Fullscreen HTML5Med Res (??MB)HTML4

Using Electronic Modules with Tim Blythman Self-powered Wireless Switches These so-called ‘self-powered switches’ do not need a separate power source. You might have heard these referred to as kinetic switches, and seen them in wireless doorbells and remote-controlled light switches. We’ll investigate how they work and ways to interface with them. T hese are RF transmitters that do not need a battery or other power source. The accompanying receivers do require power, but as they are used to control the likes of mains-powered lights and appliances, power is readily available. They use a form of energy harvesting to send a brief transmission. The examples we tested use some interesting strategies to make best use of the limited amount of energy available. All devices mentioned in this article use the 433MHz LIPD (low interference potential device) band, which is actually closer to 434MHz than 433MHz. In the April and June 2025 issues, we presented a series of project articles for building a 433MHz Transmitter and Receiver pair (siliconchip. au/Series/439). The series includes an explanation of the LIPD band, its uses and its limitations. The power limits mean that its range is typically quite short, but useful within a typical household. In this article, we’ll look at a bare module, as well as a complete unit that has a matching receiver. We’ll investigate the energy harvesting circuitry and its operation, since we expect readers will be interested in that. We’ll also delve into the RF transmission protocol and how to receive signals from some of these devices, including sending and receiving compatible signals using our Transmitter and Receiver paired with Arduino code running on a Pico microcontroller module. The DFRobot TEL0146 Photos 1 & 2: The TEL0146 is a compact unit that incorporates an energy harvesting device and RF transmitter. It doesn’t need a battery. The rear of the TEL0146 shows the fixed coil, E-shaped core and moving pole pieces. The return spring for the lever is towards the bottom. 60 Silicon Chip We’ll start with this module since it is a bare unit with visible workings. Shown in Photos 1 & 2, it is just under 5cm long. There is a plastic frame that holds a black PCB and an assortment of other parts, like springs and coils. These modules are available from Mouser and DigiKey for around $16, excluding shipping. Information on the module is available from DFRobot at siliconchip.au/ link/ac84 Pressing and releasing the white lever triggers the transmission, and an onboard LED flashes briefly. The Australia's electronics magazine action is quite firm and has a satisfying click. Interestingly, the transmission occurs on the upstroke, as the lever is released. The lever travel is about 3mm at its outer end. The page noted above mentions that the lever should not be pressed more than three times per second, and that at least one of the configuration DIP switches must be selected. Fig.1 is the circuit diagram for the electronics on the module. It is based around U1, a Cmostek CMT2156B, which is an OOK (on-off keying) RF transmitter IC with integrated energy harvesting. Unlike our Transmitter module, this chip also includes circuitry to modulate the output RF energy and apply encoding. In 2021, Cmostek was bought by HopeRF. Apart from the addition of the extra voltage regulator circuitry at upper left, the circuit design closely matches an application note circuit in the CMT2156B data sheet. The regulator allows the module to be powered by a low-voltage power source like a battery. We applied 5V to the DC INPUT connections and found that this activated the transmitter in much the same fashion as the lever. So this module can be used as a conventional RF transmitter, too. Pins P2N and P2P of U1 connect to the coil, which is mounted behind the PCB. The data sheet appears to show a magnet moving near the coil, hinting that the energy harvesting is based on electromagnetic induction. The snappy action of the lever is siliconchip.com.au Fig.1: the CMT2156B includes internal rectifier and regulator circuitry to harvest energy from the coil connected to pins 10 & 11. When triggered, it sends out an RF signal, encoding the value set by DIP switch SW1. reminiscent of some piezo devices, but it is a simple mechanical spring here. The data sheet for the CMT2156B shows that the V5N and V5P pins have an absolute maximum rating of 6.5V. Based on E = ½CV2, the two 47μF capacitors can store around 2mJ each. The chip contains dedicated AC-DC and DC buck (step-down) circuitry using external inductor L1 to produce a regulated 2.4V at the Vout pin, and this is used to provide power for RF transmissions. This allows the IC to operate longer, as the higher voltage generated from the coil isn’t wasted; effectively, the initial current drawn from the reservoir cap is reduced until it partially discharges. The remainder of the circuit is for selecting and generating the appropriate RF codes. The chip supports so-called 527, 1527, 2262 and 2240 data encodings; it also has an internal EEPROM that can be programmed. The DFRobot page indicates that the 1527 encoding is used