Fullscreen HTML5Med Res (??MB)HTML4

SIGNS DON’T WORK! YOU NEED THIS JUNK MAIL REPELLER! Is your letterbox full of junk, even though you have a NO JUNK MAIL sign? If so, you need to build our Junk Mail Repeller. It might not completely prevent junk mail from being shoved in your box. . . but it should at least help. And you’ll have some fun watching the reactions of the would-be junkmeister! by Allan Linton-Smith L et’s face it, the people who deliver junk mail must be completely blind (or no comprehende Engrish!) because they can’t seem to understand the “NO JUNK MAIL” or “NO ADVERTISING MATERIAL” sign in giant letters on your letterbox. But hopefully they aren’t deaf, too; that’s where this gadget comes in. For a little over two dollars (plus a few bits and pieces that you probably already have), you can build this junk mail-triggered digital audio recorder/ playback device. Just imagine, as they cram yet another flyer into your letterbox, a voice yells back at them: “HEY YOU! The sign says NO JUNK MAIL!” That’s just one of its many uses! But fundamentally, it’s just a fun project that you could probably think of a thousand uses for. Maybe a switch on your bedroom door and a voice saying “sisters not welcome!”? By the way, even with a “NO JUNK MAIL” sign, it isn’t illegal for a business or individual to put a flyer in your letterbox (even if it is against the industry code of practice). The problem lies with psychology 101: the junk they’re delivering to you 48 Silicon Chip isn’t junk – it’s a vital message that you would be most upset not to receive. Therefore any sign doesn’t apply to them. Only to the next bloke with junk! So people who stuff junk mail in your box can’t be arrested! But you can discourage (and probably annoy) them with this device. If you actually like and want junk mail (and that is about the only mail you get these days), do not attempt this project. The project isn’t just based on the ISD1820 module . . . it IS the project! Australia’s electronics magazine Or maybe you should build it and use it to say “thank you” to the people delivering you free catalogs. How does it work? Every time a flyer or catalog goes into your letterbox, the extra weight should be enough to trigger a microswitch – and they’re greeted with a message – eg, “No junk mail please – Australia Post only......we are watching you!” Then have some fun watching their reactions! (Tough luck if it is your Australia Post postie delivering the junk, as they sometimes do!). You can put any message you like, in any language. We discourage the recording of a tirade of swear words, although that would of course but possible, as it may land on inappropriate ears. There isn’t much to it; it’s made from a pre-built, low-cost digital voice recorder which is installed in a plastic box, along with a microswitch and a battery. It then becomes a junk mail repeller! Description The voice recorder/playback module we’re using is based on an ISD1820 siliconchip.com.au We built our Junk Mail Repeller into a UB3 Jiffy box but just about any enclosure will do, as long as it fits inside your letterbox. The microswitch glued to the outside of the lid is the secret: it triggers the voice message whenever anything heavier-than-an-envelope (eg, junk mail!) lands on it. The switch on the end is optional – it changes the length (and quality) of the voice recording which YOU make to suit the situation. IC, which can record up to 11 seconds of audio. We chose this one because (a) it’s a nice, small unit, measuring just 38 x 42.5mm; but (b) more importantly, it’s cheap and really easy to get; you can get it from our online shop (see the parts list for details) or you can wait a few weeks after ordering from eBay or AliExpress, etc. Search for “isd1820”. We’d suggest being just a little careful on line – the best price we found was US$1.74 including postage. But another supplier was asking AU$10.58 PLUS $73.83 postage! Whew . . . we thought Ned Kelly was Australian. . . Ours came ready-made, complete with a tiny loudspeaker. The speaker would cost you more than we paid for the whole thing if you bought it locally! The module can be powered from 3V (its stated maximum is 7V) from two AA cells in series. The standing current drain is 220µA, and it consumes about 38mA during playback. The cells should last for months, depending on your junk mail load! Note that there is a slightly different module available than the one we used, which has a 10-pin header and two slide