Hello,

I have been trying to test and calibrate an LVDT sensor, but have found some unusual noise. I am using a signal generator to inject a sinewave with the excitation frequency of 3 kHz (specified by the dataset) and 3 vrms amplitude. The cables are somewhat long (8 feet) and a 5-wire model is used. I was planning to use delta/sum to read the position.

Using an oscilloscope, I checked that the input sinewave is exactly as expected, but when I try to read the output from one of the secondaries, there is a noisy component over the 3 kHz sinewave. Doing a spectral analysis revealed a spike at around ~150 kHz on the output signal. What could be the cause of this?

 

Is there a switched-mode power supply/inverter nearby?

 

Is there a switched-mode power supply/inverter nearby?

No, just the LVDT, oscilloscope, and signal generator. It is also strange because the signal looks really clean coming out of the signal generator.

I also use a terminal block to connect the LVDT and the wires from the output of the signal generator.

 

No, just the LVDT, oscilloscope, and signal generator. It is also strange because the signal looks really clean coming out of the signal generator.

I also use a terminal block to connect the LVDT and the wires from the output of the signal generator.

It is coming from somewhere and there are only two basic methods of coupling that are reasonable in this configuration, since it does not appear that conduction is a possibility.

 

The first thing to try is connecting the body of the LVDT package to the common side of the oscilloscope input. Next, try altering the position of the LVDT and see if that changes the amplitude of the noise signal.
The noise may be entering the cable, is the cable shielded? And if shielded, what is the shield conductor tied to??

Do you have a data sheet for the LVDT that specifies the intended amplitude of the excitation signal??

 

Doing a spectral analysis revealed a spike at around ~150 kHz on the output signal. What could be the cause of this?

Long-wave radio transmission? What country are you in?
Lamp dimmer?
Class D audio amp?

 

My fault completely.

I mixed up the oscilloscope power input cable and had it connected with a cable without the ground pin. When I changed it to a grounded cable, the noise disappeared.

It was very strange, because the LVDT cable was shielded and I tried measuring magnetic interference using a loop with the probe ground to its tip and the spectral analysis did not show any significant disturbances below the frequency of 1 MHz.

How come the lack of a ground connection to the oscilloscope made such a difference?

 

Noise is everywhere, and a missing "green wire" ground allowed what was probably switching power supply noise to create a differential voltage where it should not have been. An exact analysis gets rather complex, and fighting electrical noise is part art and part science for even those good at it. Glad that you were able to solve the problem.
Sometimes opening that "ground" connection reduces the noise. This time it was the fix.