by the TEL0146. It also mentions that 600μJ of energy is generated, which sounds reasonable given that the 6.5V rating above is an absolute maximum. The 1527 encoding includes 20 identity bits, giving just over one million unique transmitter IDs, and four data bits, which correspond to the four DIP switch inputs on the TEL0146 Switch Module. There is no checksum for error detection. siliconchip.com.au E1 is a pair of unoccupied solder pads on the PCB. Bridging them causes the device to transmit on both strokes of the switch (press & release). It’s unclear whether there’s any benefit to that configuration, but as that is not the default, we doubt it. Zener diode D1 appears to be the part that clamps the generated voltage to a safe level. Note the interesting connection of crystal Y1, between pin 9 of U1 and GND, rather than between two pins as is commonly seen (Pierce oscillators). We suspect the crystal is being used in parallel resonance mode. and effectively has a two-way bistable action. Coil voltage We were curious what kind of voltages were present around the coil and other parts of the circuit. Scope 1 shows the voltage across the coil during a lever actuation. As the coil is connected to the P2P and P2N terminals of U1 on the PCB, the voltages may be different (and probably higher!) under open-circuit conditions. As expected, there are two spikes of opposite polarity, and the voltages appear to be clamped near to the Coil and mechanism Fig.2 shows the arrangement of the coil and mechanism. The fixed coil is in the centre of an E-shaped core with many turns of fine enamelled wire. The moving part has two pole pieces separated by a magnet. The magnet causes the pole pieces to be attracted to the core, so moving it requires some force. When the lever moves as shown by the arrows, the magnetic field in the core reverses, inducing a current in the coil. In the TEL0146 module, a spring is fitted. This returns the pole pieces to their original positions when the force is removed. Otherwise, the magnet causes one or the other of the pole pieces to remain stuck to the centre of the core. Later, we’ll look at another device that lacks the spring Australia's electronics magazine Fig.2: a moving magnet induces a current in the windings of the coil. The TEL0146 unit includes the spring shown here, and the mechanism returns to the lower position after each actuation. The rocker switch mechanism is bistable and is held in place by the magnets after operation. March 2026  61 Scope 1 (left): the voltage across the coil in a TEL0146 module. It appears there are internal clamps in the CMT2156B chip that keep the voltages within its 6.5V limits (or D1 clamps the voltage; possibly both). Scope 2 (right): the red trace shows the voltage on C1 and the blue trace on C5 (from Fig.1). The green trace is the RSSI signal from a nearby Receiver and shows when the chip is actively transmitting. By waiting for the upstroke, the chip harvests energy from both the down and up actions of the mechanism. 6.5V limits noted earlier. The timing of the pulses depends on the time between the lever being pushed and then released. Scope 2 shows the voltages on the two 47μF capacitors relative to circuit ground. The voltage on C1 (red) rises first, followed by the voltage on C5 (blue). The green trace is the RSSI (received signal strength indicator) voltage from a nearby 433MHz Receiver, from our project series noted earlier; this trace’s height roughly corresponds to the average RF energy received. We can see that the CMT2156B only starts transmitting when the second coil pulse arrives, and the RF is sent in packets. The small dips in the green trace correspond to the changes in the slope of the capacitor voltages. About three packets were sent in this case. Based on our calculations, the circuit draws around 5mA during transmission. The voltage levels out at about 1.8V, after which the resistors slowly bleed off the remaining charge over the course of seconds. The data sheet mentions that the minimum operating voltage of the CMT2156B is 1.8V, so presumably the chip shuts down when it detects this low voltage and stops drawing current. Other devices We also found a complete wireless switch system that appears to be based on the same principle. It includes a large rocker-style switch and a 230V wireless receiver module. The two units are paired, and when the rocker is actuated, the output of the receiver module toggles on or off. As a set, the switch and module worked quite well before we disassembled them. Photo 3 shows the transmitter and receiver set, while Photos 4-6 show how the switch unit comes apart. The main rocker simply pulls off. It is held only by small clips that also Photo 3: This wireless kinetic rocker switch works similarly to the TEL0146 but includes a simple enclosure and mains relay unit. The enclosure (left) measures 8.6 × 8.6cm, while the relay unit (right) measures 4.8 × 5cm Source: www.ebay.com.au/ itm/405115817334 62 Silicon Chip Australia's electronics magazine allow it to pivot on its axis. There is an enclosed transmitter unit that clips onto the rear plate. There are also versions that incorporate