Unexpected Smith chart behavior

[wasedadoc]

0. edit the first post, and fix the typo in the title 

Thanks, fixed!

You did not fix all the instances of 'smith'.

[Randy222]

1) you don't have a cable in the test, you have a transmission line.
2) on that TL you have some uncharacterized adapters on the front side. Each insertion changes effective VF of the TL.

I suggest you calibrate the nano using BNC open,short,load, leave all the S11 adapters there, calibrate with those attached to S11.

And finally, as stated, the cable in pic is not a "cable", it's a TL (cable + connectors), aka cable assembly.

[RoV]

You can also establish vf by measuring the capacitance of a known lenght of cable of known impedance.
Starting from well known equations Z0=sqrt(L/C) [cable impedance] and v= 1/sqrt(L*C) [propagation velocity],
where L and C are inductance and capacitance for unit length [metric],
you can use the known Z0 [50 ohm] to obtain L from the first equation and substitute it in the second, getting:
v = 1/(C*Z0), or vf = 1/(3e8*C*Z0).

With 24.8 pF/ft, or 81.36 pF/m, you get vf = 0.82.
vf = 0.66 would require C = 101 pF/m.

Clearly, a good measurement requires a few meters of cable, so you can ignore the effect of connectors, but you can get a reasonable idea of your cable characteristics with your existing cables, if you have a capacitance meter. Remember to compensate open probe capacitance, which generally ranges several pF.

[AJ528]

Clearly, a good measurement requires a few meters of cable, so you can ignore the effect of connectors, but you can get a reasonable idea of your cable characteristics with your existing cables, if you have a capacitance meter.

interesting. What kind of tool would you use to measure the capacitance? I've never heard of a capacitance meter. Is that just the capacitance measuring function on an LCR meter?

Remember to compensate open probe capacitance, which generally ranges several pF.

How would you do that compensation on an LCR meter? On my nanoVNA, it's pretty easy to perform a calibration, but I don't think my LCR meter can be calibrated in the same way.

[TimFox]

Measuring capacitance is a native function of an LCR meter.
LCR meters (as opposed to the rough function in a DMM) need to calibrate the fixture by measuring the impedance with the fixture open-circuit and shorted.
After removing the short, the measured capacitance includes anything connected where the short and open used to be.
This is technically compensation for the leads between the box and the measurement connections;  the actual calibration usually depends on a precision resistor in the box.
Which LCR meter do you have?

[RoV]

interesting. What kind of tool would you use to measure the capacitance? I've never heard of a capacitance meter. Is that just the capacitance measuring function on an LCR meter?

That would be fine

How would you do that compensation on an LCR meter? On my nanoVNA, it's pretty easy to perform a calibration, but I don't think my LCR meter can be calibrated in the same way.

Many RLC meters have a "zero" button to get a relative reading, otherwise you can read the open circuit capacitance and later subtract it from the measurement. Not perfect, but good enough.

[AJ528]

I have an GW Instek LCR-916 (https://www.gwinstek.com/en-US/products/detail/LCR-900)

Looking at the manual, it turns out there is a calibration button on it. Cool. That's good to know!

[TimFox]

For LCR meters, it is mandatory to compensate the connections using the calibration procedure in the manual.
I do it at the start of any measurement set, after connecting the test fixture or cables.

For coaxial cable transmission lines, with normal construction, the velocity factor is a function of only the dielectric constant between center conductor and outer shield.
Common cable dielectrics include air (vf = 1.0), solid polyethylene (vf = 0.66), foam polyethylene (vf = 0.80), ptfe (vf = 0.69), and foam ptfe (0.85 to 0.90).
"Foam" means that there is a substantial amount of air within the solid, decreasing the dielectric constant.
Also note in the RF literature that the velocity factor does change at high frequencies due to higher-order effects in the transmission line.

[Randy222]

I would rethink the testing.

Calibrate the VNA by itself.

Attach all the adapters and cable.

Obtain/calculate VF from the info the nano is showing you. Note this is not actual VF of the cable, it's a VF of the TL.

Continue on to obtain measured TL length using the calculated VF. This is the measured or "effective" electrical length.

Then compare those results to the Smith chart, it should make sense. The Smith chart should be matching the measurements of the TL because a TL has an effective electrical length and VF, which are different than actual length and VF of just the RG cable.

Get VF of the TL --> https://youtu.be/aWvPB299U60

Get length of the TL --> https://youtu.be/9thbTC8-JtA   

Also worth a read --> https://www.eevblog.com/forum/rf-microwave/vna-for-cable-characterization/

For LCR meters, it is mandatory to compensate the connections using the calibration procedure in the manual.
I do it at the start of any measurement set, after connecting the test fixture or cables.

For coaxial cable transmission lines, with normal construction, the velocity factor is a function of only the dielectric constant between center conductor and outer shield.
Common cable dielectrics include air (vf = 1.0), solid polyethylene (vf = 0.66), foam polyethylene (vf = 0.80), ptfe (vf = 0.69), and foam ptfe (0.85 to 0.90).
"Foam" means that there is a substantial amount of air within the solid, decreasing the dielectric constant.
Also note in the RF literature that the velocity factor does change at high frequencies due to higher-order effects in the transmission line.

It's also well documented how the changing dielectric materials impact VF, since insertions are often impedance mismatch to the next thing it connects to.
In highly critical TL's, the connectors need to match the cable very closely.
Once all those oddities are accounted for, the Smith chart will make sense. To see what an adapter does, just add another or two on the S11 side, see how that changes the Smith chart.
And just for clarity in practical applications of the TL, nobody just measures a cable length and relates that to wavelength of the applied frequency. A TL is always an "effective" measurement, which can be tuned, but rarely is the math using just coax physical length.