About switching transient issue ?

[mckinson]

Hello . These days , I fonud the switching transient failed .
But the fiailed was not caused for the ramping up or ramping down too fast .
It is at the useful portion of the power ramp .Any guys know what maybe cause these issue ?
Best regards !

 

[vfone]

Is first time when I hear something like this. GMSK is a constant envelope modulation, and during the useful part the burst should be flat. Check with the scope to don’t have some glitches on the ramping voltage. But even like this, conforming theory, 95% of the transient spectrum is due to ramping up and down.

 

[mckinson]

Vfone ,many thanks for your comments !
Do you mean check the signal control the PA power ramp ? Yes ,you are right ,the most switching transient is due to the ramping up and down .
The issue happen when the high power level , and that is radon .Do you think maybe the noise interfere the Tx signal ? The failed the spectrum is +/- 400Khz .

 

[vfone]

No, the noise cannot make transient spectrum to fail.
Play with falling edge ramping. Definitely this will fix switching at 400kHz.

 

[mckinson]

Thanks!
But the maximum value for the spectrum is not at at the fallling edge .
Look into the waveform for 400Khz spectrum ,I find the maximum value is at the middle , not at the up and down portion .
And at first view , the switching spectrum value is low. Wait a moment , it get bigger and bigger .

 

[xinxin]

mckinson , which way can you find the waveform for 400Khz ? is it found on spectrum analyzer ? I think the problem you met may occurs on power supply to PA , pls check the layout for it , the trace should be shorter asap. from power to PA supply .
I have a question about it , is there anyone that tell me why the transient spectrum get bad when Low Power and Low temperature?

 

[mckinson]

Hi ,xinxin . You can see the waveform for 400KHz if you use the CMU200 with the latest version software . Also you can use the spectrum analyzer ,trigger with the CMU or 8960.

 

[Wilson_yu_chen]

Mckinson,

I thought Vfone's thought on remedy for S/T 400KHz by adjusting falling edge of ramp would meant the time domain burst waveform from low pedestal to totoally TX off (Vramp close to TX chain off) is too deep !!

Since the max value of S/T you saw is NOT located on ramp up/down, some possible causes you may have to check !
Any periodic frequency spurs ride on Vramp or PA supply trace, like harmonics of symbol rate/4, due to layout interference. Is this degraded 400KHz occured at only
High power control level ?

Pls kindly feedback me your updated status, Thanks much !
With Regards,

 

[mckinson]

Wilson ,thank you kindly comments .
Yes , the 400k occured at only high power level .
The maximum point is not constant location at useful portion .