Backup and UPS power

A few years ago, I visited a satellite transmission facility in a remote location that was designed to operate “off grid” if necessary for a considerable amount of time. In the event one of two sources of power failed, the UPS would take over and generators would immediately start. That might be too simplistic. If power dropped, a UPS system with literally thousands of lead acid batteries — occupying a room better suited for a Kmart — would support the facility for about 20 minutes at 3MW of consumption. At the same time, two turbine-powered generators would fire up and synchronize, and either could support the facility's full load. Just in case one of the generators failed to fire, or synchronization could not be achieved, a third turbine-powered generator was being heated by hot exhaust from the first two so it could start faster. In the ground, enough fuel was stored to keep the whole shop operating for a month, which was not a bad idea since skiing in this location is less dangerous than driving a fuel truck in the winter.

This serves to illustrate that we are highly dependent on the continuous supply of power our facilities require. But failure of the power to a facility is actually not a technology problem. It is, in fact, an economic problem. Our facilities do not exist to protect the public, and despite the introduction to the Communications Act of 1933, our facilities do not exist for the purpose of informing the public. They exist to make money, which is measured in minutes and seconds as much as in dollars.

It therefore follows that protecting power is in the economic self-interest of corporations involved in providing “broadcast” content, however you wish to define that in today's multiverse of delivery options. It falls to the technologists, those who generally read this magazine, to figure out a way to best protect the economic interests at stake by using available technology to keep power flowing.

Keeping the lights on

In general, like layered storage systems, power systems are intended to be both redundant and scaled in capacity relative to the intended usage and frequency of use. UPS, an option for power that is inherently dependent on input power, is a first option but cannot sustain the facility for a long period. UPS can be online at all times, leaving no latency to switch from one source to another. Generators can supply a full facility, but they cannot inherently come online until they “spin up” and stabilize. That stabilization period is dependent on the internal combustion engine warming enough to be reliable, which is the principal source of latency and the need for tertiary sources of power. The simplest redundancy, which is sometimes not possible, and often not considered, is to ask the local power company to supply power from two substations, greatly lowering the likelihood of simultaneous failures killing the input power.

I think the task of the technologist crafting a solution is to evaluate all options and make a recommendation on how to lower the risk. If the ownership has no ability to accept some outage, as was the case with the satellite facility I referenced, then the job is bounded mostly by dollars. If the chances of outage are related to the value of minutes lost in some defined time period, the engineering calculation is bounded by mean time between failures (MTBF) and mean time to repair (MTTR). In the real world, cases of budget insensitivity are few and far between. The usual question is, “What would it cost to provide reasonable protection?” The sane response might be, “What do you consider to be a reasonable cost?” I often tell clients contemplating a project to define either the problem to be solved or the budget, but not both at the same time. (See Heisenberg's Uncertainty Principal.)


People tend to think of UPS as a ubiquitously available option these days. The ability of small UPS systems to locate spots in a plant that need to be protected with high reliability has been used in many facilities. One common motive is to protect software-controlled systems that would take a considerable time to boot up after power failure. A modern video server with a large RAID array can take tens of minutes to return to life after a power failure, for instance. The maintenance of many small UPS systems is a headache, however. A centralized UPS with distribution to “spots” where it is needed solves that problem while leaving less flexibility for incremental change being facilitated.

The easy solution, and the one most likely to be understood by managers who are technology challenged, is using generators. Environmental requirements today make generators considerably harder to implement without referring to zoning restrictions and building code requirements for the management of fuel, noise control and exhaust dissipation. When possible, a generator is a great solution because it is relatively inexpensive and requires only one connection to the switch gear where “shore power” lands. But a large generator capable of holding the entire load of a large facility can still be expensive, and it begs for a redundant generator as an intellectually consistent recommendation. Obviously, the turbine generators I referenced would provide a highly reliable solution, but at an astronomical cost.


Other solutions deserve mention, of course. Battery technology has been rapidly evolving in no small measure due to research into solar and wind power. A UPS system with both shore power and solar (or wind) inputs may well become an attractive option when considering the movement to make facilities “green.” The entire goal of a UPS is to store energy for later use, so it matters little where the energy comes from, except when considering cost.

No discussion of storing energy would be appropriate without discussion of storing mechanical energy. Today, even race cars use rotating mass to store energy and then release it to electric motors to both reduce fuel consumption and power past competitors. In our world, a number of companies make systems that use shore power to spin up rotating mass energy storage systems. In the event of (certain, eventual) power failure, the stored kinetic energy of the considerable rotating mass runs a generator that replaces the lost shore power just like a battery-based UPS. There may be fewer environmental concerns with this approach than large battery-based UPS systems that require proper ventilation, at a minimum, by code. In general, rotating mass systems supply short-duration power replacement. Both battery-based systems and rotating mass systems claim better efficiency. Sounds like a lot of claims in our media industry doesn't it? Nothing is certain except that power will eventually be interrupted.

John Luff is a television technology consultant.

Send questions and comments to: