[SI-LIST] Re: Quasi Static Assumptions

Insofar as problems from using quasi-static models, a big 
problem you may face is the one that bites you in the rear 
when you're not looking.

Suppose you you have a transmission media separating a source
and a load.   Let's suppose, for the sake of argument,
that you are running a 10 GBps system with a 20 ps rise/fall time.

Fknee = 0.35 / 20 ps  = 17.5 GHz

So most of your high frequency content is below 17.5 GHz.
Well, that's true, but not all you need if you're doing broadband.

Lets hone in on that information even more using the basic
Fourier representation of the wave.  Suppose
you're information content is a simple 101010... pattern @ 10 GBps.
That puts a fundamental at 5 GHz and a 3rd harmonic at 15 GHz.

Send in different patterns, and your signal will have harmonic "lobes"
centered around 5 Ghz & 15 GHz respectively with nulls centered
at 10 GHz and 20 GHz respectively.  

The entire 3rd harmonic "lobe" spans from 10 GHz to 20 GHz, 
with the bulk of the 3rd harmonic info centered at 15 GHz.   (If your rise 
time is slower, then you probably have attenuation of that 3rd harmonic 
lobe-- but it's still centered between 10-20 GHz, peaking at 15 GHz.)

THE PROBLEM:
------------
Now, suppose that your transmission media is right on the borderline
of being electrically long with respect to a wavelength-- say, 
(1/20)*wavelength at 1 GHz.    

That's fine for lower frequencies (fundamental lobe), but for your third
harmonic lobe @ 10-20 GHz, the transmission media appears to be
1/2 wavelength (@ 10 GHz) and 1 wavelength (@ 20 GHz).   

If your load/source is mismatched at all to your transmission media, then
this will introduce problems.   Low frequency content will reflect, but the
reflection will not have a significant phase mismatch when it bounces from
the load, back to the source and back to the load again.   That is NOT
true for the higher frequency content.   Reflections of the higher frequency
content will be significantly out of phase which can cause slowing of
rise times or outright bounces in your rising/falling edge-- not nice.

If you just assume that the quasi-static model is good enough (if, in fact,
it is not), you may not fully represent this transmission media phase delay 
and your high-frequency content will look significantly better than it
will in the lab.   

The sum of it all is that you have to make sure it's electrically small with

respect to that full harmonic lobe-- which I would argue (from the
standpoint of 
Fourier construction of digital signals) spans the entire 3rd frequency
lobe.   

Whether you use 1/10 or 1/30 wavelength as the cutoff point really depends
on the quality of signal you're trying to obtain at the load.

   - Bart McCoy






-----Original Message-----
From: Moeller, Merrick [mailto:mmoeller@xxxxxxxxxxxxxxx] 
Sent: Thursday, July 31, 2003 10:44 AM
To: Si-List (E-mail)
Subject: [SI-LIST] Quasi Static Assumptions


When does a Quasi Static 3D simulation tool become obsolete? B. Young's
book describle the limitation by maximum circuit dimension d < wavelength/10
or d < wavelength/30.
With wavelength of interest being calulated from the standard digital
frequency of
interst equation f = .35/Tr, and the maximum circuit dimension or effective
electrical length
to be, SQRT(Er)*Mechanical Length, Is this a correct approximation? Why the
large range of wavelength ratios?
Are there tell tale signs that the Quasi Static assumption has broken down?
Thanks,
Merrick




------------------------------------------------------------------
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
  
------------------------------------------------------------------
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: