Is it okay to plug a power-strip into a UPS?

[Benjamin Scott]

   This is a surprisingly complicated topic.

Power Taps
----------

   First, the easy one:  The technical term for the "octopus" that Bill Sconce 
was describing is "relocatable power tap" or "RPT".  This just takes one plug 
and turns it into multiple plugs.  No protection of any kind is implied. 
Many "power strips" are nothing more then an RPT.  The presence of an on/off 
switch, over-current protector ("circuit breaker"), and/or indicator light 
does not guarantee any additional protection, either.

   As everyone else has been saying, plugging a plain RPT into a UPS is fine, 
as long as you don't cause an over-current (over-load) condition.

   Beyond that, things get interesting.  I've done a bit of research on this 
topic.  I am somewhat hampered by the fact that I'm not an EE and don't 
understand the truly technical stuff about electrical theory.  Here's what I 
know...

Power Protection
----------------

   One of the few things I can say for sure is that most of the information 
written for people without an EE background is usually trying to sell 
something.  It often boils down to "Everybody else's technology is crap, but 
we use $TECHNOLOGY which will solve all your power problems".

   TVSS = Transient Voltage Surge Suppressor.  This is what most people are 
talking about when they say "surge protector".  One of the most common types 
of power threats is a brief but large over-voltage -- spikes and surges.  (A 
spike is basically just a shorter-duration surge.)  TVSSes protect against 
these.

TVSS Specifications
-------------------

   The UL standard for TVSSes is UL 1449.  Note that exactly how useful this 
standard is is open to debate.  UL 1449 is mainly concerned with clamping 
voltage, which is the highest voltage that the TVSS will let though before it 
kicks in and "clamps" the power.  So lower is better.  The best (in terms of 
UL 1449) is 330V; almost any TVSS will meet this requirement.

http://www.ul.com/consumers/surge.html

   Other specs one sees are the response time and the energy absorption. 
Response time is how quick the TVSS can clamp an over-voltage.  Pretty much 
everything is <1 nanosecond, making this part uninteresting.  Energy 
absorption tells you the largest power transient the TVSS can handle before it 
fails.  Measured in joules, with larger numbers generally being better. 
However, I've also seen literature stating that joule ratings alone are 
misleading.  See above about selling stuff; I'm not sure who to believe at 
this point.

TVSS Mechanisms
---------------

   Next we can talk about TVSS mechanisms -- how they actually do the job.

   It's a fundamental law that you cannot make the excess voltage that is a 
surge just disappear.  You have to do something with it.  There are two 
approaches to this.  One is to shunt the excess into the grounding conductor 
("third prong").  In theory, this is fine, since all US power ends up grounded 
to earth eventually anyway.  However, dumping all this excess power into your 
grounding line might (or will, depending on who you ask) cause other problems. 
The other approach used is to store up the excess energy, and then gradually 
let it back out over time.

   When it comes to the shunt method, there are multiple mechanisms for that as 
well.

   The most common type, as everyone has been pointing out, is the MOV 
(metal-oxide varistor).  As used in a TVSS, MOVs are basically a switch which 
triggers on voltage.  Once the voltage goes over the clamping point, the 
switch closes ("turns on"), dumping the surge into the grounding conductor.

   It is true that some kinds of MOVs will burn up if subjected to a prolonged 
over-voltage.  However, it is possible to design it so this doesn't happen, 
and UL 1449 requires this, so this is a kind of FUD.  It is also true that 
MOVs wear out.  It is also true that *everything* wears out, given enough 
time.  Just how much of a threat "MOV death" is again depends on who you ask.

   Another method used to shunt over-voltages is the gas tube.  It does the 
same thing as a MOV -- diverts excess power to the grounding conductor -- but 
via a different mechanism.  They use a gas which is normally non-conductive. 
During a surge, the gas ionizes and becomes conductive, closing the shunt.

   There are other mechanisms as well.  All of these, shunt and store, MOV and 
gas-tube, foo and bar, apparently have their advantages and disadvantages, and 
their advocates and detractors.  I think the best advice I've seen suggests 
using multiple mechanisms together.

Other Protection
----------------

   All of that is just about TVSS -- protecting against over-voltages.  Beyond 
that, you have noise filtering, voltage regulation, isolation transformers, 
and other neat stuff.  You can worry about back-feed on grounded (neutral) and 
grounding (third prong) conductors.  You can worry about data line surge 
protection.  You can get a UPS for battery backup.

   Some TVSSes are designed to fail safely -- they disconnect the load if they 
are are overwhelmed or have an internal fault.

   There is also non-technical protection.  Look at warranty, for example. 
I've got a Tripp-Lite IsoBar that has a lifetime warranty.  If it ever stops 
working, for any reason, Tripp-Lite will replace it.  They've done so once in 
ten years, so far.  A lot of power protection equipment also provides 
reimbursement if the device fails and your equipment suffers resulting damage.

Lightening
----------

   Finally, let me state that just about nothing will protect against a 
lightening hit.  I've the results of a direct building strike, and it ain't 
pretty.  You worried about your MOVs?  This hit caused light switches, 
outlets, light fixtures, various wiring, several TVSSes, miscellaneous 
equipment, and a mains transformer to burn up and/or blow apart.  It fried all 
the network hubs and several NICs.  It killed a couple of air conditioning 
units.  It really, really sucked.  All the normal rules go right out the 
window when you're dealing with lightening.  We're talking hundreds of 
millions of volts.  At that kind of potential, *EVERYTHING* is a conductor.

The Original Question
---------------------

   So, is it safe to plug a TVSS into a UPS?  The only sure answer I can give 
is "I don't know".  Given that it's trivial to get an RPT without TVSS, I 
would say you're safest going that route.

-- 
Ben <dragonhawk at iname.com>