Direct YouTube link: https://youtu.be/Fgqc3V7aEUI
Question: Why does my data have 60 Hertz (or 50 Hertz) peaks and harmonics (Figure 1)?
Figure 1: Same measurement location acquired on test object at two different times. The red curve had ground loops present while the blue curve did not.
Answer: Congratulations! You might have a “ground loop”!
The electricity in a wall outlet operates at 60 Hertz (in the United States) or 50 Hertz (in Europe). If your measurement system and transducers are not properly isolated from electrical power sources, these frequencies can contaminate the measurement data.
For example, electrical contamination can be caused by electromagnetic interference when a power cable is laying close to an accelerometer cable. Or fluorescent lights might induce electrostatic interference on poorly isolated wires.
But another cause is something called a “ground loop”. In a ground loop, the measurement system and sensors are grounded at multiple points. Because there is more than one ground point, electrical disturbances can flow between the two grounds and contaminate the measurement signals with electrical noise.
An example of a ground loop is shown in Figure 2.
Figure 2: Ground loop is caused because there are two paths to ground in measurement system.
The measurement and sensor system has two paths to ground:
The accelerometer sensor has direct metal to metal contact with the grounded test object. This allows electricity to flow to and from the structure ground.
The SCADAS frontend is plugged into ground of the electrical mains power (ie, the wall outlet).
To avoid the ground loop, one of the paths must be disconnected from ground. There should only be one path to ground for the entire system.
What can be done?
There are two things that can be done (in order of importance) to avoid a ground loop:
The best solution is make sure the sensors are isolated from the test structure, so that there is no possibility of electrical flow between them. This can be done by inserting an isolation pad between the test object and accelerometer as shown in Figure 3.
Figure 3: The ground loop is broken by using an isolation pad to separate the accelerometer and test object.
Most accelerometer manufacturers also make isolation pads as shown in Figure 4.
Figure 4: Accelerometer with isolation pad (top), accelerometer without isolation pad (bottom)
This ensures there is only one path to ground for the complete measurement system.
Isolated Measurement System
If many sensors were mounted directly with metal to metal contact to the test object, it might be time consuming to install isolation pads.
A quick test to determine if a ground loop condition exists is to take advantage of the built-in battery of the SCADAS hardware. The SCADAS can be unplugged from the wall outlet/mains electricity and run on battery. This disconnects the SCADAS from ground as shown in Figure 5.
Figure 5: By unplugging the SCADAS from the wall outlet (electrical mains), one path to ground is removed from the measurement system.
When doing to, if the 60 Hertz (or 50 Hertz) electrical noise goes away, then there is a ground loop issue. If the battery charge is adequate, the measurement can even be performed. Otherwise, the sensors need to be isolated from the test object.
For more information, see the Knowledge Article "How Long Can Simcenter SCADAS Run on Battery?".
A bit unconventional method to break the path to ground for the SCADAS would be to isolate or eliminate the grounding plug (Figure 6):
Figure 6: Eliminate grounding plug.
This allows the SCADAS unit to still receive power but eliminates the path to ground. This might be useful for sensors that cannot be isolated from the test object.
Troubleshooting a Ground Loop
Note that troubleshooting such a ground loop can be difficult.
For example, if using bee’s wax material to mount the accelerometer on the test structure, depending on how tightly the accelerometer is pressed, the bee’s wax might provide a certain degree of isolation (light pressure), or no isolation (pressed firmly). This can be puzzling as the electrical noise can “come and go” over the time duration of a test.
Even the number of accelerometers can be key – the more accelerometers directly mounted on the test structure, the better the grounding and electrical flow.
Hopefully, this article gives some overall guidance on how to recognize this issue and troubleshoot it.