I'm not sure if there is a device which can give you a single value directly, but you may extract it from scattering parameters measured using a network analyzer. Convert the scattering matrix to an ABCD matrix, and Zo = sqrt(B/C)

I think you can use in CST, Macro>>> calculate>>> line impedance,

and solve it.

In order to measure the characteristic impedance of the transmission line, if it is long one such that its length L is much greater than the wavelength, its input impedance will be equal to characteristic impedance Z0. In you case your transmission line length is not long. In this case you need to terminate your TL by an impedance equal to its characteristic impedance which results in input impedance equal to Z0. Then all what do you need it to measure the input impedance. As hinted before in microwave frequencies you can a network analyzer to measure it by measuring S11. Assuming 50 ohm internal impedance of the analyzer:

S11= (Z0-50)/(z0+50)

Then you can calculate Z0.

Best wishes

Dear Mr. Farid Hosseinzadeh

The impedance of a device can directly be measurd with the help of a network analyzer. A network analyzer allows you to display the Smith Chart of the impedance, with the frequency as a free parameter. You can move the cursor to the desired frequency 1.5 GHz, and the analyzer will display the real and imaginary part of the impedance at that frequency. There is no need to do any calculation at all.

Just make sure to perform a proper SHORT, OPEN, and LOAD calibration prior to taking the measurement. This calibration must be made at the exact contact points (e. g. the end of coaxial cable connected to your analyzer) at which you connect your device under test. This is necessary to compensate for the impedance transformation resulting from the cable. If you are unsure of how to do it, consult the reference manual of your analyzer. By the way, it is often possible to hire such analyzers from RF companies or institutes, so that you don't need to buy one if you only need it occasionally.

Best wishes

Dr. Hubert Aebischer

If you have designed a 50 ohms transmission line, you can measure S11 and S21 using Vector network analyzer. If the insertion loss is minimum then you will have very good vswr. This way you can confirm the impedance. You can plot Impedance in Smith Chart also.

All the above tips are useful for matching microstrip transmission line with antenna. But they do not allow you to measure the characteristic impedance of the microstrip line at high frequencies due to the fact that in the value of the measured S11 can not be divided contributions from the diffraction at the junction of the two transmission lines and the characteristic impedance of the measured line.

Best wishes.

Just keep in mind that there are several possible definitions for the microstrip characteristic impedance (e.g. voltage/current or maybe voltage/power etc)A microstrip line is not a clean TEM line ..it acrried a hybrid mode since the field in the different regions of the cross-section would like to propagate with different velocities. Thus we get dispersion both for the char. impedance AND the propagation constant. Of course everything can be measured nicely with a properly calibrated VNA...there are MANY methods..but the question is about precision...and there may also be some radiation (antenna) effects.

You can measure your transmission line with VNA. Just plot the measured impedance on the Smith chart, add the marker and see the complex impedance of your MS line. Make sure your VNA is properly calibrated.

You can use VNA. In fact, the obtained S-parameters can be converted into Z-parameters.

Please look at this paper:

http://www2.electron.frba.utn.edu.ar/~jcecconi/Bibliografia/04%20-%20Param_S_y_VNA/Parameters%20Conversion%20Table.pdf

Some engineering opinions, in addition to the many valid points that have already been made concerning the use of a Network Analyser, NWA, may possibly be helpful:

* An S11 measurement will include reflections from: the input connector; the nominal 50 Ohm track; and the GNSS chip input.

* As previous contributors have mentioned, with knowledge of the GNSS chip S11 and the S parameters of the input connector interface it would, in principle, be possible to calculate the S parameters of the nominal 50 Ohm track and from these find the track Characteristic Impedance. Unfortunately:

** GNSS chip manufacturing tolerances will cause its S11 to depart, perhaps significantly, from the nominal S11 listed in the chip's data sheet. Additionally the following will affect measured GNSS chip S11: how closely it has been possible to replicate the manufacturer's recommended PCB layout, including vias, and also the PCB material; the fact that there will be tolerances on PCB material Dielectric Constant and the track width.

** There probably will not be good knowledge of the S parameters of the input connector interface.

* Given the uncertainties and difficulties mentioned above, practical approaches include, but are not restricted to, the following:

** Make a test PCB with 50 Ohm tracks and measure this with the NWA. However, there may not be time or money for this.

** Measure, using the NWA, S11 magnitude looking into the input connector of the 'powered-up' GNSS PCB and see if this is 'respectable'. Standard RF practice, in the absence of other special considerations, tends to be that 14 dB Return Loss, corresponding to 1.5:1 VSWR, is considered 'respectable'; apologies for mentioning an engineering 'rule of thumb' without a reference. However, the GNSS chip Return Loss may be lower than 14 dB, in which case Return Loss, looking into the PCB input connector, one or two dBs lower, in the GNSS frequency band, than the GNSS chip data sheet value would be 'respectable'; engineering judgement again. Return Loss looking into the 'powered-up' GNSS PCB is a useful performance metric that should, if possible, be measured.

* On the subject of VNA use, it is prudent to strictly follow all your institution's procedures and, if possible, have an experienced VNA user show you what to do. This recommendation is made because VNA Calibration Kits are expensive and easily damaged.

Prof Almekhlafi has, most succinctly, provided a solution that enables the effect of track characteristic impedance to be isolated from the effects of the connector interface reflection and the GNSS chip input impedance, which is more problematic with the NWA, for reasons previously explored.

RF engineers automatically think of equipment that excites a unit under test, UUT, over its operating frequency band, with sinusoidal test signals - and many of us went down that path, with all the complications that resulted. Perhaps, also, we had in our minds the equipment we would have in our own labs. A key message, here, is that, in order to measure just the track, wide bandwidth excitation of the GNSS PCB is required, which a TDR machine can achieve. A NWA equipped with Time Domain option operated over a frequency span of some GHz can, in principle, gate out the connector interface and GNSS chip; the frequency span required would mostly be determined by the length of the track. However, this would probably be an exotic VNA and the TDR machine seems a much more practical option, if available.

As usual, it should be checked that the GNSS chip input damage level will not be exceeded by the TDR machine output pulse.

I failed, in a previous post, to state some conditions for which the 14 dB return loss rule of thumb would be reasonable, which could include moderate, rather than multi-octave bandwidth. The input return loss of the GNSS PCB, within the GNSS band, could be a candidate for the 14 dB rule of thumb. As usual with rules of thumb, it is not a hard and fast rule, but more a guide to whether, or not, colleagues might be thinking that one could have tried a bit harder to obtain 'respectable' Return Loss.

Denis:

My understanding of the 'exam question' that Farid Hosseinzadeh has set us is:

Given the PCB with GNSS chip installed, what piece of test equipment should be connected to the input connector of the PCB in order to find the characteristic impedance of the track connecting the input connector to the chip?

The picture I have in my head is that the track is in the order of a few cm in length and I am additionally assuming that no modification of the PCB to facilitate the measurement would be welcome.

From the assumption about track length I was then thinking along the lines of section 9.2 in Keysight Technologies, 'Time Domain Analysis Using a Network Analyzer' http://literature.cdn.keysight.com/litweb/pdf/5989-5723EN.pdf, which is saying that the TDR NWA should be operated with wide frequency span to resolve down to disconuity separations in the cm region.

If you could provide details of the measurement you thinking of, which answers the requirement as stated in this post, without using a wideband test pulse or a wide frequency span NWA test signal, I will happily be educated.

Phew! It was with some trepidation that I resumed looking at this thread after a few days off; I, too, had a concern that I might be in need of someone with enhanced didactic abilities. On the face of it, though, this is not required.