RE>Conducted Emissions, average method >> ...with today's goodies, that certainly isn't the path of choice!
Right you are! But many folks do NOT have the computer control even yet! >> I tried to get him to do a stand-alone Peak plot, which >> would have shown him his worst case frequencies. Then, >> he could revisit each high peak, without sweeping, using >> the external QP adapter to accurately measure the amplitude. >> (That way, he wouldn't have to worry about sweep speeds.) Yup! That's what I was describing. >> Are you saying that the averaging process, handled by >> the SA firmware, averages by display units, and not by the >> absolute levels represented by the display? For instance, >> point #1 is displayed at 40dBuV and point #2 is displayed >> at 80dBuV, so the average (the wrong way, if you are using >> a logarithmic display) is displayed as 60dBuV? That does seem to be the case for video averaging. If the video has been passed through a log amp. you will be averaging dB's, not unprocessed amplitude. >> The firmware does not know that 40dBuV is 100uV and that >> 80dBuV is 10,000uV, and that the average of 100 and 10,000 >> is 5,050uV, or 74dBuV? Well, the firmware DOES know that 40 dBuV is 100 uV, and it does know that 80 is 10 mV. After all, you will get the right figure if you select dB while in linear mode. What is does NOT know, is the absolute value of one pixel, which MAY be 40 dBuV, and another, 51 pixels higher (say), which is 80.. .it knows only that these pixels are 40 dB apart, but not what their absolute amplitudes are. You can move them around on the screen with a twist of a knob, and the analyzer wil assign the right dB figure to each of the (is it 256 vertical?) pixels it shows... but it seems not to do the antilog(.05*db) calculation. >> Oh no!? Uh-huh. >> But if you are observing the output of a QP detector, >> you already have to be sweeping very slowly. You don't know >> when you will encounter a peak until you actually hit it. That is true. I have seen someone try try to take a QP of the entire 30 to 200 Mhz range. Would take ALL day. The differnece in techique I mentioned is that some folks QP a narrow range around the peak they noted on a faster, peak-only, scan. Others tune that range manually (in QP). Amounts to the same thing. >> It seems to me that you really should use a Peak detector, >> sweeping fast, and then let a computer or operator decide >> to go back and remeasure, either at a certain peak, or over >> a small frequency span around a peak, ... Yes, we agree. >> using the CISPR BW and QP detector. The CISPR bandwidth is specified for peak readings, too. >> Yes! I had to retire an HP-3325 sig gen, primarily because the >> tech using it had the strength of a gorilla and a strong >> preference for positive tactile data inputs! Did we have the same tech? (Just joking.) It didnt' seem to matter which tech took the readings, though one I knew was into martial arts. The other was just as punishing on the buttons. A co-worker who is still there (I think) got to be expert at replacing the switches. >> Actually, it's a waste of a human to stick someone in that >> kind of a situation. After a few thousand :-( button pushes, >> the average human has begun to glaze over and go into robot >> mode. This mode is then usually terminated with the utterance >> of "Oh no!". It does happen. I have been tremendously impressed by the techs I've known, who were serious enough about their work to tolerate the repetitive button pushing and still catch things that a less patient person might have skipped. Maybe I've been lucky. And that's THAT. Cortland --------- This message is coming from the emc-pstc discussion list. To cancel your subscription, send mail to [email protected] with the single line: "unsubscribe emc-pstc" (without the quotes). For help, send mail to [email protected], [email protected], [email protected], or [email protected] (the list administrators).
