Fullscreen HTML5Med Res (??MB)HTML4

Using Electronic Modules with Jim Rowe PM (particulate matter) “Dust” Sensors In this last article on low-cost air quality sensors, we look more closely at particulate matter (PM) sensors, also called “dust” or “smoke” sensors. A s mentioned in the first of these articles, PM sensors fall into three groups based on the size of the particles they are designed to detect: less than 10μm (PM10), less than 2.5μm (PM2.5) and less than 1μm (PM1.0). Currently, PM2.5 types are the most common in the low-cost section of the market, so we’ll concentrate on modules that support it. The basic principle of the most common type of PM sensor is shown in Fig.1. This was described in the first article but we’ll briefly go over it again. A small fan pulls air from the surrounding environment into a channel which passes through a sensing chamber. A laser sends a focused beam of light through the chamber, and any particles in the air scatter the light towards the sides of the chamber. One or more photodiodes detect this scattered light on the sides of the chamber. Any light not scattered by particles passes through the chamber to be absorbed by the ‘beam dump’. By controlling the fan speed and thus moving the air through the sensing chamber at a known rate of volume and measuring the photodiodes’ output, the concentration of particles in the air can be calculated. The result is in terms of μg/m3 (micrograms per cubic metre), because the traditional and most accurate way of measuring PM is the ‘gravimetric’ method. This involves using a preweighed clean filter to collect particles from the air over a 24-hour sampling period, then weighing the filter again to determine the total mass of the accumulated particles in micrograms. The concentration is then obtained by dividing this figure by the total volume of air that passed through the filter during the 24-hour sampling period. Available PM modules There are several low-cost PM sensors currently available, including the Grove-Laser Sensor module, based on the Seeed Studio HM3301 sensor from Shenzhen, China, and the SN-GCJA5 sensor made by Panasonic Photo and Lighting Co in Osaka, Japan. The first is a fan-type sensor, as shown in Fig.1. But other types of PM sensor modules do not have an internal fan, including the Panasonic SN-GCJA5 and the XC3780 from Jaycar, The Grove-Laser air sensor module is based on the Seeed HM3301 particulate matter sensor. The sensor itself measures 38 x 40 x 15mm and the module comes with a suitable cable. Fig.1: the basic operating principle of a particular matter (PM) sensor. Air is drawn through a chamber with a laser beam, and any laser light scattered by particles in the air is picked up by one or more photodiodes. 78 Silicon Chip Australia's electronics magazine siliconchip.com.au Fig.2: the components of the HM3301 sensor. The part at left is basically identical to what’s shown in Fig.1, while the section at right shows the electronics that pick up the scattered light level and turn it into a digital measurement. based on the Sharp GP2Y1010AU sensor. We will look at all three of these sensor modules in this article. The Grove-Laser module The Seeed Studio HM3301 sensor comes inside a compact plastic and metal case measuring 38 × 40 × 15mm. In addition to the fan, laser and photodiodes, it has built-in electronics that provide fan control, photodiode signal amplification, filtering, multi-channel data acquisition and an MCU (microcontroller unit) for data processing. The output is via a two-wire I2C serial interface. In the Grove-Laser module, the HM3301 sensor is mounted on a PCB measuring 80 × 40mm, with a four-pin connector at one end for connections to a 3.3-5V power supply and the I2C lines for connection to a PC or an external MCU. The effective PM2.5 measuring range of the module is 1-500μg/m3, although it can measure up to a maximum level of 1000μg/m3. This module is available from Australian distributor Pakronics in Rosanna, Vic for $46.06 plus shipping and GST, totalling $62.07. Fig.2 shows a functional block diagram of what’s inside the