phase of series LC impedance

[willdoit]

Hi

I understand that the impedance looking across series LC must have the phase component varying as an ideal step function.

I have been trying to replicate this in spice but I always see a gradual transition from 90deg to -90deg... I modified the gm for my simulation as well..

Any suggestions on which spice parameters need to be tweaked or if there is some problem in my understanding.

Thanks

 

[FvM]

Assuming you have ideal L and C in your simulation, SPICE apparently has a better understanding of the problem.

 

[willdoit]

FvM,

SPICE has poor understanding of high-Q circuits.. I was looking for tweaks to get it reasonably fixed..

I hope you also agree that the phase must be either 90 or -90 in non-lossy circuit or else power dissipation is implied..

 

[LvW]

FvM,

SPICE has poor understanding of high-Q circuits.. I was looking for tweaks to get it reasonably fixed..

Hi willdoit,

I have simulated a LC-circuit with SPICE: L with a series resistance of 1E-9 ohm and C with a parallel resistor of 1E12. The phase jump from 0 deg to -180 deg was like an ideal step . Do you want anything more ?

 

[willdoit]

Hi LvW,

I see something like the attached response.. i don't know why it is not as desired and as in your simulation.. which simulator are you using?

I used 1e-9H inductor and 1e-5F cap.. simulated from 150KHz to 15MHz

 

[LvW]

Willdoit, at the moment I have no explanation for your simulation result.
Check the characteristics of the parts used by the program from Linear Tech.
Perhaps they are lossy.
Obviously, the LC-circuit is a lowpass and the phase must start at 0 deg - contrary to your graph.
As mentioned before, my simulaltor (ORCAD/PSpice) works perfectly and presents correct results: Lowpass with heavy peaking around 160 kHz and a phase response between 0 deg and -180 deg.

Correction: peaking at 1.6 MHz

Added after 31 minutes:

Willdoit, do you know that the inductance in LTspice has a 1mohm series resistance as a default ?

 

[willdoit]

Obviously, the LC-circuit is a lowpass and the phase must start at 0 deg - contrary to your graph.

Thanks LvW for confirming that there is a problem in simulation only.

Just a point to note - The phase of current must be -90deg at low frequency (-j/wC dominates) and +90deg at high frequency (jwL dominates) and never 0 deg w.r.t. voltage. Its in lowpass RC filter that phase is 0 deg at low freqs..

 

[LvW]

Thanks LvW for confirming that there is a problem in simulation only.

I don´t see any problem. The simulation engine works properly, does it not?
(General comment: In most - if not in all cases - the simulator works correct and the user makes interpretation errors. Example: Bias point calculation of an opamp with POSITIVE feedback. The simulator calculates a stable bias point. Question: Who is wrong? The simulator or the user who expects instability?).

Just a point to note - The phase of current must be -90deg at low frequency (-j/wC dominates) and +90deg at high frequency (jwL dominates) and never 0 deg w.r.t. voltage. Its in lowpass RC filter that phase is 0 deg at low freqs.

Yes, if you speak about the phase of the current you are right. I was referring to the phase of the voltage of the LC divider circuit which acts as a lowpass.