[SI-LIST] Re: Hspice simulation in relates with capacitor

Some care is needed when doing transient simulations of AC-coupled
signals.

In a .TRANsient simulation, all time-varying signals are effectively
turned off (not switching) for t<0, then turned on at t=0.  There may be
a DC or bias shift after it turns on.

If you turn on a sine wave at t=0, its average value when on (averaged
over N cycles) is the same as it was when off.  No bias shift.

If you do this to a cosine wave, where the value for all t<0 is what it
would be at t=0, the average value is not the same.

If you pass the sine wave through a high gain buffer or logic gate, then
the output probably sits either high or low when the sine wave was off.
Think of a square wave that was 0V for all t<0, then starts switching
between 0V and 1V.  The average value was 0V when it was off, but
becomes about 0.5V for t>0 when it is on.

If you AC couple this signal, the bias voltage across the coupling
capacitor (or transformer if you use one) will change, between t<0 with
the signal off, and t>0 with it on.  Typically this takes time.  You
need to allow time for the coupling capacitors to charge to their new
bias voltages with the signals switching.  If you look too early in the
simulation, they won't have done this yet.  There are ways to avoid this
wait, using .IC statements, but you probably haven't done that.

The 10000pF capacitor takes 25 times as long as the 400pF capacitor to
charge toward its new bias point.  You may not have waited long enough.
Hence the apparent difference between the two capacitors in case 1.

In case 2, perhaps the average value of that output signal doesn't
change, so the capacitor doesn't need to charge to a new bias point.
Like that switched sine wave, above.

Does it make any difference, outside of simulations?  That depends on
whether your signals need to turn on and off.  If these are clocks that
become active as soon as the power supplies ramp up, and the clocks
never turn off, then SPICE is showing you something that would never
happen to you.  You'd want to look at what the waveforms look like as t
approaches +infinity.

Regards,
Andy



------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:
http://www.freelists.org/webpage/si-list

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field

List archives are viewable at:     
                http://www.freelists.org/archives/si-list
or at our remote archives:
                http://groups.yahoo.com/group/si-list/messages 
Old (prior to June 6, 2001) list archives are viewable at:
                http://www.qsl.net/wb6tpu

Other related posts: