Inverter RS 48/6000 injects DC offset in mains

Q&A and troubleshooting
, ,
21
22

Yep. Thanks!

I installed a type B anyway, (back then when we built the house) because there are small children etc.

23

Hello!

Sorry for the late reply,

Did some tests with the scope. However, I’m not very sure if it was set up correctly. I’m not a big scope guru and there are some contradictions on the web how to set up a scope properly for mains DC offset measurement. We used an isolation transformer to power the scope, and measured on the mains. As I stated before, my system consists of 2 pcs Inverter RS 48/6000 Smart Solar, in parallel AC output configuration.

  • one test was with (almost) no load, the offset is around 27mV
  • one with max nominal load (total mains consumption 13kW - this is about 6500VA / inverter - mostly resistive). The offset was around 42mV.
    **
    What is worth observing, that even on full load, the sine wave remains perfect, no harmonics nor distortions. (Once I checked with the scope the sine waveform from the national grid (in Romania), and it looked way more shitty :slight_smile:
    **
    So, it seems @alexpescaru , you were right, the standard DMM measurement seems to be way off (350mV), probably it also measured some noise?

no load
no load1600×1200 144 KB

full load
full load1600×1200 165 KB

1 Like
24

Putting a large AC voltage into a DVM set to a low voltage DC range may give unreliable results because of the way the internal ADC in the DVM works – it may actually be clipping the peaks off the AC waveform so you’re not measuring a real DC voltage at all, just a very small duty cycle distortion in the sinewave. And trying to measure such small DC voltages in the presence of large ones with an oscilloscope is likely to be even less reliable.

To get a reliable result you need to remove the AC first, meaning a passive lowpass filter (multi-stage not 1st order) with plenty of rejection at 50Hz, so maybe with a cutoff down in the Hz region, which means high-value R and C components (RC time constant of 20ms or so) with negligible leakage – then you can use a DVM set on DC to measure the output.

25

Many thanks for getting back with the pictures. :+1:

That for sure, but depends on what’s working in the area and at which hour during the day, because large inductive loads (motors and so) during working hours will give that distorted wave.