Floating Vpulse Influence Simulation Result

Dear Went,

Speaking for myself, I can't provide any definite insight into what you are observing as there is not anywhere enough information on your circuit. I understand your question, but there are multiple sources of DC offsets in a direct conversion receiver and there is not enough information about your circuit and your inputs (not only the"pulse" you mention). For example, you can get mixing between the RF and LO ports if the isolation is not sufficiently high that is a source of DC offsets. In your case, if your increased amplitude refers to a signal applied to the RF or LF port input, the added amplifier may result in enough coupling to change the DC offset.

Shawn

It's probably not a bug - but hard to know precisely what it is without more context (and seeing the issue). If you're concerned, you should contact customer support.

Several possibilities:

1. Altering the time steps causes some numerical differences - this tends to be more noticeable when you have oscillation in the system as it can cause the oscillator behaviour to change if it's not been adequately captured by the normal tmesteps without the additional source
2. The difference is small and within the tolerances - this is normal and a consequence of the fact that the matrices being solved are now different and both answers are "correct" within the specified tolerances. Just because they are different doesn't mean they are wrong...
3. The circuit has multiple stable operating points and a small change to the conditions pushes you into a different operating region. 

However, this is all guesswork based on very little information as Shawn mentioned. Customer support would be best so that we can look at it with you.

Andrew.

Dear Shawn

Thanks for your reply. I do not think it is the circuit that cause the mismatch, cause the existence of a vpulse which only connects to ground is the only difference of two circuits. It means that even if I get a bad designed DCR, I will get bad result twice. And they should be the same, I think.

Went.

Dear Andrew

Thanks for your reply. I use vsin to drive my mixer, so there is no oscillator in my circuit,. But I see ripples in the output. I mean the output waveform will vibrates for a while before convergence. Is this the oscillation you said ?

Went

Hi Went,

Possibly - you may have some near instability in the circuit and the behaviour of this is affected by the time steps introduced by the additional pulse source. You didn't say what changes when the frequency of the pulse is altered, so we're having to do a lot of guessing here.

Andrew.

This is my circuit. I used vcvs to model VGA, its' output is above 100 V. while the difference between the two branch is less than one hundred millivolts. 

Hi Andrew

Thanks for your help. Your analysis helped me find out the cause of the different simulation results. I think it is the time step and tolerance, cause I get a "good looking" result when I use conservative. And before that no matter whether I change the rise time, fall time or period of the vpulse, the result remains the same. So the vpulse is not the source of different results I got before.

Went.