Analyzing Mixers Using Intermodulation Distortion Analysis
Mixers are widely used in RF applications. Their main function is to simply multiply two signals
together, typically the LO (local oscillator) signal and the RF (radio frequency) signal. Multiplication
is accomplished by passing the signals through a nonlinear stage. The resulting output waveform
contains signal content at the sum and the difference of the LO and RF frequencies.
One of the many mixer topologies is based on two diodes. This type of circuit is known as a single
balanced diode mixer. Because the inputs are balanced it provides good rejection of the input signals
at the output.
This type of mixer is singly balanced because it does not isolate the two input ports. Signal from the
LO node may leak onto the RF node (or vice versa), producing intermodulation distortion. However,
for many applications this circuit operates quite well. A double balanced mixer can be used instead if
input leakage is a problem.
In order to assess the degree of IM distortion we can run Intermodulation Distortion analysis from
the Analysis menu.
Intermodulation Distortion analysis is a type of transient analysis. It applies two sinusoidal signals
to the named input source with user-specified non-harmonic frequencies and amplitudes, and then
measures the resulting second and third order distortion in the specified load resistor using the
HARM function.
The figure below shows an example of such a diode mixer.
To assess its intermodulation distortion, select the Intermodulation Distortion analysis option
from the Analysis menu. Its analysis limits menu looks like this:
The system will apply the specified two frequencies, 400MHz and 450MHz, to the source named VRF. Its voltage
will be log stepped from 50mV to 1.0V. The VLO source will use the frequency and voltage level specified in
the schematic (1 Volt and 500MHz)
Press F2 to start the run and in a few seconds the following plot will appear.
This shows a plot of the H1 harmonic in blue and IM3 (3’rd order intermodulation) in green. The
program will find the region on the H1 plot where the slope is closest to 1.0 and also on the IM3
plot where the slope is closest to 3.0. It will then draw lines of slope 1.0 and 3.0 respectively, and
compute where they intersect. That point of intersection is the IP3 point. It also computes and
marks the point where the H1 curve drops by 1dB from the ideal slope=1 line. Both of these are
important figures of merit for a mixer.
In this case the values are P1dBm=2.6 and IP3=18.4.
The second plot window holds plots of the waveforms which are used to compute the distortion
and harmonic variables. The program applies the FFT to these basic waveforms and computes the
necessary distortion plot values. This "working plot window" looks like this.
Here we've plotted the following waveforms.
HARM(V(RL)) The frequency spectrum of the output voltage waveform V(RL)
V(RL) The output waveform
HARM(I(VRF)) The frequency spectrum of the input current waveform V(RL)
HARM(V(VRF)) The frequency spectrum of the RF input voltage waveform
HARM(V(VLO)) The frequency spectrum of the LO voltage waveform
Not all of these are necessary. The later two are for information only to show what spectra are actually
applied during the run.
