Image Protection

The Conversion tab lets you control image protection parameters in your VSA hardware.

You can choose one of the following frequency conversion options:

Single conversion modes have better DANL, but are not protected from frequency images. To avoid frequency images, the input signal must not have components near frequencies of 2 x IF Frequency (see the hardware data sheet for what IF Frequency the hardware uses, a typical IF frequency is 300 MHz Megahertz: A unit of frequency equal to one million hertz or cycles per second.) typically. If your input signal does have components near frequencies of 2 x IF Frequency away from the center frequency, then use Auto Image Protect or Prefer Double Acquisition to avoid frequency images. Single conversion modes have faster maximum sweep rates compared to the double acquisition or auto image protect modes. Auto image protect usually has faster maximum sweep rates compared to prefer double acquisition.

You can also enable IF dithering for additional image and spur protection:

The sections below describe in detail the frequency-conversion parameters and image-protection methods mentioned above.

Spurious and Residual Responses

Two types of internally generated errors are especially troublesome in any signal analyzer: spurious responses and residual responses. Both are unwanted artifacts in a spectrum display, and both represent errors in the measurement. Residuals are unwanted responses that are unaffected by the presence of an input, meaning they are present with or without a signal at the input. In contrast, spurs are unwanted responses that show up only when an input is present. They scale along with the input's amplitude, though not necessarily in a one-to-one relationship.

The noise-correction methodology described earlier can be used to significantly reduce residual responses in the signal chain. Because the residual responses remain the same in the S+N and N measurements, the noise-correction algorithm will effectively subtract them out of the final result (N represents the analyzer's noise floor and S represents the Signal present at the input).

Image Rejection and Spur Cancellation

Some hardware does not provide hardware pre-selection. This is especially true at higher frequencies (see the hardware data sheet for information about whether pre-selection is supported and at what frequencies). This, combined with the FFT Fast Fourier Transform: A mathematical operation performed on a time-domain signal to yield the individual spectral components that constitute the signal. See Spectrum. stitching methods of Power Spectrum measurements, means images from either a high side or low side downconversion will be readily seen if not suppressed.

The 89600 VSA software provides another category of signal processing when used with some hardware that allows for the suppression of the images and spurs present in a non-preselected receiver. This is accomplished by performing two measurements to identify both the real and image components, and then applying math operations to produce a single, correct result.

The first measurement performs a low-side mix, where the LO is lower than the input signal. The second measurement performs a high-side mix, where the LO is higher than the input signal (the LO is twice the IF of the first sweep). The signal in common between the two measurements is the real signal, and all other signals are images that are rejected.

This combination of low- and high-side mixes is also effective at dealing with internally generated spurious signals. While there are many sources of spurs, one of the most common is low-level modulation on the LO. If the LO has a small sinusoid located 1 MHz away, the front-end mixer will produce two outputs: the input signal translated down by the LO frequency and a smaller, undesired copy of the input 1 MHz away.

Normally, LO-generated spurs are dependent on the LO frequency and will vary as the LO frequency changes. As a result, the unwanted spurious artifacts will appear at different places in the low- and high-side mixes and can thus be rejected, as is done with image responses.

IF Frequency Dither

Another technique available in some VSA hardware to reduce internally generated spurs is IF Frequency Dither.

In an FFT analyzer, an unwanted spur can be created with any pair of tones separated by four times the IF and roughly centered on the center frequency of the measurement span. This problem can be mitigated by using a varying or dithered IF. If the IF used for downconversion is subject to slight random variations from sweep to sweep, then only one problematic condition still exists: when the random four-times-IF separation exactly matches a pair of continuous-wave signals at the input.

When IF dithering is enabled, the false response that is always present in this situation mostly disappears. A false result appears only when the random four-times-IF is at (or very near) the spacing of the two input signals.

The figure below shows the results of dithering for a 2.5 GHz Gigahertz: A frequency measurement which equals one billion hertz. span centered on 5 GHz. From top to bottom, the traces show the following: the captured signal after standard image rejection has been applied, the low-side measurement, the high-side measurement, and the final result after dithering and image rejection. Although it is impossible to completely eliminate the chances of a false response on the display, enabling randomized IF dithering and minimum-hold across a few display updates will, for all practical purposes, ensure correct results.

IF frequency dithering also incorporates averaging to further improve spur protection.  Because of this, measurement speed is slower when IF frequency dithering is enabled.

See Also

Achieving Excellent Spectrum Analysis Results Using Innovative Noise, Image and Spur-Suppression Techniques (Application Note)

Conversion Tab