switches instead of a 12pin header. This one is also suitable for use in this project, but you have to make a few slight changes. These are simple enough that we’ll leave them to you. That alternative module is quoted as working from 2.4-5.5V, which is fine since our battery is around 3V. And speaking of Australian, Jaycar have a similar module which is not that much more expensive but doesn’t come with a speaker. (Cat XC4605). siliconchip.com.au The circuit The circuit of our module is shown in Fig.1. The ISD1820 (IC1) is responsible for all audio recording and playback tasks. A 100nF capacitor bypasses its 3V supply (from two AA cells). During recording, it samples audio from onboard electret microphone MIC1, which is AC-coupled to its pin 4 and 5 differential inputs. MIC1’s power supply voltage is filtered by the 1kΩ resistor and 220µF capacitor, while the 4.7kΩ series resistors provide suitable biasing. A 4.7µF capacitor sets the time constant for IC1’s internal automatic gain control (AGC), used during recording to automatically provide a suitable gain for the microphone. Recording is initiated by the REC pin (pin 1) going high and continues as long as it stays high. During recording, the RECLED pin (pin 13) is held low, so LED1 lights. The RECLED output is also pulsed low at the end of playback, causing LED1 to flash briefly. IC1 has a small internal audio amplifier, allowing it to drive the 8Ω speaker directly, via pin header CON2. The module is supplied with a suitable cable to connect the speaker to this JST header. Playback is initiated by bringing either pin 2 (PLAYE) or pin 3 (PLAYL) high. The difference is that the recorded message will continue to play until EXTERNAL MICROSWITCH TO TRIGGER PLAYBACK VCC 2xAA or 2xAAA 1 3 FEED-THROUGH ENABLE JUMPER S4 CON1 2 4 FT 5 6 PLAYL 7 8 PLAYE 9 10 REC 11 12 S3 S2 S1 LED1 RECORD 13 12 3 1 S5 S1: PRESS TO PLAY S2: HOLD TO PLAY S3: HOLD TO RECORD SC  20 1 9  100k 1nF 10 RECLED 220 F PLAYE MIC PLAYL AGC REC IC1 ISD1820 SP+ SP ROSC VSSA 8 4.7k 11 VCCA MICREF FT 1k 100nF 1k 2 PLAYBACK LOOP ENABLE JUMPER VCC 5 100nF MIC1 4 6 100nF 4.7k SPEAKER SPK1 + 9 7 – VSSD 14 4.7 F CON2 OPTIONAL SWITCHED RESISTOR TO ADJUST SAMPLING RATE JUNK MAIL REPELLER (isd1820-BASED MODULE) Fig.1: the circuit of the voice recorder/playback module, with IC1 providing all of the recording and playback functions. This diagram includes the three extra components you will need, ie, a two-cell battery to power the unit, a microswitch to trigger playback of the recorded audio and optionally, a resistor connected across the onboard 100kΩ resistor to provide better sound quality. Australia’s electronics magazine August 2019  49 Fig.2: the internal workings of the audio recording and playback chip. The external resistor from ROSC to ground sets the oscillator frequency which determines the sampling rate. When recording is activated, the output of the microphone preamp feeds into the storage array via an antialiasing filter. And when playback is activated, the contents of the storage array are fed to the output amplifier, which is capable of driving an 8Ω speaker at a reasonable volume. the end even if PLAYE goes low again, whereas PLAYL must be held high for playback to continue. In other words, PLAYE is edge-triggered while PLAYL is level-triggered (hence the names). If pin 12 (FT) is held high, audio from the microphone is fed through to the output. The module has on-board tactile pushbuttons which pull the PLAYE, PLAYL or REC pins high when they are pressed. These signals are also fed through to pins 7, 9 and 11 of CON1 where they can be connected to external buttons, microcontroller outputs etc. FT is fed to pin 5 of this header, while power and ground appear on pins 1 and 3 respectively. The other half of CON1 is intended so a jumper can be placed across pins 2 & 4, permanently enabling feedthrough, or between pins 4 & 6, in which case no connection is made and feedthrough is controlled by pin 5. Bridging pins 10 & 12 causes the RECLED output to be connected to the PLAYE input. Since RECLED is pulsed briefly low at the end of playback, after playback finishes, this will cause playback to start again, as there is a lowhigh transition on the PLAYE input. Therefore, playback will loop forever, or at least until the bridging jumper Fig.3: we varied the value of ROSC and measured the recording/playback time. As expected (based on what it says in the data sheet), the sampling rate is inversely proportional to