly looking to shape the pulse going in using a bandpass of approx 100Hz - 20kHZ. This should eliminate high frequency components since there is a high level of ringing as too much energy is trying to get through the channel. However, because of the associated phase shift in a non linear filter, I will have to compensate at the receiver. The transceivers already introduce a slight phase shift, which I am hoping to eliminate with an equalizer. The eye diagram at the receiver is poor so, especially with 4 level signalling, I hope to improve it.
Cheers
Steve