Phase detectors noise

What phase detector?

What system?

Why isn't reducing the bandwidth allowed?

Where's the end of the last sentence?

(1) Reduce the noise on the input.

(2) Use better transistors in the phase detector.

(3) Change the type of phase detector. The single flip-flop is about the worst for noise.

But it sure has a lot of gain.

John

Andy replies:

Assuming you are talking about something like a frequency synthesizer application :

You might try filtering the B+ to the phase detector . If there is any B+ noise, it will translate into phase jitter .

Also, assuming you are using a 4044 type, which is a type of three state phase detector, you might consider using a faster logic family.

If you are talking about "dead zone" types of noise, you may be better off putting a slight leakage on the input which will raise the sideband level, but is easier to filter out because it is well above the loop bandwidth.

If you are using an op amp for an integrator, you should use a low noise op amp. If you are using ANY active devices between the phase detector and point of useage, it should be low noise, and the B+ heavily filtered.

Sometimes using a differential phase detector helps. Depends on where the noise is coming from...

These are just a few ideas I have used over the years. Some help, some don't. Good luck.

Andy in Eureka, Texas

depends what phase detector you have, an analogue PD such as a mixer/demodulater has less noise compared to a digital one, generaly increasing the signal into it should increase the SNR, this means a digital one will benefit from faster edges as noise can cuase premature triggering on a slow edge, an analogue one will benefit from larger output with the same noise but better SNR, but you must consider effects of distortion. usualy the noise contribution is greater from previous stages, however lossy phase detectors can worsen the overall noise performance with weak signals.

Colin =^.^=