[SI-LIST] Re: Minimum eye opening of 8b10b encoded signals
- From: "Charles Hill" <chuck@xxxxxxxxxxx>
- To: <tkjeon@xxxxxxxx>, <dave.instone@xxxxxxxxxx>
- Date: Wed, 23 Mar 2005 18:49:25 -0700
Hi TK, Dave, Michael, et al,
A good discussion. I have a few comments:
1) SATA II is intended to replace SATA 1.0a. A lot was learned in the
process of developing the new SATA II spec. There was a concern that simply
changing the 1.0a spec would create problems (render components
non-compliant) where no real problem (BER was OK) existed. So, the 1.0a
requirements for 1i/1m were copied into SATA II.
2) Eye diagrams are not part of 1.0a nor SATA II. With an objective of
ensuring a link performance of BER<1E-12 (actually frame error rate since it
is measurable in the protocol) there are 3 major problems with eye diagrams:
a) too small of a sample size to be statistically significant to prove with
95% confidence BER<1e-12 b) eye display subject to undersampling effects and
missing timing information c) eye mask tests are deterministic leading to
problems when a single sample falls outside the allowable mask. These
concerns bear on compliance tests; eye diagrams are useful quick tests to
show a system is unacceptable (its easy to show a problem, difficult to show
performance).
3) The timing reference point is defined by the "reference clock". Thus, if
one wanted to construct a statistical "eye mask", the center of the mask is
already defined.
4) Chris is right, most receivers do operate below the minimum receive level
in SATA. Some receivers may exceed the spec requirements. However, relying
on 150mV receiver sensistivity, for example, for the link to work is relying
on an unspecified parameter and interchangability is not ensured.
5) There is no SATA II specification for the signal coming out of an
interconnect into a matched load (Lab Load). This is due to the impedance
mismatch uncertainty of the receiver as well as no knowledge of the
interconnect inside the receiver. The specification is concerned with
measurable quantities at the compliance points. The mismatch of the
receiver and its effect on minimum amplitude and jitter is part of the
design burden of the receiver designer; he knows the mismatch loss and
attenuation loss characteristics inside the receiver, that is, inside the
box the spec defines as the receiver made up of chip, package, PCB traces,
SATA connector etc.
The SATA II spec isn't perfect, but is an improvement over 1.0a. We
certainly know of some deficiencies. We welcome inputs to improve the spec.
Regards,
Chuck Hill, consultant
SATAIO Phy committee chairman
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx
[mailto:si-list-bounce@xxxxxxxxxxxxx]On Behalf Of T.K. Jeon
Sent: Wednesday, March 23, 2005 3:35 PM
To: dave.instone@xxxxxxxxxx
Cc: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Minimum eye opening of 8b10b encoded signals
Hi Dave,
Speaking of Tx min vtg measurement interval, although the Table 4 does
not say the interval for Gen1i/m, Clause 6.4.2.1 explains how to measure
the min amplitude and it says the measurement interval is in [0.45UI,
0.55UI]. I guess that there should be a note for clarification saying
whether the clause applies to both Gen1i/m and Gen2i/m or not.
And, SATA II Clause 6.4.4(page 117) states 'The minimum amplitude value
is measured 0.5UI after a reference clock edge' for Tx amplitude not
specifying Gen*i or Gen*m or Gen*x, which may also need some clarification.
Regards,
TK
Dave Instone wrote:
> Hi TK,
> according to the SATA II spec Table 4, for Gen1i and Gen1m there is
> no Tx minimum Voltage measurement interval specified, in fact there are
> many holes in that table for 1i and 1m and my interpretation of 6.1.1
> of SATA II is that it states, somewhat unclearly, that the existing SATA
> 1 spec applies to these variants. SATA 1.0a does show an eye although
> it states ' The EYE is more of a qualitative measurement than a
> spec.........Nonetheless this method is useful and easy to set up in the
> lab.'
>
> Regards
> Dave Instone
> T.K. Jeon wrote:
>
>
>>Hello Michael,
>>
>>I'm not sure if you make measurements on SATA 1.0 or SATA II. But,
>>SATA II specifies 'min/max amplitude', but not the eye opening.
>>Therefore, there is no eye mask specified in the standard. I'm saying
>>SATA II because I guess it's the latest spec.
>>
>>On SATA II, the min amplitude measurement is based on statistical
>>distributions from n samples collected between 0.45UI and 0.55UI using
>>three different patterns, which are HFTP(high frequency pattern),
>>MFTP(middle freq. pattern) and LBP(lone bit pattern).
