LDO PSRR for Low VCO jitter

[captab]

Hi,

I am using a VCO delay cell whcih swings from VDD to VDD-Vgs in process of generating a delay.Now If I want to use a LDO as supply for this VCO, how do I decide upon the PSRR of the LDO?

Thanks.

 

[dick_freebird]

You ought to begin with simulating the VCO while applying a
triangular (or sinusoidal) ripple which is asynchronous, then
plot the VCO output as a 'scope would (I place a "raster"
pulse source with a cycle-delta risetime, and plot output
against that). Then you will see clearly the jitter (to the
extent that your models are accurate and complete) and
you can assess its dependence on the supply activity and
set your input ripple requirement. That, and the guaranteed
worst case ripple of the raw supply will give you the PSRR
(at frequency=) required.

 

[saro_k_82]

A PSS+PXF analysis will tell you the supply sensitivity on an open-loop VCO over which the applying the appropriate transfer function would let you arrive at the noise transfer function from the VCO supply to the loop output. This will provide a frequency plot required that needs to be bettered by the LDO + the supply decoupling capacitors. In addition to this, the flicker and thermal noise from LDO output could translate and create problems by folding on too (If PFD_CP is linearity is not too good) which is most likely seen only at the lab as noise simulations are not performed at the top-level mostly. In most cases, it is only the decaps on the VCO supply that help as the VCO frequency would be many times the UGB of the LDO- for noise reasons