[SI-LIST] Re: Need advice on basic 6-layer stackup
- From: steve weir <weirsi@xxxxxxxxxx>
- To: GrahamDavies@xxxxxxxx, si-list@xxxxxxxxxxxxx
- Date: Mon, 02 May 2005 12:04:54 -0700
Graham, The fact that your company's "slow" "1999 era" product doesn't
work, is testament that something in your shop is sorely lacking. Your
board house has handed you a massively improved and as Dr. Zheng has
verified, a 50 ohm stack-up. That you still have questions and doubts
should tell you to seek education or qualified help sooner rather than later.
If you are looking for a handbook presentation on SI material you might try
Lee Ritchey's "Right the First Time". It will at least give you issues to
discuss with this consultant you use and or your boss. You can obtain his
book only from his web site www.speedingedge.com for $125.00 plus tax and
shipping. I caution you that a handbook improperly used can be
trouble. But for the kind of design that you describe, you are likely to
find a number of useful topics in Lee's book, and you can certainly look
further for more information.
If you want to understand the subject matter in more depth, I would go with
either Dr. Bogatin's "Signal Integrity Simplified" or Dr. Johnson's "High
Speed Digital Design". Tom Granberg's book "Handbook of Digital Techniques
for High-Speed Digital Design..." provides an excellent survey of recent
and current signaling technologies.
Steve.
At 04:27 PM 5/2/2005 +0000, Graham Davies wrote:
>--- In si-list@xxxxxxxxxxxxxxx, steve weir <weirsi@xxxx> wrote:
>
> > ... I suggest ... search through
> > the ... information on the WWW ...
>
>Well, that's how I got to this list. I couldn't find anything
>anywhere that addressed my question.
>
> > ... purchase ... books ...
>
>I was hoping that someone could just fire off an answer to my
>question based on their experience without my having to wade through
>an entire book, much of which will be familiar to me, and without any
>assurance of the answer being in there anyway.
>
> > ... series termination ...
>
>I understand that this is fine for non-clock signals even when
>bussed. On the other hand, it seems to me that for slow, non-clock
>signals you might OK without *any* kind of termination.
>
> > If you are the PCB designer ...
> > ... pushing back on the design
> > engineers ...
>
>I was hired to manage the engineering department. The designers left
>the company five years ago. PCB design is done by a consulting firm
>but the guy is just a technician and has zero appreciation for the
>electrical side of what he's doing. For example, he loves to line
>vias up in nice long rows because it's pretty. The plane layers are
>therefore covered in slots.
>
> > ... hire a consultant ...
>
>It could definitely come to that, but, again, I just thought that for
>such a low-tech board there ought to be some simple rule of thumb for
>selecting a good trace impedance.
>
> > ... slow busses ... get away
> > with murder ...
>
>Well, exactly.
>
> > ... 10 inch trace ... failing
> > to model that trace as a
> > transmission line will give
> > you the wrong answer ...
>
>As I said, this trace does give us trouble. Empirically, slowing down
>the edge with a series resistor masks the trouble. I am only going
>ahead with this because the previous PCB design is so much worse.
>
> > I strongly disagree with SPS SGS
> > as a signaling improvement.
>
>The person who suggested this (off-list) did not say it would be an
>improvement from a signal integrity point of view, only that I might
>be over-engineering due to a failure to clarify my assumptions. I
>have grown quite fond of your suggestion, Steve, and since I know
>this board can be routed on three layers I'm inclined at the moment
>to go with your SGP SGS (where P is grided, not solid).
>
>I think this is a wrap, so I thank everyone, especially Steve, for
>their help. I've learned a lot but also verified that at least I'm
>not a total newbie. Unfortunately, I don't have an answer to my
>question, which is how to select the best separation of signal and
>ground layers. Even if I go with a 50 ohm impedance, just because
>everyone else does, I need to spend some serious money to find out
>what that is as a layer separation. Because I know I'm having trouble
>driving traces at 3.2 mil, I think that I'll just back off a bit to 4
>mil and see where that leads me.
>
>Graham.
>
>
>
>
>------------------------------------------------------------------
>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
>
------------------------------------------------------------------
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:
