开云体育

Hi Herb
Great, I am pleased to hear the progress. Actually if you male as sweep of the test fixture without the clamp and in the NanoVNA-saver display r+jx then you have a graph for the rod impedance to print. Doing the same with the clamp fitted you can subtract the two readings directly and plot it manually.
If the NanoVNA-saver had a feature for saving data files (r + jx) then a spreadsheet quickly could draw the graph by subtracting. Else it is not difficult to do the S to Z in a spreadsheet
Kind regards
Kurt

-----Oprindelig meddelelse-----
Fra: [email protected] <[email protected]> P? vegne af hwalker
Sendt: 16. november 2019 19:12
Til: [email protected]
Emne: Re: [nanovna-users] NanoVNA-Saver 0.2.0 #Ferrite_Impedance

Kurt,
After a good nights sleep I and a second mulling over your post it became even clearer that I had deceived myself into thinking I could calibrate out the test fixture by treating it as an extension to the coaxial line and calibrating at the outside side. The reference sweep using a 50 ohm load looked convincing, but one of the gotcha's with a VNA is that if you perform a bad calibration it will look great compared against itself.

I re-cal'd this morning at the end of the connecting cable and used Rune's software to take a reference sweep of the fixture terminated with a short. I took another sweep with the ferrite under test installed. I visually subtracted the reference trace impedance values at 10 MHz, 25 MHz, 100 MHz and 250 MHz from the live trace values and compared them to the corresponding manufacturer's table (see attachment).

The comparison data was much better than the previous run. The data points at 100 MHz and 250 MHz were especially close to the manufacturer's table.

Thanks for lending your keen eye to this. As they say, "There is no substitute for experience". Now I wonder if I could get Rune to add some trace math to his program …

- Herb