VCO initial condition

[EHY]

i made TOP block of CDR circuit.

if i don't set initial condition of VCO CK, VCO does not oscillate. (get in stuck)

but i made free-running VCO (control volatage=0)

so, do i make a circuit that sets initial condition such as switch to operate properly when i make real-chip(fab-in)???

help me~~

 

[erikl]

if i don't set initial condition of VCO CK, VCO does not oscillate. (get in stuck)

In simulation, do the same as in reality: start your power supply from 0 and ramp up to VDD !

 

[dick_freebird]

Depending on an externality such as supply ramp or even
noise, leaves you exposed to the possibility that this will
fail you at some point. If you "own" the application, then
this call is yours to make. For someone designing a broad-
market general purpose part, expect the unexpected - a
customer will surely one day call you about their board
failures at low temp which bench testing fails to replicate
and come from the end user equipment having a very
leisurely soft start in the power supply, or something.

It's tough to fully test out a VCO's "will to live" in a
simulation, usefully.

A band-aid such as a timeout watching the output for
failure-to-oscillate, and a "kicker" for the loop, ought to
cost not much but provide some idiot-proofing (and the
esteemed Mr. Murphy is a tireless and dedicated idiot
in his own right, with an army of followers you might
call customers).

 

[EHY]

i tried approach you mentioned but vco does not oscillate in top block

 

[D.A.(Tony)Stewart]

Oscillators are a combination of positive feedback for Barkhausen condition and negative feedback for DC bias to get into the previous condition of AC phase shifted by 360 deg & Av>=1 condition.

Noise is inevitable as is any slewing DC.
Which do you think applies to the lack of oscillation?

Hint: much more details are need this side to help any further.

 

[EHY]

In the same Top block,

in case of differential odd stages VCO can oscillate without any initial condition set.

but, in case of differential even stages VCO can not oscillate.

Oscillators are a combination of positive feedback for Barkhausen condition and negative feedback for DC bias to get into the previous condition of AC phase shifted by 360 deg & Av>=1 condition.

Noise is inevitable as is any slewing DC.
Which do you think applies to the lack of oscillation?

Hint: much more details are need this side to help any further.

 

[D.A.(Tony)Stewart]

Oscillators are a combination of positive feedback for Barkhausen condition and negative feedback for DC bias to get into the previous condition of AC phase shifted by 360 deg & Av>=1 condition.

Noise is inevitable as is any slewing DC.
Which do you think applies to the lack of oscillation?

Hint: much more details are need this side to help any further.

So when you analyzw feedback loop of even stages, are both conditions above true?

 

[EHY]

So when you analyzw feedback loop of even stages, are both conditions above true?

Well, when i disconnect the output of VCO to Phase Detector input(disconnect feedback loop) , VCO can oscillate without any initial condition or supply step.

So How can i solve them?? i must use 4-stages VCO to get 0, 90 degree phases CK

So, How can i test two conditions in CDR loop???

AC analysis and How can i get gain and phase??

 

[D.A.(Tony)Stewart]

By logic, Gate to Drain inverts 180 deg.

To get 90 deg;

One must use quadrature phase shift method using an XOR logic to frequency double to get 90 deg.

**broken link removed**

 

[EHY]

By logic, Gate to Drain inverts 180 deg.

To get 90 deg;

One must use quadrature phase shift method using an XOR logic to frequency double to get 90 deg.

**broken link removed**

I just want 4-stages differential ring VCO in CDR loop

what about making switch to set initial value of VCO output such as TG switch.