<Safety Instruction> Chapter 4 | Ensuring the reliability and quality of power (2)
Part Two – Power For Critical Equipment
Temporary replacement for utility power was discussed in the preceding paragraphs. But no generator can pick up the load from a failed electric utility immediately. In actuality the sequence of activities looks something like this, although the time associated with each step or process can vary considerably.
- Utility power fails
- Failure of primary power is sensed by the transfer panel (milliseconds)
- A timer is started and after a pre-set time instructions are issued to start the generator (varies; perhaps 15-30 seconds)
- Generator warms up a bit and voltage and frequency are stabilized (15 seconds to one minute)
- The transfer panel switches the load from the dead utility power supply to the generator.
So, what happens between the time the first and last events occur? Bad things! Computers crash and as the data communications infrastructure loses power, IP telephones go dark and all communications with the outside world is lost. So, this points up the requirement for the survivable enterprise to have two forms of back-up power. The first must carry the load for hours, days or even longer, and is most likely to be a generator, as discussed above. The second is an uninterruptible power supply. One is generally never a substitute for the other.
Essentially, one UPS device is required for each piece of equipment that cannot be permitted to fail. This would typically include communications and data communications systems, routers, switches and the racks that contain them. Equipment throughout the building such as servers, wireless access points and IP telephone must have their own UPSs if not powered by POE, power over Ethernet. (This, by the way, is a strong argument for the use of POE.)
As is with the case with generators, uninterruptible power supplies are available in many varieties. Here are the most important variables:
Type of output?
The decision as to whether sine wave output is required or modified sinusoidal wave is acceptable is based on the requirements of the equipment to be powered. The former is more expensive, as may be expected. (Generally, Peplink equipment runs fine on most good quality UPSs – even if the output is not a perfect sine wave.)
The purpose for a UPS in this environment is only to power the critical equipment between the time utility power is removed and the load is transferred to the generator – not for use for an extended period. So, “theoretically” the UPS battery(ies) need only last for a very few minutes. However, the tendency to increase the load on UPSs over time is well known and over time the UPS batteries become less efficient. So, it’s best to provide a bit of margin vis-a-vis run time.
So, the most obvious and well-recognized rationale for employing an uninterruptible power supply is to provide power for essential equip – ment until the main back-up system, typically a generator, comes on-line. However, there’s an additional rationale which is less commonly recognized. While the loss of utility power may be the most dramatic event and the one that catches one’s attention the quickest, it’s not the only threat. The quality of utility power is highly uneven around the world. A well-de – signed UPS can protect against voltage sags, high voltage, and certain transients. Indeed, many of the well-known manufac – turers exhibit confidence in the quality of their products that they provide insurance for the equipment behind the UPS if the former is damaged. So, a UPS provides protection not just for short disruptions in primary power but also for certain abnormalities in that power.