As promised.......the Phase Noise Files.....

[Jocko Homo]

Here are some recordings of the phase noise of the oscillator in the SAA7220.......one of my all time favorite parts.

[Snicker, snicker.........]

As I have been ill the last few days, and don't feel like thinking up another clever April Fool's Joke, this will have to suffice. But make no mistake.......these recordings are for real.

Now......if we can get Phred to upload the X-Files theme........

First off.........a recording of the SAA7220 in a CDB-472, stock configuration. The reference oscillator is one of the low-noise jobbies that I have posted here. Note all the line related crap, as this player has tons of ground loops.

[Jocko Homo]

Now......the same chip, but with a 7805, being fed by a 7815 to get rid of the nasty ground loop. Quieter, but you can hear the phase noise on this one.

[Jocko Homo]

As above, but with the 7805 replaced by a low-noise regulator of my design. A little bit quieter, but not much.

[Jocko Homo]

Here is the phase noise of one of my oscillators.

It is recoreded at 10 dB higher in level.

Moral of the story:

Use a good oscillator for your clock.

And a quiet regulator.

Get rid of ground loops.

Make a separate distribution scheme. Do not use the internal buffer in the filter chip to distribute the clock signals.

I'm really sick today.......so "forget you guys.....I'm going home......"

Jocko

[HpW]

red : 472_a
aqua : 472_b
yellow : 472_c
green: Joko

enjoy....

HpW

[ALW]

Jocko,

Thanks for taking the time out from your sick bed to post those - excellent stuff.

Were the phase noise measurements made using readily available methods?

I've been thinking about comparing to a known 'good' oscillator (by mixing, and gathering the data via a sound card), but haven't got my head around the effects of non-ideal behaviour in the reference clock on the output signal (mathematically). I think it's simple and will just limit the ultimate measurement noise floor?

I know self-triggered measurements (e.g. on a 'scope) give misleading results based on timebase settings.

Andy

[Jocko Homo]

The graphs may be somewaht misleading, as the tape recorder used intorduced some noise, and I had to bandlimit things to get the file size under 100k.

The real noise of my oscillator is not even on the scale on my RTA, whereas the SAA7220 stands out like a sore thumb. Remember, I amplified mine 10 dB so that it would not be drowned out by the noise floor of the cheap tape recorder.

The way it is done........a known good oscillator is phase locked to the unit you want to test. Both are then fed into a double-balanced mixer.

I might make some plots on my H-P FFT, but I don't think that they would scan OK.

Yes, I am really am sick today, but I could not let April 1 pass without doing something out of the ordinary. Feel free to ask any questions, but I may not answer right away.

Jocko

[Jocko Homo]

Yeah, the tape deck has really skewed stuff. The noise hump around 180- 500 Hz on the stock unit doesn't show up as well. I'll have to come up with a better way of displaying this stuff.

Jocko

[Jocko Homo]

The top plot is the SAA7220, with the low-noise regulator.

Bottom plot is my oscillator. The hump at 41 Hz seems to be the noise contribution of the PLL, as that is close to the natural frequency that I calculated.

Basically, mine is 20 dB lower noise.

Jocko

[mrfeedback]

I downloaded the 'jocko.zip' file and I could not hear anything, so therefore you are a fraud Jocko.

Eric / - Only jocking.

What sonic differences do you hear with your arrangement ?.