Volume 13 Issue 1
Appendix: Detailed Methodology
Occupancy SensorsThe performance of a PIR sensor can largely be captured through two characterizations: the spatial sensitivity pattern and the absolute sensitivity of the infrared detector. The testing used a Peltier cooler device, which had a 24°C temperature difference between its front and back sides; flipping it from one side to the other simulated human movement in a room-temperature environment. For this test all sensors were set to their maximum sensitivity settings. As shown in Figure 5, NLPIP tested spatial sensitivity to only 30° from nadir, but the sensors are able to detect motion across a wider field.
Figure 5. NLPIP tested occupancy sensors to 30° from nadir. The sensors are able to detect motion outside of the tested sectors.
The spatial and absolute sensitivity testing was conducted by mounting the sensor on a bar goniophotometer and oriented so that the nadir direction was aligned along the length of the bar. A 5 cm × 5 cm Peltier cooler device was mounted 2.5 m away from the sensor. The hot side of the Peltier cooler device was approximately 37°C and the cool side was approximately 24°C, which corresponds to the temperatures of a human body and room-temperature surroundings, respectively. A stepper motor flipped the Peltier device by 180° (taking ~300 ms to do so), then paused for 0.5 seconds, and then flipped it back. During the flip and for 2 seconds after the flip the state of the PIR signal was monitored for a detection response. If the sensor registered motion, it was allowed to reset before the Peltier device was flipped again. This test setup allowed NLPIP to make sensitivity measurements more cost effectively than by following the National Electrical Manufacturers Association (NEMA) Occupancy Motion Sensors Standard NEMA WD 7- 2011. In order to measure the spatial sensitivity pattern, the goniometer was set to scan elevation angles from 0 to 30° in 1° increments and azimuthal angles from 0 to 355° in 5° increments.
To measure the 50% sensitivity threshold, the Peltier cooler device was moved farther from the occupancy sensor until only half of the flips were detected. Since detection of low-level signals is determined probabilistically, three to six trials were conducted at each distance to determine the detection percentage. The threshold distance corresponding to a 50% detection rate was calculated. Threshold distance was calculated for two angles of view: nadir, and at the approximate center of one of the sensitivity sectors next in elevation angle from nadir (typically 16° elevation angle). For this test, all products were set to their maximum sensitivity settings.
PhotosensorsTo compare the performance of wired and wireless photosensor systems, NLPIP purchased Lutron and Leviton wired and wireless photosensors that were appropriate for controlling the lighting in one room. Three of the systems (excluding the Leviton wireless photosensor system) could control multiple lighting zones separately, but NLPIP did not test this capability. While WattStopper occupancy sensors were investigated, their photosensor products were not tested because they did not offer a wireless photosensor at the time of the study, and the goal of the testing was to compare wireless photosensors with wired photosensors from the same brand.
The Lutron wired controller was connected to a Lutron ECO10 0-10V dimming ballast. The Lutron wireless controller was connected to a Lutron digital Ecosystem H-Series dimming ballast, one of the Lutron products it was designed to work with. Both the wired and wireless Leviton systems were connected to a Universal Triad dimming ballast.
The Leviton photosensor can be powered by either its integrated PV module or a replaceable battery, and NLPIP installed a battery for this test. The Lutron photosensor can be powered with only a replaceable battery.
The photosensors were tested in the Lighting Research Center (LRC) Daylighting Controls Simulator per the methods described in the NLPIP Scale Model Bench Test (NLPIP 2007). This device is a box with two openings that simulates an empty room that has a rectangular floor plan and a window. A high-power white LED is used to simulate daylight entering through the opening. It can provide over 6600 lux at the opening. This apparatus has a controller that measures the analog voltage (0-10V) from the photosensor, ballast power, illuminance in the simulator and relative light output in the LED box outside the opening. A custom software program changed the LED light output to mimic changing daylight levels over one day. The lighting from pre-dawn to post-dusk was simulated over a two-hour period. The LRC modified the simulator from that described in the NLPIP Specifier Reports (NLPIP 2007) in the following way:
Wireless CommunicationsNLPIP tested the maximum distance the sensors and controllers could be separated and still communicate reliably. The sensors were powered with new batteries and moved progressively further from the controllers. The testing was done in an office environment. Line-of-sight was maintained between the sensors and controllers.
To test the maximum communication distance, the following procedure was used:
To test EMI, the sensors (powered with batteries) were placed 15 ft (4.6 m) from the controllers. A signal generator (TPI Synthesizer Version 5.0; RF-Consultant, Austin, Texas) was used to generate electromagnetic energy. The generator was uncalibrated, but the signal frequency and strength was measured using a handheld spectrum analyzer (RF Explorer ISM Combo; Nuts About Nets, Bellevue, WA). The power of the emissions was increased in steps until the receiver no longer acted on signals sent by the transmitter. The signal generator was set to the same frequency as the wireless communication signal for each system, shown in Table 2. Testing was done in a laboratory and office environment, without protection from EMI. Background RF energy was measured at the transmission frequencies used by the three systems and found to be between -100 dBm and -94 dBm.
To measure the presence of electromagnetic energy, the RF Explorer spectrum analyzer was used. The frequency accuracy is ±10ppm and the absolute power accuracy is ±5dBm (RF Explorer 2015).
The signal generator (TPI Synthesizer) was also used to generate electromagnetic energy to test for interference with communication. The device is uncalibrated. However, the testing did not rely on the generator to determine the RF power. Instead, the electromagnetic power and frequency were measured with the RF Explorer spectrum analyzer.
To test EMI, the following procedure was used:
Energy Generation and StorageTo calculate battery life, NLPIP made the following assumptions:
The initial capacity level of each sensor’s battery is as shown in Table 4.
The battery run time was calculated using the following discharge rate equation:
where Q is the charge (in coulombs) available in the battery or capacitor, k is the rate at which charge is used by the device (coulombs/day) and r is the self-discharge rate (%/day). The discharge rate due to the load consists of three parts: 1) the steady state, non-transmitting operating current (Iop), 2) transmitting charge/day during unoccupied time periods (Ivacant), and 3) transmitting charge/day during occupied time periods (Ioccupied). In equation form:
where η is the fraction of time the space is occupied. For these calculations occupancy was assumed to be 50% of an 8-hour work day: η = 0.50(8h/24h) = 0.165.
The solution to the first-order differential equation describing the remaining charge is:
Q0 is the initial battery or capacitor charge (in coulombs) at time = 0. Life is given by solving this equation for time when Q(t) = 0 (i.e., no charge left).
To measure PV electricity generation, the super capacitor voltage was set to 3.0 volts.