Computer rooms - EPO's

[mshields]

Sanity check - the requirement for the EPO to shut down all electronic equipment in a computer room (Article 645) does not include lighting correct? And if it does does that include emergency lighting?

Thanks,

Mike

 

[dereckbc]

Mike you have it correct.
However as a fellow PE who designs data centers I strongly urge you to re-consider taking the leniencies associated with 645, EPO's are a really bad idea.

 

[mshields]

An interesting response

An interesting response

And much appreciated though you leave me in suspense. Can you elaborate as to why they are a bad idea? Is it a concern for the data or a safety concern?

And regardless, wouldn't this be more than a leniency but rather a flagrant violation of the NEC?

Sounds like interesting stuff. Do fill me in!

Mike

 

[dereckbc]

Sure Mike I can tell you all kinds of horror stories. The bottom line is they can loose billions of dollars in revenue. All it takes is one disgruntled ignorant employee, or a dumb visitor to push the button. Let's not forget about dumbo the technician checking voltages for something unrelated and gets into the circuit, or my favorite a failure and false trip.

Think about this; an installer sets a cabinet near one of the relays. He accidentally drops the cabinet, or something on the control panel and jars the trip relay. The room goes quite, then you hear OH #%$^. I can go on if you need more. Bottom line is false unnecessary trips. It will happen

Anyone of these real life scenarios will trip the whole data center off-line. If it happens to be a large one say like Verizon, ATT, etc these facilities run millions of dollars of revenue per minute, so even a short re-start event of say an hour can run into billions, not to mention the losses of customers as a result.

There is no code requirement that dictates you have to use 645. It is optional by design. All reputable phone companies avoid them like the plague. In my 30 years only two places demanded them and was later removed by court, King of Prussia, and San Jose.

Maybe Ron or Sam will chime in as they also have Data Center design experience.

 

[don_resqcapt19]

Mike,
Article 645 is an "optional" article. There is nothing that says you must build a "computer room" to that article. The only time you must comply with the rules in 645 is when you are taking advantage of the leniencies that are permitted in that article.
Don

 

[ron]

Mike,
I agree with what was said.
The reason that 645 can be considered optional, is in the text of the code section.
645.2 (2002) or 645.4 (2005) both indicate that the article (645) shall apply if you meet several conditions, such as EPO, certain HVAC items, listed IT equipment, fire separation, etc.
So if you don't meet those conditions, then the article doesn't apply.
It is set up so that if you want to use the more relaxed wiring methods later on in that article (645), then the "prerequisites" must be met, such as the EPO, etc.
Try to avoid 645 and wire the room exactly like any other equipment room.
For casual reading, take a look at the article below.
http://www.csemag.com/article/CA6290839.html

 

[steve066]

I think I have probably seen a dozen posts saying "avoid Article 645".

But I still have two questions.

1. Exactly what do you loose by not using 645? Power cords running to receptacles below the floor? What about standard Cat 5 and other computer cables? Can they still run under the floor? Or can they run from device to device overhead?

2. I sometimes design rooms like this. But all the computer equipment and wiring is usually done by the owner or another contractor after the basic room is constructed. How can I avoid 645 when I'm not sure what the owner will install?

Steve

 

[ron]

Steve,
You only lose out on the more relaxed rules that are listed in 645.
Such as running a power strip cord through the raised floor tile and plugging into a receptacle below the raised floor. Most folks can't use this leniency anyway since they can't find a power strip that has a DP rated cord for entry below the floor.
Probably the most used leniency found in 645, is that a listed raceway whip from a PDU or RPP can be run below the floor and doesn't have to be secured.
Another that is used in some jurisdictions is that you can run non-plenum rated communications cable below the floor, even if it is a plenum. Most don't use this one either, since a lot of AHJ's require fire suppression to be installed below the floor when there are "combustibles" (non-plenum rated cable) below the floor. It is funny though, since plenum rated cable is combustible, by because of the name, plenum rated, it is allowed below the floor most times without fire suppression.
Those are a few.
You need to make clear to the client that if they don't want an EPO, then they just have to install the infrastructure and equipment like any other office or equipment room.

 

[dereckbc]

Ron touched on something I would like to expand on in reference to running cables under the raised floor. Initially with the cables under the raised floor makes for a very clean room as none of the cables are visible. However the real purpose of the raised floor is to be used for a plenum for positive pressure HVAC distribution.

Equipment installers are very lazy when it comes to running cables, and have no plans on cable routing, and spend even less time dressing them if they don?t have too. Running under the floor just gives them the excuse to be sloppy. The processing equipment requires a lot of cabling, and over time as the room is filled up, the area beneath the raised floor becomes very congested and will block the air flow causing hot spots through out the room. Now couple that with cable mining, it is not possible with the chaotic install method when ran under the floor.

