Help:GSM's phase eror

[rhjang]

Hi martinthorn:

I think the 2nd harmonic of PA output cause the VCO pulling is right.
But I still can't get why only first few bits of a burst have the worse phase error.
I have check the bits of one burst, there are many period contain five 0's, but only the beging of burst have worse phase error.
:idea:
Is it possible that at the beging of the burst, the VCO interfered by PA's 2nd harmonic, cause the VCO have abnormal phase noise.
These phase noise make PLL need more time to lock.
So only the beging of the burst have the phase error.
After the PLL locked, the phase erroe vanished.
Could this be the reason?

Thanks

 

[xinxin]

At baseband , the IQ imbanlance or interface could be adjust via software. Philip solytion can do it , but some other can't do it .

 

[martinthorn]

Could this be the reason?

Perhaps. Another cause of phase error at the start of the
burst can be the effect of the TX modulators turning on.
When this happens, just before the burst starts, the VCO
sees a different impedence (through the /2 divider). When
this happens the VCO frequency changes slightly (pulling !).

This can cause the PLL to lose lock slightly and an increase
in phase error is seen, only at the start of the burst. This is
fixed by ensuring the modulator is switched on early enough
to allow the PLL to resettle.

GSM900 modulator has a smaller effect because /4 dividers
may have more reverse isolation.

Let us know when you find the reason for the problems.

Good Luck !

 

[katykaty03]

hi guys, I also came across the problem of phase error. It happened on GSM band but not DCS/PCS. And some difference from yours is, the poor phase error at the beginning of burst only occur when battery drop to 3.4V. The situation is getting more worse when the phone operate in low temperature. Do anyone have idea about that?