Comments on Uninterruptible Power Supplies
Some material is very old and may be incorrect today
© July 2004 BigDumbDinosaur
My primary recommendation is Powerware (formerly BEST Technology, now part of the Eaton Group ) UPS's, specifically the Ferrups ferroresonant unit for difficult or lightning-prone environments. My recommendations are based on a lot of real-world experience with manufacturing environments, which tend to pose the greatest difficulties in providing computer-grade electrical power. Powerware's 9100 series is also high quality and fine for installations where heavy electrical loads attributable to integral horsepower motors are not present in the same building.
One of our installations is at a rubber goods manufacturer (No! Not that kind of rubber goods.), where some of the shop machinery is powered by electric motors big enough to sink a small ship. Starting some of these motors gives new meaning to "power surge." One of the motors, a 75 horsepower unit, draws nearly 1000 amps from the line at start-up. Needless to say, power line instability is the norm there.
To deal with the power line cruft we have their main computer system (an H-P 9000 RISC server) attached to a Ferrups 1.4 KVA ferroresonant UPS. Another Ferrups unit, an 850 VA model, supports their network gear and also protects the telephone system. On several occasions, I have 'scoped the UPS output and have never seen so much as a millivolt of junk, despite all the machinery running and the continuous power line blips.
The smallest Ferrups unit available is a 500 VA device, which is what powers the UNIX box here in my office. The largest currently available is rated at 18 KVA, which will run a heck of a lot of hardware. I don't know how long the run time is on the 500 VA unit's battery, because the longest power failure we ever experienced was only 1-1/2 hours, and the UPS was still going strong. The 1.4 KVA unit installed at the rubber goods plant has been able to keep their system up for in excess of two hours before initiating an automatic server shutdown due to low battery charge (we subsequently added a second battery to extend the run time to what is probably close to five hours -- it has yet to be tested).
Ferroresonant UPS's, in general, are larger, heavier and more costly than line interactive or double conversion units. The UPS gets its name from the special transformer that isolates the input from output. The load never sees the raw AC line voltage but instead, sees the output of the transformer.
A ferroresonant transformer is a special type of transformer that has been tuned to resonate at the power line frequency (60 Hz in North America, 50 Hz in Europe and many other locations). Because it has been tuned to pass a narrow frequency range, the transformer acts like a brick wall to anything outside of that range, resulting in very effective power line filtering -- and a nearly impervious firewall to lightning strikes.
Also, because the transformer resonates at the power line frequency, it acts like an electrical flywheel and stores energy in its massive iron core. In the event a momentary decrease or complete loss of voltage occurs the transformer will replace some of the missing AC cycles -- a momentary boost. This results in a true "no break" power source. If the line voltage increases for any reason (e.g., a spike or transient induced by a nearby lightning strike) the transformer will buck the increase, keeping the output within a fraction of a volt of where it should be.
Ferroresonant technology is hardly new. The concept was developed in the 1930's and has been widely available in heavy duty power conditioners. My first encounter with a ferroresonant device was in the 1960's aboard a U.S. Navy warship, where power problems were routine. Much of the electronic gear was protected by ferroresonant transformers, which given that this was aboard a military vessel, amounts to a ringing endorsement of the technology.
"Also, has anyone made a claim to APC for any UPS's that have been hit and got the money?"
I'm not aware of anyone making such a claim and getting a settlement. Actually, it's fairly easy for APC (or any other UPS vendor, for that matter) to weasel out of paying off the claim, simply because irrefutable proof that the UPS was to blame for damaged equipment is difficult for a non-technical person to produce. Even most computer hardware geeks are not experts on power generation and UPS capabilities, which means most will not be able to conclusively determine that the UPS didn't do its duty. Adding to the difficulty is that in many cases, the UPS may continue to function normally following the destruction of the supposedly-protected load.
I find it interesting that APC in particular has long maintained that there is no benefit to ferroresonant technology and has even gone so far as to produce "studies" that "prove" that ferroresonant UPS's are inferior to line interactive and double conversion units of the type produced by APC. Yet, almost all power-critical installations run by the military and many government operations depend on ferroresonant technology for power protection. So, who's right: APC, who has a sales ax to grind and has never achieved a high rate of success in large commercial installations, or the military, whose primary concern is maximum reliability?
Got something to add? Send me email.
(OLDER) <- More Stuff -> (NEWER) (NEWEST)
Printer Friendly Version
More Articles by BigDumbDinosaur © 2011-04-28 BigDumbDinosaur