[SI-LIST] Re: Return Path
- From: steve weir <weirsi@xxxxxxxxxx>
- To: Doug Brooks <doug@xxxxxxxxxx>, si-list@xxxxxxxxxxxxx
- Date: Fri, 05 Aug 2005 13:18:49 -0700
Doug, in the fluid model, there are two misleading elements:
1. The focus is on the source of EMF, sic the battery,
2. It implies a time lag between the foward current starting from some
point and the matching return current closing that path.
If we take the switch example you offered, one might imagine a couple of
different cases:
a. The switch is located very close to one terminal of the battery and a
say 300m wire connects it to the other through some load resistor.
b. The switch is located at the end of two 300m wires back to the battery
through some load resistor.
Now, what will each the fluid analogy, and wave propagation tell us about
each case? Where does each model show the propagation beginning and
ending? How accurate is each? Which model can explain behavior from
virtual DC to any frequency we like? I don't think it's the fluid analogy.
On a PCB with switching I/Os the time and distance scales have changed but
not the behavior. What we have is in essence case b from above. The wave
emanates from the switches in our ICs, not from the power supply. The wave
model makes this clear, as it does the propagation path. The wave model
makes clear the critical point that the return and forward currents
propagate together. The fluid analogy with its unidirectional emphasis
fails us badly.
Where has the fluid analogy brought us? How many times have you seen
people talk about bone-headed ideas like the PCB planes or bypass
capacitors supplying current to high speed edges, when the entire edge has
completed long before the wave front through power pins can reach
significant charge in the planes, much less even reach the PWB bypass
caps? Yet this kind of junk mythology sadly makes its way into books and
other publications on a regular basis. I don't like it one bit.
For my money, I find the fluid analogy terribly misleading, and a source of
much misunderstanding. One doesn't need to be able to derive Maxwell to
understand wave propagation. I think that as Eric's book demonstrates,
most SI concepts are not that difficult to understand. Even a dummy like
me gets them from time to time.
Regards,
Steve.
At 12:51 PM 8/5/2005 -0700, Doug Brooks wrote:
>A couple of people have interpreted my statement re "flow" of electrons as
>meaning electron drift. Let's kill that right now.
>
>One electron in = one electron out is the flow of electrons. One electron
>in = SAME electron out is electron drift --- not at all the same thing.
>
>Certainly I don't argue against Maxwell's equations. But I don't argue
>against the fundamental definition of one amp of current either --- the
>flow of one coulomb of charge (6.25 x 10^18 electrons) across a surface
>in one second. I spend a lot of time with engineers (and technicians) who
>never took Maxwell's equations and didn't understand them if they did. My
>goal has been to take our difficult SI concepts and explain them in terms
>that these "poor" people can understand. To suggest that you can't explain
>what happens during planar transitions without Maxwell's equations (I
>believe) is simply wrong. To say that the classical description of current
>can't explain the difference between DC and high frequency is also (I
>believe) flat wrong. To say that one description is "more accurate" than
>the other --- well I suggest that depends a lot on whose working with
>them! And while people have been misled by seminar leaders teaching
>without the benefit of Maxwell's equations, we all know seminar leaders
>whose ability to mislead wasn't one bit hampered by a thorough knowledge
>of Maxwell's equations!
>
>Don't sell these more basic principles short when it comes to
>understanding what is happening on circuit boards. They can very
>effectively explain what is happening, and why one design approach may be
>more effective than another depending on the important design
>considerations. Especially for all those board designers who have no
>knowledge of Maxwell and wave theory.
>
>Doug
>
>
>
>
>At 12:01 PM 8/5/2005, steve weir wrote:
>>Doug, well I am going to argue vehemently that until someone repeals
>>Maxwell that the wave description is fundamentally more accurate than the
>>fluid analogy. The E/M fields cause the electron drift in those
>>wires. From the time you closed the switch the changing E/M field that
>>resulted propagated outward. Marconi found a useful purpose for that
>>phenomenon.
>>
>>The fluid analogy is certainly easy to understand, but what is the point
>>when it is so misleading? I can't tell you how many times otherwise
>>intelligent engineers that I have known have been thrown off understanding
>>PCB wave guides, because they were intent on following the DC current loop
>>of the fluid analogy.
>>
>>Teaching the fluid analogy requires that we later break that teaching when
>>we want to explain what happens at significant frequencies. Consider for
>>instance visualization of return current ( which is the original subject
>>matter ) when we transition planes in a PCB. If we think about it as a
>>fluid model we are easily misled into searching out a conduction path. For
>>ready examples of this mass confusion, just look at some of the discussions
>>on splitting-up grounds in the wrong ways for the wrong reasons, with the
>>wrong results. But if we simply consider waves to begin with, then the
>>behavior is easy enough to intuit out.
>>
>>Eric does a very nice job in his book explaining signal propagation that
>>does not rely on the fluid analogy. I think his approach is very accessible.
>>
>>Regards,
>>
>>
>>Steve
>>At 11:21 AM 8/5/2005 -0700, Doug Brooks wrote:
>> >With all due respect, Steve, if I have a battery connected to a transistor
>> >through a switch, I can turn the transistor "on" and "off" with the
>> >switch. That is easy to explain using the electron flow concept (which I
>> >hesitate to call an analogy, it in fact describes the physics involved).
>> >
>> >Is your description more complete AND also easier to understand?
>> >
>> >And if it is the frequency with which I "flip" the switch that bothers
>> >you, that simply means that some of the parameters that were not an issue
>> >with slow "flipping" (inductance and capacitance, for example) start
>> >becoming more of an issue with faster "flipping!" But the basic nature of
>> >what is happening (in particular where the electrons are flowing) is not
>> >changing, just speeding up. (How the electrons are flowing is speeding up,
>> >the electrons themselves, of course, don't change speed!)
>> >
>> >Doug
>> >
>> >
>> >
>> >At 10:45 AM 8/5/2005, steve weir wrote:
>> >>Doug, I have some real heartburn with some of those representations,
>> >>particularly the fluid analogy that speaks of current as the flow of
>> >>electrons. When I grew up current was defined as time variation of
>> >>electric flux. When an E/M field impinges a chunk of metal the resulting
>> >>interaction concentrates the field forming a wave guide. All practical
>> >>wave guides leak, be they a microstrip over a plane, a stripline, or
>> >>whatever. Some, like a good semirigid coax leak only a little tiny
>> >>bit. When they leak too much creating excessive disturbance in nearby
>> wave
>> >>guides, we have cross talk problems. I hope that this is what you were
>> >>trying to convey.
>> >>
>> >>Regards,
>> >>
>> >>
>> >>Steve.
>> >>
>> >>A
>> >
>> >Check out UltraCAD's new presentation videos and new skin effect
>> >calculator at http://www.ultracad.com
>> >
>>
>>
>>------------------------------------------------------------------
>>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 FAQ wiki page is located at:
>> http://si-list.org/wiki/wiki.pl?Si-List_FAQ
>>
>>List technical documents are available at:
>> http://www.si-list.org
>>
>>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
>
>____________________________________________________________________________-
>Check out UltraCAD's new presentation videos and new skin effect
>calculator at http://www.ultracad.com
>
------------------------------------------------------------------
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 FAQ wiki page is located at:
http://si-list.org/wiki/wiki.pl?Si-List_FAQ
List technical documents are available at:
http://www.si-list.org
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:
