[SI-LIST] Re: EMC versus SI, or EMC + SI?
- From: "Chris Cheng" <Chris.Cheng@xxxxxxxx>
- To: <olaney@xxxxxxxx>
- Date: Mon, 24 Mar 2008 09:26:27 -0700
I can't continue this discussion.
If you have never heard of 10GBase-KR Ad Hoc channel. Never know multidrop
source synchronous bus analysis methodology. Never design a package. Never
design a power distribution to a highspeed chip from die,package through the
system. And yet claim the above as boys game, I can only stay on the short bus
with the rest of the boys then.
Oh yah, I am making face at you from the bus window now.
________________________________
From: olaney@xxxxxxxx [mailto:olaney@xxxxxxxx]
Sent: Sun 3/23/2008 10:48 PM
To: Chris Cheng
Cc: si-list@xxxxxxxxxxxxx
Subject: EMC versus SI, or EMC + SI?
The field of EMC engineering is evolving from the classic test-and-fix
paradigm. The test lab guys will always be needed, but the new perspective
recognizes that many EMI problems are symptoms of underlying SI faults best
addressed at the design level rather than by slapping EMI bandaids on a black
box. Really good SI practitioners have some EMC competence, and really good
EMC engineers understand SI. At that level of expertise the areas blend
together because it's the same physics, with the difference that SI deals with
the management of milliamps inside the box while EMC deals in microamps leaking
outside. The essential factor that separates the men from the boys in either
area is as stated: "adequate understanding of the underlying principles"
rather than "iterat(ing) empirical designs into something workable". Software
is supposed to be a tool, not a crutch, and a yellow pad is for conceptual
work, not "accuracy down to -50dB" (assuming that is meaningful information).
Engineers who come to me waving simulation results but cannot explain what
makes the plots the way they are, can be very hard to help. In any instance, I
presume we are agreed that digital is just a special case of analog. The
techniques and analog components necessary for high speed designs prove it.
That is also why I have to disagree that "analog is analog". Opamps and LC
filters are one thing. A thing of a different kind is the ability to inspect a
layout to discern the hidden schematic of parasitic components that govern many
EMC/SI performance issues, the ability to visualize the invisible fields that
result, and the ability to gain useful information from symptoms that change
simply from moving a hand or touching a surface. This is what makes EMC/SI
seem like a black art to the uninitiated. If you accept that RF engineering
has long been a specialty not to be taken lightly, surely the ability to design
competent audio or power supply circuitry (both nontrivial tasks, based on the
specification of competent as opposed to adequate) is not of itself
qualification for dealing with microwave rate bit streams.
Here's a thought for this forum: EMC has a recognized credential in the form of
NARTE certification for those who demonstrate relevant experience and pass a
qualifying exam. Is the SI field ripe for an equivalent? Should they be
separate?
Orin Laney
P.S. Thank you, Roy, for the reference list.
On Sun, 23 Mar 2008 19:30:02 -0700 "Chris Cheng" <Chris.Cheng@xxxxxxxx> writes:
I walked many miles in EMI and still do.
My company had never and will never have EMI only design engineers. I
am responsible for anything that is not 1 and 0 in our system. I don't care if
it is call SI or EMI. Analog is analog. Any engineer who work for me is capable
of doing both.
> As things stand today any EMI engineer can tell you that they can
> make an entire
> career on fixing prototype SI/EMI problems based on the same
> half-dozen principles
I am still waiting for you to tell me how your EMI engineer can make
their career out of my examples below.
________________________________
From: Charlotte and/or Roy Leventhal [mailto:crleventhal@xxxxxxxxxxx]
Sent: Sun 3/23/2008 8:08 PM
To: Chris Cheng; olaney@xxxxxxxx; Roy.Leventhal@xxxxxxxx
Cc: avtaarenator@xxxxxxxxx; si-list@xxxxxxxxxxxxx
Subject: RE: [SI-LIST] Re: Best Signal integrity Schools in the USA
Chris,
I'm heavily into using modeling and simulation. Check my website:
http://www.semiconductorsimulation.com
<http://www.semiconductorsimulation.com/> . I'm also a great advocate of
combining EDA tools with the yellow pad for maximum efficiency and
understanding.
I recently had/have the opportunity to do some EMI engineering. Before
either SI engineers cast aspersions on EMI engineers or vice-versa I suggest
they walk in the others' moccasins a few miles.
You are right that EMI engineers will have to be better tool users in
the future. Why don't you help them get started, as I am trying to do?
Best Regards
Roy
-----Original Message-----
From: Chris Cheng [mailto:Chris.Cheng@xxxxxxxx]
Sent: Sunday, March 23, 2008 7:52 PM
To: olaney@xxxxxxxx; Roy.Leventhal@xxxxxxxx
Cc: avtaarenator@xxxxxxxxx; si-list@xxxxxxxxxxxxx
Subject: RE: [SI-LIST] Re: Best Signal integrity Schools in the
USA
I would love to learn how to model a multi-giga bit channel
with accuracy down to -50db with a yellow pad.
I would love to learn how to predict eye openings of heavily
loaded DDR2/3 buses with multiple loads and multiple branches and driving
positions under sso and crosstalk conditions with a yellow pad.
I would love to learn how to model package interconnects that
has imperfect return reference planes with a yellow pad.
I would love to learn how to deliver power to a multi-giga
hertz IC where the power grid and via structure is inherited 2 1/2 and 3D with
a yellow pad.
Are you sure we are talking about the same SI work here ? What
does your average EMI engineer knows about the above anyways ?
________________________________
From: si-list-bounce@xxxxxxxxxxxxx on behalf of olaney@xxxxxxxx
Sent: Sun 3/23/2008 2:07 PM
To: Roy.Leventhal@xxxxxxxx
Cc: avtaarenator@xxxxxxxxx; si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Best Signal integrity Schools in the USA
The symptom we see today is that many designers are heavily
reliant upon
really big, expensive software tools to iterate empirical
designs into
something workable. The same designer, given adequate
understanding of
the underlying principles, can often do 90% of the work on a
yellow pad,
then use software for cleanup and as a sanity check. When I
see SI
related job descriptions that want work experience with a big
list of
tools, I can readily guess what the company mindset is:
substitution of
tools for competence, and bring in the consultants when they
get into
trouble.
Orin
This email and any attachments thereto may contain private, confidential, and
privileged material for the sole use of the intended recipient. Any review,
copying, or distribution of this email (or any attachments) by others is
strictly prohibited. If you are not the intended recipient, please contact the
sender immediately and permanently delete the original and any copies of this
email and any attachments thereto.
------------------------------------------------------------------
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:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List technical documents are available at:
http://www.si-list.net
List archives are viewable at:
//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:
