LMX2594: PFD-VCO-Crosstalk

Fabian Pfaeffli1

Part Number: LMX2594

Hi All

I ran into an issue I don't completely understand and I also didn't find any further information about that. I'm feeding a 100 MHz Ref CLK to the PLL and divide it by 2 in the input path to get a PFD of 50 MHz. Afterwards, I operate the PLL output (A or B, doesn't matter) at, say 10GHz, so there's no output divider enabled. At a Distance of PFD, there are two strong spurs visible in the spectrum analyzer, that are only 50-60dBc down. After consulting the PLLatinum Sim, it tells me that this spur is due to the PFD-VCO crosstalk! Can someone explain me this type of spurs or point me to further reading? Also, is there a way to reduce those spurs?

Thanks and best regards

Fabian

over 7 years ago

0 Fabian Pfaeffli1 over 7 years ago

Prodigy 180 points

After playing around with the evaluation board I found out that these spurs get MUCH!!! worse when activating the phase-sync mode. Why is it?

Pure integer mode:

Integer-N mode with Phase-Sync enabled:

0 Simon Damphousse over 7 years ago in reply to Fabian Pfaeffli1

TI__Expert 5370 points

Dear Fabian,

thank you for the detialed note. I need some time to go in lab and duplicate your measurement. I will get back to you before the end of this week. Regards, Simon.

0 Fabian Pfaeffli1 over 7 years ago in reply to Simon Damphousse

Prodigy 180 points

Hey Simon

Cool, thank you very much!

0 Simon Damphousse over 7 years ago in reply to Fabian Pfaeffli1

TI__Expert 5370 points

Dear Fabian,

your measurement aligned with mine. However, here are two important points in sync mode:

1) if you are at 10 GHz , select BOTH output to show as VCO. Your PFD spur will go back down.

2) If you are below 7.5 GHz and you are using div 2 or more: you can reduce the PFD_DEL_SEL. this will reduce the spur. You can also reduce that spur by changing the loop bandwdith.

I hope this helps.

0 Fabian Pfaeffli1 over 7 years ago in reply to Simon Damphousse

Prodigy 180 points

Hi Simon

Wow, that's great! By disabling the channel divider I was able to get the spurs down by a considerable amount, 30dB to be precise. Thanks for your help, that was awesome.

Regards

Fabian

Clock & timing

Clock & timing forum

LMX2594: PFD-VCO-Crosstalk