LMX2594: SYNC function

Hi There,

Assuming your fosc = 100MHz, fpd = 50MHz, then this is a Cat. 2 sync configuration.

Sync mode is required, so VCO_PHASE_SYNC must be =1. 

For Cat.2 sync, a non-timing critical sync pulse is required to trigger synchronization.

For of all, set INPIN_IGNORE = 0 so that the sync pin would accept signal.

Pull the SYNC pin to High to trigger synchronization.

Hi Noel,

I have already did all you mentioned (VCO_PHASE_SYNC =1 and  INPIN_IGNORE = 0 and a pulse to High) but when I resend the registers to lock again the same freq then sometimes the outputts are in phase and sometimes in 180 degrees each to other. why it is not deterministic? 

by the way, the pulse width is about 50 nsec  and it is also going through a NAND gate with the clock oscillator input, so I got 5 pulses with setup time of 2.6 ns and hold time of 2.4 ns. is it good?

Hi There,

To sync, the device needs only one sync pulse. Since in Cat.2 sync, the timing of the sync pulse is not important, you can use a GPIO to toggle the state of the SYNC pin.

After sync, if you reprogram the device or do a VCO calibration, the phase is again not deterministic, another phase sync is required. 

thanks Noel,

my reason for doing the timing is because I need also cat.3 cases.

my question is : if multiple pulses (let say 5 pulses) in spaces of 10ns (the clock period) can damage the sync process?

and another question: is the ambiguity of sometimes 0 degrees and sometimes 180 degrees an expected result or the sync function should yield always 0 degrees?