I have read that the output of the phase detector(PD) in PLL is a DC voltage proportional to the phase difference of Ref and Output. So then, why do we need an LPF when PD output is DC?
I couldn't picture the output of PD. Can anyone please explain(With a figure if possible) the function of LPF in PLL?

 

Hello,

Wich type of PLL are you talking about?
When you look at the 4046 PLL, it contains a couple of PLL circuits:



Bertus

 

The direct output of most phase detectors is not a steady DC.
The two phase-detectors in the CD4046, for example, both have pulsed outputs so you need a low-pass filter to get DC.
There then is a tradeoff between filter response speed and allowable ripple in the DC signal which affects the response time of the PLL to a change in input frequency.

 

Small phase changes over time around an equilibrium point is OK for most PLLs. Large phase changes around the same equilibrium point are not OK. The VCO will change frequency rapidly, then overshoot, then undershoot and so forth. The LPF is like a limiter on how fast the VCO can respond.

 

I have read that the output of the phase detector(PD) in PLL is a DC voltage proportional to the phase difference of Ref and Output. So then, why do we need an LPF when PD output is DC?
I couldn't picture the output of PD. Can anyone please explain(With a figure if possible) the function of LPF in PLL?

Hi,

You might "read" that because the description of a phase detector sometimes includes the low pass filter as an integral part while when you look at actual chips that do this they dont have the low pass filter built in it has to be added externally.
So in some reads it will say it already has a low pass filter, while other reads will say you have to add a low pass filter.

One way or another there will always be a low pass filter because the digital part of it only outputs a pulse and that has to be converted to smooth DC in order to run the VCO control properly.