Hi Lee,
Must designs I see are in plastic enclosures. Terabit routers are a very
specialized product. Most of what I see is everything else like automotive,
medical, and the like.
I agree shielding is needed for your speeds, partially because industry
standard connectors are so bad from an EMC point of view, although they are
getting better. I have measured a strong standing wave along one of these
connectors because the shield transfer impedance was very poor. Great SI does
not mean good EMC as the tinest bit of energy extracted from a signal, not
enough to affect SI, can cause EMC problems, one of my favorite experiments in
my classss. Because of this I agree your specific designs need shielding. Most
electronic designs these days do not.
It is amazing that many, if not most, designs still use 2 and 4 layer PCBs. But
they can be made to work. Every household has many 2 layer boards in it
(appliances, DVD players, etc.).
I personally prefer 6 layer boards as a minimum for a quiet design, but if an
embedded processor works on a 2 or 4 layer design, then that is the best
solution.
Doug Smith Sent from my iPhone IPhone: 408-858-4528 Office: 702-570-6108 Email:
doug@xxxxxxxxxx Website: http://dsmith.org
On Sun, May 13, 2018 at 11:01, leeritchey@xxxxxxxxxxxxx
<leeritchey@xxxxxxxxxxxxx> wrote:
WOW! That is a pretty broad statement! Want to tell that to engineers designing
terabit routers?
-----Original Message-----
From: Doug Smith <doug@xxxxxxxxxx>
Sent: Saturday, May 12, 2018 8:15 PM
To: leeritchey@xxxxxxxxxxxxx; CHARLES.GRASSO@xxxxxxxx
Cc: joel@xxxxxxxxxx; gurushankara@xxxxxxxxxxx; si-list@xxxxxxxxxxxxx
Subject: Re: [SI-LIST] Re: Logic ground and chassis/safety ground
Again the lowest cost solution is best and often shielding is the most
expensive solution! A Faraday cage should not be necessary for a good design.
Doug
University of Oxford
Department for Continuing Education
Oxford, Oxfordshire, United Kingdom
--------------------------------------------------
Doug Smith
P.O. Box 60941
Boulder City, NV 89006-0941
TEL/FAX: 702-570-6108/570-6013
Mobile: 408-858-4528
Email: doug@xxxxxxxxxx
Web: http://www.dsmith.org
--------------------------------------------------
On Sat, 12 May 2018 09:24:51 -0700, wrote:
The massive shielding is what we call a good Faraday Cage. Best solution if a
product allows it.
-----Original Message-----
From: Grasso, Charles
Sent: Friday, May 11, 2018 7:12 AM
To: doug@xxxxxxxxxx; leeritchey@xxxxxxxxxxxxx
Cc: joel@xxxxxxxxxx; gurushankara@xxxxxxxxxxx; si-list@xxxxxxxxxxxxx
Subject: RE: [SI-LIST] Re: Logic ground and chassis/safety ground
Sun Microsystems (many yrs ago now) adopted a DO160 /MILSTD approach to noise
abatement for their high end graphics products . That is: *no*
interconnection(between logic and chassis) technique combined with massive
product shielding and were very successful doing that. By successful I mean
that Sun passed all of its EMI/Immunity specs
with flying colors. Costly - but back then margins were not an issue!!
Thanks
Charles Grasso
(w) 303-706-5467
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On ;
Behalf Of Douglas Smith
Sent: Friday, May 11, 2018 12:29 AM
To: leeritchey@xxxxxxxxxxxxx
Cc: joel@xxxxxxxxxx; gurushankara@xxxxxxxxxxx; si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Logic ground and chassis/safety ground
But Lee I am talking about solving operational and ESD problems for over 40
years now, having been in engineering for iced 55 years.
Solutions are not unique but carry different costs. Multiple grounds result in
the least cost solutions most of the time and the highest PCB to chassis
resonant frequencies (hopefully beyond what would be a problem. All the boards
and chassis members are connected whether you want or not by impedances on thd
order of 30 Ohms at high frequencies. Single point ground guarantees strong and
low frequency resonance. Only with multiple grounds do you have the option of
killing the Q of the PCB/chassis resonance by adding loss. I demonstrate this
effect in all of my classes. It is easy to measure and prove. I demonstrated
the advantages of multiple grounds at AT&T Bell Labs about 40 years ago.
