Question on SAS-1.1 Sinusoidal Jitter Limits

Bill Lye lye at pmc-sierra.com
Thu May 12 10:27:34 PDT 2005


* From the T10 Reflector (t10 at t10.org), posted by:
* Bill Lye <lye at pmc-sierra.com>
*
With reference to sas1r09c.pdf, notes c and d in Table 50 (PDF page 201,
document page 161) say:

  c  Applied sinusoidal swept frequency: 900kHz to the minimum of 5MHz and
     (3 x 2^(generation-1) MHz) (e.g. 5MHz for 1,5Gbps and 7,5 MHz for 3Gps)
  d  Applied sinusoidal swept frequency: 1 800kHz to the minimum of 5MHz and
     (3 x 2^(generation-1) MHz) (e.g. 5MHz for 1,5Gbps and 7,5 MHz for 3Gps)

This text was fed into sas1r08.pdf via 04-370r3.  In previous iterations
of SAS-1-1, this text was ">5MHz".  04-370r2 had the text ">5 MHz or
the highest frequency trackable by the receiver device".  In Las Vegas
we agreed that we did not like this text, and that we wanted to increase
the upper limit with the bit rate and wanted to place a minimum of 5MHz,
thus Rob created 04-370r3 with the current wording.

For generation=1, min(5MHz,3MHz*2^(generation-1))=3MHz (not 5MHz),
and for generation=2, min(5MHz,3MHz*2^(generaion-1))=5MHz (not 7.5MHz).
This is clearly not in line with what was intended.

A small modification to the equation, making it max(...) not min(...), 
gives 5MHz and 6MHz, not 5MHz and 7.5MHz.

Rob, was your intent to create an equation that would map Gen1 to 5MHz,
Gen2 to 7.5MHz and Gen3 to 15MHz (in which case the equation is wrong),
or was it your intent to map Gen1 to 5MHz, Gen2 to 6MHz, and Gen3 to 12MHz
(in which case the examples are wrong)?

Either way, this probably needs cleaning up.

-Bill
-- 
"I don't prefer Verilog over VHDL; I hate them both.  The
 one I hate least is whichever one I'm not currently using."
*
* 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