K. Hall

Unspecified LISN parameters can cause emissions to appear greater than 10 dB high

Experimental findings of the testing of two different products using FCC/VDE LISNs from several manufacturers are summarized. The initial problem, inconsistent measurement of emissions, was noted on one UUT; while measurement of emissions on another UUT were consistent. Modifying the grounding of LISN VA Hash 2 allowed more reproducible measurements compared to LISNs VC or VD which meet the impedance specifications and had a low internal ground impedance.<<ETX>>

An update on comparison of site-to-site measurement reproducibility using HP site reference source

In this paper, the authors report the results to date of the current round of reproducibility testing that started in August 2000 and is expected to finish by July 2002. A total of 33 sites have participated in this study, which includes nine 10 M semi-anechoic chambers, sixteen 10 M OATS, and eight 3 M sites. Of the eight 3 M sites, five are 3 M semi-anechoic chambers, and three are 3 M measurements on 10 M sites. The reference site chosen for this year was Liberty Labs in Kimbleton, Iowa, USA. Of the 33 sites, 25 performed at least 4 measurements, with two completing 20 measurements over several days. Statistically, 95% of the standard deviations of the 10 M sites was within +/- 1.6 dB as a maximum from 30 MHz to 1 GHz, with greater than 80% falling in the 0.3 to 0.6 dB deviation for measurements made on the same site over a period of days. Test result differences in repeated EMC tests are partially caused by physical characteristics of the measurement sites and changes or differences in the measurement chain. This program meets the requirements of A2LA and NVLAP for site-to-site comparisons.

Tests with different IEC801.2 ESD simulators have different results depending on product sensitivities

Electrostatic discharge simulators, (ESD), from different manufacturers yield different results while testing products to the same standard, IEC801-2 (1991). Several different ESD simulators were used to illustrate the different results that can be obtained. All the ESD simulators complied with the IEC801-2 (1991) contact discharge pulse calibration requirements when tested with a 1 GHz oscilloscope. Two products used met the performance requirements with at least one of the simulators, but failed with others. Furthermore, variances of greater than 4:1 where experienced in the testing. The paper discusses the two following results: (A) with an Hewlett-Packard 2 GHz single-shot digitizing oscilloscope and a 1.2 GHz low pass filter the high frequency differences between the ESD simulators are shown; (B) with an HP 2 GHz digitizing oscilloscope and a Barth TEM horn antenna, it is shown that differences in the radiated field from different simulators exists. If the IEC standard specifies the events and methods of measurement more rigorously, ESD simulators from different manufacturers would perform the same. Unfortunately, the radiated fields methods of measurement and performance are completely unspecified and different.

NSA using EXCEL '97 with GPIB allows making measurement and analysis less tedious

Performing normalized site attenuation (NSA) measurements and reporting the results is a tedious process. Usually the data is obtained manually or with a program developed using an instrument control language. Analyzing, comparing and reporting the data usually requires importing the current and past data into a spreadsheet to perform the needed calculations. The theoretical NSA values were normally taken from tables buried in different standards. Up to now the tables in various standards had errors which required mitigation. Calculating NSA on the fly removes the potential data entry and interpolation errors. With the release of Excel 97 the process is streamlined without the need for additional add ins. Microsoft /spl reg/ Excel 97 includes Microsoft Visual Basic that can be used to control GPIB, serial and general purpose input output devices.

Report on an inter-laboratory comparison of EMC antenna factors on antennas used above 1 GHz

An inter-laboratory comparison of greater than 1 GHz antenna calibration was started in September 2002, managed by Hewlett-Packard. This study was started because data from six different calibrations of a horn antenna varied more than expected by National Physical Laboratory (UK). The variance was also more than that obtained in the very well defined and controlled study of an X-Band Horn antenna (Stubenrauch, 1996) study piloted by NIST. Initially three antennas were being circulated by ETS-Lindgren, Liberty Labs and NIST (US), Schwarzbeck, Rolf Heine, Rohde-Schwarz and CONFORMITAS (Germany), Austrian Research Center (Austria), NPL and Communications Research Laboratory (Japan). Four laboratories have completed the calibration. Preliminary results show worst case deviations between laboratories to be greater than 5dB. Some of variations appear to be abnormal spikes that do not follow the normally expected antenna trend and need to get resolved before supplying factors to a user/customer. One of the laboratories provided antenna pattern data that supports findings by Michael D, Foegelle (2002) and Christian Burns et al. (2003).

CISPR 22: 1997 conducted emission on telecom ports products pass/fail depending on test method

This paper analyzes test result differences for emissions testing in accordance with CISPR22: 1997. Emphasis is laid on unshielded twisted pair cables. It shows potential errors due to imbalance in coupling of the capacitive clamp to the wires of a differential pair and the influence of the common mode source impedance on current clamp and ISN measurements.