Intermodulation distortion (IMD) measurements are frequently used to characterize the non-linearity of RF components, including ferrite circulators and isolators. Measuring the IMD products of a ferrite circulator or isolator can be difficult. So, care must be taken to avoid the introduction of any unwanted IMD products.
Our analysis revealed a reliable test set up for IMD products where two continuous wave (CW) tones ( f1 and f2) are combined and fed into the device under test (DUT). The resulting output is measured on a spectrum analyzer. In the case of IMD analysis, the third order (and occasionally the fifth order) products are the critical unwanted frequency products to analyze and measure because higher-order IMD products are negligible due to their smaller amplitudes. These can be expressed as:
3rd order products at: (2 x f1) " f2 and (2 x f1) " f1
5th order products at: (3 x f1) " (2 x f2) and (3 x f2) " (2 x f1)
Forward IMD Test Set Up
Figure 1 is a typical block diagram of a forward IMD test stand into a spectrum analyzer. This set up uses Agilent 83712B signal generators and Amplifier Research 100W1000M1 amplifiers. The dual isolators provide >50dB of isolation and K&L's six cavity tunable low pass filters remove system harmonics. A 3-dB hybrid coupler (QH32-0018-N)gives further isolation between the tones.
The directional couplers must be terminated with a broadband cable load with superior VSWR (<1.1:1). A six-cavity notch filter attenuates the fundamental frequencies (f1 and f2) to provide greater dynamic range. It is important not to overdrive the internal mixer of the spectrum analyzer because this could cause clipping. As a result, the input level into the analyzer should be <-23dBm.
Typical analyzer settings are:
- Span = 2kHz
- Resolution Bandwidth (RBW) = 3Hz
- Video Bandwidth = 3Hz
Note that cables, connectors, and test fixtures can all affect the IMD measurement. N-type connectors that are not nickel plated are good choices, and they must be clean and torqued correctly. Cables with superior phase stability are recommended. For example, the test set up in Figure 1 uses M/A-Com's B19X036-0-8585N cables. In order to insure adequate shielding from stray radiation, the test fixture must fully enclose the DUT. It is important to filter out harmonics with tunable low-pass filters before measuring the power. Recognize that spectrum analyzers should be used to measure relative power only (dBc), not absolute power (dBm). Finally, since power meters are only accurate at their characteristic impedance (50 ohms), keep in mind that poorly matched DUTs will introduce inaccuracies.
1. Forward IMD test set up.