HM3301 sensor. The actual PM measuring section with the fan, laser, detection chamber, and photodiode detector is on the left. On the right is the electronics section with its filter/amplifier, multi-channel acquisition and internal MCU for digital signal processing and the I2C data communication interface. Since the HM3301 sensor operates from a 3.3-5V DC supply and has a standard I2C interface, connecting the module to an Arduino module or similar is relatively straightforward. A sample connection scheme is shown in Fig.3. Note that although the HM3301 sensor itself has no internal pull-up resistors on the SDA or SCL lines, the Grove-Laser module provides pull-up resistors plus logic-level converters on its PCB. That’s why the connections shown in Fig.3 are so straightforward. Of course, wiring the module up is only part of the story. You also need software that can communicate with it and display the results. So if you want to use it with an Arduino, you’ll need both a matching library and a sketch designed to communicate with the HM3301 sensor using it. When I went to the “Reference” section of the Arduino website and scrolled down through the Libraries/ Sensors list, I found a library that had clearly been produced to do the job: “grove-laser-pm2.5-sensor-hm3301”. And when I clicked on “Read the documentation” on its page, it took me to GitHub, where I found both the documentation and a link to download the library (v1.0.2). After downloading and installing the library, I found that it came with an example program called “basic_demo. ino”. After verifying and uploading that program ‘sketch’ to an Arduino Uno connected to the Grove-Laser Fig.3: connecting the Grove HM3301 module to an Arduino is simple. All it needs is a ground connection, a 5V DC supply and the SDA and SCL pins connected to an I2C bus. Fig.4: HM3301 sample output siliconchip.com.au 08:46:39.046 -> sensor num: 0 08:46:39.046 -> PM1.0 concentration(CF=1,Standard particulate matter, unit:ug/m3): 404 08:46:39.046 -> PM2.5 concentration(CF=1,Standard particulate matter, unit:ug/m3): 850 08:46:39.046 -> PM10 concentration(CF=1,Standard particulate matter, unit:ug/m3): 1356 08:46:39.046 -> PM1.0 concentration(Atmospheric environment,unit:ug/m3): 266 08:46:39.046 -> PM2.5 concentration(Atmospheric environment,unit:ug/m3): 524 08:46:39.046 -> PM10 concentration(Atmospheric environment,unit:ug/m3): 776 module as per Fig.3, the Arduino IDE’s Serial Monitor (set to a baud rate of 115,200) sprang into life. I immediately saw the text shown in Fig.4, with two sets of PM1.0, PM2.5 and PM10 measurements appearing every five seconds. The example output shown in Fig.4 is higher than normal (it should be just above zero). That’s because I struck a match and blew it out just before that, blowing the smoke towards the HM3301 sensor. The readings jumped up quite quickly but went back to normal after about 10 seconds. So while it’s not particularly low in cost, the Grove-Laser PM module is easy to use and seems quite sensitive. Panasonic SN-GCJA5 sensor similar to the HM3301 sensor innards shown in Fig.2, apart from not having any internal fan to move the air through the detection chamber. Since it has an I2C interface, it connects to an MCU like the Arduino in much the same way as the Grove-Laser module, as shown in Fig.6. But there’s one small but significant problem: connections to the SN-GCJA5 sensor are all made via a tiny 5-pin ‘pico’ connector at one end, but a connection cable with a matching plug is not supplied with it. So if you want to use – or even try out – the sensor, you first need to obtain a matching cable. Panasonic’s data sheet for the SN-GCJA5 sensor states that its connector is made by JST (Japan Solderless Terminals) Manufacturing Company, and has the type number SM05B-GHS-TB(LF)(SN). I had a lot of trouble finding any compatible cables – most cables I found with similar connectors turned out to have pins either 1.0mm or 1.5mm apart, not the 1.25mm of the JST SM05B-GHSTB(LF)(SN). Just as I was on the brink of concluding that I would not be able to try out the SN-GCJA5 