Keeping your world up and running.®

Mice, transients and power drops: three things that make traffic lights go bad

Picture a northeast Texas town. There are 48 traffic signals in the town's 45 square miles, plus 74 school zone flashers and about 16 other flashers, all under the care of the Traffic Operations Manager, David, the traffic signal supervisor, Gary, and two signal technicians.

Traffic signals are safety-critical items, and one fails it's at least an inconvenience. Some failures - green in all directions, for example - are dangerous.

To make sure that doesn't happen the Naztec signal controller for each intersection is paired with a monitor mounted in the same cabinet that watches all its inputs and outputs. "If the monitor detects anything erratic or nonstandard it will automatically put the signal on flash," says David. The monitor latches the controller into the flash mode and notifies headquarters via wireless; a technician is then sent out to correct the problem and reset the controller.

The controller's 24 V dc inverse logic (ground = on) output goes to optically-coupled load switches with a triac output that control the 120 V ac to the LED banks in the signal (the town changed its traffic signals from incandescent bulbs to LEDs, which use less power and are warranted for five years, rather than the two-year maximum for incandescents). There's also 12 V ac for pedestrian pushbuttons, plus lines for vehicle detection sensors and the like.

Any number of things can cause a traffic signal to malfunction. Aside from internal failures the incoming voltage can go out of bounds or be noisy, traffic vibration can loosen internal connections, and sometimes a small creature will get into a cabinet and either short itself across two wires or gnaw through some insulation.

For some time the tools for checking out traffic signal problems were a Fluke 179 multimeter, the technician's experience, and the diagnostic indicators inside the signal cabinet. The tech would open the cabinet, check the status indicators on the controller, the load switches and the monitor, take some voltage and current readings with the 179 and have the answer. "We use all Fluke meters throughout the department," says David.

Once in a while things aren't so simple. Loose neutral connections, for example, are tough to track down, but there was one problem that defied explanation: One traffic signal would go on flash, but after a few seconds would resume normal operation. At certain times of the day this would happen repeatedly, causing no end of confusion for drivers. The problem had been going on for two or three weeks, says Gary, and "we were quite sure that it was some kind of power problem, but we were never there when it was occurring."

David suspected an intermittent power problem, but had no way to check it, so he bought a Fluke 43B Power Quality Analyzer. The 43B trends voltage, current, frequency, power harmonics and captures voltage sags, transients and inrush current. It measures resistance, diode voltage drop, continuity, and capacitance and can calculate 3-phase power on balanced loads from a single-phase measurement. Most importantly for this problem, its monitoring functions help track intermittent problems and power system performance and it records two selectable parameters for up to 16 days.

David took the 43B out to the malfunctioning traffic signal controller. When he arrived it was on flash, but as he watched it went in and out of flash several times. He set the 43B to record the 120 V ac input voltage; while he was watching it the cabinet's cooling fan came on, and the 43B registered a huge voltage spike. The controller went into flash immediately, then recovered.

Each time the fan came on the spike was repeated and the signal went into flash. But where was the spike coming from? Hidden up at the top of the cabinet next to the fan was an RC network that was supposed to suppress transients; David pulled it out and found it charred and useless. He replaced it and the problem was solved, less than 24 hours after getting the 43B.

Since that time David and Gary have found the 43B useful for any number of things, and its long-term recording feature is one of the reasons.

"I generally will put it on an intersection," says Gary, "and check it every 24 hours. I'll check it initially more often, but then if it's a random problem I'll check it about once a day, and I'm looking for power drops or current spikes, a lot of times." As the signals operate, he explains, they create a noticeable current pattern - when the signal goes to all red it uses more current, for example - "but then if you all of a sudden have a huge spike," he continues, "I'll look for the times that those occur, and then try to find something in the controller that matches up." The controller keeps a log of all fault indications that it sees, "so sometimes you can match that up to what happened."

The 43B's recording function is also useful for verifying brownout conditions that the power company tends to insist do not happen. "I particularly like the sags and swells," says Gary. "I like to use the transient part of it." He also appreciates the FlukeView software. "I can pull the data off easily onto my laptop," he says, "and then I take it back in the shop where I can compare the data of the controller to the data of the 43B."

Another feature Gary finds useful is the 43B's battery backup. It's often necessary to take measurements that for 24 hours, so standard practice is to operate the meter on AC power. If the power goes out during the recording period the meter will continue to operate on its internal battery and take an uninterrupted record.

David and Gary seem to be sold on Fluke equipment. "I've never had to buy more than one each time," says David.