Electrical noise is an inherent characteristic of all electronics and power supplies, often stemming from sources such as improper grounding, power supply irregularities, or electromagnetic interference from nearby equipment. As noise levels increase, the instrument's ability to distinguish the desired signal from unwanted interference becomes compromised. This can result in inaccurate velocity measurements, reduced range, and unreliable data processing, as the noise obscures the true signal and diminishes the instrument's effectiveness.

A spectrum analysis is a valuable diagnostic tool for assessing the performance of an instrument by providing a visual representation of its bandwidth and any electrical noise affecting its operation. By plotting the frequency components of the signal, the analysis can highlight noise as distinct spikes in the data. These spikes are particularly problematic when they fall within the instrument’s operational frequency range, as they can interfere with signal detection and reduce measurement accuracy.

The operational frequency bandwidth refers to the frequency range within which an instrument operates effectively, as shown in the Figure 1. The highlighted region represents the bandwidth for a specific signature (500 kHz), with a clear peak in amplitude indicating the primary operating frequency. Noise peaks inside the operating band should be removed, as this will bias the velocities. Outside this range, signal amplitudes drop, signifying it’s out with the operating frequency of the instrument, noise within these regions will not bias the velocities, but the interference might introduce noise that can reduce the instrument range.

Figure 1: Operational frequency bandwidth of a Signature 500 kHz.

This technique also enables the identification of localized issues. For example, if one beam shows an elevated noise floor compared to others, it may indicate electrical noise specific to that beam. Such interference could stem from nearby equipment, improper cabling, or grounding issues. Recognizing these anomalies is essential for diagnosing problems and implementing targeted solutions.

By using spectrum analysis to isolate noise sources and mitigate their effects - such as improving grounding, shielding, or adjusting the instrument's placement - users can enhance the instrument’s performance, ensure reliable data acquisition, and maintain the overall quality of measurements.

Below are a few examples of spectrum plots from Signature instruments:

Figure 2 depicts some examples of spectral profiles (Signature 55, 500 and 1000 kHz), highlighting good spectra displaying no electrical noise within their operational frequency bandwidths in normal-range mode.

Figure 2: Ideal Spectrum for Signatures: (a) 55 kHz; (b) 500 kHz; (c) 1000 kHz.

Figure 3 depicts spectral profiles of the instruments ranging from 55 kHz to 1000 kHz, illustrating some common features of interference and electrical noise within the operational frequency bandwidths (highlighted yellow) for normal-range mode. Electrical noise within these bandwidths can significantly bias velocity measurements and should be removed before deployment of the instrument.

Figure 3: Spectra of signatures with interference features: (a) Frequency peak within the 55 kHz bandwidth; (b) Elevated noise floor in beam 3 of the 100 kHz signature; (c) Multiple peaks across the 250 kHz spectrum; (d) Frequency peak within the 500 kHz bandwidth; (e) Spectrum indicating improper grounding in water; (f) Frequency peak within the 1000 kHz bandwidth. Yellow circles indicate areas of electrical noise.

How to identify and eliminate electrical noise?

The spectrum plot is a powerful diagnostic tool that helps visualize specific noise frequencies and can often provide clues about the type of interference affecting the instrument.

Identifying and eliminating the source of electrical noise can be time-consuming and requires meticulous troubleshooting, especially in complex systems. However, doing so is essential, as electrical noise cannot be filtered out during post-processing.

Common Noise Sources to Check

• Power supply noise: Switching power supplies are frequent culprits, often introducing interference at distinct frequencies. Pay close attention to any periodic spikes that may align with switching activity.

• Grounding issues: Ensure the instrument’s grounding plate is in direct contact with seawater to maintain proper electrical grounding.

• Electromagnetic interference: Nearby transmitting devices (e.g., radios, sonar, or acoustic modems) can emit signals that the instrument inadvertently picks up.

• Mechanical sources: Engines, pumps, and other machinery may introduce broadband noise, which appears as an elevated noise floor across a wide frequency range.

Eliminating the Noise

It's critical to identify and eliminate the root cause of electrical noise, as interference within the instrument’s operating frequency band will bias velocity measurements and compromise data accuracy. During the design phase, it is essential to test each component introduced into the system and conduct thorough bench tests in a low-noise environment before deploying the full setup in the field. During the troubleshooting process, be sure to physically relocate or shield potential sources of interference, and verify that all cabling and connectors are properly shielded and grounded. Additionally, some instruments are equipped with internal noise filters, typically located inside the endbell, that can help reduce specific noise components before data is recorded.
If you're unsure whether your instrument includes these filters, contact your Nortek sales representative or technical support for guidance.