When diagnosing conducted emission problems, it is helpful to
distinguisg between the common- and differential-mode emissions. This
is desirable, because different power supply components affect
differential-mode noise than those that affect common-mode noise.
Similarly different components of the power-line filter suppress
differential-mode noise, while others suppress common-mode noise.
Knowing which mode is predominate in a product's conducted emission
spectrum basically divides the problem in half. Performing the
conducted emission test as specified by the various EMC regulations,
only measures the total noise and provides no clue as to whether the
predominate noise is common- or differential-mode.

The figure to the left (Ott, H. W., Electromagnetic Compatibility Engineering, John Wiley & Sons, 2009, page 705) shows a product's power supply connected to a LISN (represented by two 50 Ω resistors), with both common-mode and differential-mode noise currents emanating from the power supply.

The noise voltage produced on the phase side of the LISN will be

Vp = 50 (ICM + IDM).

The noise voltage produced on the neutral side of the LISN will be

Vn = 50 (ICM - IDM).

Adding the phase and neutral voltages gives

Vp + Vn = 50 (2 ICM) = 2 VCM.

Subtracting the phase and neureal voltages gives

Vp -Vn = 50 (2 IDM) = 2 VDM.

Therefore, one can determine the common-mode and differential-mode noise voltages individually by adding or subtracting the two LISN voltages respectively.This addition and/or subtraction of the two LISN voltages, however, cannot be done after the measurement has been made, because the measurement is just a magnitude, and does include the phase information. The addition and/or subtraction of the two voltages must be done before the phase information is lost. Therefore, what is required is a network that adds or subtracts the two voltages before they are measured–and, therefore, before the phase information is lost.

One simple network for doing this, often referred to as a LISN MATE (Nave, M. J. Power Line Filter Design for Switched-Mode Power Supplies, Van Nostrand Reinhold, 1991), consists of only five resistors as shown in the figure to the right (Ott, H. W., Electromagnetic Compatibility Engineering, John Wiley & Sons, 2009, page 706). This network adds the two input noise voltages together so that the output represents only the common-mode noise voltage. Therefore, it is referred to as a differential-mode rejection network. To achieve sufficient differential-mode rejection, the resistors used must be 0.1%, and laid out carefully. A picture of such a network is shown below.

In use, the conducted emission test is done first without the differential-mode rejection network, to obtain the total conducted noise voltage (common-mode plus differential-mode). If the product passes, you are done.

If not, the test can be repeated with the differential-mode rejection network placed between the two LISN outputs and the measuring device, to obtain the common-mode conducted noise voltage. The difference between the two readings must then be the differential-mode conducted noise voltage.

** ** **© 2011 Henry W.
Ott
Henry Ott Consultants**

Return to Top of Page |
Home Page |
Search Web Site |

Henry Ott Consultants 48 Baker Road Livingston, NJ 07039 Phone: 973-992-1793, FAX: 973-533-1442

August 29, 2011