Power room close to server room

No, just use decent surge/transient protection on the power to the servers.

Your kidding, high level of voltage? define please.

Server rooms are typically fed from PDU's. Do you know what they are and/or how they work?

Service equipment areas typically may have elevated levels of EMF's (electro-magnetic fields) which are difficult to shield against. The level might be comparable to standing under power lines or slightly greater depending upon the loading.

Also, unless there is some sort of open switchboard, the electric field component, is well shielded and effectively minimized to almost zero.

Neither of these components associated with AC current should cause any problems with servers being in close proximity.

If you are worried about your hard drives becoming erased, it is highly unlikely that you would suffer any ill effects, from these low level EMF's. Erasure levels need to be many orders of magnetic greater field strength and concentrated such as with a commercial tape degausser.

Beachcomber

Concepts that every server room should address are demonstrated by Montandon and Rubinstein in their 4 Nov 1998 paper entitled "Some Observations on the Protection of Buildings Against the Induced Effects of Lighning" in IEEE Transactions on Electromagnetic Compatibility. The same techniques used to make a direct strike to building (or incoming utlities) irrelevant also make other lesser problems (such as nearby fields) irrelevant. Figures demonstrate how to run cables between various server room components AND how all incoming cables should be routed. As Sun Microsystem Planning Guides discuss, your connection must include low impedance (not just low resistance) grounding to avoid voltage differences within the room. That means routing of power and of signal cables.

Many electricians understand low resistance but have little idea what is low impedance.

Not > I have a question that maybe this group can awnser for me.

snipped

Sure would like to see a design that can handle a building lightning strike. Last one I worked in every wire in the building as well as 90% of conduit was toast. There was an installed U.L. listed lightning protection system. This was 20 years ago, all I remember was replacing all of the electrical conduit and wiring. Even today's with modern surge protection devices and constant voltage transformers, I do not think a server room would survive a building hit.

Much appreciate the information. The server room is about ten feet from the buildings' main power supply.

Thanks

It is routine even before WWII to suffer direct strikes and not suffer damage. A telephone switching computer connects to overhead wires everywhere in town. Do they replace that $multi-million computer every year? Of course not. Electronics atop the Empire State Building would suffer about 25 direct strikes every year without damage - even before WWII. Technology is that old and that well proven.

We only do same with even better devices.

Modern surge protector devices and constant voltage transformer implies a protector will somehow stop, block, or absorb. Not for one minute. Well proven protection is not even about protectors nor is it about stopping or absorbing surges. Well proven protection is about earthing - shunting - diverting. The protector is not protection. Earthing is the protection. A protector is sometimes used to make that connection to earth. Sometimes a wire does same earthing connection. That is all that a protector does - make a temporary connection to protection.

Meanwhile where some install protectors to stop or block surges, then damage results. Of course. Many are sold on myths so often promoted on retail store shelves: that shunt mode protectors somehow stop surges.

It is routine for emergency response centers to suffer direct strikes and not suffer interruptions - as well as no damage. Server rooms are supposed to be installed to do same. Direct lightning strikes earthed without damage is routine when facility is properly installed. Why does the telco not replace its switching computer every year? Protection from direct lightning strikes must be routine to achieve service that is routine in every town. Server room must do same.

BTW, UL does not mean the protector works. UL only means a protector does not threaten human life. In fact, a protector gets UL approval even if protector totally fails during testing. This because UL approval says nothing about transistor protection. Whether protector works is irrelevant to UL approval. But some retailers may hype that UL approval to promote a myth that UL says protector does something effective.

Citing a *UL protector* is a misguided assumption that the protector provides protection. We install effective protectors where a direct, hardwire connection cannot be made to earth. Earthing - not the protector with or without a UL label - is protection.

Where humans assume the protector is protecti> Sure would like to see a design that can handle a building lightning strike.

I don't mean to be harsh but you are not qualifed to install this "server room". Literally, you don't have even the basic fundamental knowledge to do it. Have someone qualified do this and it will be safer, and possibly more reliable which is what a server room is typically all about.

Usually the voltages on the building power breaker are not the problem on such situation.

What could be problem could be the magnetic fields from the high power carrying cables and power rails in the power breaker panel.

Typically the magnetic fields are not strong enough to harm hard disks etc.. that are inside the PCs. But the magnetic field could be string enough to make traditional CRT display unuseable (typically image shaking etc. problems). I have seen that personally happen around 15 years ago on one site where the house main distribution panel (230V/400V AC power at few hundred amperes) was on the other side of one of the wall in computer room. The monitors near that wall had very much shaking images in them, point of text being unreadable on screen. Installing some iron or steel plates (few mm thick if I remember right) on the computer room wall helped situation somewhat, but did not completely cure the problem.

Moving the monitors to other location (few meters away) and doing operation with help of KVM switch solved the problem. Servers worked nicely there without any special problems.