sensor, Silicon Chip’s Editor emailed me to say that he believed he had found a supplier of compatible cables on AliExpress (www.aliexpress.com/ item/33005797784.html). I quickly checked them out and then ordered a pack of 10 (the smallest quantity). These cost $18.20 including postage and GST, and they took quite a few weeks to arrive. But they did finally arrive, and I used one (or half of one, to be precise) to hook up the sensor to an Arduino and try it out. It was again necessary to find a suitable Arduino library to communicate with the SN-GCJA5. Luckily, I found one in the Reference section on the Arduino website, under siliconchip. au/link/abep When I downloaded this library and installed it, I found that it again included some example sketches. The first of these was called “Example1_ BasicReadings.ino”. When I verified and uploaded this sketch to the Arduino Uno connected to the SN-GCJA5 sensor, as shown in Fig.6, it finally sprang into life. Once again, I had to set the Arduino IDE Serial Monitor to 115,200 baud. You can see the output of the sketch ► The Panasonic SN-GCJA5 sensor is again mounted inside a compact moulded plastic box that measures 37 × 37 × 12mm and weighs 13g. As with the HM3301 sensor, it includes electronics to control the laser and amplify and filter the signals from the photodiodes, plus an MCU for data processing. The output is via either an I2C or a UART TX terminal. The effective measuring range of this module is 0-2000μg/m3. The Panasonic SN-GCJA5 sensor is currently available in Australia from element14 for $33.56 plus delivery and GST, giving a total of $53.42 (less if you buy it along with enough other stuff, such as a second sensor, to get free delivery). Fig.5 shows what is inside the SN-GCJA5 sensor. As you can see, it’s External connections to the SN-GCJA5 sensor are via a tiny 5-way JST connector with 1.25mm pin spacing. No matching cable is supplied, which is a bit of a problem as they are hard to find! Fig.5: the Panasonic SN-GCJA5-PM sensor does not use a fan. It instead relies on passive diffusion of air through its sensing channel. Otherwise, its structure is similar to the HM3301 shown in Fig.2. The Panasonic SN-GCJA5 particulate matter sensor is in a small moulded plastic case measuring 37 x 37 x 12mm. In addition to the laser and photodetector, it contains all of the electronics and provides both I2C and UART digital outputs. 80 Silicon Chip Australia's electronics magazine siliconchip.com.au Fig.6: connecting the Panasonic SN-GCJA5 module to an Arduino is again simple. All you need to do is connect a 5V DC supply, a ground connection and the I2C bus via the SCL and SDA pins. Pin 1 isn’t used for anything, nor does it have any function. Fig.7: SN-GCJA5 sample output 08:03:23.189 08:03:23.189 08:03:23.189 7.5, 10, 08:03:23.236 08:04:18.209 08:04:28.238 146,4, 08:04:48.249 08:05:58.231 in Fig.7. It gives three PM readings (1.0, 2.5 and 10) at the start of each sample line, followed by six Count figures (labelled 0.5, 1, 2.5, 5, 7.5 and 10). The first three figures are the ‘mass densities’ for the three main particle categories, while the later figures are ‘particle counts’ for all six particle size categories. Looking at Fig.7, the first values outputted are pretty low, they then shoot up to much higher levels after I lit a match about 150mm from the sensor and then blew it out, blowing the smoke towards the sensor. So the Panasonic SN-GCJA5 sensor does work, and even works quite well, once you manage to find a suitable cable to connect to it. It would be -> Panaosnic SN-GCJA5 Example -> Sensor started -> PM:1.0, 2.5, 10, Counts: 0.5, 1, 2.5, 5, -> 2.79,3.12,3.50,5,32,4,0,0,0, -> 57.99,135.45,448.29,39,598,801,13,71,2, -> 1370.39,1730.99,2392,60,440,3824,3759,37, -> 139.76,154.51,173.83,513,1210,153,1,2,0, -> 62.35,73.86,83.09,200,591,120,1,0,0, a lot easier if they supplied a matching cable! The Jaycar XC3780 sensor As mentioned earlier, Jaycar’s XC3780 dust sensor module is based on the Sharp GP2Y1010AU fanless sensor. The sensor itself is pretty compact, measuring 46 × 30 × 17.5mm, and the XC3780 module