Problem while testing Phase modulated carrier signal

[afz23]

Hi Guys,

I am facing an unusual problem while testing a analog phase modulated carrier generation system.

The system has following PCBs:

1.TCXO PCB(reference source,20MHz)
2.PLL PCB(2GHz carrier generation,loop bandwidth 20KHz)
3.Buffer amplifier(to provide isolation between VCO o/p and diode phase modulator)
4.Phase modulator(reflection type diode phase modulator) to phase modulate 2GHz
carrier with audio frequencies.

All these PCBs are fed with 15V dc.

While testing its observed that,if the dc line to PLL board is (15V) is disturbed,the side band
level of phase modulated 2GHz signals(fluctuates) changes .
This happens only for audio frequencies within and just beyond loop bandwidth,for audio frequencies much higher than loop bandwidth,this phenomenon is not observed.

What could be the possible reason and the remedy?

Following exercise has already been done to isolate the problem:

1.Agilent Signal source was used as reference ,instead of 20MHz TCXO,the side band fluctuation could be minimized.
2.Buffer amplifier and phase modulator were tested independently with Agilent signal source,and the side bands were rock steady.
3.VCO 15V supply wire was disturbed,but no fluctuation was observed in PM signal.

Above exercise narrows down the problem area to TCXO or PLL board.

What more can be done to isolate the problem and how to correct it?

 

[deba_fire]

Hi,

Did you try changing the pll loop bandwidth and checking this phenomenon?

I wonder when the audio frequencies are on the 15 V dc supply, they some how find a path to the PLL control voltage, and thus affect the sidebands. Is it a integer or fractional PLL?

Thanks

 

[afz23]

Hi,

Did you try changing the pll loop bandwidth and checking this phenomenon?

I wonder when the audio frequencies are on the 15 V dc supply, they some how find a path to the PLL control voltage, and thus affect the sidebands. Is it a integer or fractional PLL?

I haven't changed the loop BW yet,as it is chosen for best phase noise for the locked carrier.
Its a integer-N PLL.