DC UPS distributed throughout a critical infrastructure facility can reduce initial capital costs and lower overall O&M expenses.
By Gary Mulcahy, TDI Transistor Devices
Uninterruptible power supply (UPS) issues are often solved at the facility level with unnecessarily large, inefficient, expensive and complex AC UPS systems. While this provides an easy demarcation between the facility and end equipment – with each focusing on a different part of the problem – it makes it difficult to determine overall operating efficiency and total cost of ownership. It also makes it hard to optimize performance.
Initial costs as well as operating costs can generally be reduced, and system reliability increased, by incorporating a relatively simple DC UPS as an integral part of the data processing or communication equipment.
The newest generation of data processing equipment provides ever-increasing speed and bandwidth. Internet-based communication and commerce spur the need for more performance. A result is the amount and concentration of electrical energy being ported toward core data processing infrastructure. This has created new challenges regarding energy and resources used on inefficient power systems.
Data processing equipment is generally segregated into channels built around microprocessor cores. In the past, these microprocessors typically used less than 100 W of energy per channel. But modern equipment often requires power in excess of 200 W per channel. Future equipment is predicted to need even more. As these channels are brought together in parallel clusters, the total energy being utilized presents challenges.
Current industry practice typically relies on AC distribution systems to power data processing equipment clusters. In addition, critical infrastructure installations generally use systems with an uninterruptible source of electrical power. Uninterruptible power is generally viewed at the facility rather than the equipment level. While this provides an easy demarcation between the facility and data processing equipment, with each focusing on a different part of the problem, it also results in operating efficiencies and ownership costs being difficult to ascertain and optimize. Facility-level AC UPS systems are often built around models similar to the one shown in Figure 1.
The central AC UPS converts incoming raw AC power to DC, then converts it back to AC. Moreover, since data processing equipment generally operates from 208/220VAC line voltage, facility-level entry power usually must be transformed from 480VAC down to 208VAC, either before or after the UPS. A battery is then connected as an alternate input to the DC-AC converter. Should incoming raw AC power be interrupted, the system switches to battery power. Alternately, a rotating flywheel can store energy and provide uninterruptible AC power. Uninterruptible AC power is then passed through an AC power distribution grid to individual data processing apparatus.
Once inside data processing equipment, AC power is again converted to DC. Off-line supplies provide power factor correction as well as load isolation from the incoming power line. This generally involves at least two power conversion stages. In addition, most modern microprocessors require very low voltages at fairly high currents, such as 1.1V at 100A. Moreover, the precision of the voltage required is such that voltage regulation circuitry must be directly next to the microprocessor. To effectively realize this circuitry, most processors require that an intermediate DC voltage, such as 12VDC, be delivered to the processor/local regulator combination. (In ATX compatible servers, off-line, “silver box” power supplies provide this directly to the motherboard. In blade servers, the off-line supply usually provides a higher voltage and an “intermediate bus converter” is included on the micro-controller board to provide 12V.) Thus, six or more power conversion stages lie between facility power entry and the microprocessor. They include:
- 480VAC to 208VAC (Step down transformer)
- 208VAC to 400VDC (ACUPS front end)
- 400VDC to 208VAC (ACUPS back end)
- 208VAC to 400VDC (Silver box PS PFC front end)
- 400VDC to isolated 12VDC (Silver box PS back end)
- 12VDC to 1.1VDC (DC-DC converter on motherboard)
Some of the limiting factors on power conversion efficiency are voltage ratings of semiconductors, along with their corresponding conduction losses, as well as power converter package size limitations and economics. All combine to limit available efficiency improvements. Power conversion efficiency for best-in-class AC UPS equipment ranges up to 92 percent for isolated battery-based systems. In the case where an ATX computer grade AC-DC power supply is used, efficiencies generally reach no higher than 75 percent. Non-isolated converters that provide final processor power conversion range as high as 88 percent efficiency. In a typical ATX-based solution, the net power efficiency from facility entry AC to the processor is 59.5 percent. For every watt of power used to process data, another 0.66W is required to support power conversion. (For air conditioning, up to another watt of power will be required for each watt utilized to cool the power conversion equipment.)
In the case of a cabinet-level 48V Bus central architecture, best-in-class AC-to-48VDC front-end power supplies run at 92 percent efficiency, while 48VDC to 12VDC intermediate bus converters will provide up to 94 percent efficiency. Net power efficiency from facility entry AC to the processor is 68.6 percent. For every watt of power used to process data, another 0.45W is required to support power conversion. (In the case of air conditioned facility cooling, up to another 0.68 watts of power will be required for each watt utilized to cool the power conversion equipment.)
While not especially a problem at the individual microprocessor level, this inefficiency can have a significant effect when overall data processing activities reach power usage of 500 kW or more, which is not uncommon in large internet hub installations.
Uninterrupted AC power distribution constitutes another area of complexity in large data processing installations. Facility-level AC UPS systems usually have a centralized circuit breaker panel with power monitoring and bypass provisions for UPS and battery servicing. Quite often these are located away from the equipment being powered, leading to possible operator confusion regarding what equipment is fed by which breaker and raising the possibility of inadvertent data processing equipment shut down. Furthermore, AC power distribution must be installed in anticipation of eventual data processing equipment usage, often requiring upfront capital expenditures larger than necessary for planned phase in. In some cases the nature of the UPS being used is sensitive to load current harmonics, requiring care to ensure compatibility between data processing equipment and UPS. Facility-level AC UPS can also expose a single point of failure for large amounts of data processing infrastructure.
Centralized battery plants used in facility-level UPS systems may be a compromise between what’s needed in a successful system and component limitations. Many times the battery run time needed for acceptable system operation is a fraction of a minute, representing the time required to switch between alternate utility power feeds or the time to bring an auxiliary power source on line. Correctly sizing batteries for centralized AC UPS systems can be complicated and may cause delays and wasted energy capability. Likewise, the lower efficiency provided by the off-line power supplies creates an extra load on batteries that only produces more heat.
A Better Way
System design can be optimized using DC uninterruptible power systems distributed throughout the data processing facility. Unlike traditional central office telecommunications installations that have one large DC power plant feeding an entire facility, these systems are made up of smaller power plants spread throughout the facility. They are sized to power clusters of end-use equipment with optimization possible up to 200 kW per cluster. And since an abundance of telecom equipment is built around -54.5VDC, this becomes a logical distribution voltage. At 200 kW, the equivalent current for this voltage is approximately 3670A. A distributed DC UPS approach is depicted in Figure 2. In this system, each data processing cluster has a dedicated DC UPS system feeding it. Variations on this solution are possible to address nearly any data center situation.
Steadily growing data processing power demands require a fresh look at how things are done as well as the total cost of ownership. DC UPS technology developed in the telecommunications industry can provide a reliable, cost effective alternative to traditional AC UPS systems, especially when applied in a non-traditional, distributed manner.
Gary Mulcahy is executive vice president of the Commercial Products Division of Transistor Devices, Inc. (TDI). He received his BE-EE from New York University followed by graduate study at the Polytechnic Institute of New York and is a recognized authority in power conversion technology and the design, development and production of power systems for maximum performance and reliability with minimal life cycle cost of ownership.