FU: UPS on an i5 520 - Not Your Father's Attachment

All,

Back in February, I posted a number of messages regarding UPS attachment to
a i5 520.  I think my particular situation is now resolved, but I felt I
needed to post a summary of "what is."

After being involved with a number of people at IBM on this, I find that we
have 2 completely UPS attachment scenarios:

1) Historically, UPS was attached via a J14 port.  The i5 520/550/570 models
do not have a J14 port.  Instead you must connect to a serial port.  This
serial port acts entirely different from the J14 port in a number of ways.
At least one of these, to me, is a defect.  The defect is that the FSP
_only_ recognizes UPS attachment to this serial port if an 1827 cable is
attached at the time the FSP is powered up.  After power up, it makes _not
one iota of difference_ if there is a cable attached or not.  I can unplug
it after power up and the ASMI says it's still attached.  If none was
attached at power up, I can attach one later and the ASMI will still say
none is attached.


2) The MATMATR API pays no attention to this serial port in any case.  The
MATMATR API only works if you have a J14 port.  Since I got an 0595
expansion unit with my 520, I found this out accidently because, lo and
behold, even though IBM can't spring for a J14 on a 520 system, they put a
fully functional one on the expansion unit.  If I short the various pins in
the J14 port per an IBM internal document instructions, the MATMATR works.
But the ASMI shows nothing.


3) This means IBM has 2 completely different and unassociated methods for
determining UPS attachment.  There is the serial port that the ASMI looks
at.  Then there is the J14 port (your father's UPS attachment method) that
the MATMATR API uses.                                                       

But I think IBM still has a problem with that serial port.  Be advised that
if you're using an 1827 cable (of which there were many threads here), you
do _not_ have the same level of communication from your UPS to the i5 that
you used to have with the J14.  The serial port connection signals are 1)
UPS attached, and 2) battery low.  That's it.  No 'on bypass' indication.
No 'on battery' indication.  You read that right.

I have both HW and SW PMRs still open at IBM on this, and I don't believe
until just very recently, that there are many (if any) people in IBM that
know and understand the entire situation regarding UPS attachment and what a
problem it has become. 

Though I'm not an electrician by any stretch, I don't consider myself a
dummy.  (Comment welcome here. <g>).  I just don't understand how IBM got
the whole UPS attachment thing situated this way. 

HTH someone else.

(To compound the above for my case, I found out from Liebert, my UPS
manufacturer, that their UPS connection signal output has changed.  It only
took 6 months and talking to 4 different people at Liebert to find this out
and is at odds with what those people told me 2 years ago when I got the new
UPS.  Believe it or not, even though I'm using the EXACT same cable as with
the old Liebert UPS, connected to the same COM1 output port, Liebert has
completely changed the signals generated.  To get to what I used to have,
for connecting to that J14 port, I must put another card into my UPS and
have a cable custom made by Black Box.  Both of those items are now
installed.  This now puts me back to where I was 2 years ago before this all
started.)