Is it okay to plug a power-strip into a UPS?
Benjamin Scott
dragonhawk at iname.com
Fri Jun 17 10:00:01 EDT 2005
On Jun 16 at 11:47am, Bill Sconce wrote:
> Ben didn't mention that shunt TVSS isn't the only way.
Actually, I did, it was just buried in the paragraph about shunt TVSS. I
don't understand much of the technology involved in series mode protectors, so
all I can do is describe it in very general terms, which only takes a few
words.
> Note also that 330v is a LOT - 3 TIMES what a device sees in normal service.
> A good suppressor can be designed to clamp at a 10v rise (or at 2v) if
> you're willing to pay for it.
Actually, no. Remember, we're talking alternating current. In the US, AC
goes from zero to (IIRC) 180 volts (absolute) twice per cycle, or 120 times
per second. In other words, AC is *constantly surging*. Simply clamping at
any voltage increase isn't possible, because AC does this all the time. I've
read a technical justification for the 330 V figure but didn't understand it.
Now, I would expect it is possible to build a device that clamps if the
voltage rise is too fast or out-of-phase, but I'm really just guessing here.
One of the things to consider here is that wall current is not precision
signal quality and never has been. Any device expecting to perform precision
electrical signal work is going to have extensive filtering built into it.
It flat out wouldn't work otherwise. So the job of a TVSS should not be to
attempt to make the power "perfect", but to keep *damaging* surges from
getting through.
For example, on the www.brickwall.com site you linked to, they post some
oscilloscope traces.
Now, since they don't post scales, I'm already a little weary. A dishonest
company might use different scales for different measurements to make their
product appear better. But they seem like an honest company, and you're
obviously satisfied with them, so I'll give them the benefit of the doubt, and
assume all the scales are the same. But we still don't know what the scales
*are*. In other words, just how big are those surges they are graphing?
I can deduce the time scale. I know some of those products they list are UL
Listed to clamp in less then one nanosecond. So they must be using a scale of
no larger then about 500 picoseconds per vertical grid line. That's damn
small. Remember, US AC is 60 Hz. A single AC cycle is about 16.5 billion
picoseconds long, if my math is right. Is a fluctuation during 0.000003% of
the cycle really something to worry about?
I'm sure if you spend $200 on a really elaborate device from a quality
manufacturer, you're getting something for the money. The question is, does it
really matter? Are you paying for protection that simply does nothing, due to
the fundamental nature of AC?
Given the choice, I suspect it would be better to put (less) money towards a
quality voltage regulator, which not only protects against surges, but sags as
well. I know from personal experience that the brief under-voltage conditions
often encountered in a factory (when big machinery kicks in) are far more
problematic to a lot of equipment then surges are. I know this because the
test equipment kept malfunctioning when the 100 amp electric motor on the big
mill kicked in -- until we put a voltage regulator on the TE.
> MOVs are junk.
It does seem to be the case that simply slapping a MOV or six between the
line and grounded conductors is not a good way to build a TVSS. That being
said, I've seen tech stuff that alleges that MOVs can be designed for the
application, and that coupled with other components, will do a quality job of
shunting over-voltages. As I keep saying, I don't understand the EE theory
enough to make a judgment either way.
For example, here's one analysis I've found that goes into all sorts of math
about this kind of thing. Alas, it's way over my head. Link:
http://www.iaei.org/magazine/04_b/04_b_brown.htm
Now, to me, the big question is, "Just how bad is shunting power to the
grounded (neutral) and/or grounded (safety ground) lines?" It doesn't matter
if you have really great MOVs (or gas tubes for that matter) if the whole
design concept is flawed from the start.
I've seen lots of vague comments about how bad this is. In effect, "The
surge is just going to go from line to ground, in through the ground to the
equipment, and cause Bad Things To Happen!" While that seems like a
reasonable concept to me, what little I know about electrical engineering
includes the knowledge that "common sense" is often misleading. I'd like to
some unbiased (i.e., not trying to sell me something), technical analysis on
this.
As I said, surprisingly complicated. :-)
--
Ben <dragonhawk at iname.com>
More information about the gnhlug-discuss
mailing list