Long-term (hours+) solution for power outage: PBX, lights, and computers

Ted Roche tedroche at tedroche.com
Fri Sep 1 16:01:01 EDT 2006


On Aug 31, 2006, at 5:39 PM, Ben Scott wrote:

>  As I know you know, a large battery plant is a non-trivial thing.
> In addition to everything Bill Sconce already pointed out, there may
> be fire code requirements as well.  I know the battery plants in telco
> COs are truly scary in their potential for destruction.

We've seen what a couple of amp-hours in a laptop battery is capable  
of recently.

You should have seen the one I used to manage onboard a submarine!

>  So regardless of whether you go with a generator, a large battery
> plant, a flywheel, or a giant mutant hamster in a wheel, I think you
> are going to face problems in the area of physical plant.

There's definitely a TANSTAAFL aspect to this.

>  All the UPS vendors I am familiar with say "Don't do that" --
> generally very loudly.  It's also rather inefficient.  Inefficiency is
> the enemy of economy and compactness, which you have give as goals.

Well, when has a client not asked for something for nothing? Fast.  
cheap and efficient. It never hurts to ask.

The desktop UPSes and a full UPS  are almost solutions to two  
different problems. The desktop UPSes are primarily line conditioners  
and for one minute or less blinks of power. The big battery is for  
long-term power outage.

> So you're probably best off getting rid of the small UPSes and going
> to one big solution.

One concern there is that the power off one of these big boxes is  
ugly AC and may not be suitable for powering computers. I've heard no  
horror stories pro and con.

I'm not really their IT guy as much as their lead developer. I was  
asked if I knew of a better solution. While the space is rented, I  
suspect the client might be there for a while, but infrastructural  
improvements might not be that welcomed.

Their power distribution is wired in the walls, and their computers  
go right into outlets next to their fans and space heaters, loads you  
wouldn't want on the UPS. Short of rewiring the place with a special  
feed for computer power, I'm inclined to recommend they beef up their  
desktop UPSes for a half-dozen workstations. Maybe a couple of  
workstations with long-term use would be sufficient for them. But if  
they go the other way, I'll caution them to consider removing the  
desktop UPSes.

>  From their site: "The systems do not produce heat..."
>
>  Neat!  I wasn't aware the Second Law of Thermodynamics had been  
> repealed.  ;-)

It's not just a good idea, it's the law. I-squared-R resolves to zero  
when I is zero. But "negligible heat" is probably a more honest  
answer. I'll bet it's warmer at full discharge current, though!


Ted Roche
Ted Roche & Associates, LLC
http://www.tedroche.com





More information about the gnhlug-discuss mailing list