two switch paddles, and the backplate clearly has room to carry two transmitters. The transmitters have two arms. Their internals are a little different from the other module, but they appear to use a similar coil and magnet arrangement. Our investigations also revealed that they use the same 1527 protocol as the TEL0146 modules. The set (switch mechanism, relay and tape) cost $20 from eBay, including delivery. That particular item is now out of stock, but other items that appear identical can be found with a search for “kinetic switch”. That search brings up some other items that appear to work in a similar fashion, but we have not tested them. There also appear to be different sets available with dual and triple switch mechanisms and multiple relays. These devices are not supplied with a circuit diagram, although there is a small instruction booklet including details of how to pair other transmitters to the relay. Operation We thought that the TEL0146 modules took a substantial amount of force to actuate, while the rocker switches were easier to toggle. To quantify this, we placed the switches onto a digital siliconchip.com.au Photos 4-6: The front rocker of the switch pulls off to reveal a smaller module attached to the back plate. This module is self-contained and could be incorporated into a 3D-printed enclosure if you didn’t like the appearance of the original. The smaller module contains a similar coil- and magnetbased energy harvesting circuit and RF transmitter. scale and noted how much extra force had to be applied to actuate them. The TEL0146 modules took around 900gf (grams of force) to actuate, while the rocker switches required about 240gf. Given that the TEL0146 modules have a return spring, it makes sense that their operating force is much higher. As a comparison, miniature tactile switches, like those we use in many projects, have an operating force around 100gf. We tried the transmitters over different distances and found that they did not seem to have the same range as other battery-powered transmitters, although they were still capable of working from a few rooms away. Transmission protocol Scopes 3-5 show the RSSI (red) and data (green) traces from a 433MHz Receiver while receiving signals from various transmitters. Scope 3 shows a transmission from a TEL0146 module, Fig.3: the timing of the 1527 encoding is based around a fixed timer period, with the sync pulse being one period of RF on followed by 31 periods with it off. The longer on-period of the ‘1’ bit could also be viewed as the RF being on at the half-way point (after the rising edge) of each bit. siliconchip.com.au while Scope 4 shows a transmission from the rocker-style switch. Although it uses a different encoding, we also recorded a waveform from the transmitter in a Jaycar MS6148 Remote Controlled Mains Outlet, shown in Scope 5. The Jaycar Mini Projects series (siliconchip.au/ Series/417) includes a few projects that interface with this system, including the Arduino Clap Light and the RF Remote Receiver. With the knowledge that the TEL0146 uses the 1527 encoding (seen in Fig.3), we found a couple of Arduino libraries that claimed to be able to receive and decode that protocol. However, it did not report any codes when the module was triggered. Comparing Scope 5 with Scope 3 and Scope 4 gave us a clue. It turns out that this version of the protocol is sent at a much faster rate than other protocols we had seen previously. Importantly, the self-powered modules were clocking their data faster than the libraries were expecting. By tweaking some of the library timing parameters, we were able to see results that corresponded with codes that we found by manually decoding the scope grabs. This was unexpected, but not surprising, given the strict power requirements. Clearly, a faster transmission means less energy is needed! With this in mind, we noted some other aspects of the design that are Scope 3: this waveform is from a TEL0146 module; the green trace is the signal from a 433MHz Receiver, while the red trace is its RSSI signal. The third packet is truncated, probably because the capacitors discharged before it was finished. Australia's electronics magazine March 2026  63 Scope 4: the output of the rocker switch module shows a much faster transmission. Six packets have been transmitted, but the first has not been received correctly, possibly due to the Receiver AGC not settling in time. The last packet has also been truncated due to the harvested power running out. useful in a low-power situation. For example, the 1527 encoding has quite a large gap after its synchronisation pulse (compared to the sync pulse itself). This reduces the duty cycle of the RF transmitter and thus the average power requirement. The receivers work by comparing the instantaneous RF energy to the average, so a 50% average duty cycle provides the best contrast between the RF on (100%) and RF off (0%) states. The codes we saw favoured ‘0’ bits over ‘1’ bits, reducing the average to around 35% duty cycle. For example, the narrow peaks in Scope 3 correspond to ‘0’ bits, which outnumber the wider ‘1’ bits. Unfortunately, the libraries we tested