Re: [Amateur-repairs] Discreet built PLL query HELP please?
|
Be sure to check the cleanliness of your ultimate 1152 MHz LO.? Remember LO noise will also degrade receive performance in direct proportion to the amplitude of the noise pedestal.? In addition, the width of the multiplied noise pedestal will be multiplied by the SQRT of the multiplication factor, in this cast: by 3.5.? -50 dBc is pretty noisey for a quiet LO.?
I set my signal generator to 1152 MHz and took pictures with both the HP8568 and the TINY SA with identical measurement parameters (you can read those from the images).? See the attached images.
Interesting (especially for Eric) that they both show about -70 dBc to the noise pedestal.? I'm impressed with the TINY SA performance in this respect!?
Hi and thanks for all your considered replies, they are most appreciated.
?
It was the fact that the way the operation of the EXOR gate as a phase detector seems to be explained in theory, such that when 'locked' the output pulse should be extremely narrow or non-existent as the system should only need to 'slightly nudge' the filter capacitor to keep it 'topped' up to the desired control voltage.
I was therefore confused to see a 120nS pulse when the system is locked.
I couldn't find any data regarding the input impedance of the VCXO which, if significant, may be discharging the filter capacitor causing the PFD to constantly 'pump' it back up?
Both inputs of my EXOR are driven by digital signals, one from the divided 96MHz and the other from the GPS so both fairly 'fast' square waves.
As far as I can see it doesn't seem to matter if this 120nS pulse exists. The PLL locks securely and to the correct frequency, it was just that I was surprised not to see the EXOR output disappear as the target frequency was achieved.
?
I have looked at the output on the TinyVNA and the noise is >50dB down on the signal, though the 2nd and 3rd harmonics are only 20/30 dB down. I'm not sure if this is due to the buffer amplifier at 96MHz that converts the output to 50 Ohms or a 'dirty' initial signal? Will investigate.....
As I now need to feed this through several multiplier stages up to 1152MHz, it will be interesting to see if these harmonics are suppressed by the time we get to 1152MHz?
?
Thanks for your comments.
Bests
Dave
?
|