is only a little larger, at 62 × 35 × 19mm. 7.5mm diameter holes in the top and bottom of the sensor (and the PCB) allow air containing any particulate matter, dust or smoke to diffuse through the sensor. At the time of writing, the XC3780 module is available from Jaycar stores for $23.95 or their online Techstore for $31.95, including delivery. Because the sensor’s mini six-pin SIL connector is on the top of the case, the XC3780 module comes with a short six-wire cable connecting it to a matching mini SIL connector on the end of the module’s PCB. There are some passive components at the same end of the board plus a four-pin SIL header with standard 0.1in/2.54mm spacing, to simplify connection to an external MCU. Fig.8 shows the components inside the GP2Y1010AU sensor itself, and as you can see, it’s similar to Fig.5 apart from not having a microcontroller to digitise and process the output signal. In this case, the analog output signal “VO” is simply made available at pin 5. Note that the centre amplifier The Jaycar XC3780 module is based on the Sharp GP2Y1010AU dust sensor. Being fanless, it relies on air diffusing through 7.5mm diameter holes in the top and bottom of the sensor’s case. It has a varying DC voltage output rather than digital outputs, so conversion into a dust density figure is done by software running on the controlling MCU. siliconchip.com.au Australia's electronics magazine November 2022  81 Fig.8: the main difference between this GP2Y1010AU ‘dust’ sensor and the Panasonic sensor shown in Fig.5 is that this one lacks any digital control electronics; it only includes analog signal processing. Therefore, the driving microcontroller module must power the LED via pins 1-3, measure the voltage at output pin 5 and convert that into a particle level. Fig.9: this curve shows the transfer function between the output voltage of the GP2Y1010AU sensor and the corresponding dust density in mg/m3. A table (or similar) representing the points in this plot needs to be loaded into the microcontroller to perform this conversion. Fig.10: there aren’t many components on the Jaycar XC3780 module besides the Sharp sensor. All they do is filter the power supply to the module, provide a power-on indication via LED1 and route the necessary signals to a standard four-pin header for connection to an MCU. 82 Silicon Chip Australia's electronics magazine section has a small adjustable resistor or trimpot to adjust the sensor’s effective sensitivity. But the Sharp data sheet for the GP2Y1010AU sensor warns that this trimpot is set to make the sensor conform to its specification before shipment. As a result, they advise against further adjustment of the trimpot. This specification is summarised in Fig.9, which shows how the output voltage (VO) varies with dust density. VO is close to 0.9V with zero dust in the air, rising relatively linearly to about 3.25V at a dust density of 0.4mg/ m3 before flattening off at about 3.55V for a dust density of 0.53mg/m3. It then rises very slowly to about 3.6V for a dust density of 0.8mg/m3. Note that 1mg = 1000μg. The complete circuit of the XC3780 module is shown in Fig.10, and there are only a few passive components on the PCB apart from the GP2Y1010AU sensor itself. The 150W resistor and 220μF capacitor provide decoupling and smoothing for the supply to the sensor’s internal LED, while the 1kW resistor and LED1 indicate when the module is powered up. Connecting the XC3780 module to an Arduino is quite straightforward, as shown in Fig.11. The GND and VCC pins of the module can be connected to the GND and +5V pins of the Arduino. The LED pin should be connected to the IO3 (D3) pin of the Arduino while the VO/OUT pin goes to the Arduino’s ADC0 (A0) input. These are the connections needed to ensure that the XC3780 module works correctly when a specific sketch runs on the Arduino. That sketch uses a particular library to control the LED inside the GP2Y1010AU and convert its DC output voltage into the equivalent dust density. I found this library on the Arduino website in the reference → libraries → sensors section. Called