[SI-LIST] Re: About the chassis gnd and logic gnd

Peter,

If you design the chassis for a specific customer,
please check with him what he exactly wants. The grounding
scheme of the boards inserted and the chassis must match.
If not, you are likely to have bad surprises during radiated
emissions testing.
The ATCA specification says pretty clearly that Logic Ground
and Shelf Ground shall be separate with the possibility to
connect them together.
It depends also on the installation if a 2-wire or 3-wire 
scheme is used. 
Ideally your chassis performs well in all configurations, but
in real world you will have to live with a compromise.
As Shelf Ground and Logic Ground are present in the backplane, 
you will probably end up with the solution proposed by Peter
being the better one.

Personally I do not like mingling Shelf Ground and Logic Ground
within the chassis, especially the backplane. A consequent single
point grounding, when required, is no longer possible as the coupling
between the two grounds is too tight.

Compared to CPCI, the ground connection between board and backplane
could be better. 

Q: Did anybody do tests on that?

From my experience I cannot tell that multipoint grounding has a 
better performance than single point grounding. I have seen chassis
with boards in a multipoint grounding configuration which worked
perfectly in terms of emission and ESD. Speaking of the number of 
ESD issues, I saw less problems with single point grounding 
configurations. The critical point are the seams, the openings and
the quality of the connection between connector and front panel.
When the enclosure provides a good path for the discharge and the
coupling to the boards on the inside is low, you should not have any 
bad surprises.
I do not want to suggest that any of the schemes is better, I just 
have seen chassis in both worlds performing perfectly. 
The important thing is that the grounding scheme had to be followed 
with ultimate consequence! 
And here lies the weakness of ATCA in my eyes. The only
decision was to make no consequent decision on the grounding scheme.
The point is, that with an increasing number of connectors and cables
you increase the coupling between the two ground domains, so even with
a thorough single point grounding design, you end up having a badly
designed multipoint grounding configuration when it comes to higher
frequencies with all the consequences you don't want to have.

When you have a relatively small box and a small number of external
connections, single-point grounding does a great job in terms of
ESD and radiated emissions. ATCA does not belong to that category
from my point of view.
Having a lot of external connections, it is probably better to use
multipoint grounding as Steve indicated.

These are my (non religious :-) ) views on this. 

Best regards,
Michael

 Michael Nagel - Senior EMC Engineer    
 
 FORCE COMPUTERS GmbH              Lilienthalstrasse 15
 A Solectron Company               D-85579 Neubiberg/Muenchen - Germany     
 Tel: +49-89-60814-0               Fax: +49-89-60814-376        
 e-mail: Michael.Nagel@xxxxxxx     WWW: http://www.forcecomputers.com