>>
>>Regards,
>>TK
>>
>>Dave Instone wrote:
>>
>>
>>>Dear Michael,
>>> The question of where to position the eye has been the subject of
>>>many many discussions in Fibre Channel (which also uses 8b10b)and
>>>also I suspect in SATA as a lot of the work of the Fibre Channel
>>>jitter methodology appears in the SATA spec. As yet no single answer
>>>has been decided. Essentially it depends on how the clock recovery
>>>circuit centers itself. The majority view is that you determine the
>>>mean of the crossing and place the mask such that 0 UI is on the
>>>mean, this will be correct for mean following CRCs. However if the
>>>jitter distribution is skewed this can result in a lot of margin in
>>>one direction, and much less in the other, there are also
>>>oversampling type SERDES which may center themselves differently and
>>>there are also SERDES which used undisclosed methods of
>>>'compensating' for skewed distributions. To be safe you should use
>>>the mean of the crossing to position the mask.
>>> Regarding your particular violations you don't say if you are
>>>looking at the Rx end or the Tx end. If at the Rx end it seems to me
>>>that your link has too much HF loss, alternatively the TX risetime
>>>could be too slow and/or its amplitude too low.
>>>Regards
>>>Dave Instone
>>>
>>>Michael.Kurten@xxxxxxxxxxxxxxxxxxx wrote:
>>>
>>>
>>>
>>>>Hello all,
>>>>
>>>>I have a question with regard to the minimum eye opening. I'm examining
>>>>an SATA link which has 8b10b encoding. The measured eye opening does
>>>>not
>>>>exhibit a lot of jitter. Therefore the double trapezoid, describing the
>>>>minimum eye opening, can be shifted along the time axis. The eye height
>>>>instead violates massively the minimum eye opening if the minimum
>>>>eye is
>>>>centered or placed at the left corner of the measured eye. Shifting the
>>>>minimum eye opening to the right corner results in an only slightly
>>>>violated minimum eye opening.=20
>>>>To date my understanding is that if the minimum eye opening can be
>>>>placed anywhere along the time axis and not violating the eye opening
>>>>the link will work fine. Otherwise the minimum eye opening would not be
>>>>the minimum eye opening. But now I am looking for a more hardware
>>>>oriented explanation of this matter.
>>>>Any help is highly appreciated.
>>>>
>>>>Kind regards
>>>>Michael Kurten
>>>>------------------------------------------------------------------
>>>>To unsubscribe from si-list:
>>>>si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>>>>
>>>>or to administer your membership from a web page, go to:
>>>>//www.freelists.org/webpage/si-list
>>>>
>>>>For help:
>>>>si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>>>>
>>>>List FAQ wiki page is located at:
>>>> http://si-list.org/wiki/wiki.pl?Si-List_FAQ
>>>>
>>>>List technical documents are available at:
>>>> http://www.si-list.org
>>>>
>>>>List archives are viewable at:
>>>>//www.freelists.org/archives/si-list
>>>>or at our remote archives:
>>>> http://groups.yahoo.com/group/si-list/messages
>>>>Old (prior to June 6, 2001) list archives are viewable at:
>>>> http://www.qsl.net/wb6tpu
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>>
>>>
>>>------------------------------------------------------------------
>>>To unsubscribe from si-list:
>>>si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>>>
>>>or to administer your membership from a web page, go to:
>>>//www.freelists.org/webpage/si-list
>>>
>>>For help:
>>>si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>>>
>>>List FAQ wiki page is located at:
>>> http://si-list.org/wiki/wiki.pl?Si-List_FAQ
>>>
>>>List technical documents are available at:
>>> http://www.si-list.org
>>>
>>>List archives are viewable at:
>>>//www.freelists.org/archives/si-list
>>>or at our remote archives:
>>> http://groups.yahoo.com/group/si-list/messages
>>>Old (prior to June 6, 2001) list archives are viewable at:
>>> http://www.qsl.net/wb6tpu
>>>
>>>
>>>
>>
>>
>>
>>
>
>
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> //www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
> List FAQ wiki page is located at:
> http://si-list.org/wiki/wiki.pl?Si-List_FAQ
>
> List technical documents are available at:
> http://www.si-list.org
>
> List archives are viewable at:
> //www.freelists.org/archives/si-list
> or at our remote archives:
> http://groups.yahoo.com/group/si-list/messages
> Old (prior to June 6, 2001) list archives are viewable at:
> http://www.qsl.net/wb6tpu
>
>
>
>
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List FAQ wiki page is located at:
http://si-list.org/wiki/wiki.pl?Si-List_FAQ
List technical documents are available at:
http://www.si-list.org
List archives are viewable at:
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List FAQ wiki page is located at:
http://si-list.org/wiki/wiki.pl?Si-List_FAQ
List technical documents are available at:
http://www.si-list.org
List archives are viewable at:
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts: