Can excessive jitter lead to lol in Si5338? Please refer to the figure below. If the input step always changes before acquisition time always, then Si5338 will lose lock. Therefore if

Can excessive jitter lead to LOL in Si5338?

Please refer to the figure below. If the input step always changes before acquisition time always, then Si5338 will lose lock.

Therefore if:

1. 
   Jitter amplitude is always greater than or equal to 20,000 ppm AND if the jitter frequency is in the band between 16 Khz – 1.6MHz. {(the lower freq = one-tenth the loop bandwidth , typical acquisition time = 10*(1/Loop Bandwidth))}
   
2. 
   If the noise freq < 16 KHz , the Si5338 will track the step change and if the noise freq > 1.6MHz, the loop filters out the step changes and the average excursion becomes much less than 5000 ppm
   

Even assuming that the jitter does not affect the lock status of Si5338 which is usually the case even if the input clock jitter is high:

1. 
   Si5338 being a wideband PLL and does not attenuate jitter up to jitter frequencies of 1.6MHz.
   
   1. 
      Therefore, it passes jitter “as is” for the noise frequencies < 1.6MHz.
      
   2. 
      Above 1.6 MHz, the jitter from input is attenuated.
      
2. 
   Therefore the input jitter requirement for Si5338 will follow the output jitter requirement.
   
   1. 
      For example, if you need an RMS jitter of < 1 pS in the 12KHz – 20MHz band from Si5338 output, then the input has to meet the same specification or be better.
      
      1. 
         Si5338 will pass all the jitter in the 12KHz – 1.6MHz band as is
         
      2. 
         It starts attenuating in the 1.6MHz and therefore, it becomes effective only at offset frequencies  > 16 MHz
         
3. 
   There is finite jitter generation by Si5338 which will add to input jitter.
   
4. 
   Our datasheet specifications for Si5338 have an implicit assumption:
   
   1. 
       The input jitter is much less than datasheet specification for Si5338 output jitter.
      

Therefore excessive jitter highly unlikely to cause LOL but will definitely degrade the jitter at Si5338 outputs.