Basic Metrology ( Measuring Techniques)

This is to show how the type(s) of measurement equipment to be used when measuring the parameters of your circuit , it depends not just on what you wish to measure but also on the magnitudes of these signals to be measured.

The picture below shows a range of measurement devices combined to measure a 20 watt power amplifier for 3.4 GHz , most measurements for the basic S11 ,S22 ,S21 & S12 can be done using a Vector Network Analyser ( or a Spectrum Analyser with it's own Tracking Generator & Return Loss Bridge ) provided the circuit under test is either passive or it's output is less than 10mW ( +10dBm ) with NO DC present on the connections to it .

To measure at any greater power will require the use of attenuators or directional couplers to bring the magnitude(s) of the signals into the range of the test equipment , this also leads to making errors by the wrong choice of techniques .For instance you cannot simply add an attenuator to the input of a VNA to measure the S22 of an amplifier capable of more than +10dBm O/P , because the VNA will effectively only see twice the value of the attenuator and not the real value of the stage under test . eg for a VNA limited to +10dBm input measuring the output of a 30dBm amplifier you will need at least 20dB of attenuation , so from 1st principles the port connected to the attenuator will see a return loss of 2 x 20dB ( Assuming the attenuator is a good quality item ) with either O/C or S/C applied to it . The only way that you can measure these magnitudes of power is INDIRECTLY using a directional coupler and a termination of system impedance & suitable power rating , then comparing the incident and reflected power to obtain the Return Loss ( Match ) the same technique can also be used to measure input match for a power amplifier when driven at correct ( Large ) signal levels rather than the usual 0 to + 10dBm ( Small ) output from the VNA , this technique also allows the phase delay of the circuit to be measured at the desired power level not just at small signal values ( But it's always a good idea to test at lower power levels first ! ) The test setup above shows the main 20 watt amplifier ( Working at 3.4 GHz ) fed into a precision power attenuator capable of 30 Watts continuous dissipation with a frequency vs attenuation graph showing it's performance ( DC - 12GHz ) the output from the power attenuator is fed to a splitter to enable the power to be measured on a power meter , along with the spectrum analyser ( off screen ) to measure the output spectrum ( harmonics etc ). The input power & return loss for the power amplifier is measured using a pair of directional couplers one for each channel , again the power is measured , this time using a pair of Spectrum Analysers to ensure the signals from the driver amplifiers are " clean " so that any spurious signals seen on the output must be generated in the high power amplifier section . The signals supplied to the driver amplifiers are capable of being adjusted in phase using a pair of adjustable transmission line phase shifters ( two are required to get the correct range ) again these signals are sampled and fed to the large Digital Storage Oscilloscope , this is used to measure the phase angle between the two signals applied to the driver stages . When measuring power it is essential to measure all the components used , each cable will have an insertion loss , each connector pair will have a loss ( small but measurable ) and most importantly when dealing with high power measurements using attenuators is to measure the attenuator at it's operating temperature ( Yes it does vary ,often upto 0.5db , that's an error of 10%) so if you are for example making efficiency measurements ( DC + RF in to RF Out ) be warned ! The failure to correctly tighten up the connectors to the specified torque can also lead to intermittent fluctuations in power being measured , so the use of the correct torque is neccesary . The use of a " Breakover " spanner with SMA connectors ( and other precision connectors ) ensures this value without damaging the connector thread & nut which can so easily happen with ordinary spanners ( DO NOT USE a 250mm Adjustable spanner on an SMA !) Yes I've seen students doing this rather than ask for the correct spanner .

This leads to ensuring that ALL the test equipment to be used is not only CALIBRATED to a known standard ( Certified & traceable ) but that it is also powered up to reach it's normal operating point ( Typically 1 hour from switch on is specified by the manufacturer )

Pictures of typical Test Instruments used in the measurements described on this page

Vector Network Analyser

Spectrum Analyser ( Note NO Tracking generator fitted )

Portable Spectrum Analyser with Tracking Generator & Return Loss Bridge

Return Loss Bridge

The Directional Coupler used to sample the RF power from between the driver amplifier & the power amplifier .

Typical Low power Attenuators 1 or 2 Watt in SMA connectors

The Power Attenuator used to dissipate the output from the amplifier under test and reduce the signal amplitude to a level that the Spectrum Analyser can handle if it is inconvenient to connect the items under test straight to the measuring port of the instrument then a useful tip is to use ODD ELECTRICAL 1/2 wave length leads to minimise any reflections due to discontinuities in cables & connectors etc.

A simple LC Band pass filter being tested for S11

An Alternative instrument for measuring the reponse

Insert Picture of Passive Circuit under Test here ( 145MHz BPF & HP VNA )

Insert Screen shot of trace here

A simple LC Band pass filter being tested for S21

Insert Picture of Passive Circuit under Test here ( 145MHz BPF & FSH3 & RL Bridge )

Insert Picture of Passive Circuit under Test here ( 145MHz BPF & HP VNA )

Insert Screen shot of trace here

A simple transistor amplifier using a single GasFet ( Link to project ) can be measured using either the VNA or a Spectrum Analyser with a tracking generator ,however , in order to avoid saturating the amplifier when measuring the S21 it is advisable to turn the O/P of the tracking generator ( VNA port 1) down or reduce it using a fixed attenuator of 10dB ( nominal value ) on the input to the amplifier .The use of a signal generator and RF milliwatt meter can enable the user to test at a single frequency for the gain and get a better idea of the S21 with respect to stage gain & saturation point ( P1dB ) where the forward gain falls off due to saturation of the device not readily accomplished using a swept system such as the VNA with it's limited range to set output amplitude .

When measuring the harmonics of an amplifier this can be done in more than one way either using the Spectrum Analyser with a suitable attenuator on it's input or by using a directional coupler and termination , there are limits to both ways !

With the power attenuator , the frequency response of the attenuator is reasonably flat across the range of the harmonics to be investigated ( +/- .5dB is quite typical ) the coupler however will definitely NOT have a flat frequency response ,however by measuring the coupling coefficent at these respective frequencies an accurate measurement can be made ( often at higher powers than an attenuator is capable of ) note the insertion loss of the connecting cables will also change across the range so these too must be checked if the power out from the coupler still exceeds the input rating of the Spectrum Analyser then the use of a low power (1 or 2W rated ) attenuator will be neccesary ( note this also improves the directionality of the coupler when using a flexible cable ) another point to remember when using a high power attenuator is the actual value of attenuation varies with temperature so it is advisable to check the actual value having warmed it up to the operating temperature there may be upto 1dB error in the quoted value of attenuation ( that's an error of 20% ) .

This page will be added to shortly when time permits

Back to Homepage

This page last updated 11/05/2011 6