Jens,
1) there is the odd Doll on this group as well as [far too many, /pro rata/] Guys
2) for the technical question, imagine that there are *2* oscillators active, one at f?, one at f?+IF, where IF is an intermediate frequency and f? is the measurement frequency. Then, as square waves, the measurement oscillator will produce frequencies f?, 3f?¡, and the local oscillator will produce frequencies f?+IF, 3f?+3IF¡ Then at the fundamental frequency, the difference frequency is IF, but when the third harmonic products mix, the difference frequency is 3IF. The products at 3IF will be rejected by the post mixer bandpass filter.
3) similar arguments apply when harmonic mixing is employed, such as for the 300?MHz - 900?MHz range. Note in this case, that there will be an unwanted product from the fundamental frequencies at ?IF (hence the use of a bandpass, rather than lowpass, filter) and the 5^th harmonic at 1?IF.
Also note the reducing power in the harmonics, thus doubly reducing the mixer products to be measured. The fact that any of this works is a triumph of human ingenuity and DSP.
4) Identifying all the frequencies and corresponding mixing products for n^th harmonic operation is left as an exercise for the interested reader!
HTH, 73, Stay Safe,
Robin, G8DQX
PS: For this sort of hybrid analogue & digital machine, often the block diagram is more informative than the circuit (schematic) diagram. The actions of the software are also key components.
toggle quoted message
Show quoted text
On 20/12/2021 11:11, Jens M?ller wrote:
Hello Guys,
I've one question which puzzles me, and was unable to find information neihter in the forum nor in the Wiki.
Since the NanoVNA outputs a rectagular signal which's harmonics are not neglectable and has no selective input (what I can see from the schematic is broadband), how can the S12 and S21 measurements be accurate as all the harmonics of the output will go into the mesurement at the input.
What do I have overseen?
Thanks for bringing me on the right track
Jens DF9HJ
