A Brief Introduction to Ultrasound Measurement

scroll icon Scroll Down

Let’s talk about measuring ultrasound.

First of all, when we measure ultrasound, we measure the amplitude using decibel microvolts, or dBμV, and we’re also looking at the nature of the signal and the dynamic data. What does the pattern look like?

There are many ways we can characterize and quantify the readings. Some of the words and phrases we can use to describe a signal include continuous tone, crackly, intermittent clicks and pops, repetitive content, and much more. This is vocabulary we can use to describe what we actually hear when we heterodyne signatures down to the audible range.

We can also describe the signal statistically through numeric values, particularly root mean square, a.k.a. RMS; the Peak; the crest factor; and kurtosis. These can indicate conditions.

Acquisition Time for Rotating Machinery

Always remember the golden rule of measurement: The way you take the measurement determines the result.

Make sure you take good-quality data. In this context, the duration of the measurement, or the acquisition time, is an important consideration.

If you have slow-moving equipment, make sure there are at least 3–5 revolutions of the shaft so that you have enough repetition to provide meaningful information from the data you capture.

Quantifying Ultrasound

We need a way to gather and trend data that provides an indication of the nature of the fault instead of only using dBμV. RMS, Peak, and crest factor provide a way to describe the signals we are seeing.

For a sine wave, as in the diagram below, the relationship between these three factors is very simple.

RMS = 0.707 x Peak

Crest factor = Peak / RMS = 1.414 (with the Peak and RMS in dB, not in dBμV)


This is, of course, a very simplistic sine wave. The signal we tend to see looks more like this.


In these complex signals, the relationship between RMS, Peak, and crest factor becomes more useful.

The RMS value can be calculated in a number of ways, but the two most common are digital and analog.

With digital, we digitize the signal and calculate the RMS value. With analog, the circuitry rectifies the voltage and makes everything positive, so all of the values that would have been below the baseline are flipped to the top of the baseline. It is therefore essential to understand your instrument and be consistent in how you collect data.

In the signal below, the RMS represents the broad energy in the waveform. The occasional peaks do not contribute greatly to the RMS value.


Peak value, on the other hand, is the highest amplitude recorded in the sample, and it is about 1.414 x RMS in sine waves – but you will almost never measure a sine wave when you are collecting data.

The highest amplitude can occur above or below the baseline, depending on where you find the highest amplitude. In the example below, A is greater than B; therefore, the Peak value is A.


Crest factor refers to the relationship between the Peak value and the RMS ratio:

Crest factor = Peak / RMS.

This is essentially a measure of whether at least one large peak exists. There are a few general rules about this:

  • If the crest factor is around 3, it is just random noise.
  • If it is equal to or less than 4, friction is likely.
  • If it is between 4 and 50, there are impacts.

When performing ultrasound-based lubrication, you usually hear rubbing or friction until you start applying grease. As the grease reaches the bearings, the level of friction will drop. If there is no damage to the bearings, the noise will eventually drop to a “random noise” level. But if you have a severely damaged bearing, you will still have impacts, which will show in the signal.

Assuming the RMS is the same, both of the waveforms below will generate the same crest factor. But the upper waveform has more peakiness while the bottom only has periodic peaks.

ultrasound analysis

If the crest factor is around 4.07 (in this case, RMS: 44 dBμV and Peak: 56.2 dBμV), you will hear what sounds like background noise. There are not many big peaks.


But if the crest factor were to increase to 9.89 (RMS: 51 dBμV and Peak: 70.9 dBμV), you would hear pops and clicks. There are more peaks in this time waveform.

ultrasound analysis

Peak Value

Peak value is important when discussing bearing defects, which generate spikes in vibration and increases in ultrasound. The detection of the Peak amplitude and the use of crest factor are very helpful here. For example, every time the balls in a bearing pass over a spall in the inner race, it will generate an impact, giving you a clear signal that there is a bearing defect.

Measurement Time

How long will you record the time waveform?

When you use an ultrasound device, you can initially scan without recording. But if you hear something, or you want to trend bearing wear, you need to place the sensor at the right location and record the sound so it can later be analyzed. You can listen for changes with your ears, and/or you can trend using RMS, Peak, and crest factor.

But we need to make sure we collect the data for a long enough period of time. For example, the two data windows in the waveform below look like they came from two different machines. Analyzing them will give you very different results.

ultrasound analysis

The data-collecting window was not long enough to capture all of the noise and ultrasound that machine was generating. We need to record long enough that the data we get represents the true health of the machine.