The real solution is an engineered structured cabling system complying with Bellcore Network Equipment Building Systems (NEBS). This evolves installing multi level cable racks and raceways to facilitate the various cable groups used. This typically includes three ladder type cable racks or optical troughs installed centered above and running parallel to the equipment line-ups segregated by signal type (alarm, telemetry, signal, optical, dc power, and ground). Further more installation practices should be put into place and strictly enforced as to how and where cables are installed. Basically the cables have to be formed and laced in such a manner you can place your finger on the cable and run it along the entire length of the route with out any cross-overs, twist, or divers.

 

[mshields]

Very interesting stuff

Very interesting stuff

For some reason I stopped getting instant notification on this string. I want to thank all for this very interesting conversation. Sounds like I should not only look at EPO's but also the wiring method we use for under the raised floor. We call for UL listed as an assembly, Liquid Tight Flex. I believe, this application requires the EPO. Incidentally, the one thing that we do to protect against accidental triggering of the EPO is to provide a cover like those that are used over FA pull stations where there is a concern for kids pulling them for kicks. In this way, it would take more than someone leaning on it, or something falling on it, etc. Still, the point is taken.

thanks,

mike

Thanks again,

Mike

 

[iwire]

Incidentally, the one thing that we do to protect against accidental triggering of the EPO is to provide a cover like those that are used over FA pull stations

Mike using a cover is common and makes sense.

But as a tradesman I hate working in a data center that someone else has wired.

I can never tell how the EPO system is set up and what may accidental trigger it. I knocked out a large PDU simply by closing the door on it. I called Liebert and they explained they where having an issue with the factory installed EPO button on certain units. They came out and rewired it, at least that got me out of hot water with the customer, but they still lost a couple of weeks of testing time.

 

[sandsnow]

I go through this dog and pony show all the time with designers and even some inspectors.

Someone somewhere started the myth that EPO's are required. We had people putting them in closets with a PC.

Another tack:
People just don't understand or realize what they unleash when they choose to comply with 645. Fire resistive Walls FLOORS and CEILINGS. All that crap leaving the room has to be firestopped. Huge problem if you don't plan for it.

So as a plan checker I caution people sort of to be carefull what you wish for.

 

[mshields]

One more question

One more question

Where does it say in the NEC that only chapters 1 through 4 are mandatory.

Furthermore, doesn't each state need to adopt the NEC to make it law? e.g. here in MA, we go by the MEC which is a modified version of the NEC.

Mike

 

[ron]

Mike,
If your jurisdiction adopted Article 90 of the NEC, look at 90.3.

 

[DHkorn]

"Article 645 is an "optional" article. There is nothing that says you must build a "computer room" to that article. The only time you must comply with the rules in 645 is when you are taking advantage of the leniencies that are permitted in that article."

I've heard the same thing at an NFPA seminar. Our instructor skipped the whole article as a waste of time. The benefits were so meager, and the requirements so thick.

 

[Dave58er]

I'm still a little confused as to whether the "problem" with 645 is down the road for the customer or with the EC doing the wiring.

If the difficulties will be dealt with by the customer, and 645 makes the job easier and cheaper, how can the EC not do it that way when everyone he is bidding against will?

I have not worked in a lot of these rooms. Every computer room I've worked in (5 or 6) has been wired under 645 but not by my choice and I wasn't in a position to have to learn many of the details.

I just kind of assumed it was a big advantage and that they were all done that way.

I recently saw plans and specs for a 911 dispatch center. They called for a 3 double duplex receptacle boxes fed from under floor wiring methods at each station. There were over a dozen stations in each of two rooms. How would you feed these if it were a 645 room? How about if it weren't?

 

[dereckbc]

If the difficulties will be dealt with by the customer, and 645 makes the job easier and cheaper, how can the EC not do it that way when everyone he is bidding against will?

That is an oxymoron because to build a room compliant to 645 is far more expensive than conventional methods. Besides this isn’t something an EC decides, the client requires the AE to design it, bids it out, and an EC builds it to design plans.

 

[ron]

Dave,
If the "whips" under the floor were NOT going to be listed per 645.5(E), then either way, 645 or Article 300, you would have to install them the same way. Securing them per Article 300.

The only small benefit from 645 would have been found if the "whips" were listed, then they wouldn't have to be secured if it was a 645 compliant room. The types of raceway permitted are also slightly different when you compare 645.5(D)(2) and 300.22(C)

 

[Dave58er]

... 645 is far more expensive than conventional methods.

This is the upshot I was trying to get to. Most every time 645 comes up I here how much it sucks, but it usually sounds like it's from the customers perspective.

I have limited experience in these rooms. I wasn't on the planning and/or management side at the time. I was paying more attention to doing my job well than I was, sometimes, to why I was doing that job. Also we were following specs so there wasn't much discussion of this way versus that.

I know that many of you are experts with this stuff and it's probably irritating to talk to someone so ignorant on this subject. Thank you for taking the time to help enlighten me.:smile:

 

[mshields]

Real life examples of EPO's wreaking havoc

Real life examples of EPO's wreaking havoc

This thread has been a real eye opener. I'd like to prepare a training session for our department. A great attention getter would be some true disaster stories. Can anybody point me in the right direction.

Thanks,

Mike