You may have designs that work but I can show they cost more than they should
to achieve good ESD performance. This is easy to show which I do every six
weeks at my courses. A single pF of capacitance will couple Amps of ESD current
at very low Voltages.
There are about a half dozen articles on my website ( http://emcesd.com ;) that
prove my point with measurements and analysis.
I am not saying your designs donât work, only that single point ground is not
achievable above tens on MHz and attempting single point ground raises system
cost to make it work.
Doug Smith Sent from my iPhone IPhone: 408-858-4528 Office: 702-570-6108 Email:
doug@xxxxxxxxxx Website: http://dsmith.org On Thu, May 10, 2018 at 8:50, ;
leeritchey@xxxxxxxxxxxxx wrote:
Sorry Doug. Been doing this a long time and have had very good results. Have
fixed many EMI problems by removing all but one connection between logic ground
and the case or Faraday cage- which many seem to want to call chassis ground.
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx On Behalf Of Douglas Smith
Sent: Wednesday, May 9, 2018 9:33 PM
To: leeritchey@xxxxxxxxxxxxx
Cc: joel@xxxxxxxxxx; gurushankara@xxxxxxxxxxx; si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Logic ground and chassis/safety ground
Hi All,
I respectfully disagree with Leeââ¬â¢s comments. It almost guarantees an ESD
problem. I have fixed hundreds of ESD problems by connecting the two grounds in
as many places as possible.
Also doing so significantly reduces the cost of switching supplies operating
from the mains, lowering induced noise by 30 dB on average for a given supply.
I have lost count on this kind of fix.
One connection also guarantees the lowest resonant frequency of the chassis-PCB
resonant circuit (which WILL exist) and prevents the simple solution to kill
the Q of that resonance. Multiple connections raise this frequency but the best
is multiple connections and then to de-Q the resonance you will always have
between the PCB and the chassis.
After 40 years of fixing chassis-PCB interactions I have developed quite a base
of solutions.
Doug Smith Sent from my iPhone IPhone: 408-858-4528 Office: 702-570-6108 Email:
doug@xxxxxxxxxx Website: http://dsmith.org On Wed, May 9, 2018 at 10:12, ;
leeritchey@xxxxxxxxxxxxx wrote:
There is no electrical reason why you need to connect these two. If you do
connect them, do it in only one place and do it with a DC connection. I make
that connection on the side of the PCB where the unshielded wires exit as a way
to minimize potential EMI escaping on those unshielded wires.
Don't know where all of the rules of thumb come from about capacitors and
multiple connections.
Lee RItchey
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx On Behalf Of Joel Brown
Sent: Wednesday, May 9, 2018 9:47 AM
To: gurushankara@xxxxxxxxxxx
Cc: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Logic ground and chassis/safety ground
Unless you have a requirement or other reason to keep them isolated It would be
best to connect them together.
If you decide to keep them them isolated then add multiple capacitors between
signal and chassis ground at each chassis connection point that can be
populated with different value capacitors or zero ohm resistors. Another option
is to have a chassis ground plane layer in your PCB that is connected to signal
ground at multiple points with capacitors. If you don't have a solid RF
connection between signal and chassis ground then there can be an RF potential
between the two grounds that can radiate if you have antenna structures such as
slots in your enclosure or external cables.
Joel
On Wed, May 9, 2018 at 9:16 AM, Guru Shankara
wrote:
Hi Experts,chassis.
I am working on a modem design that is powered by an AC/DC supply
(isolated).
The design includes a base board that hosts a mezzanine/daughter card
and both the baseboard and the mezzanine cards are powered by 12V DC.
The chassis walls and misc. mechanical hardware have a conductive
coating on the inside and will be connected to the
chassis/safety/earth ground of the AC/DC power supply.
As the name implies, the chassis/safety/earth ground provides for
safety of the operator if there is a situation where the higher
voltage side line of the AC/DC supply rips off and makes contact with
the
So, I prefer NOT to connect the board logic ground to the
chassis/safety/earth ground of the supply.
Any thoughts on if I should connect them together or not?
Also, if they need to be connected, is there a recommended method?
Really appreciate your help.
Thank you,
Gurushankar
------------------------------------------------------------------
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 forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
