List members, This may be the first time I have disagreed with Mr. Price. I think it's just s simple oversight on his part, maybe mixing up transfer impedance and insertion loss.
Insertion loss is, as he states, a measure of the efficiency of the injection clamp. It is usually measured in a calibration jig, just as he says. The error in his calculation is comparing current to potential. That is a transfer impedance measurement that describes a current probe, as opposed to an injection device, and you would do it backward from the way he described it. That is, you would put a known current through the window and measure the output potential of the probe connector loaded appropriately (most of the time 50 Ohms). In this case the impedance of the current-carrying conductor penetrating the probe window is unimportant, it has no effect on the measurement. Using Mr. Price's scenario, with 0.1 Ampere flowing through a 50 Ohm load induced by the injection device clamped around a conductor, the power would be 0.5 Watts. If the potential from a 50 Ohm source applied to the injection clamp that causes that current to flow is 100 dBuV, that corresponds to -7 dBm, which is much less than 1 Watt, so the example doesn't work. But if instead the power flowing into the clamp were 1 Watt, then the insertion loss would be 3 dB. I picked that number for a reason. If instead of the test set up that Mr. Price used, we use the more usual set up of the coaxial test fixture terminated in 50 Ohms at each end, then we have a dilemma as to how we define insertion loss. Do we define insertion loss as the ratio of clamp input power to the power dissipated in or both of the 50 Ohm loads terminating the fixture? By convention the ratio is taken with respect to only one of the 50 Ohm loads. Therefore, BY DEFINITION, insertion loss as measured in the coaxial test fixture as defined in DEF-STAN 5941, ARP 1972, RTCA/DO-160D, and MIL-STD-461E is bounded to be a minimum of 3 dB or higher. If the above is not clear I have slides from old seminars I used to teach that illustrate the set up and the measurement. Ken Javor on 8/20/03 7:20 AM, Price, Ed at [email protected] wrote: >-----Original Message----- >From: Luke Turnbull [mailto:[email protected]] >Sent: Wednesday, August 20, 2003 12:05 AM >To: [email protected] >Subject: Insertion loss of an injection probe > > > >Dear group, > >I have a BCI testing standard that states the insertion loss >of an injection probe must be less than 7dB. Does anyone have >an idea about what they mean by the insertion loss? > >Thanks, > >Luke Turnbull Luke: The insertion loss is a measure of the efficiency of the probe. To test this value, inject a signal into the current probe coaxial port. Let's assume the injected signal level is 100 dBuV (50 ohm coax system). Now put a loop of wire, terminated in a 50 ohm resistor, through the current probe window. Finally, put an oscilloscope across the resistor terminals. If you read 5 volts across the 50 ohm resistor, that means you have a current flowing, in the resistor, of 0.1 amps, or 100 dBuA. 100 dBuV input - 100 dBuA output = 0 dB insertion loss More typically, if you read 2.5 volts across the resistor, that would be 0.05 amps, or 94 dBuA. This would yield an insertion loss of 6 dB. Obviously, you need to specify the input impedance and the load impedance. And measurements are more accurate and repeatable if you use a current probe calibration jig, typically sold by Solar or others. Using the calibration jig, you can calibrate probes in both directions (for current probe factor for emissions & insertion loss for immunity) using a spectrum analyzer with a tracking generator or just discrete test gear. Ed Ed Price [email protected] WB6WSN NARTE Certified EMC Engineer & Technician Electromagnetic Compatibility Lab Cubic Defense Systems San Diego, CA USA 858-505-2780 (Voice) 858-505-1583 (Fax) Military & Avionics EMC Is Our Specialty -- Ken Javor EMC Compliance Huntsville, Alabama 256/650-5261
