Didier skrev: > Chuck, > > I am quite familiar with how to calculate a voltage or power ratio in dB, > but refering to the first issue, when you combine two oscillators, does the > noise improve by 3dB?
If you look back at my previous post, when combining the two outputs, the noise raises by 3 dB, but the signal strength raises by 6 dB, giving a net effect of lowering the noise by 3 dB relative the signal strength. You will find this described in many places for transistors. It occurs for instance in the MAT-0x series of datasheets among several other places. Cheers, Magnus > Didier > >> -----Original Message----- >> From: [email protected] >> [mailto:[email protected]] On Behalf Of Chuck Harris >> Sent: Wednesday, December 24, 2008 10:02 AM >> To: Discussion of precise time and frequency measurement >> Subject: Re: [time-nuts] New topics (was Re: He is aTime-Nut >> Troublemaker....) >> >> Didier wrote: >> >>>> Square root of 2 is about 1,414 or about 3,01 dB. >>> I am always confused when considering noise, is it 10*log(p1/p0) or >>> 20*log(p1/p0)? >> A moment's reflection on why the 10 log vs. 20 log, might help. >> >> The conversion from a power ratio to dB is: >> >> dB = 10 log (P1/P2) >> >> Remember that Power = VxV/R, so: >> >> P1/P2 = (V1xV1)/(V2xV2), the R's cancelling. >> >> So, >> >> dB = 10 log [(V1^2)/V2^2)] or, 10 log[(V1/V2)^2] >> >> If we want to express this as a ratio of voltages, rather >> than a ratio of powers (there's a pun in there somewhere ;-), >> we need to take the square root of (V1/V2)^2 outside of the log. >> >> To do this, we need to remember that log[X^2] = 2 log X, so: >> >> dB = 10 log[(V1/V2)^2] = 20 log[V1/V2] >> >> A couple of things to note: >> >> 1) dB's are dB's. 3dB represents the doubling of a power ratio, >> 6dB represents the doubling of a voltage ratio. >> 2) Convention says that if -dB's are loss, and +dB's are >> gain, but that >> is just convention. >> >> -Chuck Harris >> >> _______________________________________________ >> time-nuts mailing list -- [email protected] To unsubscribe, >> go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts >> and follow the instructions there. > > > _______________________________________________ > time-nuts mailing list -- [email protected] > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
