Spurs visible on my spectrum analyzer

[Harry Zachrisson]

The sidebands might come from Spectrumanalyzer LO. It does use a PLL.
Although 50Hz seems to be kind of a low frequency to be a PLL sideband.

 

[D.A.(Tony)Stewart]

Radiated Line spikes should be reduced with shielded wire and CM choke for starters. Perhaps be from cap surge currents from mains rectification filter. I assume lid is ON and chassis is grounded. It appears to have external line filter. Also quality of coax or test method used to measure 30Mhz may be a likely cause with a ground loop on probe acting as an antenna and mixing with source to produce sidebands.

 

[biff44]

the first thing I would suspect is a measurement error. For instance, try putting the RF signal thru and Inside/Outside DC block, and see if these 50 Hz harmonics magically go away.

 

[hafrse]

The sidebands might come from Spectrumanalyzer LO. It does use a PLL.
Although 50Hz seems to be kind of a low frequency to be a PLL sideband.

You are right, I also discovered other spurs at 300Khz from mesured CW at specific input frequencies, I suspected the YIG, I used another spectrum analyzer to check the YIG output at 0 span and no input signal to the spectrum analyzer and there it was, spurs are coming from the LO! so now I suspect that the PLL circuit of the LO is picking the noise from the power supply (?) , I have also replaced all the electrolytes on the PLL synth board and the analog control board for the YIG,YTF,mixer, etc... did not help.
so it should be a filtering problem in the power supply, bad electrolytes (?)

- - - Updated - - -

the first thing I would suspect is a measurement error. For instance, try putting the RF signal thru and Inside/Outside DC block, and see if these 50 Hz harmonics magically go away.

The spurs happen to be coming from the YIG output... Visible also at 300Khz on sepcific YIG output frequencies.
Attached is a picture of 300Khz purs at 20Ghz input cw

 

[D.A.(Tony)Stewart]

You already know it is radiated interference, so find out a better way to attenuate the noise from source or target.

" moved power supply from the chassi about 10 cm, all spurs gone"

I suggest a CM choke on the generator coax. (clamshell type.)

 

[hafrse]

You already know it is radiated interference, so find out a better way to attenuate the noise from source or target.

" moved power supply from the chassi about 10 cm, all spurs gone"

I suggest a CM choke on the generator coax. (clamshell type.)

Unfortunatly discovred more spurs at 300khz (see picture on previous post) on specific YIG frequncies (no generator signal involved here ), the 300Khz spurs will not change if the power supply out of the chassi.
It seems that I am having 50hz spurs radiated through air and 300kjz spurs radiated through the power supply lines (?)

 

[Harry Zachrisson]

You are right, I also discovered other spurs at 300Khz from mesured CW at specific input frequencies, I suspected the YIG, I used another spectrum analyzer to check the YIG output at 0 span and no input signal to the spectrum analyzer and there it was, spurs are coming from the LO! so now I suspect that the PLL circuit of the LO is picking the noise from the power supply (?) ,

300kHz spurs on the LO is definitely coming from the LO PLL. You will probably have to live with them.
I don't know the specs on your Spectrum analyzer but it might actually be with in the specified limitations of the instrument, they are nearly 60dB down from the carrier and at 20GHz this might be as good as it can be.
If it's out of spec and it should be lower I don't think you can fix it yourself without the service manual and a lot of time so I suggest you live with them and learn to navigate around them.
You know they will be 300kHz around any signal that is 60dB or more above the noise floor.

The other 50Hz signals I cant say were they are coming from and it will be hard to find the source without more information and photos of your bench setup and device under test.

//Harry

 

[D.A.(Tony)Stewart]

There are lots of methods to eliminating unwanted power conducted, radiated and load regulated noise.

With a sniffer loop probe, scope and a good understanding of the design, I know how to trace them then fix them,

but without your big "picture" of physical and functional components and their orientation CM impedance and CM gain etc, I can only encourage more details.