[snip]What is a good uninterruptible power supply unit for the Asus
P4800E-Deluxe. I want to pick up a good one and avoid frying my P4 3.2 and
other components during the many summer brownouts around here.
[snip]I have 8 hard drives, 2 dvd drives and a 5550W PS.
w_tom said:Recommend the $500 UPS to solve what a $1 per appliance
solution does? Again, even the green bypass wire makes the
"true online type" UPS ineffective. Destructive transient
bypasses the UPS. Just one of so many reasons why a plug-in
UPS is not effective.
What does that UPS do that the power supply does not?
Nothing. The plug-in UPS does not even claim those 1000+
volts isolation. Power supplies must withstand 1000+ volts as
even required by industry and Intel specs. Robust protection
inside a power supply - that still requires a 'whole house'
protector. $500 for a true online type UPS provides no
additional protection. $500 and the 'whole house' protector
is still necessary? What kind of protection is that?
Mythical.
I'm pretty sure I remember you from a similar discussion 3 or 4 years back.w_tom said:Recommend the $500 UPS to solve what a $1 per appliance
solution does? Again, even the green bypass wire makes the
"true online type" UPS ineffective. Destructive transient
bypasses the UPS. Just one of so many reasons why a plug-in
UPS is not effective.
What does that UPS do that the power supply does not?
Nothing. The plug-in UPS does not even claim those 1000+
volts isolation. Power supplies must withstand 1000+ volts as
even required by industry and Intel specs. Robust protection
inside a power supply - that still requires a 'whole house'
protector. $500 for a true online type UPS provides no
additional protection. $500 and the 'whole house' protector
is still necessary? What kind of protection is that?
Mythical.
Lightning strikes somewhere across the street close to the
below grade West cable vault. ... The first line of defense
is the telco protection panel, but the panel must be connected
to a low resistance / inductance ground. There was no adequate
ground available in the telephone room.
Conceptually, lightning protection devices are switches to
ground. Once a threatening surge is detected, a lightning
protection device grounds the incoming signal connection
point of the equipment being protected. Thus, redirecting
the threatening surge on a path-of-least resistance
(impedance) to ground where it is absorbed.
Any lightning protection device must be composed of two
"subsystems," a switch which is essentially some type of
switching circuitry and a good ground connection-to allo
dissipation of the surge energy.
http://www.harvardrepeater.org/news/lightning.html
Well I assert, from personal and broadcast experience spanning
30 years, that you can design a system that will handle *direct
lightning strikes* on a routine basis. It takes some planning
and careful layout, but it's not hard, nor is it overly
expensive. At WXIA-TV, my other job, we take direct lightning
strikes nearly every time there's a thunderstorm. Our downtime
from such strikes is almost non-existant. The last time we
went down from a strike, it was due to a strike on the power
company's lines knocking *them* out, ...
Since my disasterous strike, I've been campaigning vigorously
to educate amateurs that you *can* avoid damage from direct
strikes. The belief that there's no protection from direct
strike damage is *myth*. ...
The keys to effective lightning protection are surprisingly
simple, and surprisingly less than obvious. Of course you
*must* have a single point ground system that eliminates all
ground loops. And you must present a low *impedance* path for
the energy to go. That's most generally a low *inductance*
path rather than just a low ohm DC path.
Lightning surges cannot be stopped, but they can be diverted.
The plans for the data center should be thoroughly reviewed to
identify any paths for surge entry into the data center. Surge
arrestors can be designed into the system to help mitigate the
potential for lightning damage within the data center. These
should divert the power of the surge by providing a path to
ground for the surge energy. Protection should be placed on
both the primary and secondary side of the service transformer.
It is also necessary to protect against surges through the
communications lines. The specific design of the lightning
protection system for the data center will be dependent on the
design of the building and utilities and existing protection
measures.
Those who say "nothing will withstand a direct lightning strike"
are very misinformed. My towers take direct lightning hits most
every big storm. So do most all tall commercial towers. With NO
damage!
The real problem is the haphazard use of common all-mode
protectors at AC outlets or outlet strips. In many cases, this
practice causes either system noise problems or hardware damage.
... The hardware damage does not usually occur in the power
supply, where you would expect it, but in the signal I/O
interfaces that connect to the outside world.
...
the ground wires may be quite long. Most transient over-voltages
are high-frequency events, having most of their energy well
above 100 kHz. At these frequencies, long wires, regardless of
their gauge, have high impedance and will develop extremely high
voltage drops when carrying the high current pulses created by
MOV clamping. ... This voltage is likely to reduce interface
circuitry in the computer, printer or both to silicon vapor.
