Hi. I am looking for answers or hints why there is such change in Uv when swithching frequency is changed.

I was using SDT270 ( Ultrasound )

Please review the spectrum and waveform soundfile

Look forward to hear some brainstorming.

All best



Audio (1)
Files (1)
Last edited by Gisli A. Gudm
Original Post

Gisli, I am sure that there is a much more technical response then what I have to offer, and I will leave that to someone else, but I will give you my non-technical reason for the change in dBuV when changing frequency.  I hope this helps.

I will start by saying that SDT does have the capability of using a variable frequency mixer from 1 kHz to 128 kHz on every 270, and that there are some limited applications that I do recommend changing the frequency for, but for most applications including bearings, there are some very basic reasons I do not recommend doing this.

One reason I do not recommend changing frequency is simply when you do change frequencies you are only changing the frequency of the receiver, not the frequency that the crystal is transmitting at. The analogy that I like to use is when listening to a radio station that is transmitting at 96.5, when you dial your receiver into 96.5 the signal is very clear. When you start changing the frequency on the dial, the signal becomes garbled and distorted. The same is true with ultrasonic’ s.

Using the natural frequency of the Piezoelectric crystal will generate the most accurate, clearest signal and highest amplitude you will be able to find with that crystal. Unless you change the frequency of the Piezoelectric crystal itself for different applications, you will be losing clarity and intensity.

One of the strongest tenants of SDT is repeatability. From your question, I can see you are aware that the amplitude on your ultrasonic receiver will vary dramatically if you are using a consistent ultrasonic source and change frequency. If you are trending equipment over time and you are not extremely accurate with the frequency that you are recording at, you may have dramatic differences in the readings, which could give you a false indication.

The bandwidth on the270 equipment is a very narrow 2 kHz, which allows the user to take advantage of the characteristics of ultrasound around 40 kHz, which is rapid attenuation. High-frequency, short wavelength signals of ultrasound attenuate very quickly with distance from the source. This is one of the things that makes ultrasound very useful. Using structure borne or airborne ultrasound, it is very easy to identify the source.  Also, the higher the frequency, the lower the penetrating power. When listening to rotating equipment, specifically a bearing or gearbox, mechanical sounds that are generated elsewhere in the rotating equipment train will not be transmitted well when contacting a specific point and listening at higher frequencies with a narrow bandwidth. This allows the operator to get information specifically from the asset that he is interested in and not parasitic signals.

We have designed our resonant frequency contact probes for optimum response to 38.4 kHz, which is the natural frequency of the Piezoelectric crystal we use. As you may be aware, the SDT handheld contact sensors are only available in three lengths to optimize the resonant frequency that we listen at. It is all about repeatability and accuracy of the sensors.

In summary, the SDT 270 has the capability to allow the operator to listen at a large spectrum of sound (1 kHz to 128 kHz) but always defaults to a frequency of 38.4 kHz, which is the peak response for the Piezoelectric crystal that SDT use. Using the same frequency guarantees accuracy and repeatability.






Thanks Paul : )

Soo....these changes are not necessarily because of friction (which produces ultrasounds ) due to the frequency converter. They are known to reduce bearing life.....

I am still confused but understand ( almost what you are saying. What's confuses me is the fact sdt only listens to ultrasounds due to friction ( but those sounds which can be heard from frequency converter in sdt are not from friction or turbulence or what ? &nbsp

Am I totally lost or should i just ignore this and say it is trouble hearing bearing when motors are controlled by frequency converters...at least in this case. ( unless the switching frequencies are kept the same through the years during CM )

Sorry for my bla bla : )

All best





Last edited by Gisli A. Gudm

Hello Gisli,

thank you for sharing this important data with the community.

Varying the switching frequency can have a significant impact on the vibration and ultrasound data from your motor. There are three thoughts which I would like to share with you initially:

1. What is the energising signal looking like?

2. What is the switching frequency doing?

3. What is happening with the bearings?

The energised signal can best be described as a chopped up DC signal. It is a series of pulses, square waves of varying shape and size. Fourier tells us that a square wave is made up of an infinite series of odd harmonics of the fundamental frequency: 1x, 3x, 5x, 7x, 9x, etc. The higher harmonics will either be directly present, or their effect will be present in your measurement.

What is this signal doing? Again simplistically, that energising signal is telling the motor to start and to stop so many thousands of times per second. It is not unreasonable to project from this that this excitation is going to generate a mechanical response, a rattle of the bars, and that such a response will have a resonant condition. In part, you will be measuring this response.

What happens to the bearings? They fail in the normal way (friction, intermittent clicks and pops, bearing defect frequencies) plus they fail in a special way due to potential defects related to earth leakage currents, capacitive discharges, etc.

In your measurements, you have all of this going on.

Looking at your excellent data, as you increase the switching frequency towards to "magical" 4kHz it appears that you are heading up to a resonant condition. The 5-8kHz data suggests to me that you have gone beyond that resonant condition into a frequency zone where the motor body cannot respond quickly enough and so the amplitude of the signal decreases. In my own work, I once reduced the acceleration signal on a motor to less than one tenth of the initial reading simply by doing what you show here.

Finally, the nature of the bearing defects will continue to develop in the same way. As long as you "Ignore the whining" (a paper I gave at an IMC(?) conference a few years ago.

The absolute value of Vibration or Ultrasound amplitude can be misleading. The RMS value of a time variant signal will change with the sample time. Many instruments do not measure true Peak. All of these parameters are taken care of in your SDT270: there is maxRMS (every 250mS), RMS (user-adjustable acquisition time), Peak and the Crest Factor as well as the Time signal to help you.

I do not want this answer to turn into a paper. If I have explained what you have observed, great. If you need more, I am always available to help either via this board or better still directly on tom@sdthearmore.com


Tom Murphy

Corporate Training Manager, SDT Ultrasound Solutions

Add Reply

Link copied to your clipboard.