Gain of an Antenna measurement

If you want to measure the Gain of an antenna, a Standard Gain Horn (SGH) is necessary. A SGH doesn't have to be a horn antenna. It just has to be an antenna with known, repeatable gain and a polarization compatible with then antenna you intend to test. The calibration procedure for a SGH is here.

Test Setup:

The  Network Analyzer could be replaced by a Signal Generator and Power Sensor if you prefer; it will make calibration harder, in my opinion. An amplifier may be necessary to bring your signal above the noise floor of the Network/Spectrum Analyzer.

Calibration Procedure:

• Find an Anechoic chamber or antenna test range that is quiet at your frequencies of interest (FOI).
• There are no background signals (i.e. cell phone tower) large enough to be picked up by your SGH.

• Acquire two SGH's.

• Space the horns apart from one another so that we measure the Far-Field Gain of the SGH's.

• Rule of thumb: distance between antennas, R, should be >> wavelength at FOI. (R > 100λ)

• Calculate the freespace loss (Lf) from Friis Equation: Approximate the distance R between the horns with a measuring tape, aperture-to-aperture. 
• Calculator is here. 

• Set the Transmission Power (Pt): Adjust the power of the Network Analyzer to overcome the Freespace loss minus the expected Gains of the transmit and receive antennas; Signal should be well above the noise floor of the Network/Spectrum Analyzer.
• Pt = Lf - Gt - Gr + Pr, where Gt and Gr are the Gains of transmit horn and receive AUT.
• Set Pr, in this equation, to an appropriate level above the noise floor of your receiver (~ -20 dBm for most Network Analyzers)

• Align the horns as best as possible so that they both have the same polarization and are facing each other.
• Make fine adjustments (up/down, left/right, swivel) to each horn until you see a maximum amplitude for S21 on the Network Analyzer. 

Photo copied from Wikipedia. Author: T. Truckle.

• Antennas typically have a main lobe of radiation, flanked by sidelobes that radiate in all directions. The purpose of aligning the horns until maximum power is received is to ensure that Gain is measured at this main lobe

• Conduct an S21 THRU calibration on the Network Analyzer at your FOI. S21 should now read 0dB.

The Measurement:

• We know the Gain of the receive-SGH because it was measured previously.
• Enter an amplitude offset into the Network Analyzer equal to the SGH gain at your FOI.
• ex: If your SGH has G = 25 dB @ FOI, the Network Analyzer should have S21 = 25 dB.
• If you do not have this option, or if you are using a Spectrum Analyzer, make a note that whatever maximum power displayed is equal to the SGH gain at your FOI.

• Replace the receive-SGH (#2) with your AUT. 
• Align the antennas as in the calibration procedure.

• Note the S21 amplitude on the Network Analyzer (dB) has changed.
• It should now display the gain of the AUT.
• If using the Spectrum Analyzer, add the change in power to the gain of your SGH to determine the gain of the AUT.

*Please Note: It does not matter if the AUT is the transmit or receive antenna. Gain is identical in either direction. Just make sure to offset S21 by whichever SGH will be replaced by the AUT.

Standard Gain Horn Calibration

If you want to measure the Gain of an antenna, a Standard Gain Horn (SGH) is necessary.

If you were to order a SGH, it typically would have a table listing the Gain vs. Frequency. But if you want to verify the Gain at each Frequency or if you want to make your own SGH then this is the procedure to make that table.

What kind of Antenna should I use for my SGH?
It must have a polarization that is compatible with the Antenna you eventually intend to test (AUT).
Polarization types:

• Horizontal
• Vertical
• RHCP (right hand circular)
• LHCP (left hand circular)

Most commercially available SGH's are Vertically polarized (and Horizontally polarized if you rotate them 90degrees). They are also made of sturdy aluminum and extruded from rectangular waveguide to form the Horn shape. This is a pretty robust construction that will not change much over time; It can be nicked and handled frequently without changing the Gain... to a certain extant. Dents WILL change the antenna radiation pattern.

Test Setup:

The  Network Analyzer could be replaced by a Signal Generator and Power Sensor if you prefer; it will make calibration harder, in my opinion. An amplifier may be necessary to bring your signal above the noise floor of the Network/Spectrum Analyzer.

Calibration Procedure:

• Find an Anechoic chamber or antenna test range that is quiet at your frequencies of interest (FOI).
• There are no background signals (i.e. cell phone tower) large enough to be picked up by your SGH.

• Acquire two identical SGH's.

