LMX2594: Spur at offset = Fosc is not simulated in PLLatinum?

david royle

Part Number: LMX2594

I am seeing spurs at offset equal to Fosc (using LMX2594EVM into spectrum analyzer), even when using input multiplication. PLLatinum identifies spurs at Fpd and its harmonics, but not spurs at Fosc.

TICS Pro setup shown. PLLatinum, in same configuration, identifies spurs at harmonics of 40MHz, but does not identify 100MHz offset as spur location. I see -73dBc spur at 100MHz on spectrum analyzer.

over 7 years ago

0 Noel Fung over 7 years ago

TI__Guru** 120710 points

Hi David,

I cannot see the picture, would you resend it again?

0 david royle over 7 years ago in reply to Noel Fung

Intellectual 585 points

Just edited it in.

0 Noel Fung over 7 years ago in reply to david royle

TI__Guru** 120710 points

Hi David,

I believe you will see some small spurs at 100MHz and 200MHz intervals. They are created by the state machine clock. You can verify this by making CAL_CLK_DIV=1 (Div/2), this will change the state machine clock to 50MHz, the spurs level will also be smaller.

0 david royle over 7 years ago in reply to Noel Fung

Intellectual 585 points

Noel,
Yes, this seems to be the case. PLLatinum software does not consider these spurs? Can I use CAL_CLK_DIV=1 as a fix? It would seem to violate the datasheet (page 43).

0 Noel Fung over 7 years ago in reply to david royle

TI__Guru** 120710 points

Hi David,

Crosstalk spurs such as this state machine clock spurs is difficult to model, if not impossible. PLL Sim does not take this spurs into account.
CAL_CLK_DIV = 0,1,2 or 3 are possible. The goal is to ensure the max. state machine clock of 200MHz is not violated.
Since the VCO calibration speed depends on state machine clock, the calibration time will be longer when CAL_CLK_DIV > 0. If you don't care the VCO calibration time, you are free to make state machine clock slower.

0 david royle over 7 years ago in reply to Noel Fung

Intellectual 585 points

Thanks

Clock & timing

Clock & timing forum

LMX2594: Spur at offset = Fosc is not simulated in PLLatinum?