开云体育

I have a NanoVNA-H: HW version 3.4 with Dislord firmware 1.2.20. I have run the SOLT calibration. When I leave CH0 and CH1 connected together and do an S21 Logmag plot, there is almost -3dB offset over the entire plot. Shouldn't there be a straight at 0dB? I know I can enter an offset value to compensate but that seems like a work-around rather than fixing the problem. What might I be doing wrong?

Some ideas to consider:
Do you have the ports hooked together in the exact same way they were
during the thru part of the calibration procedure?
I suspect not, because that last thru step of the calibration sets the 0dB
point of the thru signal, and the way to verify that the calibration took
proper effect is to hook the ports together and observe the constant 0dB
S21 logmag trace.
So maybe you are making a mistake in the cal procedure. Or does the cable
you are connecting with now have a T connector to some other monitoring
point? Splitting the signal would cause a 3dB loss. Another thing to
check is if you are really looking at S21 logmag, i.e. is the trace really
set to what you think it is. The display labels should confirm that.

Hello

You should tell us what you call "over the entire plot".

Attached are the results of my measurements on my NanoVNA-F between 1 MHz and 100 MHz with nanovna-saver. The measurement is made on the original cords each equipped with an SMA - BNC♀ adapter. The connection is made with a BNC connector♂/♂. Calibration is done on the BNC plugs♀ with an N HP 908A load and an N♂ short circuit


73
--
F1AMM
Fran?ois

Hi Stan,
Yes, I am sure that the cables I am using are the same as the ones used to calibrate the VNA. I actually left everything connected after doing the T calibration and then did my measurements. The only parts connected are the two coax cables and the F-F SMA adapter. The info at the top of the screen show S21 Logmag. I also did a Reset before calibrating.

Fran?ois, by "over the entire plot" I mean the entire trace. It is a flat line at about -3dB.

That is curious. For your reference, I just did an SOLT cal, and left the cable connected from the last through cal step, and get 0dB as expected across the whole trace.
I set my frequency span from 1-100MHz to match the range Fran?ois showed in his graph.
Here is a screen capture. Note that the trace is on the second line down on the display by default, since that is the default 0dB reference for the S21 trace. Are you perhaps misinterpreting that as being -3dB? Each trace can have its own scale and 0 reference position - so verify that you have that set correctly. The current scale/div is the number just after the label for each trace at the top of the screen, followed by a / and the current value at the active marker.

Fran?ois: I really don't understand why you have such accentuated noise and deviation from 0dB on your post-calibration S21 plot. Mine shows less than .01dB deviation from 1 to 100 MHz. See the attached plot with the vertical axis zoomed in to +/- 0.5dB.
Stan

Fran?ois: I really don't understand why you have such accentuated noise and deviation from
0dB on your post-calibration S21 plot. Mine shows less than .01dB deviation from 1 to 100
MHz. See the attached plot with the vertical axis zoomed in to +/- 0.5dB.

Hello
Indeed the curve you show (nanovna-h4_S21zoom_in.png) is perfect. I have never managed to do better than what I offered you. I don't know the cause. My box is a nanoVNA-F as attached. It has an aluminum alloy case.



I do with :(

73
--
Fran?ois

-----Message d'origine-----
De : Stan Dye
Envoyé : samedi 2 septembre 2023 02:10

You may want to check the cable-to-plug interface of the supplied coax cables.

On mine the ground sleves are not soldered or crimped but just hold in place by the shrink tubing. Depending on how I bend or twist the cables I get all sort of funny results.