[Beowulf] Re: Emergency Power Off

David Mathog mathog at caltech.edu
Fri Mar 23 08:20:52 PDT 2007

Tom Mitchell wrote:

> While thinking about grounding look at it with care
> from the UPS event point of view too.

Yes, I've been reviewing this, and the more I look the more complicated
it gets.  The one bright point is that it turns out that most of the 
UPS units in the room can be turned off by just shorting two
pins together, either on an existing RJ11 EPO jack (2 TrippLites), via
a special cable (1 TrippLite), or an external box (1 APC).  That
leaves just two that need externally supplied pull up voltages,
and that can be supplied by separate batteries at the unit.  There's
some question about whether the APC unit must have an EPO, it's 750VA,
which is the largest capacity that normally does not require an EPO.
However, it's not clear at the moment if by placing it in the room with
another 5000VA of UPS it might not cancel that exemption.  (Thought
experiment:  consider a data center with 1000 separate 750VA
UPS systems, none of which have an EPO.  Not very safe, is it?)

After consulting with the fellow on campus in charge of code
compliance apparently we can't get away with adding a second big
red button, the one big red button has to do everything.  So now I'm
leaning towards adding more contact blocks to the existing button
(if possible), or replacing it (if not).  One switch would go to
the existing panel cutoff wiring, and the rest would be wired up inside
the wall to an RJ11 keystone jack panel close to the wire
tray near the ceiling. These switch blocks are rated for largish
currents and voltages, here they would only be switching signal level
electricity.  From the panel, 4 wire cables with RJ11 connectors
on both ends run to the UPS units.  For the two units that need
an external positive voltage the battery, resistor, and diode
would be wired in a small box in that RJ11 cable.

Bottom line, I think you're right, better to have completely separate
circuits for each EPO line, with no electrical connection between them
at all.  The only expensive part for my proposed solution is the
replacement switch.

I've been looking for an off the shelf EPO solution, but so far the only
things I've found are for huge data centers.  Anybody know of a product
in my size range, to control roughly 10 EPO shut offs?  There's a point
where adding more switch blocks becomes a bit awkward. 


David Mathog
mathog at caltech.edu
Manager, Sequence Analysis Facility, Biology Division, Caltech

More information about the Beowulf mailing list