the resistor value, thus the recording time is directly proportional to it. The sampling rate is equal to 640,000 divided by ROSC in kilohms, which gives 6.4kHz with the default value of 100kΩ. 50 Silicon Chip is removed (it can be kept on pins 8 & 10 when not used). Finally, the 100kΩ resistor from ROSC to ground sets the audio sampling rate to 6.4kHz, which means the maximum length of the audio recording is around 10 seconds (we measured it at 11). This can be changed either to give a longer recording time with worse quality, or a shorter time with better quality. Chip internals Fig.2 shows the internal block diagram for the ISD1820 IC. It comprises a microphone preamplifier, oscillator, audio sample storage array, audio amplifier, filters, power conditioning and control logic. The storage array is quoted as retaining the saved audio data for up to 100 years, or until the next time you press the REC button! The power amplifier can deliver about 80mW into 8Ω, which is sufficient to give quite a reasonable volume when the speaker is mounted in a Jiffy box (ie, using it as a baffle). A more powerful amplifier could be hooked up to the output, along with a larger speaker, but this may annoy your neighbours! Recording quality vs time We tested various values for the resistor from ROSC to ground and plotted the results in Fig.3. As you can see, it’s very close to being a straight line. Australia’s electronics magazine siliconchip.com.au Fig.4: the measured frequency response of the unit from microphone to speaker, when the recommended 390kΩ resistor is connected across the 100kΩ onboard resistor from ROSC to ground. This gives a sampling rate of around 8kHz and an audio bandwidth of just over 3kHz. The Nyquist limit (ie, highest possible frequency) when sampling at 8kHz is 4kHz, but the filter’s transition band reduces the usable bandwidth to around 3/4 of that figure. This gives eight seconds of playback time and we deem the audio quality to be adequate. The minimum recommended value is 80kΩ, giving a sampling rate of 8kHz and a maximum recording time of eight seconds. But you can reduce the value down to 18kΩ, giving just under three seconds of recording time, and presumably a sampling rate of around 35kHz. The maximum recommended value is 160kΩ, giving a sampling rate of 4kHz and a maximum playback time of 16 seconds. You can go as high as 200kΩ, but the resulting sampling rate of 3.2kHz is poor, giving an audio bandwidth of just 1.3kHz. While the default rate of 6.4kHz is good enough for voice, after some experimentation, we settled on 82kΩ as the best compromise, giving a sampling rate of 8kHz and around 8.5 seconds of playback time. While the 100kΩ resistor is an SMD +3V Fig.5: if you solder a 33kΩ resistor in parallel with the existing 100kΩ resistor, you get 25kΩ and that sets the sampling rate to around 20kHz, resulting in the nearly 10kHz of audio bandwidth shown here. The sound quality is better, but the playback time is now limited to around three and a half seconds. That may or may not be enough, depending on what message you intend to convey! type, since you will probably want to lower the value if you’re changing it, you can simply solder another resistor across it. For example, connecting a 390kΩ resistor across the existing 100kΩ resistor will get you close to the 82kΩ ideal value. You can even connect this resistor via a switch, giving you two different options by merely flicking it. Note though that if you record with the switch in one position and play back in the other, you will either sound like a chipmunk or Barry White! While we mounted the switch and resistor inside the Jiffy box, this may be regarded as superfluous – once you’ve decided on the resistor you require (if any), it could be soldered across R2 and the switch could be left out. Building it Once you have gathered the items in S5 and 390kΩ RESISTOR IN SERIES (OPTIONAL) 0V 8Ω SPEAKER (VIA CONNECTOR) TO MICROSWITCH siliconchip.com.au Australia’s electronics magazine PARTS LIST – JUNK MAIL REPELLER 1 ISD1820-based voice recorder module with a small speaker and speaker wires (SILICON CHIP Online Shop Cat SC5081) 1 microswitch 1 UB3 Jiffy box (eg, Jaycar Cat HB6023 or Altronics Cat H0153) 1 2xAA or 2xAAA cell holder 1 390kΩ 1/4W 5% resistor (other values can be used; see text) 3 M3 x 10mm panhead machine screws, flat washers and nuts (for mounting the speaker) 1 SPST toggle switch (optional, for switchable sound quality) 2 female-female or 4 male-female