SAS PHY transients during mode transitions questions

William Gintz wcgintz at ix.netcom.com
Thu Feb 3 18:05:58 PST 2005


* From the T10 Reflector (t10 at t10.org), posted by:
* William Gintz <wcgintz at ix.netcom.com>
*
--=====================_87055703==.ALT
Content-Type: text/plain; charset="us-ascii"; format=flowed

Hello Brian

Don't see that anyone has answered your pleas for interpretation, etc
so heres a few tips:

1. Intel has advised SATA working groups that breakdown damage from
spikes or continuous voltage exceeding the breakdown of low voltage i/o
will AGGREGATE!...Eg, all spikes will add up to deterioration.

2. Early SATA limits were derived primarily to avoid saturation, non
linearities
of receivers and transmitters, not the above (higher) breakdowns.  the
actual 
breakdowns are a function of esd clamps which vary according to design
of 
i/o cell. 

3. Early SATA didnot incorporate, nor authorize hot plugging nor voltage
level
switching. Transient evaluations were restricted to recovery time to
threshold
levels, not breakdown issues. 

4. Swithing, backplane pluging etc will create transients. As far as I
know, 
only D. Huber  (Ex DEC I now think Compaq or Maxstor?) has published any
simulations to predict transients. This was presented at a T11 meeting
in 04,
should be in documents in T11.  VALUES PREDICTED WERE ~2 Times Vs!
USER BEWARE OF DAMAGE, RECOVERY TIMES MAY VARY. 
(such simulation work is labor intensive, not likely freely offered!)

5. Bill Bissonette is also in draft stage of a document on this
topic..try this:
http://www.t10.org/ftp/t10/document.05/05-019r1.pdf
<http://www.t10.org/ftp/t10/document.05/05-019r1.pdf>  



regards




At 09:57 AM 1/31/2005, Day, Brian wrote:


* From the T10 Reflector (t10 at t10.org), posted by:
* "Day, Brian" <Brian.Day at lsil.com>
*
I've been trying to understand more of the issues being discussed on the
transients during mode transistions.  I have a couple basic questions
regarding the various numbers in the SAS 1.1 rev 7 draft spec,
particularly
related to the common mode voltage.

Table 28 lists +/- 1.2V for transmitter transients during a mode
transistion.  As long as Vp and Vn are within the 1.2V range on the test
circuit, is there a range that the common mode voltage must remain
within as
well?

Also in Table 28, shows the Receiver AC common mode specs for voltage
and
frequency.  Do these apply during mode transitions in OOB, or only
during
"normal" data transfer?  I would think not during mode transitions.

Table 29 lists transmitter OOB common mode delta of +/- 50 mV.  The
notes f
and g make me believe that this is not required to be met during a mode
transition, but only during the valid OOB burst signalling.  True?

Table 30, for the maximum noise during OOB idle time of +/- 120 mV.  I
think
this also would mean the allowable noise during the "real" idle time
(and
negation time) of an OOB signal.  A mode transition farther away in time
of
an OOB signal may go beyond 120 mV, and doesn't really matter?

Thanks in advance for helping a digital guy understand these analog
specs
more.

Brian Day
LSI Logic










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

regards,

W.C. Gintz
SEUS, Inc 
wcgintz at ix.netcom.com
office    650 494 0335
mobile  650 799 6622 


--=====================_87055703==.ALT
Content-Type: text/html; charset="us-ascii"


 Hello Brian

 Don't see that anyone has answered your pleas for interpretation, etc
 so heres a few tips:

 1. Intel has advised SATA working groups that breakdown damage from
 spikes or continuous voltage exceeding the breakdown of low voltage i/o
 will AGGREGATE!...Eg, all spikes will add up to deterioration.

 2. Early SATA limits were derived primarily to avoid saturation, non linearities
 of receivers and transmitters, not the above (higher) breakdowns.  the actual 
breakdowns are a function of esd clamps which vary according to design of 
i/o cell. 

3. Early SATA didnot incorporate, nor authorize hot plugging nor voltage level
 switching. Transient evaluations were restricted to recovery time to threshold
 levels, not breakdown issues. 

4. Swithing, backplane pluging etc will create transients. As far as I know, 
only D. Huber  (Ex DEC I now think Compaq or Maxstor?) has published any
 simulations to predict transients. This was presented at a T11 meeting in 04,
 should be in documents in T11.  VALUES PREDICTED WERE ~2 Times Vs!
 USER BEWARE OF DAMAGE, RECOVERY TIMES MAY VARY. 
(such simulation work is labor intensive, not likely freely offered!)

 5. Bill Bissonette is also in draft stage of a document on this topic..try this:
 http://www.t10.org/ftp/t10/document.05/05-019r1.pdf 



regards

 


At 09:57 AM 1/31/2005, Day, Brian wrote:
 * From the T10 Reflector (t10 at t10.org), posted by:
 * ;Day, Brian; <Brian.Day at lsil.com>
 *
 I've been trying to understand more of the issues being discussed on the
 transients during mode transistions.  I have a couple basic questions
 regarding the various numbers in the SAS 1.1 rev 7 draft spec, particularly
 related to the common mode voltage.

 Table 28 lists +/- 1.2V for transmitter transients during a mode
 transistion.  As long as Vp and Vn are within the 1.2V range on the test
 circuit, is there a range that the common mode voltage must remain within as
 well?

 Also in Table 28, shows the Receiver AC common mode specs for voltage and
 frequency.  Do these apply during mode transitions in OOB, or only during
 ;normal; data transfer?  I would think not during mode transitions.

 Table 29 lists transmitter OOB common mode delta of +/- 50 mV.  The notes f
 and g make me believe that this is not required to be met during a mode
 transition, but only during the valid OOB burst signalling.  True?

 Table 30, for the maximum noise during OOB idle time of +/- 120 mV.  I think
 this also would mean the allowable noise during the ;real; idle time (and
 negation time) of an OOB signal.  A mode transition farther away in time of
 an OOB signal may go beyond 120 mV, and doesn't really matter?

 Thanks in advance for helping a digital guy understand these analog specs
 more.

 Brian Day
 LSI Logic

 








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

 W.C. Gintz
 SEUS, Inc 
wcgintz at ix.netcom.com
 office    650 494 0335
 mobile  650 799 6622 

--=====================_87055703==.ALT--





More information about the T10 mailing list