The analysis of RLC in series is inconsistent with simulation results.

[xxgeneral]

Hi everyone,

I simulated a simple circuit with RLC in series under Cadence Spectre. L and C resonate at 1MHz. A sinasoidal source with an amplitude of 600mV stimulates RLC at 1MHz. And, R=50Ohm, L=25.33mH, C=1pF.

My thought is, LC gets its minimum impedance at 1MHz. So in steady responses LC has a very low voltage, wheras R has a large one reaching 600mV.

However, the simulation result is reverse. That is, LC has a very large voltage, nearly 600mV. R has a small one, only 20mV.

I did another simulation with L=25.33uH and C=1nF. Then, LC has a very low voltage, and R has a large one reaching 600mV.

This confuses me a lot.
Anyone can help me?
Thanks a lot!

 

[keith1200rs]

Phase.

If you look at the phase of the voltage across the inductor and capacitor at resonance you will see they are perfectly in anti-phase and the same amplitude so they cancel out. So, the result is a purely resistive load.

Scaling your values changes the Q. The higher inductor has a lower reactance but you have left your resistor the same so the Q goes down. The voltage across the resistor therefore goes up.

Try changing the frequency slightly and see what happens to the phase and voltage of the various signals.

Keith.

---------- Post added at 10:20 ---------- Previous post was at 10:02 ----------

By the way, when doing your simulations you need to ensure a very small timestep. At 1MHz I would suggest 2ns maximum with the sort of Q you are using. This may seem extreme but with resonant circuits it will make a huge difference. In a transient analysis the resonance will appear to be in the wrong place and all sorts of voltages/currents will be wrong. For example, your resistor voltage should be 600mV not 20mV. If I run a simulation with 2ns step I get the correct voltage. With a 2ns step I get a considerably lower voltage.

Keith.

 

[xxgeneral]

Hi Keith,

Thanks for your great help.I've tried again with a smaller step of 2ns. And correct tran simulation results were obtained. Thanks.

By the way, why does a higher Q make a spurious tran simulation result? Thank you very much.

Regards,

Haoran Hou

 

[FvM]

Seriously speking, I don't see the purpose of simulating a linear RLC network with sine excitation in a transient analysis. AC analysis will do the same without any timestep problems. Of course it's a suitable test case to learn about the limitations of transient analysis.