I received quite a response to last issue's (TNPC #4.02) article on power spikes, sags, and blackouts and their effect on your computer. A number of TNPCers wrote in with questions and requests for additional information and I heard from several who work for power utilities that shed some light (no pun intended) on what really goes on when power is restored after a blackout.
Last issue I talked about surge protectors and the inherent weaknesses of the MOV technology that provides the surge suppressor its ability to suppress surges. I concluded that the best protection from the three main types of power problems-- surges, sags, and line noise--is provided by connecting your computer to an uninterruptible power supply (UPS).
However, there is some confusion about what constitutes a true uninterruptible power supply as opposed to a "standby" power supply and as TNPCer Sol S. pointed out, I added to the confusion by not spelling out the differences. I have been known to use the terms interchangeably when they really are not the same thing. A true UPS provides constant power from the battery to the connected equipment. There is no switching from incoming AC current coming from the wall outlet (line power) to the battery when the power fails because the battery is always the power source. The AC power is used by the UPS to keep the battery charged. A standby power supply (SPS) feeds line power to your equipment and it may or may not condition the line power against line noise depending on the model. When there's a sag or power failure the SPS "switches" from line power to battery power.
The rule of thumb is that a UPS provides better protection than an SPS. During that split second when the SPS switches from line power to battery power, things could go wrong. At best if the switch doesn't happen fast enough your equipment could reboot as though you turned the power off, then back on. Worse case would be not switching and letting a power surge hit your equipment. The good news is that in the past few years SPS devices have gotten a lot better at switching when the line power is interrupted. It's good news because a lot, if not most, of the devices that are advertised as being UPSs are in reality SPSs.
It used to be that you could look at the specs and if you found a "switching time" rating the device was a SPS. A true UPS doesn't switch between line power and battery so there is no switching time to measure. But manufacturers just stopped quoting the switching time in the specs. If your device has indicators as to when it's providing line power as opposed to battery power it's probably a SPS. Ditto if one of the features is that the battery can be swapped while the device is running and equipment is plugged in. An SPS is not a bad thing if it also provides line conditioning. Don't panic if your "UPS" labeled device is really an SPS. I use an APC BackUPS Pro 500 and it's really an SPS. The price was right and APC offers a guarantee against equipment failure, although I'll confess I've never had the occasion to try to collect on it.
The main thing is to be sure that your device provides enough
power for the equipment you want to protect. Each piece of
equipment you are subject to plug into the UPS/SPS draws a finite
number of watts. You need a device capable of feeding all the
watts necessary. American Power Conversion (APC) has a Web form
for calculating what size UPS (most of which are really SPS
devices) you need depending on what you plan on hooking up.
http://www.TheNakedPC.com/t/403/tr.cgi?ups1
Keep in mind that APC has a vested interest in selling you as big a unit as you'll spring for so use their form as a guideline not gospel.
Another common question was how can a computer run off a battery, which provides direct current (DC) when it requires alternating current (AC) to run. Yes, the computer plugs into an AC source for power, but the AC current is then converted to DC before it hits the internal components. Therefore, when battery power is provided by a UPS/SPS the power device converts the DC power from the battery into AC power to the outlet where the computer is plugged in, which then goes into the computer where it is converted back into DC power for actual usage.
A number of readers wanted to know what "line noise" is all about. Line noise results from static, and electrical interference caused by electromagnetic interference (EMI), radio- frequency interference (RFI), or inductive loads caused by machinery (usually electrical motors) and not just your machinery. The guy with a woodworking shop in his garage next door can cause you to experience line noise. Electrical wire acts just like an antenna and can pick up interference from fluorescent light fixtures, telephone lines, etc. By filtering out line noise you get what is called conditioned power. Be careful though, not all the outlets on a UPS/SPS may provide conditioned power.
Two TNPCers have enlightened me about what really happens when the power company shuts down the power and then switches it back on. TNPCer Mike S. who works at a power company admonished me for thinking that power companies boost the power when trying to restore power. He says, "...often the initial demand exceeds the settings of the protective relays at the power station" and this can require the power utility to try several times to get the power to stay on. The problem is exacerbated by everyone turning on every switch and appliance in their house to see if the power is out throughout the home. They leave these turned on and this increases demand on the returning current. So it would appear that my claim that the utilities upped the amperage when restoring power is incorrect. The most they do in this regards according to Mike is push the voltage "as much as load, voltage maximum limits, and generator capabilities allow." Mea culpa.
TNPCer Oscar Desmonteix at Macrotel International Corporation put
it this way, "The power grid is full of inductive loads (wiring,
transformers, electric motors, relays, etc.) and inductive loads
have a very nasty reaction when power is applied to them, they do
not only resist the current flow but also return a voltage peak
to the power line. This returned voltage may easily duplicate the
voltage applied to the load and sometimes it goes even above
that. This voltage peak is what really damages appliances and
equipment connected to the same power circuit if they are not
protected." Oscar has provided a most detailed and technical
explanation of the problems involved with protecting an entire
building against power problems. You can find this illuminating
discourse here:
http://www.TheNakedPC.com/t/403/tr.cgi?ups2
A final word on MOVs (metal-oxide varistors). TNPCer Lou B. points out that "it takes three MOV's to properly protect a circuit. One from the Load to Ground, one from the Neutral to Ground and the third between the Load and Neutral. Another consideration is that after one 'blast' they should really be replaced." The trick is to know when you've had a "blast" or surge because as Lou correctly points out the second time around chances are that you might not be protected. This is why I don't like to have critical/expensive equipment reliant on MOVs for power protection.
You can reach T.J. Lee at:
mailto:tj_lee@TheNakedPC.com
