RF Amplifier Module Test Setup

By Thomas Rottach, RIGOL Technologies EU GmbH

Buzzwords like Industry 4.0, the Internet of Things (IoT), Mobile Computing, and Cloud
Computing can currently be found in many headlines in magazines. The common theme
throughout is the development and rapid expansion of modern communication technologies
rooted in radio frequency (RF) communication. Wireless connectivity is everywhere you look.
You will find it in the manufacturing, in the warehouse, in the office, in the car, at home, at
sporting events, and in medical technology. Even our pets wear RFID chips for identification.

Behind all these applications you will find the technologies like LTE, 3GPP, Wi-Fi,
Bluetooth, ZigBee, RFID, NFC, and DECT. Regardless of the protocol, you will find similar
hardware; a receiver, a transmitter, or the combination of both - the transceiver. Input
(receivers) and output stages (transmitters) typically utilize an amplifier module to help
increase the signal amplitude.

With respect to this it makes sense to have a closer look at the measurements of the most
important parameters of RF amplifier modules. Each application demands different
specifications and requirements. For example, the amplifier which is located in the receiver
path has to be frequency selective and have low noise (LNA). Alternately, a power amplifier
in the transmitter path has to be optimized for maximum output power. Regardless of the end
application, all amplifiers have the same basic parameters like frequency range, gain, output
power, linearity, matching and stability.

Operating Frequency Range and Gain

RF amplifiers have performance that varies with the input frequency. Two of the most basic
parameters are the frequency range and the gain (amplification) for defined frequencies or
over a defined frequency range. In general, a spectrum analyzer with a built-in tracking
generator like the Rigol DSA800-TG or the DSA1000-TG is ideal for these types of
measurements.
Figure 1: Setup for testing an amplifier gain.

The first step is to normalize the output power of the tracking generator. First, connect the
generator output to the spectrum analyzer input using a cable and any adapters that will be
used with the amplifier and use the dedicated normalize function of the instrument.
Normalization is the process of mathematically subtracting the output spectrum of the
tracking generator, through the cables and adapters, from itself. This effectively removes
cable and adapter losses as well as nonlinearities in the tracking generator output. The
resultant curve is a flat line centered on 0dB. Now, changes in the measured RF input will
represent the characteristics of the device-under-test (DUT).

The next step is to connect the amplifier between the generator output and the spectrum
analyzer input. The following figures show the frequency-response curve of an RF amplifier.
With this information, it is possible to determine the frequency range that provides the best
performance as well as the maximum gain.
Figure 2:RF amplifier spectrum output. Note how the amplified signal (purple curve) is above
20dB up to around 1.2GHz, and then the value drops and remains lower out to 3GHz.

Figure 3:A closer look at the range between 900MHz and 1GHz. In this frequency range, the
gain variance (max to min) is measured with 0.4dB.
1dB Compression Point

Data sheets for amplifiers typically report the output power measured at the 1dBcompression
point. This is the point where the measured output power deviates1dB from the theoretical,
ideal output power. This deviation originates from the saturation of the amplifier. Above the
1dB compression point, the gain decreases significantly. Adding additional input power will
not increase the output power. See Figure 4 for a graphical representation of the 1dB
compression point.

Within the small signal range the Output Power[dB] is equal totheInput Power [dB] + Gain
[dB], equivalent to the linear function Y = mX + b. Where:
Y = Output Power [dBm]; X = Input Power [dBm]; m = 1; b = amplifier Gain [dB]
If we display this equation graphically we get a linear line with the gradient 1 which is shifted
up from the origin (0dB/0dB) by the value of the gain.

Figure 4: Typical representation of the 1dB Compression point for an amplifier.

The 1dB compression point can be determined by sweeping the output power at a specific
frequency and then measuring the resultant output of the amplifier. A typical test setup
requires an RF source with a wide dynamic range, like the Rigol DSG3000 series, and a
spectrum analyzer like the Rigol DSA800 series. The output signal of the amplifier is fed into
the spectrum analyzer as shown in Figure 5.

NOTE: Depending on the gain of the amplifier and the maximum allowed input power of the
analyzer, it might be necessary to add an attenuator between the amplifier and the spectrum
analyzer to decrease the input power.

The analyzer center frequency is set to the test frequency and then configured for Zero-Span
mode. In Zero Span mode, the analyzer displays amplitude with respect to time, very much
like an oscilloscope. To achieve a stable and clear display, you can also use the trigger output
of the RF generator to synchronize the RF output with the sweep of the spectrum analyzer.
Alternately, the video trigger of the analyzer could be used.
Figure 6below shows a 1dB compression point measurement performed using aRigol
DSG30303 GHz RF generator and aRigolDSA832 3GHz spectrum analyzer.

