Designer’s Guide Community

Ryan Cheung

Junior Member

Offline

Posts: 22

About the phase noise of the buffer...
Apr 22^nd, 2008, 5:07am

The input of the buffer is a ideal clock with ideal power supply, and its rist/fall time is constant. If we change its input frequency to half of the original one, a 6dB improvement is achieved as I have expected. However, if the power supply is given from a noisy regulator, the phase noise is degrade about 20dB. This time, if we change its input frequency to half, the 6dB improvement is unreachable. There is only 2dB or even no improvement. I fail to figure out why it is like this.

The simulator used here is pss+pnoise.

Thank you!

Back to top

IP Logged

rfmems Senior Member Offline Posts: 121	Re: About the phase noise of the buffer... Reply #1 - Apr 23^rd, 2008, 12:28am try to seperate the AM noise and PN noise, I think AM noise won't improve 6dB if you halve your intput frequency
Back to top	IP Logged

Ryan Cheung Junior Member Offline Posts: 22	Re: About the phase noise of the buffer... Reply #2 - Apr 24^th, 2008, 6:01am rfmems wrote on Apr 23^rd, 2008, 12:28am: try to seperate the AM noise and PN noise, I think AM noise won't �improve 6dB if you halve your intput frequency Thanks for your reply but the noise curve is phase noise. I don't think AM noise is included?
Back to top	IP Logged

rfmems Senior Member Offline Posts: 121	Re: About the phase noise of the buffer... Reply #3 - Apr 24^th, 2008, 7:34am the pnoise simulation gives a combined result of phase noise and amplitude noise. Back to your question, I think when the supply is noisy, AM noise can be dominant. That is probably why 6dB improvement is not seen
Back to top	IP Logged

pancho_hideboo

Senior Fellow

Offline

Posts: 1424
Real Homeless

Re: About the phase noise of the buffer...
Reply #4 - Apr 24^th, 2008, 7:35am

Ryan Cheung wrote on Apr 24^th, 2008, 6:01am:

rfmems wrote on Apr 23^rd, 2008, 12:28am:

try to seperate the AM noise and PN noise, I think AM noise won't �improve 6dB if you halve your intput frequency

Thanks for your reply
but the noise curve is phase noise. I don't think AM noise is included?

Result of PSS/Pnoise is no more than small signal mixing noise. So AM noise appears in Pnoise results.

Back to top

Kita━━━━━━(ﾟ∀ﾟ)━━━━━━ !!!!!
http://www7.plala.or.jp/ungeromeppa/flash/kita.html
http://www.youtube.com/watch?v=mjIxGh55bMM&feature=related

IP Logged

Ryan Cheung

Junior Member

Offline

Posts: 22

Re: About the phase noise of the buffer...
Reply #5 - Apr 25^th, 2008, 10:52pm

pancho_hideboo wrote on Apr 24^th, 2008, 7:35am:

Ryan Cheung wrote on Apr 24^th, 2008, 6:01am:

rfmems wrote on Apr 23^rd, 2008, 12:28am:

try to seperate the AM noise and PN noise, I think AM noise won't �improve 6dB if you halve your intput frequency

Thanks for your reply
but the noise curve is phase noise. I don't think AM noise is included?

Result of PSS/Pnoise is no more than small signal mixing noise. So AM noise appears in Pnoise results.

Thanks again for all of your replies! Wink

Does the AM noise that you mentioned here is the AM-PM noise part? There is no pure AM noise because of the amplitude limiting mechanism. Did you mean that the part of phase noise originating from AM noise will not achieve a 6dB improvement? Why is it?

What I think is the noise could only disturb the circuit at rise/fall time, and this lead to a fixed amount of jitter. If the rise/fall time is fixed, the jitter is fixed. So if the period is double, its phase noise should achieve a 6dB improvement. Am I right?

Thanks!

Back to top

IP Logged

rfmems Senior Member Offline Posts: 121	Re: About the phase noise of the buffer... Reply #6 - Apr 26^th, 2008, 2:11am I think when your power supplye is noisy, assumption of hard limit is not valid. Claims like there is only phase noise is no better than an approximation
Back to top	IP Logged

Ryan Cheung

Junior Member

Offline

Posts: 22

Re: About the phase noise of the buffer...
Reply #7 - Apr 29^th, 2008, 4:12am

rfmems wrote on Apr 26^th, 2008, 2:11am:

I think when your power supplye is noisy, assumption of hard limit is not valid.
Claims like there is only phase noise is no better than an approximation

I have got your idea!

But if I want to achieve the phase noise, how should I configure the spectreRF simulator?

There is a 'noise type' option in pnoise analysis, which default to 'sources'. If I set 'noise type=modulated', the simulator could give a PM noise and an AM noise. Does this PM noise is the phase noise?

If I set 'noise type=jitter', the simulator give a 'tdnoise'. What the relationship between the 'tdnoise' and the phase noise?

Any anwser is appreciated!

Back to top

IP Logged