Universal SCSI backplane stackup

Ovidiu Mesesan omesesan at kaparel.com
Fri Jul 27 06:14:53 PDT 2001

* From the T10 Reflector (t10 at t10.org), posted by:
* Ovidiu Mesesan <omesesan at kaparel.com>
Hello everybody.
I have a question relating to the "universal" backplane construction chapter
in SPI-4.
Say I choose broadside stripline offset coupling because it looks like it is
the only one that will satisfy both LVD and SE requirements for impedance
(i.e. 125-Ohm diff for LVD and 85-Ohm for SE to match requirements in table
9 in 6.3.3 of SPI-4). 

Stackup looks like this: 
1=Top 0.5 oz (Dallas SMT terminators placed here) 
2=+5V 1oz 
3=GND (DC return) 1 oz 
4=Signal 1oz 
5=Signal 1oz 
6=GND (DC return) 1oz 
7=+12V 1oz 
8=Bottom 0.5oz. 

"Positive" side signals of LVD pais are routed on layer 5 while "negative"
side on layer 4. 
Trace width is drawn at 7 mils on both layer 4 and 5. 
Offset "S" between edge of trace on layer 4 (negative) and edge of trace on
layer 5 (positive) is 3.5 mils (on horizontal direction). Distance "H1"
between layer 4 and 5 is 21 mils (core), between 4 and 3 is 24.5 mils ("H2"
prepreg) - same "H2" between 5 and 6.

---------------- Lyr 3 GND Plane

  /      \       Lyr 4 Negative Side trace
          |S |     H1
      /      \   Lyr 5 Positive side trace


---------------- Lyr 6 GND Plane

My question is: does this stackup and parameters satisfy both LVD and SE
impedance requirements? For LVD (using Polar Calculator) I get 124 Ohm diff
and for SE, considering reference plane of traces on layer 4 as being layer
3 and 6 I get 84 Ohm. For SE I ignored the proximity of traces on layer 5.
Is this correct?

Thank you.
Ovidiu Mesesan
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at t10.org

More information about the T10 mailing list