Understanding displayed average noise level (DANL)

Displayed average noise level (DANL)

DANL is the average noise level that is displayed on a spectrum analyzer when the input is terminated with a matched load. Since noise power is a function of bandwidth, DANL is normalized to a bandwidth of 1 Hz. Typical DANL values for modern spectrum analyzers are usually in the range of about -120 to 160 dBm and will vary by both frequency and measurement settings.

DANL limits the analyzer’s ability to measure low amplitude signals, as any signal that is lower than the DANL cannot be measured. You want an analyzer with a low DANL because this means that you can detect lower amplitude signals.

DANL is a function of several factors:

• Analyzer’s inherent noise figure
• Selected resolution bandwidth, input attenuation and reference level
• Use of a preamplifier
• Noise cancellation

Keep reading to learn more about each of these topics and understand how they affect DANL.

Noise figure and DANL

Noise figure is used to quantify how much noise is added to a signal when it passes through a device or component.

If a signal entered a spectrum analyzer and then passes through one of the amplifiers along the analyzer’s internal signal path, the amplitude of both the input signal and input noise will be increased by the gain of this internal amplifier. However, the amplifier will also add some of its own internal noise to the signal. And it’s not just the amplifiers that are adding noise; spectrum analyzers contain numerous other components that add noise to a measured signal. The noise of these components contributes to the analyzer’s overall DANL.

Noise figure is largely a function of the spectrum analyzer’s architecture, and an analyzer with a lower noise figure will typically have a lower DANL.

• Learn more about noise figure

Resolution bandwidth and DANL

Resolution bandwidth refers to the filter or “window” that’s used when measuring signals with a heterodyne-based spectrum analyzer.

This is the user setting with the greatest impact on DANL. A narrower resolution bandwidth reduces the amount of noise energy in the measured signal, and this in turn lowers the DANL. Decreasing the resolution bandwidth by a factor of 10 reduces DANL by approximately 10 dB.

Although reducing the DANL is generally good, narrower bandwidths will increase sweep time, and this increase can be very substantial at especially narrow resolution bandwidths.

Input attenuation and DANL

In most spectrum analyzers, the input signal passes through an input attenuator before being downconverted to an intermediate frequency (IF). This downconverted signal is then amplifier by an IF amplifier. The IF amplifier gain is coupled to the input attenuator setting to ensure that the signal level is properly reported. If input attenuation is increased, the gain must be increased by the same amount. This higher gain increases the amount of added noise and thus also increases DANL, while reducing input attenuation reduces DANL.

The relationship between input attenuation and DANL is usually 1:1. For example, reducing input attenuation by 10 dB will typically reduce DANL by 10 dB.

Although it’s possible to set input attenuation manually, it is often automatically adjusted when reference level is changed. Even when the reference level does not change input attenuation, DANL may still change by up to several dB.

Preamplifiers and DANL

The preamplifier is a component that increases the input signal level. It is usually located after the input attenuator. In general, a higher preamplifier gain lowers the DANL.

You should take care when choosing the preamplifier gain. Increasing the preamplifier gain too far can push the analyzer into “compression,” resulting in accurate measurement results and/or unwanted distortion products. The preamplifier should also have a low noise figure. In other words, it should add minimum noise to the signal.

Noise cancellation and DANL

Noise cancellation is another way to reduce DANL. When noise cancellation is enabled, the spectrum analyzer internally terminates its input and measures its own internal noise before switching back to the input connector. This initial measurement allows the analyzer to subtract its own internal noise from the input signal, thus reducing DANL and increasing measurement accuracy.

Noise cancellation can also reveal signals that would otherwise remain unseen because they are below the noise floor. This improvement can be up to 12 dB in modern analyzers. Of course, for noise cancellation to be effective, a spectrum analyzer must first be able to make an accurate measurement of its own internal noise.

Why isn’t the DANL “flat”?

DANL is usually relative “flat” across the span, but this is not always the case. Depending on the frequency and the span, “steps” or “ramps” may be present. These can be caused by:

• Different hardware paths at different frequencies or bands
• Frequency response correction or calibration

A “non-flat” DANL tends to be more common or noticeable on older spectrum analyzers.