The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> How to time-average capacitance by RF simulator
https://designers-guide.org/forum/YaBB.pl?num=1168140990

Message started by neoflash on Jan 6^th, 2007, 7:36pm

Title: How to time-average capacitance by RF simulator
Post by neoflash on Jan 6^th, 2007, 7:36pm

When designing LC-tank based oscillator, we need to analyze the tuning curve of frequency.

Instead of directly running LC-tank with active circuits, I hope that I could have a look at the tuning curve of the varactor.

Instead of the DC tuning curve which is easily available, I would like to see the effective capacitance of the varactor at certain voltage waveform.

Can eldoRF do this? Can SpectreRF do this? Which document refers to the setup and theory?

Thanks,
Neoflash

Title: Re: How to time-average capacitance by RF simulato
Post by Ken Kundert on Jan 6^th, 2007, 7:56pm

Spectre can do this (you don't need SpectreRF). Just apply an AC voltage source directly to the varactor (try to eliminate any series resistance) and perform an AC analysis at a fixed frequency while sweeping the control voltage on the varactor. The imaginary part of the current equals ωC. Just divide through by ω to get C.

-Ken

Title: Re: How to time-average capacitance by RF simulato
Post by neoflash on Jan 6^th, 2007, 10:02pm

Ken:

The average capacitance I want to observe is just like what is mentioned in Emad's book.

The capacitance obtained in your method is the small signal capacitance at certain DC biasing point. However, I thought there is also some way to get average capacitance. When we run that simulation, we need to specify input voltage amplitude and DC voltage. Thus, simulator will plot us a curve of average capacitance.

I see this mentioned in one recently pulished jssc paper, but i can not remember the title. I will check when I was back in office.

Thanks,
Neoflash

Title: Re: How to time-average capacitance by RF simulato
Post by Ken Kundert on Jan 6^th, 2007, 10:46pm

Whoops. I overlooked the words "time-average" in the title. I think you can follow the same procedure using PSS and PAC. I believe the 0^th sideband would correspond to the average capacitance.

-Ken

Title: Re: How to time-average capacitance by RF simulato
Post by neoflash on Jan 7^th, 2007, 5:03am

Ken:

Thanks for the information. But I can not locate any reference doc or application notes on this.

I guess I do not like got some results without knowing its physical meaning. Maybe I can try to acess Cadence AE, however, I doubt supporting groups in my region knows what I want.

It could be nice if you have some demo there.

Title: Re: How to time-average capacitance by RF simulato
Post by Ken Kundert on Jan 7^th, 2007, 11:15am

I don't have a copy of Emad's book, so I don't know exactly what you are trying to do. But I can tell you what the PSS/PAC set up is doing.

You apply the range of control voltages that you are interested in in the form of a large signal periodic stimulus to the varactor. Then in addition you apply a small sinusoidal test tone at a particular frequency. That test tone will generate responses at that frequency and related sidebands. Examining the 0^th sideband is equivalent to attaching a spectrum analyzer to the circuit and examining the response at the same frequency as the test tone. The magnitude of the response will be related to the average value exhibited by the capacitance weighted by the shape of the large stimulus waveform. You should be able to work through the mathematics to prove that to yourself.

-Ken

Title: Re: How to time-average capacitance by RF simulato
Post by rf-design on Feb 4^th, 2007, 2:35pm

The possible background of your question is to estimate the shift of the oscillation frequency of a tank dominated circuit. That requires that you calculate the effective capacitance for the first harmonic current. If the tank capacitane is primary current driven by the tank inductance you have to use a current source. Use a sine current source and extract the first harmonic voltage amplitude and phase. If the capacitance is not ideal you will observe phase differing from -90°. If you change the current amplitude you will observe a nonlinear change of the cophase component voltage. That indicate that the oscillation frequency will change with amplitude. Also known as AM2FM conversion effect.

I also observe that there is also bias shift effects in the circuit which are difficult to differentiate from the above effect. These DC bias shift too modify the oscillation frequency via the cap nonlinearity. Therefore the circuit is studied by selectively switching in/out the effect in the model and simulating the hole circuit instead of only the components.