RE: [SI-LIST] : What's your favourite Screwy SI Concept? More on Bends/Miters

About this list Date view Thread view Subject view Author view

From: LATOURRETTE,JEFF (HP-SanJose,ex1) ([email protected])
Date: Fri Jan 14 2000 - 18:11:29 PST


All:

It may sound a bit strange to think of electrons slowing down, but another
way to look at it is current density. The inside of any bend has much more
current density and hence more associated field concentrated on the inside
compared to the outside. This produces a capacitive effect, and a
simplified model is a lumped shunt C. Mitering can be thought of as either
reducing the capacitance at that point (or even compensating the lumped
capacitance at the bend with higher impedance (more inductive) segments of
T-line.

It is correct that if your signal's frequency content is very low, the
impact of the parasitic capacitance value on return loss (VSWR reflection)
is minimal and mitering isn't necessary. Once you hit a frequency where
that cap value is significant, you'll see measurable degradation in return
loss.

The paper, previously mentioned, Douville and James from '78 MTT
Transactions is one of the better ones I have on the subject of microstrip
bends. According to the paper, as you get higher in frequency, mismatch due
to right angle corners increases from almost none at 500 MHz to 1.5:1 or
even as much as 3.0:1 at 2.5 GHz, depending on the impedance of the line.
Lower impedance lines perform worse without a miter. The paper also
describes an optimum mitering percentage for different types of substrates
and bend angles, and has descriptions of Bend Equivalent Electrical Length,
which in some applications may be more important than presence of
reflections.

In my own experience working with broadband microwave circuits, I'd consider
mitering or avoiding bends on traces above 2 GHz for best results, but
absolutely do it up at 10 GHz. In high-speed signaling, it may be difficult
to specify a general rule on what level of reflection is acceptable. In my
opinion (without supporting data), EMC impact due to bends alone should be
minimal until you reached unusable levels of reflection.

The substrates used are a bit dated and the above article was geared at the
practicing Microwave Engineer of that era. Anyone have more modern
references with PC Board materials currently used on high-speed designs ?

Regards,

Jeff La Tourrette
  SJ Applications Manager
  Agilent Technologies Fiber-Optics Components
  370 West Trimble Rd. M/S 90TH
  San Jose, CA 95131
Voice: 408 435-4083 FAX: 408 435-4915

-----Original Message-----
From: Jian Zheng [mailto:[email protected]]
Sent: Friday, January 14, 2000 11:10 AM
To: [email protected]
Subject: RE: [SI-LIST] : What's your favourite Screwy SI Concept?

Hi:

In fact, the flow of electric current is very similar to the flow of water
current.

    | |
    | |B______
    |
    |A__________

When you have a bend, the current will take a short cut at the corner B. The
current at the corner A will be slowed down. For the electric current, the
slow down at corner A will create reflection from the corner. The short cut
at the corner B will make the electric path shorter than measured along the
center line. For standard PCB, the reflection from the corner is very small.
However, I understand that the electrical length is very important in clock
synchrinization. Accumulation of the shortening of electrical path at angles
may have effect to SI. Thanks!

Best regards,

----------------------------------------------------------------------------
------------
Jian-X. Zheng, Ph.D
Zeland Software, Inc., 39676 Mission Blvd., Fremont, CA 94539, U.S.A.
Tel: 510-797-8109, Fax: 510-797-8241, Web: http://www.zeland.com
----------------------------------------------------------------------------
------------

> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]]On Behalf Of Jeff Seeger
> Sent: Friday, January 14, 2000 10:05 AM
> To: [email protected]
> Subject: Re: [SI-LIST] : What's your favourite Screwy SI Concept?
>
>
> Tom Dagostino wrote:
> >
> > Bending a wire keeps a uniform cross section. Putting a corner
> in a trace
> > adds extra C at the corner from a larger area.
> >
> And herein lies the arguement, I believe.
>
> Perhaps those who argue against "right angles" envision a
> geometry where the line has a squared corner both on the
> inside of the elbow and the outside.
>
> I would wager that those who claim no material effect are
> working with the usual CAD system/photoplotter geometry of
> a squared corner inside the elbow and a radiused one on the
> outside.
>
> An attempt at ascii art:
>
> "Right Angle" "Standard Practice"
>
> \ \ \ \
> \ \ \ \
> > > > )
> / / / /
> / / / /
>
> HTH,
> --
>
> Jeff Seeger Applied CAD Knowledge Inc
> Chief Technical Officer Tyngsboro, MA 01879
> jseeger "at" appliedcad "dot" com 978 649 9800
>
> **** To unsubscribe from si-list: send e-mail to
> [email protected]. In the BODY of message put:
> UNSUBSCRIBE si-list, for more help, put HELP.
> si-list archives are accessible at http://www.qsl.net/wb6tpu
> ****
>
>

**** To unsubscribe from si-list: send e-mail to
[email protected]. In the BODY of message put: UNSUBSCRIBE
si-list, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
****

**** To unsubscribe from si-list: send e-mail to [email protected]. In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
****


About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Thu Apr 20 2000 - 11:34:40 PDT