EMI filter performance in your system hardware can be difficult to predict using standardized 50-ohm insertion loss measurement. Discrepancies result from the difference between the actual impedance characteristics of the hardware versus the 50-ohm resistive characteristic of a test fixture. It is generally accepted that 50 ohms is too low an impedance for accurate common mode (L/G) performance prediction and too high an impedance for accurate differential mode (L/L) performance prediction. This variance is further aggravated by the common industry practice of using 50-ohm "power splitters" for differential measurements. Differential measurements made with 50-ohm power splitters result in the measurement being taken at an actual 100-ohm impedance level. Measurements taken at 100-ohms can add as much as 6 db between measured differential insertion loss and actual performance in your hardware. To minimize this disparity, FCI uses a true 50-ohm impedance level for all measurements.

In addition to the absolute impedance level encountered, a reactive component further complicates the selection process. It is usually impractical to characterize these complex source impedances in your product. As an alternative, we suggest the following course of action:

First, call on FCI's application experience for filter recommendations to meet your requirements.

Consider evaluating at least two filter configurations and/or impedance levels (SF and LF Series, or SE and SF Series, for example) for each application.

If at all possible determine the optimum filter configuration first and then the performance level required.

Remember that FCI's filters often perform better than the published data indicates. They were engineered for the application and not for a 50-ohm insertion loss fixture.