The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> RF Simulators >> spectreRF pstb simulation with multi-loop circuit
https://designers-guide.org/forum/YaBB.pl?num=1268385779

Message started by rfmems on Mar 12^th, 2010, 1:22am

Title: spectreRF pstb simulation with multi-loop circuit
Post by rfmems on Mar 12^th, 2010, 1:22am

Description of the circuit

1, LDO+push pull VCO
2, LDO output biases VCO, inductor center tap feedbacks to LDO inputA, inputB is reference voltage
3, LDO: differential current mirror load amp.+miller compensated pmos

Objective

To simulate the LDO stability when the VCO is running in steady state.

Problem

The circuit is mulit-loop, LDO feedback, miller compensation. There is no way to cut all the loops at once. So I put the probe at LDO feedback path (from inductor center tap to LDO inputA). But I got strange results from simulation. Simulated gain drops to 0dB from DC to 100Hz, and then slowly increases to peak 40dB at 50KHz. Phase margin also started from 0 at DC and increases to 90 at 50KHz.

Comparing dc/tran+stb VS. pss/pstb, the simulation results are similar when freq>100KHz, but below that, pss+pstb makes no sense to me.

Question

1, I guess multi-loop circuit stability simulation is out of pstb's capbility, but did anyone have luck with that issue?

2, If it is not possible to simulate with pstb, is there a solution to this kind of stability simulation?

Title: Re: spectreRF pstb simulation with multi-loop circuit
Post by pancho_hideboo on Mar 12^th, 2010, 2:30am

rfmems wrote on Mar 12^th, 2010, 1:22am:

2, If it is not possible to simulate with pstb, is there a solution to this kind of stability simulation?

Use "Periodic Steady State (Harmonic Balance) Based Stability Analysis" of Agilent GoldenGate.
http://www.designers-guide.org/Forum/YaBB.pl?num=1243078889
http://www.home.agilent.com/agilent/product.jspx?cc=JP&lc=jpn&ckey=1738199&nid=-34269.917946.00&id=1738199&NEWCCLC=USeng

Title: Re: spectreRF pstb simulation with multi-loop circuit
Post by rfmems on Mar 12^th, 2010, 2:57am

Thank you very mucn pancho for your response.

Your comments on 'Stability of Total Circuits' certainly are interesting and enlighten.

However, my circuit, LDO+VCO, in its nature is instable since it oscillates. And what I try to figure out is whether the LDO feeback loop is stable. In another word, I am looking for instabilities of my circuits except the oscillator itself. I suppose that is where spectre has problem with.

Having looked at intriduction of Goldengate, I found it really attractive specially it is fully integrated into cadence ADE. Sadly, most engineers have to live with the simulators provided to them, which in my case, don't include Goldegate.

Title: Re: spectreRF pstb simulation with multi-loop circuit
Post by pancho_hideboo on Mar 12^th, 2010, 3:00am

rfmems wrote on Mar 12^th, 2010, 2:57am:

However, my circuit, LDO+VCO, in its nature is instable since it oscillates.
And what I try to figure out is whether the LDO feeback loop is stable.
In another word, I am tring to figure out instability of my circuits except the oscillator itself.

All eigenvalues are listed in Agilent GoldenGate.
So you can distinguish undesired oscillations(eigenvalues) from desired oscillations(eigenvalues).

http://cp.literature.agilent.com/litweb/pdf/5989-9448EN.pdf
This nonlinear stability analysis was introduced in Xpedion day.
But it was deleted from GoldenGate when Agilent purchased Xpedion.

It has been reimplemented in GoldenGate 4.4.0.