LMX2594: What is the output frequency accuracy of the LMX2594?

Hi Hosoki-san,

If it is an integer channel, the frequency accuracy (in ppm) will be equal to the frequency accuracy of the reference clock.

If it is a fractional channel, the frequency accuracy will depend on how do you set up the fractional denominator. If the denominator equals to the integer multiple of fpd, then the freq accuracy will also follow the reference clock. 

You can use TICS Pro to do an estimation.

Hi Hosoki-san,

Based on your input, you can see the calculated VCO frequency from TICS Pro, so you know what is the frequency accuracy.

Hi Hosoki-san,

You currently have an integer PLL. The tool does not calculate the ppm error for you.

What Noel is suggesting is to use the PLL fraction and Fvco frequency to determine the frequency error.

If you use a fraction that contains a ppm error you will see that the N divider fraction will not be equal to what you want.

Thanks,

Vibhu

Tatsuya-san,

The ratio of output frequency to input frequency of the LMX2594 as shown by TICSPro is exact and there is no error.

So the frequency error is zero.  So in the picture you show, if our input frequency is exactly 50 MHz, the output frequency is exactly 2250 MHz-- no frequency error.

Even in fractional mode, this ratio is perfect and has no error.

Sometimes we get this question because some of our competitor's parts do not do exact fractions.  For example, if you want a fraction of 1/10, you have to do something with a forced denominator of 2^32.  This is NOT the LMX2594.  We have programmable denominator.  

Now if the 50 MHz input is 1ppm too high, then so will be the output.  So there is nothing more for TICSPro to calculate here.   The VCO exactly tracks the input.

Regards,

Dean

BTW,  disregard below,  I cound not erase.

Now if the 50 MHz input is say exactly 1 ppm higher than it should be, then the output frequency will be exactly 1ppm higher, or 2250 MHz + 1 ppm.

Even in fractional mode,