Several of my articles recently have been about low-voltage direct current (DC) powered equipment and telecommunications systems vulnerabilities to power quality phenomena, especially with regard to transients, noise and other types of electromagnetic interference (EMI). It would seem that alternative communication media, such as fiber optics, would be the panacea for telecommunications, but this method of thinking ignores a critical aspect of such devices—they still need a power source. While it is true that some types of power quality phenomena may minimally impact the transmission media itself, it is the system’s weakest link that will still bring the system down.

For internal local area network and telecommunications, unshielded twisted-pair (UTP) cable, which ranges from telephone wire to Category 5 Ethernet cable, is the most common media used in North America, though shielded twisted-pair is used throughout most of Europe. The shield, which can be around individual pairs and/or entire cable as shown in Figure 1, below, provides extra protection from EMI, which includes everyday sources such as fluorescent lights. Like any protection, it’s only effective if the shields are properly terminated and grounded.

Fiber optic communications use either plastic or glass fibers to contain a digital signal composed of photons. The wave-like and particle-like properties of the photons make up the light that travels down the tube. The nature of photons makes the digital signals of light largely uncompromised by electromagnetic phenomena that affect electrical signals. However, the development of an “all photon” or “light computer” is not close as a practical device. So, the photon is pretty much confined to the transmission of data and must be converted back into the electronic realm to have any use.

The semiconductor devices that transmit and receive the electrons or the photons through the cable or fiber need a power source, which is usually a DC power supply that has its inputs from the alternating current (AC) power system—the weakest link. These power supplies have the same potential problems and needs of similar protection that I’ve described previously, but they often are neglected when designing and installing the system, particularly as the system is modified or expanded over time. What good is the computer or programmable logic controller (PLC) if it can’t communicate with the rest of the system that it controls?

For example, a private branch exchange (PBX) telephone system was relocated and installed in a newly wired telecommunications closet. The PBX wasn’t connected to an uninterruptible power supply at the new location since “we never had a problem in the old locations.” Every 20 minutes or so, the PBX would turn off and back on. This interrupted the phone calls for a couple of minutes while the system rebooted; obviously, this wreaked havoc on the business. A replacement PBX unit was installed but didn’t fare any better. Finally, someone put a power quality monitor on the power source and found significant sags occurring periodically, which corresponded to the reboots. An investigation of equipment that was on the same circuit found a soda vending machine on the other side of the wall in an adjacent room, wired directly to the same outlet from which the PBX was powered. When the compressor turned on to restart the refrigeration cycle, the inrush current caused a momentary voltage drop in the wiring back to the breaker panel, and all of the equipment, including the PBX, experienced the sag. In addition, the neutral-to-ground voltage swelled by half of the depth of the voltage sag at the same time, which is a common phenomena in single-phase circuits.

Another weak link in a process-control system can be the sensors or transducers that feed the information into the PLC for it to decide what sort of action to take. Given the length of the signal leads from the sensors, these can be vulnerable to EMI, especially if there are large horsepower adjustable speed drives (ASDs) nearby. But the power sources to the sensors also can be affected by sags and swells and transients.

Figure 2, above, shows an example from a fiber-winding system, where a typical sag of six cycles to 80 percent of nominal resulted in the output of one of the five sensors being corrupted. The takeup reel tension sensor output indicated too much tension, which caused the PLC to tell the ASD to decrease the speed of the motor when it shouldn’t have. This produced too much slack in the fiber. When the sag was over, the sensor indicated such, causing the drive to speed up too much and snap the fiber. The source of the problem turned out to be that the regulated voltage of the power supply to the sensor could not ride through even this relatively short sag.

Though one should check the obvious first, sometimes a different perspective helps find the system’s weakest link. Often, fixing that can be more cost-effective than a large-scale redundant power system.


BINGHAM, a contributing editor for power quality, can be reached at 732.287.3680.