#114 

You are most welcome Bruce Morgan, I'm glad you like the videos. Nice to hear from you.

Thanks for that, Bruce Peterson, nice to hear from you.

You are most welcome, ward procter, I'm glad you like the video. Nice to hear from you.

You are most welcome Sukapura Tech, I'm glad you like the video. Nice to hear from you.

Thanks for that, James Searle, glad you liked my (sort of) original thinking! Nice to hear from you.

You are most welcome Asger Vestbjerg, I'm glad you like the video. Nice to hear from you.

Wow! You remembered that, Bob? No prompting? Ha Ha! Those transistors would do well in a low voltage Arduino environment too. They have a 100 gain even at very low collector currents. Admittedly, we are limited to 50mA collector current. Thanks for sharing, Bob!

Glad you like it!

Ha ha! A cold Guinness, hey? Actually, I had a Murphy's this evening, just as good. And thanks for the good wishes. And I'm glad you like the simple circuit, I was somewhat satisfied with the performance too. I must now build it on stripboard whilst it is fresh in my mind. Nice to hear from you, Andy.

Was that the "rude photos" post? I think RU-vid zapped it pretty quickly (before I could delete the post here anyway) and I've blocked the username. Thanks for reporting this Jago, that's the last thing we need here.

Glad you liked it!

Thanks, Hans.

Thanks for that, VINESH V, nice to hear from you.

Thanks for that, Maxx B, nice to hear from you.

Indeed, it should be absolutely safe! The stick I get about mains safety! Thanks for posting.

@@RalphBacon I know, I know, it's my background as an electrical fitter that makes me squirm because I know the dangers of mains all too well. You do a good job of drawing attention to the dangers I just feel it doesn't do any harm to reiterate the risks involved! I won't mention the "m" word again, I promise! ;-)

I've only been "partially retired" from my old day job for a week now, and I'm already sleeping better and feeling more energetic, William, so I can only assume that you are totally correct. Glad you liked the video and thanks for your supportive words, appreciated.

Thanks for your post, Drex, good to hear from you.

Glad it helped!

They do different things, Russ. The cost of split transformers (split core current transformers) is quite high as you point out but then again it does allow the measurement of the actual current flowing rather than just the _presence_ of voltage as per my video (but which is all I wanted or needed). The cost is much lower from Far Eastern suppliers: Bare split core: www.aliexpress.com/item/Free-shipping-1PCS-5A-Sensor-Range-of-Single-Phase-Module-Ac-Current-Sensor-Module-For-Arduino/32877238653.html Split core with relay switch (to indicate overload): www.aliexpress.com/item/1PCS-5A-Overcurrent-Protection-Sensor-Module-AC-Current-Detection-12V-Relay-Module/32858449954.html So if you're only detecting the presence of voltage my adapted design should be fine; if you need to measure the current then you'll need to stick to split transformers. I'm glad you liked the video, Russ, and thanks for posting your experiences on this topic, nice to hear from you.

Thanks for the info! Glad it's making you think!

You certainly have your work cut out for you! Thanks for the sub and keep tuned! More Pico stuff this Friday.

An interesting Real World implementation you have there, Ricky, thanks for sharing. And if a picture paints a thousand words think what a video would do! Great to hear from you.

Indeed. It sure wouldn't be quality and may be difficult to follow with all the systems involved as well as the fact I have never done one before would be painfully obvious. May be a good idea since this heat precludes me from doing much else.

You are most welcome Pekka Grönfors, I'm glad you like the video. Nice to hear from you.

Certainly winding a few turns of the Line wire around a core, together with the sensor wire is how Earth Leakage Circuit Breakers work (aka RCDs) so this idea would doubtless work too, but the focus of this experiment is that it would be totally non-intrusive. Given that the range of the detection is in the area of a few mm I'm not sure stray fields would affect it, but I haven't done the experiments. Now here's an idea, Mike: it takes about 15 minutes to build so why not build one and experiment away? Then report back, of course, and share your findings. By 9am tomorrow morning would be good 🤣 😂😹 (that last one was Benny, by the way). Good to hear from you, as always.