More frequent low-voltage spikes (down to the low-current MOV
clamp of 300 V or so) will still cause high-current pulses to
flow in the same loop. These smaller noise spikes between the
grounds will cause errors or lockup.
...
The absolute best place to guard against incoming spikes and
surges is at the service entry panel or a sub-panel that
powers everything in an interconnected system.
... Square D, GE, Leviton, Intermatic, Cutler Hammer,
Siemens, Furse, etc.
[snip]A UPS from Mercury's first group is more than sufficient for
a computer.
[snip]In fact, that "when a fail occurs, the relay
simply switches in a power converter to run off the batteries"
type UPS is also called a "computer grade" UPS.
[snip]Those other
'group 3' UPSes cost on the order of $500+ to provide nothing
useful.
[snip]Learn what a power supply does.
[snip]First 120 VAC passes
through line filters.
[snip]Then 120 VAC is converted to 300 volts
DC.
[snip]Virtually all noise, spikes, and surges are eliminated.
[snip]But your power supply provides more layers of protection. An
oscillator converts that hundreds of volts DC to an AC
voltage.
[snip]Then voltage is filtered through a transformer -
galvanic isolation.
[snip]The transformer output is again converted
to DC.
[snip]Then that DC output is filtered but again.
[snip]Then an
overvoltage protection circuit further eliminates excessive
spikes and noise.
[snip]With so many layers of protection standard
in power supplies, then what does a UPS do? Typical UPS
provides less protection than what already exists inside a
standard power supply - even 30 years ago.
[snip]After so much hype about UPS protection, the bottom line
remains: a UPS only protects data from blackouts and
brownouts. Anything on power cord that protects hardware is
required and already inside that power supply. Look at those
so many layers of protection.
[snip]In summary: For effective protection of computer components
so that even a power supply failure does not damage
motherboard, disk drive, Ram, etc; the power supply must
contain functions that were even standard 30 years ago.
Functions so often missing in discounted power supplies.
[snip]For effective protection of data from blackouts and
brownouts, we install a plug-in UPS. Blackouts and brownouts
do not harm hardware - except where myths are widely promoted.
Maybe, but most often not. For the current to "pass through"
a filter, that filter must be composed (at least primarily) of
inductive components (i.e., a choke). But chokes are relatively
expensive and bulky;
Excuse me! That alone says one is soaking wet behind theA transformer has *nothing* to do with "galvanic isolation" ...
Jay T. Blocksom said:Maybe for *your* computer(s), but certainly not for mine.
The el-cheapo standby-type supplies that "Mercury" (rather charitably)
described as "Basic" offer only minimal "protection" against the simplest (and
rarest) types of total power failures, and little else. They can do
absolutely nothing about the (often abysmal) *quality* of the power coming out
your wall outlet. Hence, their ability to shield the load from the plethora
of short-term fluctuations (so-called "brownouts", noise, frequency and
waveform aberrations, etc.) that have become increasingly common as power
utilities scrimp ever more on their QOS (and chronically fail to ensure that
generating capacity keeps pace with demand) is sorely lacking. The worst of
them also often throw all sorts of self-generated noise onto the line (and
into the air), which can (and does) interfere with the operation of *other*
equipment (especially RF-based equipment, such as a wireless LAN -- or your TV
set).
...
Only by the marketing mavens trying to sell then to the Great Unwashed, and
the gullible fools who have bought into their line of bull.
...
This is flat-out wrong.
A proper double-conversion UPS will offer MANY advantages over the various
low-end pieces. Among these are effective isolation between the load and the
line, true sine-wave output with minimal THD (sometimes also provided by the
better line-interactive units, but definitely not by the low-end standby
supplies you claim above to be "more than sufficient"), more robust batteries
and inverters (they _have_to_ be beefier, as they are in use all the time as
opposed to "just occasionally") and usually better build quality in general.
As for cost... Well, that can (and will) be all over the map, depending on
the manufacturer, the vendor, and especially the load capacity. You can spend
a few hundred, or you can spend several thousand -- but this also applies, to
very nearly the same degree, to standby & line-interactive models. As I
stated in my earlier article, a proper double-conversion UPS will generally
cost something on the order of 1.5-2X what an equivalent line-interactive
model will run. Here are a few examples for units in the 1KVA range (which
would be quite adequate for the OP's application):
TRIPP LITE OMNI SMART 1050VA
(Line Interactive / Stepped-square output) $277.95
<http://www.axiontech.com/prdt.php?item=34427&PRICECOMPARISONSID=7010b72b1aad24a0238d7ba547832d20>
TRIPP LITE SU1000XL
(Double Conversion / Sine-wave output) $413.95
<http://www.axiontech.com/prdt.php?item=61114&PRICECOMPARISONSID=7010b72b1aad24a0238d7ba547832d20>
The latter is exactly 1.489296636086 times as expensive as the former.
