Bench PSU output not floating for AC

[Rolo]

I have a Delta powersupply on my bench that I use for all my projects. It has two caps (10nF and 100nF) from minus out to ground.
See attached section of schematic. With this configuration the PSU output is not really floating for AC.
I was measuring with my scope on a VFD filament driver that switches at 100 KHz, the wave was distorted, because the AC found the path to ground via the scope probe.
Trough these caps I assume. The circuit was not connected to USB or other devices. I did see Dave's tutorial on grounds and scope measurements.
So what's the function of these caps and does every PSU has (or need) them ?
Should a bench PSU not be really floating to avoid these issues ?

[David Hess]

If it is a switching power supply, then those might be included for noise filtering.  My floating bench supplies include no capacitors from the outputs to ground.

Even without capacitors to ground, the common mode capacitance will be 10s to 100s of picofarads through proximity of the circuits and the transformer.  Special "high isolation" designs are used to lower this.

[Rolo]

It's linear with thyristor pre-regulation. 30V/3A.

Verstuurd vanaf mijn ONEPLUS A3003 met Tapatalk

[Rolo]

This is the CUT (Circuit Under Test  )


This is the measument using the Delta PSU with caps :



This is the measument using a 5V floating PSU  :



For me the Vrms value was important and you can see that there is a big difference in the results.

[Ian.M]

No one experienced expects a bench supply to be fully floating at higher frequencies.  Fully floating, low capacitance isolated supplies are $EXPENSIVE$ and typically lack many of the features one expects from a bench supply.

You should have used a differential probe, or a 'poor man's differential probe' (waveform arithmetic, Y1-Y2 using identical probes, with Y2 gain and compensation trimmed with both probes on the 'hot' switching node to null out the common mode error as much as possible), or temporarily moved the center tap ground to one end of the filament or used a battery supply, with the whole circuit sat on an upturned plastic box to reduce the capacitance to ground.