Conducted & Radiated Noise

Hi,

I ve got a requirement for a PLC module such that the noise should be close to -150dBm/Hz over the frequency range 2-100MHz

checking the EMC test procedures and guide lines .. i feel that upto 30MHz conducted noise is measured & above that radiation noise is problematic

Also all the graphs for conducted noise are showing dBuV Versus Frequecny & for radiated noise it is shown as dBuV/m versus frequency.

1) how can i convert the -150dBm/Hz value to refer in the dBuV versus frequency graph..? simply add 107?

You cannot convert spectral power density to usual emission power limits without additional assumptions. For discrete frequency interferences, it doesn't seem to make sense at all.

I'm not aware of spectral density numbers used in a relevant EMC standard, but this may be my ignorance. Can you clarify where the -150 dBm/Hz specification comes from?

Thank you very much for the detailed reply.Actually my intention is to design an AC-DC power supply for some Power line circuits. The same AC power input connector to the AC-DC converter will be used to transmit the PLC data by some circuits which is powered from this AC/DC Converter.My requirement just mentions the power supply noise should be -150dBm/Hz in the 2-100Mhz band. I'm not sure whether it means AC input side conducted /radiated noise , but I'm most probably assuming it to be the output side noise power spectral density, since there will be some analog sections there..and also the power line transmitting power must obviously be high to get affected...(though i've no idea how to verify the same for the time being) .Any comments are most welcome..

Hi there,

The dBuV can be calculated from the power dissipated on the resistor at the end of the line. Most of the time, unless something extreme required, GTEM cells are used and they have commonly 50 ohm matching. For conducted noise the same applies. The reason for using dBuV is that the measurement setup for Conducted Susceptibility and the Conducted Emissions test setups provide dBuV outputs. For Radiated Emissions or Radiated Susceptibility if you use an antenna you can get dBm results also. And you can get all of them with some calculation .

You are right that at lower frequencies the conducted EMI or EMC is more important but for military applications they have lower frequency radiation tests also. Like RE101 test described in MIL-STD-461F (I'm not sure if there's a new revision. This is the last I've read.)

Also one last point is that if you do not have proper shielding or filtering you are most likely to fail. At least that is the case for military applications. The filters at the power lines are extensively used.

It would be most likely reasonable to refer to conducted emissions for the full frequency range, because it can expected as the primary interference path in a PLC application. As said, the power density specification is meaningless unless specifying additional conditions like minmal measurement bandwidth. Otherwise a discrete interference line has to be counted as "infinite" power density.

It would be most likely reasonable to refer to conducted emissions for the full frequency range, because it can expected as the primary interference path in a PLC application.

Click to expand...

So in you opinion the -150dbm/Hz will most likely be specified for conducted emissions right?? Can you help me little bit by specifying where these conducted emissions are likely to put more disaster..
Like whether the on the modulated transmitted(received) data that is being coupled into(and from) the power line (Line & Neutral)? or on the analog line driver(receiver) circuits (powered by this AC/DC converter) which generates(receives) the modulated signals coupled through the transformer to (from) the power line? or on both the input and output of the AC/DC converter??

As said, the power density specification is meaningless unless specifying additional conditions like minmal measurement bandwidth. .

Click to expand...

Do you mean here that we need the specification of the measuring filter noise bandwidth (RBW) of the spectrum analyser to make any useful assumptions?

Otherwise a discrete interference line has to be counted as "infinite" power density

Click to expand...

Can you please elaborate a little bit i didnt get that (may be my basics are quite bad) "discrete interference line" does it mean a single frequency component?