VSWR measurment with a Spectrum Analyzer - RF Cafe Forums

 Post subject: VSWR measurment with a Spectrum Analyzer

Posted: Thu Jan 18, 2007 8:27 am 

Hello everybody, i am new on this forum .

Say, i would like to measure the VSWR in a cable connected to an end-device. I would like to do that because we got some troubles with the front feeding device just before the cable (we can assume that's RF generator). This device goes quite often down and i think it's because of a mismatch between the impedances and therefore a part of the energy sent through the cable comes back and damage little by little the feeding device.

The thing is that i have only a spectrum analyzer to do this measurment, i know there is a way to do the measurment with this device and a coupler (or RF bridge). My problem is that i don't know how excatly we do that and how a coupler work. Can someone help me or forward me towards an internet page or a good pdf document which describes what i have to do and how it works.

Thank you.

 Post subject:

Posted: Thu Jan 18, 2007 2:47 pm 

Hi,
to do any SWR-measurement you must have some sort of generator and a coupler. It's then this reflected power you measure and calculate the SWR. A spectrum analyzer alone is just not enough. Together with a coupler you could use the analyzer as power meter but you must also have access to a point before the unit you have problems with.

Regards
Jens

 Post subject:

Posted: Fri Jan 19, 2007 7:48 am 

Thank you for your answer, but what do you mean by "having an access before the unit i have problems with"?

You mean to mesure the power that this unit will receive in order to be able to calculate the return loss of the assembly? But i can measure the feedback energy and then the energy which is provided to the device in two times no?

What i am afraid of is that if i let an output of the coupler "open" (a directionnal couple has 4 connectors, one for the generator, one for the device, one for the spectrum analyzer, but in that case the last will remain unconnected ) without any device connected to it, will it have any impact on the measure and so the results i can have?

 Post subject:

Posted: Fri Jan 19, 2007 9:45 am 

You are right. You must leave the unconnected port of a 4 port coupler correctly terminated in a matched load.
An alternative that is accurate enough in most cases is to connect the analyser probe not directly to the coupler ports (forward or reverse) but through a 6dB or even a 10dB attenuator attached to each coupled arm. This ensures that the ports are always matched, but allows a quick change from forward to reverse for the analyser connection.

Of course, this method only gives the scalar reflection coefficient, not the vector value.

 Post subject:

Posted: Fri Jan 19, 2007 10:19 am 

Sorry can you be more precise on your last remark? i dont see what you mean by vector value of the reflection coefficient The postion of the max and the min along the standing wave?

 Post subject:

Posted: Fri Jan 19, 2007 11:45 am 

Hello CyerOli,

First, what is the VSWR (RL) value you want to measure? This method may not have enough accuracy.

To measure VSWR with a spectrum analyzer and a coupler do the following:

For a three port coupler: Terminate the coupler's "input" port in a "perfect" 50 ohm load. Since one does not really exist use one with a VSWR of 1.05:1 will do. The RL is -32 dB.
Connect your source output to the coupler's "output" port.
Connect the coupled port to the spectrum analyzer input.
Setup the spec-ana to cover the frequency of interest and a reference level 10 dB above the expected RL. (Zero span will work also).
Setup your source for the frequency(s) of interest and a power level of 0 dBm. The coupler should have a diagram indicating the normal thru path so just reverse it - the output is the input.

Measure the reflected power on the spectrum analyzer. Add the coupling factor and then convert the RL to VSWR using the standard formular. Use the standard formulars to calculate mismatch uncertanity and you have your answer

There are a couple of error to watch out for. First the directivity of the coupler has to be at least 15 dB better that the RL to be measured. If not then the results will have serious uncertanity. For example, if the RL to be measured is -19 dBm then the directivity has to be better that -34 dB. For this kind of directivity tou will need a VSWR bridge (very expensive).

If you are using a four port coupler then also terminate the unused coupled port.

Good luck,
RFTEJerry

 Post subject:

Posted: Sat Jan 20, 2007 11:11 am 

I don't have any idea about the value i have to measure. That's whyI need to do it .

Thanks for your advices.

Can someone clarify my mind concerning the vector value (asked in my previous post)?

 Post subject:

Posted: Mon Jan 22, 2007 6:24 am 

Hi CyberOli
The Reflection Coefficient is actually a vector value, because it is a computation using a resistive reference (eg 50 ohms) as the source and a combination of resistance and reactance as the load to be evaluated.
You are right, being a vector does shift the position of the minimum around if you are using a probe carriage type test setup.
VSWR is a calculation made just using the magnitude of the Reflection Coefficient vector.

Jerry's answer is much more comprehensive than I have given, but I hope my point, about ensuring the coupled arms are always matched, is clear.

 Post subject:

Posted: Mon Jan 22, 2007 10:03 am 

Yeah i have well understood you in fact, but that was just to be sure. I suppose i can assume that the load is purely resistive in a first time, i don't think it has that much importance in my proble. The goal of this measure if to see if there is a reflected energy and try to avoid it.

Thank you for your help, you were all helpfull!