Designer’s Guide Community

fast locker

Guest

Question about integrated phase error.
Jun 21^st, 2003, 2:50pm

Hi, all

Hope this time someone can help me with it before I figure it out by myself. ;)

On most of the frequency synthesizer datasheets, a value called "integrated phase error" will be listed. Usually they will specify a range of integration, ie, between 100Hz and 100KHz. My question is why the phase noise contribution from range below 100Hz is excluded?

Suppose we will see an increasing phase noise when the offet frequency decreases around and below 1K, which is usually the case. If we have -80 between 100Hz and 1KHz, -70 between 10Hz and 100Hz, -60 between 1Hz and 10Hz, then we will have equal integrated contribution from all these 3 bands. Why one band is included and the others are excluded?

Or maybe in some applications, the phase noise contribution from offset below a certain value are not important so that they just choose to neglect it?

Thanks
fast locker

Back to top

IP Logged

August West

Community Member

Offline

I'll get up and fly
away

Posts: 71

Re: Question about integrated phase error.
Reply #1 - Jun 24^th, 2003, 1:16am

I see several reasons why one might want to ignore phase noise at low offset frequencies. First, such close in phase noise might not affect the application. For example, with various types of transmitters, slow variations in the phase of the transmitted signals will be tracked by the receiver and the effect of these variations eliminated. In such case, there is no value in reporting the close-in phase noise.

For synthesizers, the close in phase noise contributed by the reference might be higher than the phase noise added by the synthesizer, in which case the phase noise contributed by the synthesizer is of little interest.

Finally, the close in phase noise is often due to flicker noise, in which case the integrated phase error becomes unbounded.

-August

Back to top

IP Logged

fast locker

Guest

Re: Question about integrated phase error.
Reply #2 - Jun 24^th, 2003, 4:08pm

Hi, August,

Thanks for replying. I agree with all of your points.

However, here comes another question, admitted that the phase noise integration should have a lower limit, then what value should it be? 100Hz? 1K? or 10K?

As we all know, transceivers in all application have some accumulated phase error system spec, i.e., 2 degree or less for 11g maximum rate. Among all the phase error contributors, LO and mixer are the two biggest one. So we must have a definate method and spec for measuring LO integrated phase error. Stating integrated phase noise without stating the lower limit is simply not a fair comparison. Integrating from 1K will have definately better numbers than integrating from 100Hz.

Then, is there any industial standard for the lower limit to integrate phase noise?

Back to top

IP Logged