Originally Posted by PikeStaff
The purpose was due to noise induced by the power supply to the power head; that noise can be picked up by the spectral analysis and can be interpreted as yet another source of vibration that you're trying to eliminate.
And testing with a power supply is indeed wonderful for bragging rights, and/or absolute results. They're completely worthless for testing comparative to real-world applications, as there isn't a battery in the universe that can hold constant, unwavering voltage and amperage output.
Actually there is a battery that gives constant voltage for a time period. It's a thermal battery. I designed a switcher thats variable from 5V dC to 80V dC 400A, 1% regulation. The switching frequency is 10K hz. and control loop bandwidth is 8khz. Thats pertty high frequency to translate to vibration at any reasonable amplitude. But it will be there. Ballancing a rotor can be tricky. I'm sure that you have ballanced a rotor as close a you can get it. Then in actual running of the fan there is an RPM in which the rotor virbrates, but it get faster and the vibration quits. Usually caused by the natural ressonante frequency of the rotor assembly, You can not get rid of it, but you can move it to a frequency above an RPM you will never see.