X-MimeOLE: Produced By Microsoft Exchange V6.5
Received: by onstor-exch02.onstor.net 
	id <01C83319.AB00E93D@onstor-exch02.onstor.net>; Thu, 29 Nov 2007 22:24:31 -0800
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C83319.AB00E93D"
Content-class: urn:content-classes:message
Subject: Update on Rev 2 hardware
Date: Thu, 29 Nov 2007 22:23:09 -0800
Message-ID: <BB375AF679D4A34E9CA8DFA650E2B04E523480@onstor-exch02.onstor.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Update on Rev 2 hardware
Thread-Index: AcgzGXn916nNc9pOTwq5Y1kPnmBBHg==
From: "Brian Stark" <brian.stark@onstor.com>
To: "dl-Cougar" <dl-Cougar@onstor.com>

This is a multi-part message in MIME format.

------_=_NextPart_001_01C83319.AB00E93D
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Hi,
=20
We've been making progress the last couple of days in the lab, and I =
thought that I'd share an update on what we've found on the Rev 2 =
motherboards:
=20
- As I mentioned in the dev meeting on Tuesday, the most serious problem =
is a failure with TXRX memory controller 1.  After adjusting several =
memory controller parameters like drive strength, timing, slew rate, =
etc., the number of errors has been reduced from several hundred to just =
a few and there has been definite improvement in signal integrity.  =
We've also discovered that there is a dependency on power supplies, =
which could indicate either a grounding issue or a problem with the low =
voltage rails that supply the memory circuits.  We're not yet sure why =
this only affects one memory controller on one of the 1480s, and we'll =
be looking at this further.
=20
- The TXRX PHY issue was due to misalignment of the BGA to the landing =
pattern on the board.  Also, the wrong type of BGA was installed given =
the solder process that is used.  The boards will be reworked with the =
correct BGA and then xrayed to make sure alignment is correct.
=20
- A few other boards had isolated problems, and the issues have been =
identified on 2 of these.  One board could not access one of the =
QLogics, and this was due to a shorted capacitor on one of the PCI-e =
lanes.  Another board had an issue with the FP 1480, and a clock driver =
that is supposed to output 100MHz was instead outputing 400MHz.  The =
clock driver is being replaced and hopefully no other issues are found =
on that board.  Finally, one other board has an issue with the SSC =
booting that is related to memory, and this board is still under debug
=20
We'll hopefully make more progress on the TXRX memory tomorrow, and I'll =
send out an update as soon as we know more.
=20
=20
Brian
=20
=20
=20

------_=_NextPart_001_01C83319.AB00E93D
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<HTML dir=3Dltr><HEAD>=0A=
<META http-equiv=3DContent-Type content=3D"text/html; charset=3Dunicode">=0A=
<META content=3D"MSHTML 6.00.2900.3199" name=3DGENERATOR></HEAD>=0A=
<BODY>=0A=
<DIV><FONT face=3DArial color=3D#000000 size=3D2>Hi,</FONT></DIV>=0A=
<DIV><FONT face=3DArial size=3D2></FONT>&nbsp;</DIV>=0A=
<DIV><FONT face=3DArial size=3D2>We've been making progress the last =
couple of days in the lab, and I thought that I'd share an update on =
what we've found on the Rev 2 motherboards:</FONT></DIV>=0A=
<DIV><FONT face=3DArial size=3D2></FONT>&nbsp;</DIV>=0A=
<DIV><FONT face=3DArial size=3D2>- As I mentioned in the dev meeting on =
Tuesday, the most serious problem is a failure with TXRX memory =
controller 1.&nbsp;&nbsp;After adjusting several memory controller =
parameters like drive strength, timing, slew rate, etc., the number of =
errors has been reduced from several hundred to just a few and there has =
been definite improvement in signal integrity.&nbsp; We've also =
discovered that there is a dependency on power supplies, which could =
indicate either a grounding issue or a problem with the low voltage =
rails that supply the memory circuits.&nbsp; We're not yet sure why this =
only affects one memory controller on one of the 1480s, and we'll be =
looking at this further.</FONT></DIV>=0A=
<DIV><FONT face=3DArial size=3D2></FONT>&nbsp;</DIV>=0A=
<DIV><FONT face=3DArial size=3D2>- The TXRX PHY issue was due to =
misalignment of the BGA to the landing pattern on the board.&nbsp; Also, =
the wrong type of BGA was&nbsp;installed&nbsp;given the solder process =
that is used.&nbsp; The boards will be reworked with the correct BGA and =
then xrayed to make sure alignment is correct.</FONT></DIV>=0A=
<DIV><FONT face=3DArial size=3D2></FONT>&nbsp;</DIV>=0A=
<DIV><FONT face=3DArial size=3D2>- A few other boards had isolated =
problems, and the issues have been identified on 2 of these.&nbsp; One =
board could not access one of the QLogics, and this was due to a shorted =
capacitor on one of the PCI-e lanes.&nbsp; Another board had an issue =
with the FP 1480, and a clock driver that is supposed to output 100MHz =
was instead outputing 400MHz.&nbsp; The clock driver is being replaced =
and hopefully no other issues are found on that board.&nbsp; Finally, =
one other board has an issue with the SSC booting that is related to =
memory, and this board is still under debug</FONT></DIV>=0A=
<DIV><FONT face=3DArial size=3D2></FONT>&nbsp;</DIV>=0A=
<DIV><FONT face=3DArial size=3D2>We'll hopefully make more progress on =
the TXRX memory tomorrow, and I'll send out an update as soon as we know =
more.</FONT></DIV>=0A=
<DIV><FONT face=3DArial size=3D2></FONT>&nbsp;</DIV>=0A=
<DIV>&nbsp;</DIV>=0A=
<DIV>Brian</DIV>=0A=
<DIV>&nbsp;</DIV>=0A=
<DIV><FONT face=3DArial size=3D2></FONT>&nbsp;</DIV>=0A=
<DIV><FONT face=3DArial size=3D2></FONT>&nbsp;</DIV></BODY></HTML>
------_=_NextPart_001_01C83319.AB00E93D--