• Space the horns apart from one another so that we measure the Far-Field Gain of the SGH's.

• Rule of thumb: distance between antennas, R, should be >> wavelength at FOI. (R > 100λ)

• Align the horns as best as possible so that they both have the same polarization and are facing each other.

• Calculate the freespace loss (Lf) from Friis Equation: Measure the distance R between the horns with a measuring tape, aperture-to-aperture. 
• Calculator is here. 

• Set the Transmission Power (Pt): Adjust the power of the Network Analyzer to overcome the Freespace loss minus the expected Gains of the two horns; Signal should be well above the noise floor of the Network/Spectrum Analyzer.
• Pt = Lf - Gt - Gr + Pr, where Gt and Gr are the Gains of transmit, receive horns.
• Set Pr, in this equation, to an appropriate level above the noise floor of your receiver (~ -20 dBm for most Network Analyzers)

• Remove the SGH's and connect the RF cabling together. Conduct an S21 THRU calibration on a Network Analyzer at your FOI. S21 should now read 0dB.

The Measurement:

•  Align the horns. 
• Re-attach them to the Network Analyzer cables.
• Make fine adjustments (up/down, left/right, swivel) to each horn until you see a maximum amplitude for S21.

• Determine Gain of each Horn:
• Note the S21 amplitude on the Network Analyzer (dB)
• Add the freespace loss (dB)
• Divide by 2 to yield the Gain of each identical horn.
• Gtr = (S21+ Lf ) / 2 , where is the Gain of each identical transmit or receive horn.

• It is a good idea to repeat this measurement for several values of R to ensure you get accurate, repeatable values for Gtr.

Now that we have characterized two identical SGH's, we are able to measure a single, unique antenna for Gain... in the next post.

Freespace Loss Calculator

The freespace loss between two antennas can be derived from Friis Equation.

Freespace Loss (dB):

R: Distance between Antennas (m) Frequency (GHz)

Lf: Freespace Loss(dB)

Pulsed Spike Leakage using a Crystal Detector

Lets say your are sending an RF pulse through a device. It is important to determine the maximum output power from the device before it settles to its steady-state value for the duration of the pulse.

Ideally, there would be no spike at the beginning of the output-pulse; this may damage other devices if they cannot handle the power spike.

How can we measure the power spike?
This measurement is nearly impossible with most oscilloscopes because:

• the spike is typically extremely narrow (time).
• the sine wave that comprises the pulse is difficult to lock onto.

Luckily, a crystal detector makes this measurement much easier. The detector will convert RF power to DC voltage that can be observed on an oscilloscope.

Since the detector will submit a voltage to the oscilloscope- and we want to know the RF power of the spike, a calibration procedure will be needed to convert voltage to dBm. But first, lets go over the test setup.

Test Setup:

The test setup is similar to pulsed power measurements, except now we have added two important components: Variable attenuator and the Crystal Detector.

Why do we need a variable attenuator?
Typically, a detector cannot handle much RF power. During the test, we will constantly adjust the attenuator to known values so that we see enough signal on the oscilloscope, but not so much that we blow-up the detector. After, we can subtract this attenuation out to determine the spike power out of the DUT. The variable attenuator is surrounded by isolators so that when we adjust its value, we will not significantly change the impedance-match seen at the output of the DUT and the match seen at the input of the detector.

Calibration Procedure:

• Calibrate the input arm to the DUT just like we did for pulsed power measurements.
• Characterize the variable attenuator at a number of settings:
• At your Frequency of Interest, use a Network Analyzer to measure S21 of the chain of RF components connected between the DUT and the detector.
• Make a lookup table (I): S21 vs Variable Attenuator dial reading.
• Now when we dial in a setting for the Variable Attenuator, we can know the exact attenuation that is being used.
• Calibrate the Crystal Detector: 
• Use the signal generator to inject power directly into the detector and note the voltage displayed on the oscilloscope. (careful not to put more than rated input power of detector)
• Remove the detector and measure the power you injected with a power sensor.
• Create a lookup table (II) for the detector: Voltage vs. power (dBm).

Measurement Procedure:

• For a given input power to the DUT, you should be able to measure output voltage of the spike on the Oscilloscope.
• Convert the voltage to power using Lookup Table II. This is power incident on detector.
• Subtract the attenuation of the Variable Attenuation chain using Lookup table I to determine the actual value of the power spike.

Noise Figure calculator: Y-factor method

This is a calculator to determine Noise Figure of a device using the Y-Factor method.

Test setup:

For accurate measurements, you need to know what the Net Noise Figure is from the output of your DUT to the input of your Spectrum Analyzer (LNA plus all RF cabling). Call this NFLNA.

Analysis:The Net Noise Power (linear) of the entire chain is:

where Y (dB) = |Phot - Pcold| (dBm), and ENR (dB) is listed in a table on your ENR source as a function of frequency. Phot and Pcold are the noise powers observed on the Spectrum Analyzer when the ENR source is turned ON then OFF.

By re-arranging the Noise Cascade equation, the Noise Power of the DUT, FDUT, is:

FDUT = Fnet - (FLNA-1)/GDUT   

*Note: All components of this equation are linear, not dB.

In review, this is what you need to measure:

NFLNA (dB) GDUT (dB) Phot (dBm) Pcold (dBm) ENR (dB)

NFDUT (dB)

Noise without a Noise Figure Meter: Y-Factor

Why would I want to measure Noise without a Noise Figure Meter (NFM)?

• Your Device Under Test (DUT) may have a narrow frequency response. So narrow that it is less than the Measurement Bandwidth of your NFM. Think of a narrow Surface Acoustic Wave Filter.
• You cant afford a NFM on top of all the other fancy equipment you purchased.
• You want to know how a NFM works.

Regarding point #1, the Measurement Bandwidth of your NFM is listed somewhere in the documentation for your test equipment. Find it. If it is > the bandwidth of your DUT, then you have a problem. Look at the figure above. The NFM will take some average value for amplitude since it can't zoom in on the exact frequency you are interested in; only a range of frequencies equal to the Measurement Bandwidth. Oops!

(By the way, measurement bandwidth is important when using all sorts of test equipment. Keep this in mind when measuring very narrow frequency-response devices or when searching for spurious signals)

What you will need:

• Spectrum Analyzer with a Resolution Bandwidth << Bandwidth of your DUT (not hard to find) and "lots" of dynamic range.
• Excess Noise Ratio (ENR) source.
• Low Noise Amplifier (LNA) that operates at your Frequency of Interest (FOI).
• DC power supply to turn your ENR on/off.
• Amplifier with known noise and gain. (optional).

Test setup:

For accurate measurements, you need to know what the Net Gain and Noise is from the output of your DUT to the input of your Spectrum Analyzer (LNA plus all RF cabling). This will not be a "narrow" frequency response so you can use a classic Noise Figure Meter to measure this.

What if a NFM is unavailable?
You will have to look up the Noise and Gain of your LNA on its data sheet, measure the loss of your cables, and cascade them together to figure out the Net Gain and Noise Figure; this will be less accurate.

**Use this calculator to determine Net Gain and Noise of a cascaded RF network! (coming someday)

Theory:
An ENR noise source works in the following manner. It applies a Noise reference at two noise temperatures: Thot and Tcold into our DUT. We want to measure the resultant output power into the Spectrum Analyzer when Thot then Tcold are applied; their ratio will yield Net Noise through our system (from DUT input to Spectrum Analyzer input).
From the Net Noise and knowledge of Gain & Noise of the other components in our system, we can determine the Noise of our DUT by re-arranging the Noise Cascade equation.

Why do I need an LNA?
Since the ENR source will be injecting such a small amount of power into your DUT, the LNA needs to have enough gain to bring this power above the noise floor of the Spectrum Analyzer. The Noise contribution of the LNA must be removed later when determining the Noise of the DUT.

Procedure:

• Measure the Gain (GDUT) of your DUT with a Network Analyzer at your FOI .
• Measure the Noise (FLNA) from output of DUT to input of Spectrum Analyzer.
• Connect the chain together as indicated in the Test Setup.
• Set the Spectrum Analyzer to the following:

a.       Center: FOI

b.      Span: (3dB Bandwidth of Device) / 10

c.       RBW: (Span) / 100

d.      VBW: (RBW) / 100

e.       Averaging: 30

f.      Amplitude/ div: 1 dB

g.        Attenuation: 0 dB

Note: These settings are a starting point; you will probably have to do some tweaking until you get sensible data.

• Measure output power at Thot: Turn on the DC Power Supply to stimulate the ENR source (usually 28V, but depends on model of ENR source). Re-start the Spectrum Analyzer Averaging and note the output power: Phot.
• Measure output power at Tcold: Disconnect the DC Power Supply. Re-start the Spectrum Analyzer Averaging and note the output power again: Pcold.

