How to simulate the Bandwidth of LPF(AC or TRAN?)

[justin]

I want to simulate an LPF's BW. I have heard AC simulation is not precise. So I use TRAN analysis as follows: I give a low frequency signal to the input, and measure the output signal Vout0, and take the Vout_3=Vout0 x 0.707 as the -3dB frequency. Is this resonable?
And are there any other methods?

 

[wylee]

AC simulation seems prefectly precise to me

If you are using HSPICE,

try using this example

*****************
R1 in out 10k
C1 out gnd 1n

vin in gnd ac=1 dc=1

.ac dec 100 1 1G
.measure ac DUT_Av0 MAX par('vdb(out)-vdb(in)') from 1 to 10
.measure ac DUT_fc WHEN par('vdb(out)-vdb(in)') = 'DUT_Av0-3'
.end
*****************

DUT_fc should be exactly at fc = 1/(2*pi*R1*C1)

 

[papyaki]

Hi,

If you are using pspice, try an AC simulation. Choose the start and stop frequencies two decades away from the expected LPF cutoff frequency.

Assuming you have labelled the LPF's input voltage "Vin" and the output voltage "Vout".

After simulation, run Probe. In the Probe windows, select "Add Trace" and type "DB(V(Vout/V(Vin))" in "Trace Expression". Click "OK", this will plot the curve you want.

 

[riz_aj]

I feel it is quite accurate on Pspice. However, use same method like you too at time to ensure exact number but not always.

 

[justin]

To wylee:
For some simple circuit, It may be correct. But for some complicated circuit, we should consider the operating point and slew rate and maybe other things.
To papyaki ,riz_aj
I use hspice as my tool, and I don't know about pspice.
And generally what isthe cutoff frequency? 10%? 20%?

 

[justin]

To wylee:
For some simple circuit, It may be correct. But for some complicated circuit, we should consider the operating point and slew rate and maybe other things.
To papyaki ,riz_aj
I use hspice as my tool, and I don't know about pspice.
And generally what isthe cutoff frequency? 10%? 20%?