High Frequency Noise > Next Steps

[kcs47]

We have a building at our converting manufacturing site that has had some weird low voltage signal anomalies over the past 1-2 years. It was a new building 24 years ago with similar VFD equipment but only recently started surfacing new problems. It is loaded with 480V VFD's and we upgraded from incandescent bulbs to LED's within the past 5 years. A year or so ago we started getting DH+ plus faults on the associated hardware. This past year we started getting into encoder loss faults on our 1336 Force Drives and Position error faults on our Servo Axis Cards. Last week we finally found something worth investigating a bit further:

Here is a summary of the checks we did on Lines 125 and 124 yesterday (Thursday 6 Jun 2024) and the results:

First, we looked at Line 125. We checked the Voltage and resistance between +24Vdc, Common, and Ground, as measured at the power supplies. The measurements were as expected, which are:

+24Vdc to Common : 24V

+24Vdc to Ground : 24V

Common to Ground : 0V, <10 Ohms


Next, we checked Line 125 several panels away from the Power Supplies, which displayed some Stray Voltage showing up in the Common or Ground of the System.

+24Vdc to Common : 24V

+24Vdc to Ground : 24V

Common to Ground : 0.7V, > 1kOhm


We obtained a long cable and used it to confirm that there was a good connection in the ground system between panels. We connected the cable to a ground in a remote panel, and then repeated the measurements to the Common and +24Vdc at the Power Supply. The measurements were the same.


Then we checked Line 124. Checking the Voltage and Resistance at the power supplies we measured the correct results:

+24Vdc to Common : 24V

+24Vdc to Ground : 24V

Common to Ground : 0V, <10 Ohms


Checking multiple panels further away from the Power Supplies, we saw similar stray voltages showing up just like Line 125:

+24Vdc to Common : 24V

+24Vdc to Ground : 24V

Common to Ground : 0.5V, > 1kOhm


Also note that the incidence of voltage was *not* higher the further we got from the Power Supplies. The gradient on Line 124 appeared to be going the other way, as if the leakage was coming close to the Power Supply panels, and settling out to 0 leakage the further we got from it.

We plan to take our Fluke 225C Scope Meter out and determine what may be on the grounding system causing the impedance on our grounding system. Any words of advice is much appreciated. I am a decent controls technician but have limited experience with Noise, Harmonics, power quality, etc.

Thank you,
-Tyler

 

[petersonra]

I don't see how you can have 1 kohms between the grounded conductor and ground if you have a good egc and grounded conductor run out to these lines.

You should see a small amount of voltage drop on the grounded conductor to egc. It's the voltage drop on the grounded conductor. Ohms law. The voltage drop should be approx I*R. Measure I and calculate the resistance in the grounded conductor. Do the calc. Bet it is close to half a volt.

 

[kcs47]

I don't see how you can have 1 kohms between the grounded conductor and ground if you have a good egc and grounded conductor run out to these lines.

You should see a small amount of voltage drop on the grounded conductor to egc. It's the voltage drop on the grounded conductor. Ohms law.

Your comment is specifically why I reached out to forums for help. I don’t see how it’s possible with a good egc either. Maybe someone else can shed some light on this for us. My only thought was high frequency noise saturating our grounding network once we get away from the power supply. Remember, it reads 0 ohms at the power supply. The grounding conductor busses also read 0 ohms between each other. We read as high as 5 volts dc from ground to common the day prior. The next day it was gone but high impedance between the two remained.

 

[ELA]

Erroneous resistance readings
Taking resistance readings with a DVM on an energized circuit is likely to produce erroneous results. The DVM utilizes low level voltage and current that it sources to calculate a resistance. If there are voltages and currents ( possibly due to noise), in the circuit under test, then readings can be inaccurate. High frequency noise currents can also corrupt DVM voltage readings.
Use of your scope should be more helpful.

 

[kcs47]

That was one of the theories we had, we believe an analog meter would show the solidly bonded ohms appropriately. With that being said we also pondered how the digital voltage reading devices (ie: encoder input on a drive or servo axis card input would be able to handle the same erroneous values our digital volt meter also struggles with) Headed out now. Going to try and find any and all major players contributing to the voltage and or noise causing the erroneous results…

 

[ramsy]

We checked the Voltage and resistance between +24Vdc, Common, and Ground

Sounds like typical low voltage for 24 AWG ethernet cable, which may be on isolated grounding.