Andrew Beckett
Senior Fellow
Offline
Life, don't talk to me about Life...
Posts: 1742
Bracknell, UK
|
What was happening was that in IC446, it was not checking for this situation, and as a result, your PSS solution was not properly taking into account the driven input signal when in autonomous mode.
The most common cause of this was people who used pulse sources to kick-start an oscillator - and made those sources periodic (albeit with a very long period). In most cases the solution was to make that source a one-shot by omitting the period.
If you've got an autonomous circuit, and want to simulate the effect with a driven input as well, you can't - and never have been able to (it may have given the appearance of working in the past, but it wasn't doing the right thing; usually it wouldn't have converged - so the fact that it did was probably luck - perhaps bad luck?).
If this is an oscillator with a mixer (say), and you need to have a large signal input (i.e pac or pxf won't do), then what you can do is characterize the oscillator's performance, and then replace it with a verilog-a model of the oscillator (such as that in rfLib, which models noise properly if I remember rightly). Then the oscillator becomes a driven circuit, and so you can then simulate it together with the large signal input using PSS or QPSS (depending on the frequencies involved).
Best Regards,
Andrew.
|