were not able to detect these signals consistently, so we set about creating an Arduino sketch that could receive the codes from these devices. We also wrote a sketch to transmit the same codes to further validate the receiver. Fig.4 shows the wiring diagram for a Pico connected to a Receiver and a Transmitter, respectively. Arduino code Scope 5: a single packet from a typical battery-powered transmitter. This type of unit will keep transmitting as long as the button is held down. It uses a slower data rate than the other units, which have to make the best of a limited amount of energy. Fig.4: how we wired up Raspberry Pi Pico boards to our 433MHz Receiver (top) and Transmitter (bottom) to interface with the modules in this article. The pins used (GP2/GP3 here) can be changed in the software. 64 Silicon Chip Australia's electronics magazine The two sketches are named RF_ RX_EV1527 and RF_TX_EV1527 for reception and transmission, respectively. They include simple header files with some useful functions and variables. The pins used are set by #defines, so can easily be changed. These examples use the Pico Ticker library, so they should work with any RP2xxx board. The RF_RX_EV1527 sketch looks for a sync low period of at least 700μs (adjustable), so it will sometimes confuse noise with a valid signal. It will record and report (to the serial port or serial monitor) the timer period (which is 1/31 of the sync low pulse period), since the timer period is also needed for the transmitter sketch. You can look for consecutive matching packets to filter out noise, since the transmitters should send multiple packets each time they are activated. Checking the timer period can also help to filter out invalid packets. The TEL0146 module resulted in a timer period of 82μs, while the rocker switch has a timer period of around 27μs. As you can see from the scope grabs, the rocker switch sends out about five packets, compared to three for the module. The sketch simply reports the timer siliconchip.com.au period and a 24-bit result. These 24 bits consist of the 20-bit identity value and four bits that could be changed by toggling the DIP switches on the TEL0146 module. The rocker switch does not have DIP switches, but it appears that there are four sets of jumper pads that can be set using 0W resistors. The RF_TX_EV1527 sketch requires the RF_TX_TIMER_PERIOD to be set. We were able to trigger the relay of the rocker switch to activate by copying the code and timer period from the output of the RF_RX_EV1527 sketch. We could also get the RF_RX_ EV1527 sketch to produce the same code and, as expected, the scope grabs of the module and RF_TX_EV1527 sketch match quite well. In our research, we found some reports that devices like the rocker switch emitted different codes depending on whether they were being switched ‘up’ or ‘down’. This seems reasonable, since the IC would see different pulse polarities from the coil depending on which way the mechanism was moving. But we did not find that to be the case, with our unit reporting the same code every time it was toggled. The toggle action makes sense if the relay was paired with multiple transmitters, which appears to be possible. Summary These are interesting devices, and it is handy that they work without batteries. The TEL0146 is just a bare module and takes an unexpected amount of force to operate. It could be useful if incorporated into a suitable enclosure, possibly including an ergonomic lever mechanism that reduces the amount of force needed for its operation. The rocker switch unit is complete and works well, and if you need a simple switch for a mains appliance or light, as it comes with a matching wireless relay unit. The switch is unobtrusive and needs much less force to operate. Subjectively, we also found that we were able to receive its transmissions more reliably, since it usually sent more code cycles per press. Both units appear to produce only a single code each, and we were able to interface to the RF signals for both transmitter types, so it will be straightforward to create custom projects using either. Our demo software can be downloaded from siliconchip.au/ SC Shop/6/3316 siliconchip.com.au Silicon Chip PDFs on USB The USB also comes with its own case ¯ A treasure trove of Silicon Chip magazines on a 32GB custom-made USB. ¯ Each USB is filled with a set of issues as PDFs – fully searchable and with a separate index – you just need a PDF viewer. ¯ Ordering the USB also provides you with download access for the relevant PDFs, once your order has been processed ¯ 10% off your order (not including postage cost) if you are currently subscribed to the magazine. Receive an extra discount If you already own digital copies of the magazine (in the block you are ordering). THE FIRST SIX BLOCKS COST $100 OR PAY $650 FOR ALL SEVEN (+ POST) NOVEMBER 1987 – DECEMBER 1994 JANUARY 2005 – DECEMBER 2009 JANUARY 1995 – DECEMBER 1999 JANUARY 2010 – DECEMBER 2014 JANUARY 2000 – DECEMBER 2004 JANUARY 2015 – DECEMBER 2019 OUR NEWEST BLOCK COSTS $150 → JANUARY 2020 – DECEMBER 2024 WWW.SILICONCHIP.COM.AU/SHOP/DIGITAL_PDFS Australia's electronics magazine March 2026  65