Hi Jeff, 

I have worked with single point grounds and I agree with what 
you say about single point grounds conditionally. If single point 
ground is implemented properly (as with alot of other techniques), 
it's fine. The problems I have seen resulting from single point 
grounding involve implementing it where it isn't needed or a 
designer not really understanding it. 

What appear to be the tricks with either method you choose are 
ground-loop analysis for return currents, a sort of noise 
analysis of the board, and figuring out where currents add 
together on the board.  The bottom line that I've seen is that 
once the decision is made to go with single point grounding, 
you cannot violate it in the least. If there's the slightest 
possibility that the single point ground scheme is going to 
be compromised, go with multi-point. 

I have seen problems with multi-point ground as well. Determining 
the proper placement for each of your ground points is important. 
Each ground point is supposed to have a specific function for where 
it is used and it's good to take some time to be sure where they 
are put is where they will do the most good.  Noise levels within 
a board can be a culprit. Return currents from normal usage and/or 
from transients superimposed on one another can be another problem 
source. 

All of course dependent upon your specific product. 

Regards,  Doug

----------
> From: [email protected]
> To: [email protected]
> Subject: CompactPCI
> Date: Monday, January 26, 1998 12:12 PM
> 
> 
> I am seeking advise and comments concerning EMC testing and approval of
> CompactPCI chassis.
> 
> My background is in personal computers and similar ITE equipment. The
> current practice is to tie logic ground and chassis ground together as
> often as possible. Many circuit boards tie logic ground and chassis
ground
> together at every point where the board comes in contact with the
chassis.
> The traditional notion of single point grounding has been shown time and
> again to be inadequate for preventing radiation in systems that are
running
> frequencies of 100 MHz or faster.
> 
> My concern is for the portion of the CompactPCI specification that
requires
> that  logic ground and chassis ( Frame ) ground be isolated from one
> another.  Frame ground and logic ground can be tied together at only one
> point, the power supply.  The specification even gives a value of
> 9M Ohms of impedance between logic and frame ground ( measured with a
100V
> DC source ).
> 
> We are working on a CompactPCI chassis that will include; passive
> backplane, processor board, I/O companion board and  power supply that
will
> plug into the backplane. The chassis will have 7 slots for plug in cards
> that can be provided by anyone.
> 
> Has anyone worked with a CompactPCI chassis built to this specification?
> Has anyone worked with similar constructions that employ similar
grounding
> schemes? What problems can the single point grounding cause? Any comments
> or suggestions would be appreciated.
> 
> Thanks in advance for your input.
> 
> 
> Jeff Busch
> Compliance Engineer      [email protected]
> I-Bus                    619-974-8470
> San Diego           619-268-7863     fax
> 

Reply via email to