0

The mains of my German flat has 19kHz electrical fields on hot, neutral and! ground wire. I don't have to get rid of them entirely but I want to achieve 20dB reduction while sitting in front of my computer with grounded power supply.

How to achieve this goal?

Hansebenger
  • 111
  • 6
  • 5
    If it's on ground, too, it's most likely the auxiliary carrier of FM stereo picked up by your measurement equipment out of the air. To get rid of it, move to the basement. Caveat: radon levels in the basement are often unhealthy in the long run. – Janka May 18 '20 at 22:32
  • 2
    Why do you have to get rid of them? – JYelton May 18 '20 at 22:50
  • Thanks! Actually "19kHz" is 3 peaks at 16, 20 and 24kHz, don't know why I burnt 19kHz to my brain. Didn't think it would be this important. Is "Frequency modulation synthesis" a good start point to check out what you mean with "auxiliary carrier of FM stereo"? What stands against your statement is that it is strongly depending on daylight, so I thought it would be a solar voltaic. – Hansebenger May 18 '20 at 22:51
  • 2
    "FM stereo" is your good old radio. Stereophonic radio has an auxiliary carrier on 19kHz for the second audio channel. – Janka May 18 '20 at 22:55
  • 2
    If it's dependent on sunlight, it may as well be a solar inverter on a nearby roof. But again, if you can measure it on the ground line, it's not coming by wire. The wire only picks it up better as if you held your measurement tip right into the air. To check that, connect a loose end of some metres of wire to your measurement equipment as an antenna. – Janka May 18 '20 at 22:57
  • It's on ground wire but it's not on heaters, so I still think it's coming by wire. Why shouldn't ground wires transport 16 to 24kHz. They are indeed connected to neutral at the circuit breaker. – Hansebenger May 18 '20 at 23:04
  • Is it also on cold water pipes? (Yes, probably from LED or fluorescent lamps in the building, "solid state ballast," several different ones. Put it on oscilloscope and see if it's a pulse, and synced to the power line peaks. Turn off all lights, unplug all devices in your apartment, see if any of the freq peaks vanish.) – wbeaty May 19 '20 at 01:43
  • Why is the frequency a problem? What do you use to measure it, as you seem to see it everywhere? – Justme May 19 '20 at 05:49
  • @Justme I measure it with an portable not grounded oscilloscope. It's only on power lines, not everywhere, I am doing control measurements of "the void". I simply don't like it so much that I created an account here. – Hansebenger May 23 '20 at 16:52
  • @Hansebenger would you please describe how did you perform the measurement, how the oscilloscope was connected and where to get readings? – Justme May 23 '20 at 16:57

2 Answers2

1

As others stated,the cause of the radio frequency interference (RFI) is likely separate inverters for a solar power system.

You can easily calculate values for an L-C filter for ~20 kHz region. About 10 nF and 6.3 mH, for example,should block most of that RFI coming over the wires. That said, you'd need an effective ground, e.g. a copper water pipe, to connect to the capacitors.

However, since the solar panels have comparatively high currents, you'll also have problems from the radiated magnetic field inducing 20 kHz in a length of wire, such as power cords and USB cables. You probably can't build a mu-metal shielded room, so get as far from the inverters and panels as possible. For sensitive measurements, run from battery (not UPS, which have their own inverters).

DrMoishe Pippik
  • 3,634
  • 5
  • 13
  • I wanted to avoid to tinker with 230V but maybe it's necessary. Needs a lot of preparation then. Infinite Impedance at resonant frequency sounds too nice. My heaters should make a good ground. – Hansebenger May 19 '20 at 21:15
  • @Hansebenger, what problem is the RFI (actally, AFI, audio frequency interference) causing? Shocks from touching the PC? Audible noise? Display issues? False measurements with a PC oscilloscope? Something else? – DrMoishe Pippik May 19 '20 at 22:15
0

Some lighting systems use ‘high frequency’ ballasts in that range. Also may be noise from a switching power supply, or even a variable-frequency motor drive.

Use an AC power line filter to suppress it.

hacktastical
  • 49,832
  • 2
  • 47
  • 138
  • I have tried two AC power line filters theoretically barely fitting without success, maybe you know some more special? 19kHz seems to be a tough number for normal power line filters. – Hansebenger May 18 '20 at 23:46
  • Maybe there’s a fault on ground - line filters shunt some noise back to ground. – hacktastical May 18 '20 at 23:49