Lux, Jim (337C) wrote: >> -----Original Message----- >> From: [email protected] [mailto:[email protected]] On >> Behalf Of Mark Sims >> Sent: Friday, August 21, 2009 3:06 PM >> To: [email protected] >> Subject: [time-nuts] Using cheap sound cards for measurements >> >> >> Just because the cards have the same ADC clock does not mean that they >> will be sampling at the same time. There will be differences in the >> startup characteristics, register programming, etc. that can affect >> just when the input is sampled. They may be sampling at the same rate, >> but the cards could be an indeterminate number of cycles off from each >> other. >> >> > > > I think that's unimportant, as long as the relationship stays constant (or > predictable) over the entire recording interval (many thousands of seconds, > perhaps). You're not really comparing the absolute phase of the two signals, > you're comparing the relative frequency of the two signals, so it's only the > change in phase that you're looking at. > > Which brings up yet another interesting approach. Say you were to use a > digital audio recorder (as a standalone box), sample for however many hours > you want to run the test for, and then post process en-masse. > > I suspect you might NOT want a recorder that does compression, so you're > going to be accumulating some 600 MB/hr for two channels, but these days, > that's no big deal on flash media. > > There's a lot of pretty high quality recorders around (like the Edirol R-44 > or R-4 and equivalent) > > If the 2 signals are sampled at different times then there is incomplete cancellation of the phase noise of the offset oscillator. This effect increases with the delay between the 2 samples. In practice the interchannel sampling delay can be relatively long (hundreds of microsec) with a low noise offset oscillator.
Bruce _______________________________________________ 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.
