Bob, I can't argue with your story.
The fact that some people question failing data; but not passing data is true... and it is a real ethics issue...agreed 100%. In order to "run" with marginal data, you have to question both passing data and failing data. I've worked with a few labs over the last 9 years. At some labs, I don't trust marginal data, at one I do...why? Because at that lab, if there is a frequency that is passing by a dB or so...they do everything they can in order to maximize the signal...they move cables, they raise the antenna up and down, they rotate the table. It's nerve wracking to watch as readings come up time and time again...all so close to the limit. If they are uncertain, they may back out the last modification made to see if the failure comes back. The lab has many emissions sites. I've finished a shift on one site with a unit and gone back the next day to a different site and found that they can just about always repeat the measurement at those marginal frequencies. At that lab, they have the attitude that a pass is a pass; but you really have to earn it; and their measurements at least appear to be repeatable. If the unit passes after all of that abuse...then I don't have a problem shipping it. However, I will go back to the office and try to get some changes in for future revisions...especially if there are planned upgrades to the unit. From: [email protected] [mailto:[email protected]] On Behalf Of Bob Richards Sent: Tuesday, July 19, 2005 9:48 AM To: ieee Subject: RE: EMC Ethics My point was that it failed once by 2dB and then later passed by only 1dB. There is no way, IMHO, that "measurement uncertainty" could be used as a reason to accept the passing data and reject the failing data. This was a test that required maximization of the signal, who is to say that the passing data had been properly maximized? It was the same EUT, the same measuring equipment, the same lab. The only difference was the technician running the test. IMHO, the difference should have been a flag for someone to question measurement techniques. It took me a long time to convince anyone that this was needed, and when it was done it was discovered that there was as much as a 6dB variation, by the SAME technician, making the exact same measurement several times in a row. For the person that did the best job of maximizing the signal, the variation was less than 0.5 dB. There is a general consensus that, when going for compliance data, you should shoot for some margin to cover this. Obviously, the example I gave had no margin and could very well be at risk for failing at another lab. How many times has that happened? I've been in the middle of a "but it passed at another lab" situation. And, of course, they tend to question the failing data, not the passing data, which was the whole point of my original post. Cheers, Bob Richards, NCT. --- "Grasso, Charles" <[email protected]> wrote: > Agreed John > > Best Regards > Charles Grasso > Senior Compliance Engineer > Echostar Communications Corp. > Tel: 303-706-5467 > Fax: 303-799-6222 > Cell: 303-204-2974 > Pager/Short Message: [email protected] > Email: [email protected]; > Email Alternate: [email protected] > > > > -----Original Message----- > From: [email protected] > [mailto:[email protected]] On Behalf Of > John Woodgate > Sent: Thursday, July 14, 2005 11:19 AM > To: [email protected] > Subject: Re: EMC Ethics > > > Bob Richards <[email protected]> wrote (in > <[email protected]>) > about 'EMC > Ethics', on Thu, 14 Jul 2005: > > >There were many instances of an EUT being tested > and failed by one > >technician, only to be sent back to the lab for > another technician to > >test it, on a different shift, and it would pass. > It may have failed by > >2db, then pass by 1db. They always accepted the > passing data. > > What else to do? Many EMC tests are inherently > subject to very large > uncertainty, more than any 'uncertainty-guru' will > admit. There is also > 'variability' - a new term coined to cover cases > where the test results > are **extremely** dependent on the exact test > conditions, such as the > mains harmonic emissions of single-phase rectifiers. > > If there is a 3 dB variation between two successive > tests, more tests > will very likely reveal a 6 dB variation. You don't > usually get a > *better* answer by repeated testing, you get more > variation. OK, if the > distribution is roughly symmetrical, you can take an > average, but often > you get one or two outliers. > -- > Regards, John Woodgate, OOO - Own Opinions Only. > Deadlines are 90% of deadliness. > http://www.jmwa.demon.co.uk Also see > http://www.isce.org.uk > 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 [email protected] 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: Scott Douglas [email protected] Mike Cantwell [email protected] For policy questions, send mail to: Richard Nute: [email protected] Jim Bacher: [email protected] 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 [email protected] 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: Scott Douglas [email protected] Mike Cantwell [email protected] For policy questions, send mail to: Richard Nute: [email protected] Jim Bacher: [email protected] All emc-pstc postings are archived and searchable on the web at: http://www.ieeecommunities.org/emc-pstc