Para Systems Minuteman SmartSine S1000
(Line Interactive / Sine-wave output) $325.
<http://www.bisonbusiness.com/mis1fcoshcos.html>
Para Systems Minuteman MCP 1000 E
(Double Conversion / Sine-wave output) $525
<http://www.bisonbusiness.com/mimcfcoshcou.html>
The latter is exactly 1.615384615385 times as expensive as the former.
And finally, both from <http://www.battery-usa.com/Powercom-UPS.htm>:
PowerCom HOME-1000A
(Line Interactive / Sine-wave output) $281.33
PowerCom ULT-1000
(Double Conversion / Sine-wave output) $456.00
The latter is exactly 1.620872285217 times as expensive as the former.
Please note that I am *NOT* specifically recommending any of these particular
models, vendors, etc.; I'm just using them as handy examples.
...
You should take your own advice. Your description which follows is seriously
flawed in several ways (at least presuming you're talking about the typical
"switching" type PC power supply unit).
...
Maybe, but most often not. For the current to "pass through" a filter, that
filter must be composed (at least primarily) of inductive components (i.e., a
choke). But chokes are relatively expensive and bulky; and they don't do much
for you in a true AC environment. So typically, MOVs and bypass capacitors
are used instead. The current does not "pass through" these components (tho'
it might be said to "pass by" them).
...
Uhhh... no.
It is rectified, usually (or at least hopefully) by a full-wave bridge
rectifier circuit. This produces a (very rough and noisy) pseudo-DC output of
approximately 180V.
[snip]Virtually all noise, spikes, and surges are eliminated.
Not even close!
Have you ever looked at the waveform output from one of those line-AC
rectifier circuits on a 'scope? I thought not. As noted above, at this stage
in the circuit, that waveform is *extremely* ragged and noisy (technically
referred to as "ripple") -- which is why the rectifier is (at least in any
half-decent supply) immediately followed up with more filtering caps (and
maybe even a small choke, this time, since the demands on it are so much less
after rectification).
...
It is disingenuous at best to call that a "layer of protection". It is simply
a necessary part any switching-type power supply. And the reasons for it are
not to provide any better protection or "power quality" to the load. The
near-sole reason for this type of design is *cost* -- you need a LOT less iron
to transform high-frequency (typically 5KHz to 50KHz) AC than low- (i.e.,
line-) frequency AC (60 Hz in the U.S.); and the cost savings provided by
using a relatively tiny transformer (see the next step in this saga) much more
than offset the costs of the "preliminary" rectification/switching circuitry.
It also makes for a physically smaller and lighter PSU, which saves still more
money in manufacturing and shipping costs down the line.
...
A transformer has *nothing* to do with "galvanic isolation", at least in this
context. And "galvanic isolation" in turn has *nothing* to do with protecting
a PC from power-line anomalies.
Do you even know what that term means? (Hint: Think corrosion. Or
batteries.)
...
Again, a simple rectifier circuit, which again produces a very rough and noisy
pseudo-DC output -- actually, several such rectifiers, producing several such
outputs, since the transformers used in PC PSUs necessarily have multiple
secondary windings/taps.
...
Well, of course -- we've just made a ton of noise and other crap via the
rectification process, which we now need to dispose of. But this is S.O.P.
for *any* power supply, switching or linear, in a PC application or virtually
anything else.
...
Uhhh... No, again. You are presumably referring to active voltage
regulation, which any decent PSU (for virtually any application, not just PCs)
will provide. But this is (at least typically) not so much a matter of
"overvoltage protection" (or under-voltage, for that matter) as it is a means
of enabling the designer to scrimp on the final filtering caps and/or chokes.
With an adequately responsive AVR circuit to rely on, those filtering caps
and/or chokes don't need to be nearly as effective as would be necessary if
they were being counted on to provide *all* of the output filtering; hence,
they can be smaller and cheaper.
In reality, the actual "over-voltage protection" feature provided by some PC
PSUs works *against* data integrity, not for it, since it can (and will,
if/when it trips) arbitrarily shut down the PSU (requiring a AC-off cold
start) without warning.
...
Thirty years ago, the "PC" had not yet been invented (tho' the first 8-bit
microprocessors had been developed, so the birth of the PC was imminent -- I
discount such "lunatic fringe" aberrations as the MITS Altair, which was
purely a hobbyist piece -- with a *linear* power supply, by the way), and
switching power supplies (in any application) were both relatively rare and
quite crude by today's standards. They were, quite justifiably, viewed as
rot-got el-cheapo substitutes for a "proper" linear power supply.