jumper leads light-duty hookup wire neutral-cure silicone sealant Front and rear shots of the PCB showing the modifications we made. The connections to the PLAYE switch on the back of the board could also be made on pads 2 and 9 of CON1 (indicated) or indeed to the pins themselves on the top side. August 2019  51 Three 3mm screws, with washers and nuts, hold the speaker in place, as seen here. the parts list, building it is easy. Solder the bare ends of the supplied lead to the speaker (if they aren’t already connected) and then plug this into the header on the module. Wire up the 2x1.5V battery holder to pins 1 & 3 of CON1, with the positive end to pin 1 (don’t get it the wrong way around or you might let the smoke out...) You can do this quite easily by cutting a female-female jumper lead in half, stripping and soldering the bare ends to the battery terminals, then plugging these into CON1, taking care that the right leads go to the right pins. You can use a similar technique to wire up the microswitch between pins 2 & 9 of CON1. Alternatively, as we did, you solder the microswitch wires to the appropriate pads on the back of the PCB (either method is fine!). Next, drill the holes in the Jiffy box to accommodate the speaker and the microswitch. Once again, exact positioning is not needed. For the speaker, we cut the hole using a 35mm holesaw.The microswitch depends on which type and size you have. Ours (13 x 6mm) had three pins emerging and we drilled three 2mm holes through the lid for these pins. You’ll also want to drill three holes around the periphery of the speaker mounting hole, for machine screws to hold it into place. We drilled three 3mm holes about 3mm out from the edge of the speaker hole, 120° apart. With these holes just outside where the speaker surround will sit, machine screws with flat washers and nuts will clamp the speaker onto the lid from the inside. See the photo above. If using a switch to control audio quality/recording time (as we did), also drill a hole and mount this now. Put this on one side or end of the Jiffy box – you don’t want it to interfere with the microswitch operation. Depending on the type of battery 52 Silicon Chip Here is the completed project, ready to scare off any junk mail deliverer. The AA battery holder we used is a nice friction fit in the UB3 Jiffy Box. And the switch at the end (S5) is optional – in fact, we probably wouldn’t bother fitting it once we’d decided on the length and quality of our voice recording. holder you use, you may need to make a small clamp to hold it in position with a hole drilled in the base of the box. With the holder we used, there is no need to clamp it – it slides down between the PCB guides in the side of the case and locks nicely in place. Check it twice! Check that everything is working and record your message. Make sure you are happy with how it sounds, then use neutral-cure silicone sealant to seal the gaps around the edge of the speaker and microswitch holes, and any other holes you’ve made in the case. While a Jiffy box is not waterproof, (especially with a speaker in the lid!) if you fit the lid on tight, it should survive the sort of splashes it’s likely to be exposed to in a mailbox. If you want to be sure, you can always apply silicone around the edges of the lid before attaching it to the case. All that’s left is to place the unit in your mailbox with the microswitch facing up so that anything landing on top of it will trigger the recorded message. Australia’s electronics magazine Go ahead, try it out! Then hide behind a tree and wait for an unsuspecting junk peddler to wander on by... And as we mentioned earlier, this project has plenty of other uses; eg – how about a pithy message when someone opens up your school bag? Don’t forget that most microswitches can operate in a “N-O” mode when held down and close when released – eg, when a bag is opened! How loud is it? On the workbench, the answer is “not very”. Certainly loud enough to be really annoying – but when you place the project in your letterbox, with all its resonances, it becomes surprisingly loud. Sure, it’s not enough to scare the deliverer into a quivering mess but it should be loud enough for them to hear! Speaking of placing it in the letterbox, make sure it is placed so that any junk mail (usually larger than legit mail!) can trigger the microswitch but ordinary mail might not have either enough weight or be in the right place to switch it. SC siliconchip.com.au