Siglent SSA-3021X Sideband "Angel Wings"?

[nealix]

Hi:
I am fairly new at using this instrument (Newbie Alert), so this may simply be pilot error, or lack of experience.
We were looking at three different benchtop signal generators today at 10Mhz (HP, Marconi, Siglent 2042x),
to look at the waveform, noise, and spectral purity.   In all three cases, the 10Mhz signal from each generator showed
what looks like a pair of angel wings surrounding the main carrier.  On either side of the main carrier vertical spike,
there was a dip, and then a climb to a higher noise level, and then a very gradual trailing off.   But on the Rigol DSA-815
analyzer, all three signal generators had a more traditional/expected gradual rolloff/curve away from the main carrier.
A screenshot is attached.  I tried various values for span, RBW and VBW, including 1/3 ratios for RBW and VBW.
Am I using it wrong, or is this some artifact of the Siglent SSA-3021x?  Or worse yet, is this an issue with my specific
unit and main clock jitter or phase noise?



Neal

[0culus]

What does it look like closer in on the carrier?

I don't know anything about the RF circuits inside that particular analyzer, but you could also try a *very clean* external 10 MHz reference and see if that changes anything. The phase noise of the reference signal is pretty important, for obvious reasons. You could also figure out what the synthesis architecture is like and put signals in on different bands if they exist.

[tautech]

Pretty typical when you use a wider Span.

One I did a few days back using the same fundamental but with 2 Spans after freezing (View) trace 1 (yellow) then narrowing for trace 2 (pink).
SSG3021X source and SVA1032X.

[nealix]

Hi Tautech:

When you say "fairly typical for wide span", is that because of the design of the Siglent?  We did not see this
on the Rigol DSA-815, but I think that unit has worse noise close in to the carrier?  Can you teach me why
you say fairly typical at wide span?   Yes, on my SSA-3021X, it does get better close in to the signal.
See attached photo;

Neal

[tautech]

It's the SSA phase noise, hendorog and I have been chatting about it a bit today.
Study this 2 year old thread:
https://www.eevblog.com/forum/testgear/phase-noise-head-to-head-siglent-ssa3032x-vs-hp-8566b-vs-signalhound-sa124b/

Note, this was before the SSA Plus models were released.

[rf-loop]

Hi:
I am fairly new at using this instrument (Newbie Alert), so this may simply be pilot error, or lack of experience.
We were looking at three different benchtop signal generators today at 10Mhz (HP, Marconi, Siglent 2042x),
to look at the waveform, noise, and spectral purity.   In all three cases, the 10Mhz signal from each generator showed
what looks like a pair of angel wings surrounding the main carrier.  On either side of the main carrier vertical spike,
there was a dip, and then a climb to a higher noise level, and then a very gradual trailing off.   But on the Rigol DSA-815
analyzer, all three signal generators had a more traditional/expected gradual rolloff/curve away from the main carrier.
A screenshot is attached.  I tried various values for span, RBW and VBW, including 1/3 ratios for RBW and VBW.
Am I using it wrong, or is this some artifact of the Siglent SSA-3021x?  Or worse yet, is this an issue with my specific
unit and main clock jitter or phase noise?



Neal

It is nominal to this SA. It is just this SA Phase Noise distribution shape + random noise + signal under test phase noise where is SA own PN is most dominant and small amount of other things.

Near carrier Phase Noise is better than more far from carrier until it again start drop.  If can not drop more far carrier PN but if can drop near carrier PN I think is is good. Look example some nonamed around same price class competitors where PN specifications are close same in data sheet but when you go to 100Hz from carrier  PN may masks all under noise and with this you see lot of more deep...  I can see this is advantage, not disadvantage. Of course it is best if 20kHz  and more from carrier can drop but there is price...

So it give small advantage for look  bit more deep things very close carrier due to fact that RBW shape factor is around 1:5 and so with 1Hz RBW you can go really close carrier with quite far below carrier level  for example find signal generators impurity very close carrier and get example some image about residual FM etc
Also it is well inside specs in this noise figure / and partially phase noise meaning what can see in your image.

Be careful with ExtRef related to Phase Noise. If you do not really need ExtRef for reason to keep perfect frequency and perhaps sync with other equipments, just Do Not use it except if it is real important for freq. For PhaseNoise, Ext ref, even if it is state of art class, looks like can not make it any better due to my tiny experience with this SA but it can do it more bad if it is bad. (In this image I can not see any notable rise or drop in PN from nominal.)

If you are not familiar with SA you can read very good material like example Keysight good application note  AN-150 pdf
It is nearly like material what is good to read for every single SA owner, even bit more experienced or first time user. No need understand all. Of course there they talk also about MUCH more expensive SA things but most basic fundamentals are same.
Of course available is many kind of materials but this one is also theoretically quite high quality and reliable.

https://www.ab4oj.com/test/docs/5952-0292.pdf

Your SA is of course not analog, so you need read what is told about full digital IF SA but also read these basic fundamentals even when your SA is bit different. There is lot of same what is explained this originally HP / Agilent application note
You may be amazed how many things are there explained just as it is also in your SA.