Yes, today's switching PSUs (at least the higher-quality ones) are much better
than those early travesties; but that's not the point.
...
Nonsense.
*IF* we could count on both the quality and the reliability of the AC mains
power we use to feed these supplies, at least most of them would do an
admirable job of running the mobos, memory, CPUs, disk drives, etc., that your
typical PC is composed of. But we *cannot* count on that. In fact, that
situation is getting worse every day, for the reasons already touched on above
(blame the NIMBYs, if you want; but the fact remains -- we are facing an
electrical power crisis that will surely only get worse before it gets
better). Concurrently, we are becoming more and more dependant on our
electronics -- especially our computer systems -- for our day-to-day
activities and existence. That is a prescription for a very unpleasant
wake-up call, which we *will* get, sooner or later; but I digress.
Getting back to the point...
The "hold-up time" specs quoted for typical PC PSUs range from about 10ms to
perhaps 17ms -- i.e., one AC cycle _at_best_ -- and even these are "best case"
specs obtained under laboratory conditions. All manner of power-line
aberrations (besides simple "total failures") will have event durations which
exceed this threshold. So some sort of "upstream" protection is
near-mandatory, if you value your data.
The "switching time" specs for typical low-end standby-type "UPS"s are in the
several-millisecond range (tho' again, these are "best case" lab-conditions
specs). Now, you might think that, because these figures are significantly
lower than (typically about 1/4 - 1/2) the typical PSU "hold-up time", there's
no problem... But you'd be wrong. This is only the actual *switching* time,
which by definition only starts *after* a power-line disruption has been
sensed and the control circuit tripped -- IOW, by the time the "UPS" *starts*
to switch the load, the crap has *already* been fed through to the output.
...
Ah, ha! Perhaps this explains (at least in part) your apparent fundamental
misunderstanding of the issues. It's not about "protection of computer
components", all of which are virtually always easily replaceable. It *is*
about protection of your DATA -- which in any serious application is far more
valuable than all the hardware, software, peripherals, etc., put together.
...
But here you imply that any ol' "plug-in UPS" is equivalent to any other
"plug-in UPS" -- which is most assuredly *NOT* the case. (And BTW... You're
also dead wrong about blackouts/brownouts not being able to harm hardware; and
I have the fried A/V gear from both my shore home and my main residence to
prove it.)
This raises the question of just what axe you're attempting to grind.
Out of curiosity, I just took a quick look at some of your other recent
postings via Google Groups; and I must say... Your grossly simplistic (and in
some cases just-plain-wrong -- such as was exemplified in
<http://groups.google.co.uk/[email protected]>;
hint: the voltage output of a regulated power supply will hold up just fine
until *immediately* before it falls apart spectacularly -- hence, your "cure
all" DMM tells you NOTHING about reserve capacity) views on matters related to
power supplies and your apparent obsession with PSUs as the root cause of all
computer problems (even those rather obviously caused by software or "pilot
error") seems suspect, at best.
No one here (least of all me) is claiming that a (presumably decent-quality)
switching-type standby "UPS" is totally worthless; but to suggest that they
are the end-all and be-all of UPS design is both misleading and just plain
silly.
--
Jay T. Blocksom
--------------------------------
Appropriate Technology, Inc.
usenet02[at]appropriate-tech.net
"They that can give up essential liberty to obtain a little temporary
safety deserve neither liberty nor safety."
-- Benjamin Franklin, Historical Review of Pennsylvania, 1759.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Unsolicited advertising sent to this domain is expressly prohibited under
47 USC S227 and State Law. Violators are subject to prosecution.
[snip]I'm pretty sure I remember you from a similar discussion 3 or 4 years back.w_tom said:Recommend the $500 UPS to solve what a $1 per appliance
solution does? Again, even the green bypass wire makes the
"true online type" UPS ineffective. Destructive transient
bypasses the UPS. Just one of so many reasons why a plug-in
UPS is not effective.
[snip]
[snip]Serious questions: could you provide references, for those in the US 1)
that give a quantitative analysis of you r summary of plug-in
surge-protector performance? 2) That give a summary of whole-house surge
protection?
[snip]Like others, I'm mainly interested in a UPS ability to switch to emergency
power when the grid goes down. The surge protection part of the UPS is
fairly cheap, comppared to the power protection.
I also surge-protect the
lines that go through my attic and connect distant PCs, because some have
the view that secondary induction of current in such lines is probably more
common than desttruction from direct lightning hits.
[snip]Correct. Double conversion would provide protection, but IF
a destructive transient was type of transient they claim to
protect from. A destructive longitudinal mode transient is
not seen by UPS battery and inverter as transient passes right
through UPS to damage computer.
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.