the totoal current of the VCO doesn't convergence !!! why?

[diva]

ebookee

When I doing the tran simulation(Using Spectre) I foud that the current of my VCO doesn't convergence !!! why?
anybody met this statues before.

 

[LvW]

ads vco convergince

You should give us some more information.
For example, you spoke about a VCO. Within a loop ? Probably a PLL.
What kind of PD ? Which current ? Control input or output ?
Most probably the loop is unstable; but if you want a substantial answer you should formulate the problem in detail.
Regards

 

[dick_freebird]

Transient convergence is most often a modeling error -
either unrealistically low capacitances (like you might
get by trying to model with behavioral or primitive /
ideal elements or by a less than complete set of active
device parasitics) or a timestep that is just too coarse
for the timescale of circuit activity, so the simulator
steps past the action, and can't "downrange past zero".
Ideal elements which have unlimited gain can without
a limiting mechanism, can force insane node voltage
or current onto other devices and push them into
unmodeled regions where singularities or further
blow-up ensues.

You could try a few of the usual approaches, such as
forcing more conservative solutions, using a method
like Euler that is more damped, etc. I'm almost
certain that the Cadence help docs have some of the
same advice for you.

It looks like the current shown, before it quits, is in
bursts; maybe the outer loop is unstable and you're
chopping the control voltage and seeing activity / no
activity, maybe the oscillator is killed by out of range
control voltage (that would be a bad design, or a
bad testbench).

You should be capable of probing around other parts
of the circuit and using your understanding of the
normal operation, to pick out the abnormal. Might
start with opening any closed loops and validate the
basic function forward-only.

 

[diva]

Re: the totoal current of the VCO doesn't convergence !!! wh

You could try a few of the usual approaches, such as
forcing more conservative solutions, using a method
like Euler that is more damped, etc. I'm almost
certain that the Cadence help docs have some of the
same advice for you.

It looks like the current shown, before it quits, is in
bursts; maybe the outer loop is unstable and you're
chopping the control voltage and seeing activity / no
activity, maybe the oscillator is killed by out of range
control voltage (that would be a bad design, or a
bad testbench).

.

==========================================
Thx for your advice. I change the maxstep and simulation algorithm
It seems convergence but still oscillating.perhaps it's the VCO's pull and push current