Best Practice Motor Management & Return Energy Savings, Implement Electrical Preventative Predictive Maintenance.
If an Inverter is Switching at 4kHz why is the Bearing Current in the Mhz range? An impact (IGBT Current Switching) creates a Ring Down (Yellow) and the resultant FFT (Red) is in the Mhz range. The resonant peak of these Bearing Currents depend on cable characteristics and this is why Grounding, Cable Capacitance is so important. That Impact in current hits the Motor Winding and reflects, often doubling voltage which decays Motor Insulation. (Some Motors are Inverter Duty with better Insulation). EMF Cores attenuate above 10kHz attenuating the Ring Down protecting the Motor. The Root Cause is this Ringing, so fixing grounding issues over Brushes or Insulated Bearings makes sense.
Be interesting to see the scope and results after grounding complete.
Motor chokes and sine filters have great effects on this especially for longer cable lengths seen on applications like borehole pumps and conveyors. Anything over 50 meters I would say. We also offer a solution called Motor Tamer to reduce the reflective wave effect. Not to mention the longer the cable lengths without output mitigation increases the input side EMC effects also.
Insightful
Don't forget to short the AC power cables length as much as you can, specially with fast swtching devices!
There’s another reason to slow down the edges of power conversion (or motor control) switching and that is a reduction in radiated and conducted emissions. I’ll post an article on this shortly.
In electric vehicles, the inverters are clocked at 4 kHz and even higher. The voltage is well over 100V and over 400V. The interference spectrum starts at 4 kHz and contains harmonics of up to several 100 MHz. The harmonics decrease with increasing frequency. Now preferential resonances occur due to housing openings, PCB conductor track lengths and cable lengths. They are often strong in the range from 50 MHz to 250 MHz. It is therefore not surprising that the high pulse voltage peaks at a distance of 3m or 10m are over 40-100 dBuV/m and exceed the interference emission limits...
Here is an example of power conversion system ringing and how it translates to peak EMI in the frequency domain. While this example case uses a GaN switch in a buck converter, the principle is the same for MOSFET motor drivers, which can easily switch in nanoseconds. https://www.eeworldonline.com/characterize-emi-from-dc-dc-converter-ringing/
What type and brand of scope are you using for this analysis
Simply "slowing down" sharp pulse edges would greatly reduce any such artifacts: https://1ad86048-c466-4cd0-a9d6-a305587152ec.filesusr.com/ugd/53be97_373ba42f0940415481b472d5cf9c19d0.pdf
CMO & Co-founder at Personizely | We've helped 2,500+ brands generate over $300M in attributable revenue
1moMakes total sense... fixing the resonance at the source beats band-aid solutions every time.