> -----Original Message-----
> From: steve weir [mailto:weirsp@xxxxxxxxxx]
> Sent: Friday, March 26, 2004 3:40 AM
> To: peter zhu; si-list@xxxxxxxxxxxxx
> Subject: [SI-LIST] Re: About the chassis gnd and logic gnd
> 
> 
> Peter, my best recommendation is frequent ties between logic 
> and shelf 
> ground.  Keep the -48V return fully isolated from logic and 
> earth within 
> each chassis.  The -48V return should tie back to earth at 
> the customer 
> battery plant / power supply.
> Your backplane attachment sounds OK, at least superficially.  
> There might 
> be some issues if this is a midplane or you are plugging in 
> transceivers to 
> the outside world.
> 
> Are you developing that in China or CA?  If there is someone 
> in Alameda 
> involved we could set up an hour or two consulting session to 
> review the 
> plans.
> 
> Regards,
> 
> 
> Steve.
> At 10:25 AM 3/26/2004 +0800, peter zhu wrote:
> >Steve:
> >
> >We are developping ATCA chassis and using isolated DC-DC converters.
> >In fact, In ATCA chassis, there are three gnd: shelf gnd, 
> -48v Return, and 
> >logic ground. Now I plan to tie the shelf gnd and logic gnd 
> on backpalne, 
> >and let the -48V Return separated. The 48V Return may tie to 
> shelf gnd on 
> >shelf using single point.
> >
> >On backpalne PCB, we only use one gnd stack, all logic gnd 
> and shelf gnd 
> >from connector pins are connected to this gnd stack. if the 
> screw that fix 
> >the backpalne on the margin are installed, the gnd stack 
> will be connected 
> >to shelf. do you think this is feasible?
> >
> >Thanks!
> >
> >Peter
> >----- Original Message -----
> >From: <mailto:weirsp@xxxxxxxxxx>steve weir
> >To: <mailto:peter.zhu@xxxxxxxxxx>peter.zhu@xxxxxxxxxx ; 
> ><mailto:si-list@xxxxxxxxxxxxx>si-list@xxxxxxxxxxxxx
> >Sent: Thursday, March 25, 2004 11:52 PM
> >Subject: Re: [SI-LIST] About the chassis gnd and logic gnd
> >
> >Peter, the subject has been covered many times.  There are seemingly 
> >religious views on this.
> >
> >I belong to the "reduce antenna gain" camp, which is to say 
> that at high 
> >frequencies, even an air gap does not effect isolation, it 
> just increases 
> >antenna gain and ESD susceptibility.  So multipoint bonding is the 
> >preferred strategy in my book.  Bond often, and bond well.  
> There is one 
> >caveat, and that is if you have external low voltage power 
> supplies, then 
> >you have to watch the return current path.  But since you likely use 
> >isolated DC-DC converters, this should not present an issue to you.
> >
> >This point of view is supported by practices at major successful 
> >manufacturers such as Cisco where those practices resulted 
> in behemoth 
> >boxes like the 12000 that have no EMC issues at all.  It is 
> also supported 
> >by respected authors like Mark Montrose, and Doug Smith who show the 
> >science behind why this practice works, as well as 
> successful engineers 
> >with many years experience like Chris Cheng.  My 
> recollection is that John 
> >Barnes covers this subject in his books as well.  If you can 
> afford them, 
> >and even if you can't, John brings many years experience 
> including with 
> >ultra cost sensitive consumer product.
> >
> >If you go to Doug Smith's web site, www.emcesd.com and look 
> up his April 
> >2002, May 2002, and October 2002 "Technical Tidbits" he 
> demonstrates the 
> >value of multipoint grounding very clearly for ESD immunity.
> >
> >Regards,
> >
> >
> >
> >Steve.
> >At 07:43 PM 3/25/2004 +0800, peter zhu wrote:
> >>all:
> >>About the chassis GND and the logic GND, some company tie 
> them together, and
> >>some sompanies separate them.
> >>In telecom equipments, I see many companies, such as NOKIA, tie them
> >>together, and the backplane only has uniform GND layer for 
> chassis gnd and
> >>logic gnd. We separate them in our backplane, but it will 
> result in PCB gnd
> >>layer partition. So form the PCB view, we hope the uniform gnd.
> >>What's the recommendation form the EMC/EMI view? and what's 
> your company's
> >>strategy?
> >>
> >>Thank,
> >>
> >>Peter
> >>UTStarcom
> >>
> >>------------------------------------------------------------------
> >>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:
> >>http://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
> >>
> >>List technical documents are available at:
> >>                 http://www.si-list.org
> >>
> >>List archives are viewable at:
> >>                 http://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
> >>
> >Steve Weir
> >Teraspeed Consulting Group LLC
> >2926 SE Yamhill St.
> >Portland, OR 97214
> >(503) 239-5536
> >http://www.teraspeed.com
> >-------------------------------------------------------------
> ----------------------------------------- 
> >
> >Teraspeed(SM) is the service mark of Teraspeed Consulting Group LLC
> 
> Steve Weir
> Teraspeed Consulting Group LLC
> 2926 SE Yamhill St.
> Portland, OR 97214
> (503) 239-5536
> http://www.teraspeed.com
> --------------------------------------------------------------
> ----------------------------------------
> Teraspeed(SM) is the service mark of Teraspeed Consulting Group LLC
> 
> 
> ------------------------------------------------------------------
> 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:
> http://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
> 
> List technical documents are available at:
>                 http://www.si-list.org
> 
> List archives are viewable at:     
>               http://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:
http://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

List technical documents are available at:
                http://www.si-list.org

List archives are viewable at:     
                http://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: