On Mon, Jan 20, 2025 at 03:51 PM, Kim Elmore wrote:
>Another way to check for linearity is to use pink noise as an input: all frequencies in the pass band simultaneously. A perfectly linear system would reproduce exactly the input spectrum as the output. The difference can >be used to define the lack of linearity and its nature.?
?
>Kim N5OP
Excellent idea...... and I had forgotten all? about it.? ?My Behringer DEQ-2496? ( 24 bit - 96 khz)? mastering processor has a pink noise generator in it.? ?I checked my old notes, and when fed into the XCVR ( via a ton of rack audio gear, used for ESSB)? it's easy to achieve normal, max PEP output out of the xcvr and also the linear amp.? ? On the DC input side, the DC plate current with pink? was exactly 1/2 of a dead CXR.? ?IE:? 50% duty cycle....same as 39 wpm dots.?
On a pep meter, it wavered? just a tiny bit.? ?I experimented and could easily tune up the linear with it.? ? It sounds cool over the air too.? Folks who are unaware, wonder what the hell they are listening too.? I would flip to the opposite sideband, tune up, the swap sidebands again.
Pink noise has equal level per octave.? ?White noise increases? 3 db for each octave.? ?Each octave doubles in upper freq.? ?IE: 100hz - 200 - 400 - 800 - 1600 - 3200? -? 6400 hz.? ?As you go up in octaves, the BW keeps doubling.? And with double the BW, the noise doubles, hence the 3 db increase per octave with white noise.?
To generate pink noise, they start with? white noise..... then run it through a 3 db per octave roll off filter...... dead simple.? ?The trick is, at what freq does the roll off start at ?? ?To do it correctly, it should be done at a very low freq, like 50 hz.?
IMO,? pink noise would be better vs a 2 tone test, when conducting IMD tests.? ?The pink noise uses all the passband? freqs.? On a SA, it looks like a flat line across the passband, then drops off at both ends.?
Only drawback is ( same as a 2 tone)? is everything is in a steady state.? ?But for a quick definitive IMD test,? it works superb.?
?
