1 members (Scott35),
527
guests, and
32
robots. |
Key:
Admin,
Global Mod,
Mod
|
|
|
Joined: Mar 2005
Posts: 1,213
OP
Member
|
Is there a requirement for power in a data center to be automatically secured if the sprinkler system discharges? Or just that there be an EPO button by the exits?
|
|
|
|
Joined: Jan 2005
Posts: 5,445 Likes: 3
Cat Servant Member
|
Well ... what is the relevant code?
You won't find such a requirement in the NEC. By way of illustration, for kitchen range hoods, the requirement is spelled out in NFPA standards 17A and 96.
From a technical viewpoint, the trigger mechanisms for nearly all but the smallest systems have microswitches at the trip mechanism. It's probably just as easy to ti in here, than to have a separate push button.
|
|
|
|
Joined: Mar 2005
Posts: 1,213
OP
Member
|
I couldn't think of anything like that in NEC, and I couldn't find anything in IFC either. I figured that if there was a requirement, someone here would certainly know it off the top of their head and save me hours of searching! It's unfortunately far easier to find an applicable requirement than to verify that there is no requirement.
ICC and NEC apply here.
|
|
|
|
Joined: Jan 2005
Posts: 5,445 Likes: 3
Cat Servant Member
|
Steve, I'd at least scroll through the NFPA catalog, just to see if they have a standard specific to such rooms. While such a standard might not be 'code,' it's a good place to start.
|
|
|
|
Joined: Aug 2007
Posts: 853
Member
|
Is there a requirement for power in a data center to be automatically secured if the sprinkler system discharges? Or just that there be an EPO button by the exits? To my knowledge there is no requirement for EPO. Local Authorities will and do require a button EPO at a minimum,Around here (Massachusetts). Pre-Action sprinkler: most common in Data rms is in NFPA13,also NFPA 76 (Telecom).Can find no code for Data centers,specific.
Last edited by leland; 01/27/09 01:10 AM.
|
|
|
|
Joined: Mar 2005
Posts: 1,213
OP
Member
|
NFPA-13 8.5.10 requires sprinklers in electrical equipment rooms (but all the equipment can be NEMA3, which makes this seem rather pointless), and 21.15.2.1 requires IT areas be valved separately from other areas. I found no other requirements in NFPA 13.
NFPA 76 6.5 requires a means to disconnect power be identified, but not that it be automatic. 6.6.1 requires ventilation be automatically dampered when smoke is detected. (Not upon sprinkler activation.) 10.6 requires that depowering occur only after the investigation determines that depowering will mitigate the event.
Well, THAT surprised me! I didn't find a requirement that says I have to put a shunt trip on a data center, but I *did* find a piece of code that specifically prohibits it!
NFPA 110 & 111 are silent on this matter.
|
|
|
|
Joined: Mar 2005
Posts: 1,213
OP
Member
|
Eh, curse my luck, it's a UFC Requirement specific to DoD ADP areas, communication centers, command and control systems, and other mission critical systems. http://www.wbdg.org/ccb/DOD/UFC/ufc_3_600_01.pdfUFC 3-600-01 6-8.1.1 Automatic Sprinkler Protection
Electronic equipment installations must be located in buildings protected by wet-pipe automatic sprinklers. Provide complete coverage throughout the building including electronic equipment areas. Protect electrical equipment installations by disconnecting the power upon activation of the fire protection system.
|
|
|
|
Joined: Jul 2004
Posts: 9,931 Likes: 34
Member
|
The computer room disconnect is in 645.10 It has to close the HVAC dampers but it doesn't need to be activated by the fire suppression system. The controls need to be grouped at "the principle exit doors". Usually you have a halon trip, a halon abort and an EPO all under plastic bubbles. The sequence for the halon is a klaxon goes off and x number of seconds later the halon dumps. You have those seconds to decide if this is really a fire or if you want to save the $100,000 worth of halon (the 1996 price, it is probably $300k now) for another day. I have been in a halon dump, no fun. (think tornado) I also cleaned up after another one. They might use dry pipe sprinklers in computer rooms now but I never saw it much in the 20th century. Wet pipe is definitely not recommended. (leaks) AEC/DOE used to have a CO2 system in their Germantown Md computer room. There was some talk that OSHA made then pull it out.
I am surprised that DoD wants shunt trips on "mission critical systems". When I was in the military, mission critical systems had "battle override" that shunted out overcurrent devices. The thought was we will keep fighting until this thing burns up.
Greg Fretwell
|
|
|
|
Joined: Mar 2005
Posts: 1,213
OP
Member
|
Yeah, there's always contention over EPO buttons, too! Surprise-surprise that operators don't like a giant red "push here to kill the mission" button right out in the open where anyone can push it. There's enough weasel words in 645.4 that they frequently get their way, too. Some might consider the lack of fire-rated walls a violation... others, the excuse they need to ignore other codes they don't like.
DoD instructions, MIL-HDBKs and the like are like NFPA, though, in that they're living documents that can be changed; easier, really, because there's no set time limit, they can be revised at any time, and there's usually only one person that needs convincing and an email from them is just as good as an official revision. It's just a matter of finding the person (or office) who's responsible for it and giving them a call.
|
|
|
|
Joined: Jul 2004
Posts: 9,931 Likes: 34
Member
|
The "anyone can push it" concern is why these things are usually under a bubble cover. In a big data center an unscheduled power off can cost $100,000 in data damage, lost productivity and lost revenue. It might take over an hour to get things going again and longer than that to recover any lost data. Of course some may be lost forever.
Greg Fretwell
|
|
|
Posts: 1,803
Joined: March 2005
|
|
|
|