PMsensor, it was written by JongHyun Woo, and the latest version is 1.1.0. When I downloaded this library (“PMsensor-­ master.zip”) and installed it in my Arduino IDE, I found that it came with an example sketch called “PMsensor_ demo.ino”. This sketch provides almost no information on the correct connections for the sensor’s LED and VOUT lines, or the correct baud rate to use for the Arduino link back to the PC. However, after examining the code in the sketch, siliconchip.com.au I determined that the proper connections were those shown in Fig.11, and the correct baud rate was 9600 baud. I then powered it up and got the result shown in Fig.12. I decided to adapt JongHyun Woo’s sketch into one with more helpful information in a ‘header’ section. I called this new sketch “SC_PMsensor_­ demo.ino” and it is available to download for free from siliconchip.com.au/ Shop/6/62 As you can see from Fig.12, this sketch simply pulses the sensor’s internal LED once per second, then reads its output voltage and converts it into an equivalent dust density reading. This is then printed in the lines reading “Filter : XXX.XX”. You may have noticed in Fig.12 that at the top of the listing, the readings are low. But then they started rising because I struck a match and blew it out with the smoke passing over the top of the sensor. Precisely what these figures mean is not too clear, though. They could represent the dust density in μg/m3 (micrograms per cubic metre), or they might not. So the XC3780 dust sensor can be connected fairly easily to an MCU like an Arduino, and it does work using JongHyun Woo’s library and demo sketch. But the accuracy and significance of its readings are a tad indeterminate. The bottom line Overall, I prefer the Grove-Laser module based on the HM3301 fan sensor. It is the most expensive of the three, but not unreasonably so, considering its ease of use and the apparent accuracy of its readings. I would have to rate the Panasonic SN-GCJA5 sensor as the next best; although it seems to give fairly accurate readings, it lacks a fan and also has the disadvantage of not coming with a matching cable. The Jaycar XC3780 module is only about half the cost of the other two modules/sensors and is the easiest to get. But the fact that it needs software running in the Arduino to convert its DC output voltage into dust density makes me a little less confident in the accuracy of its readings. Still, it would be fine if all you needed were relative readings, eg, to use it as a kind of smoke alarm. SC siliconchip.com.au Fig.11: connecting the Jaycar XC3780 module to an MCU is straightforward. Various pin connections could be used, but this is the routing needed for the test sketch to work. It uses one digital pin (to control its internal LED) and one analog pin (for sensing the output voltage). Fig.12: XC3780 sample output 15:15:25.825 15:15:25.825 15:15:26.762 15:15:26.809 15:15:26.809 15:15:27.793 15:15:27.840 15:15:27.840 15:15:28.824 15:15:28.824 15:15:28.871 15:15:29.808 15:15:32.854 15:15:32.901 15:15:33.839 15:15:40.962 15:15:40.962 15:15:41.946 -> -> -> -> -> -> -> -> -> -> -> -> -> -> -> -> -> -> Read PM2.5 Filter: 11.15 ========================= Read PM2.5 Filter: 30.79 ========================= Read PM2.5 Filter: 78.10 ========================= Read PM2.5 Filter: 120.76 ========================= Read PM2.5 Filter: 253.01 ========================= Read PM2.5 Filter: 396.47 ========================= We assume the readings are in μg/m3 but the documentation is a bit vague Useful links Suppliers: • www.pakronics.com.au • https://au.element14.com/3523840 • www.jaycar.com.au Software libraries: • www.arduino.cc/reference/en/libraries/grove-laser-pm2.5sensor-hm3301 • https://github.com/Seeed-Studio/Seeed_PM2_5_sensor_HM3301 • www.arduino.cc/reference/en/libraries/pmsensor/ • https://github.com/ekkai/PMsensor • https://github.com/sparkfun/SparkFun_Particle_Sensor_SN-GCJA5_ Arduino_Library Panasonic SN-GCJA5 data sheet: siliconchip.au/link/aber Sharp dust sensor application note: siliconchip.au/link/abeq Australia's electronics magazine November 2022  83