Email: Password: Remember Me | Create Account (Free)

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
Joseph Hebert
04/08/13 10:59
Read: 926 times
Okemah, Ok
United States


 
#189635 - How do they do so?
Responding to: Oliver Sedlacek's previous message
Hi Oliver,

I'm not familiar with PCIe. How do they control impedance across the edge connectors, and how closely do they match impedance? I'm basing my effort on a paper I read many years ago that reported using grounded vias placed around signal vias to control impedance. The surrounding gnd'd vias supposedly act as sort of orthogonal gnd planes allowing a path for return current.

If you've ever looked at the old parallel connectors (e.g. printer ribbon cables), they used a technique wherein they alternated signal lines next to gnd lines. They said it was for "shielding," but I suspect it was a form of (perhaps fortuitous) impedance control. I'm trying a similar arrangement wherein the entire row of the edge connector contacts are gnd'd. This way, each signal will enter and pass through the connector adjacent to a pseudo-backplane return path.

Rather than just testing to see if it works, I want to measure the realized impedance. I also want to look at the effect of the stubs left by the thru-hole pins. Finally, since I can't really vary the impedance of this geometry (beyond possibly playing with the stub length of the pins), if I can measure it I might be able to adjust the traces to match the impedance, or at least get them close enough to work.

Joe

List of 19 messages in thread
TopicAuthorDate
Measuring trace impedance with a TDR?      Joseph Hebert      04/01/13 08:53      
   Maybe there's help here ...       Richard Erlacher      04/01/13 09:23      
      Thanks. Hardware/model specific      Joseph Hebert      04/01/13 10:22      
         Beauty contest      Steve M. Taylor      04/03/13 05:32      
            Growing your own      Joseph Hebert      04/03/13 16:56      
               What information do you want to extract?      Richard Erlacher      04/04/13 10:55      
                  desired information      Joseph Hebert      04/04/13 14:04      
                     "pre-loved"      Steve M. Taylor      04/05/13 12:27      
                        Pre "loved"      Joseph Hebert      04/05/13 16:40      
                        Resolution      Joseph Hebert      04/06/13 13:42      
                           ???      Steve M. Taylor      04/09/13 16:38      
                              Plenty, I'm sure, but ...       Richard Erlacher      04/09/13 21:03      
                              cost      Joseph Hebert      04/11/13 17:31      
                                 Though this is a large and cumbersome instrument ...      Richard Erlacher      04/12/13 10:24      
                     Better than standard solutions?      Oliver Sedlacek      04/08/13 01:59      
                        How do they do so?      Joseph Hebert      04/08/13 10:59      
                           Magic?      Oliver Sedlacek      04/09/13 01:43      
                           signal integrity      Andy Peters      04/16/13 15:23      
                              Exactly      Joseph Hebert      04/16/13 18:41      

Back to Subject List