That was my post, that you could not average in log mode. It sounds as if the FCC is ignoring the issue, but others know more about that than myself. As I said, averaging log responses amounts to taking a geometric mean, as opposed to a true average, which is supposed to be an arithmetic mean. When the values to be averaged don't vary much, it doesn't make much difference. But if a signal is impulsive with a low repetition rate it makes an orders of magnitude difference.
Here is a numerical example to illustrate the difference. Suppose that one sample takes a value of unity, the next sample is zero. The arithmetic mean is 0.5, the geometric mean is zero. A spectrum analysis example. A pulse with peak amplitude 0 dBm, noise floor of -80 dBm. Pulse 10 us long, 10 kHz rep rate (~10% duty cycle). Average time constant is one millisecond. True average: 0 dBm is one milliwatt, but that is there 10% of the time, thus the average is 0.1 mW, or -10 dBm. Log average: Add the 0 dBm and -80 dBm values and take 1/2, to get -40 dBm. (equivalent to multiplying 1 mW by 1e-8 mW and taking the square root). Results are off by three orders of magnitude, and averaging the log detector makes it easier to pass. In the real world of EMI testing, there likely isn't 80 dB between a signal and noise floor, and the noise floor will set the lower bound on the averaging. If the noise floor were 6 dB below the signal, the arithmetic mean (relative to a normalized signal value of 1) would be 0.75, the average of the log value would be 0.707. This is less than 3 dB difference, if the values are taken to be power, less than 6 dB if the values are taken to be potential. > From: Ralph McDiarmid <ralph.mcdiar...@xantrex.com> > Date: Tue, 24 Aug 2004 08:30:19 -0700 > To: EMC-PSTC@LISTSERV.IEEE.ORG > Subject: RE: Average Detection > > That's my understanding as well. > > Ralph McDiarmid, AScT > Compliance Engineering Group > Xantrex Technology Inc. > >> -----Original Message----- >> From: drcuthb...@micron.com [mailto:drcuthb...@micron.com] >> Sent: August 24, 2004 8:07 AM >> To: cgrassospri...@earthlink.net; >> ralph.mcdiar...@xantrex.com; EMC-PSTC@LISTSERV.IEEE.ORG >> Subject: RE: Average Detection >> >> >> One of the earliest posts to this question (I have since deleted that >> post) mentioned the incorrectness of averaging after the >> logarithmic detector. Averaging after the log detector, I >> believe, does not correctly average the signals. However, >> this appears to be standard and accepted practice. Even the >> common technique of averaging multiple signal measurements in >> decibels, such as five different PC's, is not correct. The >> technically correct method is to convert the measurements >> into linear units, average them, then convert to decibels. >> Any comments on this? >> >> Dave Cuthbert >> Micron Technology >> >> -----Original Message----- >> From: owner-emc-p...@listserv.ieee.org >> [mailto:owner-emc-p...@listserv.ieee.org] On Behalf Of Charles Grasso >> Sent: Monday, August 23, 2004 7:38 PM >> To: Ralph McDiarmid; EMC-PSTC@LISTSERV.IEEE.ORG >> Subject: RE: Average Detection >> >> >> There is some variance as to the appropriate >> VBW setting to make an accurate average measurement. >> As I recall there is (and I have forgotten what it is) >> a ratio that must be maintaied between the RBW and the VBw >> settings. Our Spectrum Analyzer uses a 30Hz bandwidth setting >> (for conducted emissions) but one lab mentioned that a 1kHz >> VBw should be the minimum. >> >> Chas Grasso >> >> -----Original Message----- >> From: owner-emc-p...@listserv.ieee.org >> [mailto:owner-emc-p...@listserv.ieee.org]On Behalf Of Ralph McDiarmid >> Sent: Monday, August 23, 2004 9:45 AM >> To: 'EMC-PSTC@LISTSERV.IEEE.ORG' >> Subject: RE: Average Detection >> >> >> It's my understanding that the 10Hz VBW provides a filter >> time-constant approximating that of an average detector. >> >> Ralph McDiarmid, AScT >> Compliance Engineering Group >> Xantrex Technology Inc. >> >> >>> -----Original Message----- >>> From: owner-emc-p...@listserv.ieee.org >>> [mailto:owner-emc-p...@listserv.ieee.org] On Behalf Of Tom Cokenias >>> Sent: August 23, 2004 6:51 AM >>> To: Ridler, Samantha; 'EMC-PSTC@LISTSERV.IEEE.ORG' >>> Subject: Re: Average Detection >>> >>> >>> Hello Samantha, >>> >>> For analyzers that do not incorporate an average detector, >> the FCC has >>> specified settings of RBW 1 MHz and VBW of 10 Hz for average >>> measurements in other parts of the Rules for measurements >> above 1 GHz. >>> They have been accepting these settings for many years for spread >>> spectrum and now 802.11 radios operating under 15.247 of the Rules. >>> >>> best regards >>> >>> Tom Cokenias >>> >>> T.N. Cokenias Consulting >>> P.O. Box 1086 >>> El Granada CA 94018 >>> >>> tel 650 726 1263 >>> fax 650 726 1252 >>> >>> cell 650 302 0887 >>> >>> At 8:21 AM +0100 8/23/04, Ridler, Samantha wrote: >>>> Hello All >>>> >>>> FCC part 18 requires the use of an Average detector. For >>> measurements >>>> above 1GHz on a HP8593E EMC (Spectrum) analyser, would >>> using a Video >>>> Average of 100 or more be equivalent to taking a >> measurement with an >>>> Average detector? >>>> >>>> >>>> It has been suggested to me by a test house that this is the >>> case, but >>>> I have not managed to find any rationale for this. >>>> >>>> The setup for the Analyser in question during the measurent was as >>>> follows if this helps: SPAN = 2GHz, RBW =1MHz, VBW = 3MHz, >>> SWEEP =40ms >>>> >>>> Many thanks. >>>> >>>> Samantha Ridler BEng MIEE >>>> Compliance Engineer >>>> >>>> ---------------------------------------------------------------- >>>> This message is from the IEEE Product Safety Engineering Society >>>> emc-pstc discussion list. Website: http://www.ieee-pses.org/ >>>> >>>> To post a message to the list, send your e-mail to >> emc-p...@ieee.org >>>> >>>> Instructions: >>>> http://listserv.ieee.org/listserv/request/user-guide.html >>>> >>>> List rules: http://www.ieee-pses.org/listrules.html >>>> >>>> For help, send mail to the list administrators: >>>> >>>> Ron Pickard: emc-p...@hypercom.com >>>> Dave Heald: emc_p...@symbol.com >>>> >>>> For policy questions, send mail to: >>>> >>>> Richard Nute: ri...@ieee.org >>>> Jim Bacher: j.bac...@ieee.org >>>> >>>> All emc-pstc postings are archived and searchable on the web at: >>>> >>>> http://www.ieeecommunities.org/emc-pstc >>> >>> ---------------------------------------------------------------- >>> This message is from the IEEE Product Safety Engineering Society >>> emc-pstc discussion list. Website: http://www.ieee-pses.org/ >>> >>> To post a message to the list, send your e-mail to emc-p...@ieee.org >>> >>> Instructions: >>> http://listserv.ieee.org/listserv/request/user> -guide.html >>> >>> >>> List rules: >>> http://www.ieee-pses.org/listrules.html >>> >>> For help, send mail to the list administrators: >>> >>> Ron Pickard: emc-p...@hypercom.com >>> Dave Heald: emc_p...@symbol.com >>> >>> For policy questions, send mail to: >>> >>> Richard Nute: ri...@ieee.org >>> Jim Bacher: j.bac...@ieee.org >>> >>> All emc-pstc postings are archived and searchable on the web at: >>> >> http://www.ieeecommunities.org/emc-pstc >> >> ---------------------------------------------------------------- >> This message is from the IEEE Product Safety Engineering Society >> emc-pstc discussion list. Website: http://www.ieee-pses.org/ >> >> To post a message to the list, send your e-mail to emc-p...@ieee.org >> >> Instructions: >> http://listserv.ieee.org/listserv/request/user> -guide.html >> >> >> List rules: >> http://www.ieee-pses.org/listrules.html >> >> For help, send mail to the list administrators: >> >> Ron Pickard: emc-p...@hypercom.com >> Dave Heald: emc_p...@symbol.com >> >> For policy questions, send mail to: >> >> Richard Nute: ri...@ieee.org >> Jim Bacher: j.bac...@ieee.org >> >> All emc-pstc postings are archived and searchable on the web at: >> > http://www.ieeecommunities.org/emc-pstc > > ---------------------------------------------------------------- > This message is from the IEEE Product Safety Engineering Society > emc-pstc discussion list. Website: http://www.ieee-pses.org/ > > To post a message to the list, send your e-mail to emc-p...@ieee.org > > Instructions: http://listserv.ieee.org/listserv/request/user-guide.html > > List rules: http://www.ieee-pses.org/listrules.html > > For help, send mail to the list administrators: > > Ron Pickard: emc-p...@hypercom.com > Dave Heald: emc_p...@symbol.com > > For policy questions, send mail to: > > Richard Nute: ri...@ieee.org > Jim Bacher: j.bac...@ieee.org > > All emc-pstc postings are archived and searchable on the web at: > > http://www.ieeecommunities.org/emc-pstc > > ---------------------------------------------------------------- > This message is from the IEEE Product Safety Engineering Society > emc-pstc discussion list. Website: http://www.ieee-pses.org/ > > To post a message to the list, send your e-mail to emc-p...@ieee.org > > Instructions: http://listserv.ieee.org/listserv/request/user-guide.html > > List rules: http://www.ieee-pses.org/listrules.html > > For help, send mail to the list administrators: > > Ron Pickard: emc-p...@hypercom.com > Dave Heald: emc_p...@symbol.com > > For policy questions, send mail to: > > Richard Nute: ri...@ieee.org > Jim Bacher: j.bac...@ieee.org > > All emc-pstc postings are archived and searchable on the web at: > > http://www.ieeecommunities.org/emc-pstc > This message is from the IEEE Product Safety Engineering Society emc-pstc discussion list. Website: http://www.ieee-pses.org/ To post a message to the list, send your e-mail to emc-p...@ieee.org Instructions: http://listserv.ieee.org/listserv/request/user-guide.html List rules: http://www.ieee-pses.org/listrules.html For help, send mail to the list administrators: Ron Pickard: emc-p...@hypercom.com Dave Heald: emc_p...@symbol.com For policy questions, send mail to: Richard Nute: ri...@ieee.org Jim Bacher: j.bac...@ieee.org All emc-pstc postings are archived and searchable on the web at: http://www.ieeecommunities.org/emc-pstc