Analysis:

The Net Noise Power (linear) of the chain is:

where Y (dB) = |Phot - Pcold| (dBm), and ENR (dB) is listed in a table on your ENR source as a function of frequency.

By re-arranging the Noise Cascade equation, the Noise Power of the DUT, FDUT, is:

FDUT = Fnet - (FLNA-1)/GDUT   

*Note: All components of this equation are linear, not dB.

In review, this is what you need to measure:

• GDUT
• FLNA
• Phot 
• Pcold
• ENR

Check your set-up:

Replace the DUT with an amplifier with a known NF and Gain. Do your measurements make sene?

...Coming someday: a calculator for Noise Figure

How to limit curent and monitor it with a sense resistor

It doesn't get any simpler than this. But somebody asked, so here we go!

So you have an experimental device to test (DUT). It is supposed to be supplied a voltage to turn it on (lets say: 12V). The first thing you should ask is: How much current can this draw before it blows up?
Once you know the maximum current specification (ex: 200mA), How do you ensure that it doesn't go beyond this maximum value?

Procedure to deliver 12V and a maximum of 200 mA from a DC Power Supply:

• Find a Current Meter that can handle >200mA.
• Find a DC Power supply that can supply 12V.
• Place the Current Meter in series with the Power Supply (see above photo).
• Most DC power supplies have at least 2 knobs. One labeled "voltage", and the other labeled "current". Set your voltage on the power supply (PS) to 12V.
• Now set your current by:
• turning your PS current knob all the way down to zero.
• replace the DUT with a short by connecting the supply-leads together.
• observe the current-reading on the multimeter and slowly turn the PS current-knob until it reads 200mA (voltage on your PS should read 0 V since leads are shorted together)
• Disconnect the leads and ensure that PS voltage goes back to 12 V.
• Re-connect your DUT and you can now monitor current via the Current Meter.

What if you cannot find a Current Meter that can handle your required current?
Say you want to deliver 10 A instead. You can still monitor the current by using a Volt-Meter and Sense resistor instead.

Choose a small resistance value for the sense resistor: 1/2 ohm. Since you expect to now deliver 10 A, the resistor must be able to handle the higher power: P=VI = 50 W... not your typical resistor found in a test kit!

Use a Voltage Meter to measure the the voltage across the sense resistor and convert voltage to current: I=V/R. Set your current limit on the PS as before.

Notes:

• Keithley makes an all-in-one unit that supplies a voltage and measures current (or vice verse).
• Most Current Meters also function as Voltage Meters and are called "Multi-Meters".
• Its always smart to check the voltage at your DUT to ensure their isn't a voltage-drop over long supply-wires.

Pulsed Power Measurements

This is a follow-up to a previous post on CW (continuous power) Gain compression measurements. Please see that post for a more complete overview of power measurements.

Why Pulsed? A pulsed power measurement may be required if your DUT is intended to function in pulsed mode, such as radar applications. It may also be conducted to reduce heating on components within your test setup and/or the device under test (DUT). This will prolong the lifespan of your DUT.

Ideally, this is what we want to send into our DUT.

We do this by sending an RF signal at our frequency of interest (FOI) from a signal generator and modulate it (turn it ON and OFF) with an RF switch. This particular pulse has a duty cycle of 30% and a pulse width of 0.3 ms. We can change the amplitude of this pulse by simply adjusting our signal generator.

What should my pulse width be? It depends. Your DUT has a rise time - how long it takes to to reach a steady-state condition. Your pulse should be on long enough such that your DUT has had a chance to stabilize AND your test equipment has enough time to make a measurement. Your RF switch should have a rise time << then the rise time of your DUT so that your input pulse isn't distorted... yielding false measurements: AARGH!

Block Diagram: Our block diagram will be as before, except now we have added an RF switch to rapidly turn the RF on and off. A pulse generator modulates the switch and TRIGGERS your test equipment. (see note on triggering below) We have also coupled the output of the device so we can observe the output pulse on an Oscilloscope.

Note that this diagram does not include filters. These may have to be added after your amplifier (and your DUT) if it produces a lot of harmonics. A narrow-band isolator will work well for this too. Just be sure to account for the loss during the calibration.

Aside from the additional equipment listed above, your power meters should be configured for a "peak mode". This is usually buried in a sub-menu somewhere but is important to reading accurate pulsed-power. And now that they are in this mode, you need to tell the meters when to make a measurement i.e. triggering.

Important note on triggering: Use your pulse generator to trigger your power meters; now they will take a measurement at the same time that the RF pulse is sent through the DUT. If your DUT has a long delay, you may need to go into the "delay" menu to tell the power meters to wait X ns before starting the measurement. These work pretty well.

What value should my attenuators be?: You want enough attenuation to protect your expensive test equipment. Make a good estimate of the most power that could possible reach your sensors and add attenuation accordingly- But not so much that you lose dynamic range.

Additional Info: 

• Calibration is similar to the CW calibration.
• The oscilloscope will allow you to read the rise and fall time of you DUT. 
• As always, ensure that your test equipment can handle the power required for this test before you turn the RF on.
• Note that some newer Signal Generators contain an internal switch- reducing the amount of required test equipment. However, make sure that its rise time is << your DUT rise time.

High Power Gain compression Test

So you've got your hands on an RF amplifier or similar 2-port device and want to measure its Gain or Output power as a function of Input power.

An ideal amplifier will have the same Gain at every given input power, as seen by the blue curves below. In practice, there is power lost (mostly in the form of heat dissipation) as you increase the input power. The red curves are more of a typical real-world response. This post will provide you with the techniques to measure these red curves so that you can compare them with their ideal counterparts.

Before you get started, you need to know the following information to gather the appropriate test equipment:

• Frequency(ies) of operation
• Maximum input power delivered to the (device-under-test)DUT
• This is usually dependent on the anticipated Gain-compression point of your device; Most tests deliver a maximum input power when the Gain is compressed by ~4 dB.
• Small-Signal Gain of the DUT
• Pulsed or continuous wave (CW) power deliver to DUT
• You may wish to do a pulse-measurement in order to reduce over-heating of components and to prolong the life of your DUT. This will be covered in a later post.

List of Test Equipment {all must operate at your frequency-of-interest (FOI)}:

• Signal Generator that can output a clean signal with good dynamic range (> 35 dB)
• High Power Amplifier (capable of delivering the required max input power to the DUT)
• Isolators, Attenuators, RF loads, couplers
• 2 Power Sensors {with meter(s)}

Here is a block diagram of your test bench:

Choose an amplifier that will boost your Signal Generator power sufficiently enough such that it will compress your DUT. Isolator #1 is to prevent any reflected power from the amplifier from going back into your signal generator; power reflected into the Sig Gen will damage it. Isolator #2 is to provide a good output-impedance match to your DUT.

Power Meter (& coupler) #1 is to measure Pin at the input of the DUT. The coupler and attenuator will have to be compensated for via the calibration sequence (next section). Power Meter #2 is to measure the output power.

The optional Power Meter (& coupler) #3 is used to measure the reflected power from the DUT. Return loss (in dB) can be found at any Pin by measuring Preflected (#3) - Pin (#1), in dBm.

{Choose attenuation values for each power meter so that they don't get damaged. For instance, if you expect your DUT to output 45 dBm max, and your power meter is rated for 20 dBm, you MUST provide >25 dB attenuation between the DUT-output and power meter #2.}

Calibration:
Zero and calibrate all power meters.
Calibrate Meter #1 for losses through the system:

• For now, Replace the DUT with output Power Meter #2.
• Turn on your signal generator and amplifier and adjust the power until you read 0 dBm on Meter #2.
• Adjust the Offset on input Power Meter #1 until it also reads 0 dBm. (If there is no "offset" function, just note the difference in power levels.)
• You can now accurately know the input power level to your DUT for any Sig Gen setting.

Calibrate Meter #3 for losses through the system:

• For now, Replace the DUT with an RF short (or leave it open, but a short is better).
• Turn on your signal generator and amplifier and adjust the power until you read 0 dBm on Meter #1.
• For an RF open, all power should be reflected back toward the source; Adjust the offset on Meter #3 until it reads 0 dBm too. Meter #3 is now calibrated.

Calibrate Meter #2 for losses through the system:

• For now, remove the DUT from the setup and connect coupler #2 directly to isolator #2.
• Turn on your signal generator and amplifier and adjust the power until you read 0 dBm on Meter #1.
• Adjust the offset on Meter #2 until it reads 0 dBm. Meter #2 is now calibrated.

Measurement:
For each Pin observed on Power Meter #1, you can record:

• Gain in dB (Meter #2 - Meter #1, in dBm)
• Pout in dBm (Meter #2)
• Preflected in dBm (Meter #3)
• Return Loss in dB (Meter #3 - Meter #1, in dBm)

If you want to get fancy, you can set up some software to take these measurements for you. The most commonly used is Labview. But you will need someone to write the code for you.

...Next time: Pulsed power measurements.