In this post I would like to address some issues related to different sensors types, bearing condition and noise on US measurement It has to do also with my other recents posts. I am not trying to make a comprehensive statement here as just a few readings were acquired. So view it only as invitation to a discussion.

Attached is a file showing US data acquired with different sensors: magnetic and handheld . The SDT170 headset output was attached to the CSI 2130 data analyzer voltage input for recording.

Data was taken on a good and a faulty bearings. Without going into details, the plots look different. The first impression is that magnetic sensor picks up more noise, as Walt mentioned before.

I've heard from Jim that US manufacturers came up recently with better sensors, so that outlined in the file concerns may not be a problem anymore.

Dave

US_sensor_effect.ppt (338 KB, 31 downloads)

David,

I have been saying this for many years: there is NO industry-wide physical standard for ultrasound contact measurements! Not only can't two different manufacturer's products measure the same physical quantity, but different sensors from the same manufacturer (like SDT or UE) will not provide the same meter value using different sensors. SDT's meters (unless the new one is different) are essentialy a microvolt meter with a decibel display. They measure dB referenced to 1 microvolt. They simply measure the sensor voltage output. If one sensor is more sensitive than another, it will produce a higher voltage (more dB) than the other sensor when measuring at the same location.
The other issue, that is apparent in your plots, is that different sensor TYPES (magnetic base Vs hand-held stinger) have different frequency response, so both waveform and spectrum are different. I could go on about why, but that is a longer story. Welcome to the Bizzare world of ultrasound!

Walt

The outputs of some ultrasound instruments produce a very poor reflection of the underlying faults frequencies. This impacts both the sound quality that you listen to as well as the spectral results you measure with the headphone output. So, you need to consider the accuracy of the signal acquisition and processing, as well as mounting. I have done comparisons with a number of instruments, but not the SDT170. Thus, I do not know how well it will produce fault frequencies. I do know that some models that I have tested do such a poor job, they are of little analytical use. As Walt says, there are no standards.