Figure 5: Test setup for determining the 1dB compression point.

Figure 6:Input Power (purple line), ideal theoretical output power (green line),real,
compressed output power PA (yellow line), 1dBcompression point (marker 1)

What is the significance of the 1dB compression point?

With digitally modulated signals like the QAM-modulated carrier from a digital TV signal,
we are usually concerned with the average power of the channel. But the ratio between peak
and average power with a QAM signal can bebetween6-13dB.If this ratio is large enough, the
peak power could exceed the 1dB compression point. Applying such a signal to the amplifier
will lead to a distorted output that could cause. Therefore it is crucial to select an amplifier
which has enough space between maximum peak power and the 1dB compression point to
accurately amplify the entire range of the input signal.

Third Order Intermodulation (TOI/IP3)

To make this clearer we first have a look into the time domain. If we apply a pure sine wave
signal to a non-linear part (e.g. a diode) we can see that the output wave is no longer a pure
sine wave. If we now transfer this distorted sine into the frequency domain, we will see not
just one line at the frequency (f) of the sine wave. We will find also frequencies 2*f, 3*f and
so on. Adding a low pass filter with pass band bandwidth lower than 2*f will again remove
the harmonics.

It’s getting more complex if we have more than one single input frequency applied to a non-
linear part. (e.g. at multi-carrier transmission). With this we will get besides the harmonics of
the frequencies also the inter-modulation products. (see figure 7) The mixing products of
second order and also the harmonics can again be filtered out with a feasible band pass filter.
The inter-modulation products third order fall back into the pass band and could not filtered
out. These products disturb the communication channel. That is the reason why especially
these products are unwanted in communications technology.

Figure 7: Second and third order distortion and harmonics

The higher the IP3 value is the better. There is also a rule of thumb that the IP3 point is
roughly 10dB above the 1dB-compression point.
For the IP3-measurement we apply two sine wave signals (CW signals) to the input and
observe the output of the amplifier. The frequency difference lies between 1MHz to 10MHz
and depends on the bandwidth of the pass band and the communication standard and both
signals have the same amplitude level.
Figure 8: Third order inter-modulation (TOI) test using two RF sources, a combiner, and a
spectrum analyzer.

Again, within the small signal range, if we increase the input power of the two signals by
1dBwe will also see an increase of the output power of 1dB on each signal. In contrast, the 3rd
order products raise their power 3dB.

There is a theoretical point where the carrier signals show the same power level as the 3rd
order interference products. (-> see graph below) This point is called IP3 and can be found in
the data sheet of your amplifier. But there is a small issue if you want to measure the IP3
directly. As already above mentioned, we do not have an ideal amplifier and therefore we see
that the output power cannot follow the increasing input power. The reason for this is that the
amplifier is saturated.

Figure 9: Estimated TOI from a typical RF amplifier data sheet.

To determine the IP3-point, we have two possibilities: One is the graphical way (s.Figure 9)
where the IP3-point will be found due to linear interpolation of the transmission graphs. The
other method is to measure the power of the original sine wave signal and the inter-
modulation product 3rd order within the small signal range and then calculate the theoretical
point. This method is precisely the technique used with Rigol spectrum analyzers that have
been updated to include the Advanced Measurement Kit firmware option (DSA800-
AMK/DSA1000-AMK).See Figure 10 for an example of a DSA performing a TOI test using
the AMK option.

Figure 10: TOI measurements on a Rigol DSA spectrum analyzer with Advanced
Measurement Kit (AMK) option.

The next pictures show how the IP3 influences the adjacent channel power (ACP)
measurement. In the example we use a digital modulated signal with 5 MHz channel
bandwidth. If we compare the two pictures we see immediately the difference. Figure 11
shows the original signal, with no TOI interference. Figure 12 shows the standard effects of
excessive TOI: The original signal now has additional shoulders which can disturb the
neighboring channels.
Figure 11: Adjacent channel power (ACP) measurement of a digitally modulated signal with
no TOI influence.

Figure 12: Adjacent channel power (ACP) measurement of a digitally modulated signal with
TOI influence. Note the increased signal levels on either side of the original signal.
This is the effect of adding TOI from an adjacent signal.

Summary

Datasheets can help to designers select the proper elements for a particular circuit design. But,
sometimes, components can have performance outside of the design parameters. Verification
of specific part performance can be performed using a few simple tools and test setups. With
the use of the RF signal generator series (DSG3000, 3 or 6 GHz, with or without IQ
modulation option) and the Spectrum Analyzer portfolio Rigol is already able to address
applications within the communications market. To complete the test setup you can add one
or more remote controlable power supplies (DP800 series) to power up your device.
Additionally Rigol can offer a 5 ½ or 6 ½ digits multimeter to measure the power
consumption or quietescense currents.