The power distribution and control function has matured into a full-fledged design discipline.
Edelson Associates, Los Angeles, CA, USA
Power distribution and control (and the associated cable management function), while not a sexy topic, has grown into a recognizable, and important, subset of industrial equipment design. Power distribution and control subsystems (PDUs) are found in almost all application areas including medical, military, aerospace, communication, manufacturing, and test/measurement systems. The complexity of these distribution and control systems ranges from simple toggle switch line control, through multiple switching arrays with controlled power on sequencing, voltage step-down, and power redundancy by means of dual input auto-switching capability. Remote control of the power up and control/distribution function is also available. This latter function can use interconnection techniques including direct connection, modem interface (either with a computer or a direct phone line), or Ethernet (TCP/IP) connection to a LAN or WAN, or combinations of these functions (Figure 1).
Depending on the final destination of the system, it is important to ensure that the power distribution subsystem meets the latest regulations and testing standards appropriate to the country of delivery. These can include UL EN60950, Listed Approval for Information Technology Equipment (latest effectivity 04/01/03), US and Canadian bilateral standards, TUV certification, and Continental European listing. The CE marking guarantees that the products meet the more stringent EMC and safety specifications written into the European Low Voltage Directive. Manufacturing quality concerns are addressed by the assurance that production standards meet ISO 9001 and the new ISO 9001:2000 manufacturing certification.
Before considering the more complex issues of power sequencing, or remote power control, the selection process of a power control subsystem must address more mundane issues such as input/output power connectors, overload circuit protection, indicator lights, EMI/RFI filtering, and spike/surge protection. Power connectors meeting either IEC 60320, or IEC 60309 (shown in Figure 2) are the most commonly specified, as the IEC connector system is used throughout the world. Use of these standard power connectors, and appropriate cables that will mate with the controlled equipment, allows a system manufacturer to stock just a single type of PDU and still deliver equipment to almost any country in the world. The number of outlets is generally limited to 10-24 IEC 60320, type C13 receptacles with bale clip securing devices as options. An alternate cable retention system, described at the end of this article will allow a 20% increase in the receptacle count, and as important, the power plug size need not be specified at the beginning of the design cycle.
In general, most power distribution/control subsystems will provide circuit protection from overloads, with the simplest design providing manual on/off switching and a breaker trip for overload conditions. UL 489 listed electromagnetic main disconnect (used to assure life and property protection) breakers with a long time delay curve will usually provide the above function and are required for systems rated over 16 A (at either 120 VAC or 240 VAC). Some high-end systems use semiconductor load current sensing and control, thereby reducing the trip time and improving the overload/normal differential sensitivity, with an attendant increase in cost. Current monitoring integrated into the PDU overload function removes the need to use external meters during setup to accurately set the value of total unit load current. It is significantly more accurate than the alternate method of setting the trip current by using the nameplate-provided data; this technique usually assumes a worst case equipment option population and is therefore excessively overrated.
Most controllers will provide indicator lights for the main breaker power “ON” condition. This minimal indicator function can be augmented with indicators for each phase (in three-phase systems), or additional indicators for power to the switched outlets and the unswitched groups. For systems incorporating remote control functions, other indicators showing system status configuration are usually provided.
The power controller should also provide EMI/RFI filtering to reduce the conducted noise both into, and out of, the controlled equipment. In well-designed systems, both differential-mode (line-to-line) and common-mode (line-to-ground) filtering is designed in. Reasonable insertion loss numbers equal, or exceed, 50 dB at 10 MHz decreasing to around 10 dB in the 100 kHz range. Units with 10 to 30 dB greater insertion loss across this bandwidth are also available (again, at higher costs). If radiated noise is of concern, the designer must be aware that the power cord connecting the controlled equipment can act as a significant source of radiated energy. Building the controller with a metal case and installing it in a metal rack will provide additional shielding. To further suppress radiation, the entire enclosure may need to be shielded or difficult to obtain shielded power cords will have to be used.
Transient Voltage Surge Suppression (TVSS) is generally provided by a MOV (metal oxide varistor); these devices are available with energy ratings of between 150 j and 400 j for 8-20 µsec transients. Peak current ratings will usually be in the 10 kA range.
Power controllers are available for both single-phase and three-phase systems with three-phase systems generally proving higher current control capabilities. Also, three-phase systems are sometimes used where it is necessary to control devices with power “on” using different legs of the supply. The selection of a single-phase controller with sequenced power “turn-on”, and “turn-off”, can sometimes allow system power from a single-phase supply rather than a more complex and expensive three-phase input. With time delay control of the power outlets, you can set the power “on” sequence to any combination of outlets and adjust this differential timing, usually from one to hundreds of seconds. While sequenced outlet control may be selected for system operational requirements, this feature also allows the “turn-on” power transient to be significantly reduced. By sequencing high load devices so that each subsequent unit is provided power after the input current surge of the preceding unit has fallen to its steady-state value, the system “turn-on” current is limited to the “running” current, and current control demands may be reduced by factors of 2 to 10.
Some power controllers add even more intelligence to the control function, enabling remote transmission of system commands. When choosing a remote control option, the subsystem designer needs to determine whether a direct (hardwired) connection is suitable, whether transmission distances require the use of modem control, or if it is feasible to use a network LAN/WAN for connection. If modem control is specified, security concerns may require the ability to personalize the power controller name or to add additional password protection. Power controller identification and IP address become a necessary part of the Ethernet protocol if local, or wide area, network communication is used. The subsystem designer should also consider the complexity and variety of the commands available. I have already listed password and unit address creation, but other commands might include: all outlets on/off; specific outlets on/off; the sequencing setup of the outlets; and auto-reboot of a specific outlet (or outlets). Automatic indication of outlet status may be controlled by remote command as can the entire control of any “watch-dog” monitoring circuit that may be included. Watch-dog circuitry is usually included to monitor the control connection and perform an automatic reboot if the connection “locks up”. Control of this feature should include the ability to set the value of the “time-out” period which must be exceeded before the auto-reboot is enabled. If remote command access through modem connection is selected, it is usually prudent to include the capability to disable this feature from the front panel to avoid control conflicts during manual operation of the controller.
OTHER CONTROL OPTIONS
Other features which may be considered in the selection of a power control and distribution subsystem can include input voltage selection by either manual selection or automatic voltage-sensing controlling internal circuitry switching, input voltage step-down, or automatic switching between two power inputs. With automatic input switching between primary and secondary power during a brownout or blackout condition, the designer must be aware of the switching response time. This switchover time constant is usually in the range of 20 ms to 30 ms, and therefore, adequate power outage capability must be built into the controlled equipment.
And in the end, remember that inadvertently pulling the power cord from the power distribution and control subsystem is just as serious as a power outage. There are three basic cable retention systems that attack this “disconnect” problem:
- NEMA Twist Lock connectors,
- IEC plugs with Bale Clips, or
- An external cord retention and management system.
The first two techniques are non-permanent and allow insertion and removal of the power cord with no external tools, while the last technique provides a permanent method for securing the cords which can only be removed with cable tie cutters. The IEC bale clip and NEMA twist lock approaches require specification during the initial purchase of the power distribution subsystem, and require that the plug type and size be known at this point in the system design. The external cord management system features a universal cable restraint tray (Figure 3) with square openings for cable ties (which hold the plug and cable in place) and increases the maximum number of connectors available on a 1.75 in. power distribution to 12 IEC 60320, C13 receptacles.
The above discussion shows that the power distribution and control function has matured from the simple plug and switch level into a full-fledged design discipline. The intent of this article is to provide a view of the various subsystem design approaches and functions which are available to the design engineer.