One thing I find comforting. I have used the Jim Brown's K9YC tutorial on
RFI and measuring impedance of ferrite devices and comparing it with the
testing I do with the VNA. Interesting comparison. I don't use it for
analyzing ferrite material for mix, just isolation. Interesting as to the
data results over the frequency spectrum.
Mel
toggle quoted message
Show quoted text
On Sun, Nov 3, 2019 at 8:31 AM UlrichKraft <kraftulrich1@...> wrote:
Kurt,
Thanks for your detailed explanation.
I will have to read a few times to fully understand the background. But
what you said makes sense and I will try to change my calibration process
according to your recommendation.
That VNA stuff appears to be more critical to what I was aware of.
I hope my rusty knowledge from university a couple of decades ago will
help.....
Thanks again !
?TypeApp for Android herunterladen ?
Am 3. Nov. 2019, 16:51, um 16:51, Kurt Poulsen <kurt@...> schrieb:
Hi Ulrich
That is probably easy to explain. When doing the Ch0 S11 measurement I
anticipate you are using a female bulkhead/PCB SMA adaptor to which you
solder the coil. Then there are two likely errors. First the
calibration might be wrong is you calibrate at the end of the test
cable using the female female adaptor and the SMA male kit as assumed
ideal. If that is the case and removing the female female adaptor for
being able to fit the SMA female adaptor with coil your calibration
plane is "out in the blue air" as you calibrated at the end of the
female female adaptor no longer fitted. If the female adaptor and the
SMA female adaptor with the coil has same delay it is just fine but no
likely and probably shorter for the coil adaptor so you have a
residual capacity shunt to the coil lowering the resonance frequency.
If you however used my data for the female calibration kit published
when calibration at the end of test cable then the calibration plane
would be correct at the SMA male adaptor but now the full capacitance
of the coil SMA adaptor will lower the true resonance frequency
probably even more.
The NanoVNA can pull the calibration plane backward by setting a
positive delay and if calibrated with the male cal kti at the end of
the female female adaptor and the coil adaptor has a shorter delay than
the female female adaptor, then you can tune the delay with the coil
adaptor without coil fitted to a phase trace showing 0 degree phase for
Ch0. The mount the coil and measure S11 reflection.
All in all measuring coil with S11 reflection is very difficult and the
only way is to create 3 calibration standards using identical SMA
adaptors as the one used for soldering the coil to. These three
calibration standard having for the short a 2 mm thick rigid shorting
disk, for the open the center conductor grinded down to Teflon surface
and the for the load 2x100ohm SMD resistor with the center conductor
reduced to 0.mm length. Even then the solder used for the SMD resistor
represent a fringe C of some 50 to 100fF incl. the fringe C for the
center conductor and the open has a small fringe C of maybe 25-50 fF.
It all has effect on the resonance frequency.
Well the S21 measurements is far better but the condition for a correct
measurement is that the source impedance for Ch0 and the load impedance
for Ch1 must be 50 ohm clean, neither the case but still closer to
reality. You have still some impact from the two SMA adaptor to which
you fit the coil which should be modelled away. A trick is to mount a
small 2 way pin socket to the two center conductor of the SMA adaptor
and create 3 calibration pin pin adaptor with two 100 ohm SMD resistor
for load, for open a pin pin adaptor in open condition and for short a
shorted pin pin adaptor. The do a standard Refelction calibration in
"S21" serial condition and solder the coil to a fourth pin pin adaptor.
The you are done and has compensated all the fringe elements. I am
pretty sure the NAnoVNA will work in the special reflection calibration
mode I "invented" several years ago. During calibration not calibrate
isolation but thru is OK (as far as I remember) I have not tested this
with the NanoVNA.
Else you can converter the S21 measurement saved as a s1p file to S11
with the formula (3*S21-2)/(S21-2) in a spreadsheet. Remember s1p
files are pure txt file and show the S parameter data as real and imag
numbers and when the math result is present in the spreadsheet save
these two columns as a text file and rename to s1p extension. You must
create the header line manually in a text editor
That was a long story to a short question ?
Kind regards
Kurt
-----Oprindelig meddelelse-----
Fra: [email protected] <[email protected]> P? vegne af
UlrichKraft
Sendt: 3. november 2019 15:15
Til: [email protected]
Emne: Re: [nanovna-users] Measuring ferrite beads #test-jig
Hello Kurt, and others,
I still haven't build a robust fixture, but experimetally taking some
measurements on EMC bead, or right now on a wire-wound inductor with a
3mm ferrite core for EMC application.
I was assuming that the results of measuring Z (S11 channel 0) will
closely match the results on measuring Gain (S21 channel 1).
So basically I soldered the coil between two SMA connectors and have
the SMA cables connected to Ch0 and Ch1.
For some reason the frequencies of max. Impedance S11 Z (Ch0) do not
even closely match the frequency of min. S21 Gain (Ch1).
Z on Ch0 shows max. Impedance at around 140 MHZ, but min. gain is
around 260 MHz.....
When connecting Ch0 and Ch1 directly the S21 gain looks pretty good
(flat at 0db)
What am I doing wrong or let's say, what is my misinterpretation of S11
Z and S21 Gain?
Or is something wrong with the VNA ?