I found the author you mentioned (and his circuit), Vasilis, thanks for the heads up. Unfortunately, I also found the exact same circuit many, many times on the Internet too, so it's quite difficult to say where it came from originally now. But it does seem to work quite reliably (for me) it will be interesting to see what the feedback is from others who build it too. Thanks for posting.

Thanks for your post, Simon Thomas, good to hear from you.

Glad you liked it Steve, best of all it encourages safe behaviour around mains electricity!

Thanks, Tony. Yes, I often about my Dad still and what he would have said about this. He would have been fascinated for sure. Regarding monitoring several things, you could use several of these cheap, transistor monitor circuits all running a signal to a single Nano (on different input pins). Then the Nano would know which circuits were live and which were not. Use an ESP8266 instead and we have the basis of a wifi-enabled, mains monitoring station with simple phone app... sounds like we might have just completed my design! Thanks for posting, good to hear from you.

I saw you already posted later on that you had made this so I replied to that. Cheers, Ed.

Check that this will accurately detect 24v AC - in this demo we're detecting 240v AC, 10 times as much and therefore a bigger electromagnetic field. It would be interesting if you would report back whether it can detect it or not. What do you reckon?

I built the circuit (no charge pump or 4th transistor) powered it with 12 volts, ussed a buzzer instead of led. I get .4 volts and a faint buzz running the antenna along side an (18 guage, 120v us) power cord. Not quite there. Maybe a 4th gain transistor. Ill report back if I have any success. Otherwise assume I am sulking in a corner....

Oh, don't sulk, the best antidote is another transistor (and/or the charge pump). Let us know how you get on. Come on, now, out of that corner...

There is no circuit to complete here, Mike, as the sensor wire acts pretty much as an antenna or aerial, so feeds in its signal to the rest of the circuit. I hope that clarifies things for you, thanks for posting a good question.

Yes, Cir, that is the fiddly bit, and might not matter to many people. But I certainly need it to detect my lights. You _might_ get away with running it in parallel with the switched wire of the pair but it might make it very unreliable; I'd prefer to gently (and safely) split the wire and loop it around the switched wire, exactly as the demo, by YMMV. Thanks for the great question, nice to hear from you.

I did try and look at the sense wire signal but it was just a jumble of noise; after the first stage transistor, though, I got a spiky 50Hz signal OK that my 'scope could resolve. Regarding Australian light switch wiring, they looked to me pretty much the same as UK ones: diy.stackexchange.com/questions/28013/why-is-my-australian-light-fixture-wired-this-way Logically, this is the way to do it. There is no more efficient way, and bringing the neutral to the light switch serves no purpose at all (unless the light switch is also a PIR, in which case it will need it). I've also seen some low voltage remote switches (eg for bathrooms but can be used anywhere) where there is a control unit in the ceiling where the ceiling rose is and a couple of low voltage wires run to the switch - now THIS is a nice way to do it as it reduces risk and hum and just feels right. It keeps the high voltage stuff where it needs to be and protects consumers from 230v (or whatever it is in your part of the world) when removing light switches when decorating, for example. Unfortunately, this circuit would not then work! Thanks for your post, Tom, nice to hear from you again.

Well, it's good it worked but it sounds very sensitive. Put the aerial/antenna via a very small capacitor (eg 100pF) and see if that helps. If you disconnect the aerial does it turn off (ie it definitely turns on because it has detected something via it)?

If you build it and use it in a non-hobby environment then you are not covered by your work insurance as you are using a non-approved item. But if it is battery powered (this could be), in a plastic box, then the risk must surely be very low? But your employer will have different ideas, I'm sure.

That sounds like something is not connected up correctly. The idea is that the LED should light when something is ON not OFF!

​@@RalphBacon Sorry, my mistake. I had the resistor 100ohms instead of 100Kohms by error. Fixed it. Now, the LED lite up near the switches which are ON and drop from 4V high (using a 18650 cell) to around 1.5V for ON and back to 4V for OFF. More wraps and wire strands also improved the detection. Will try it over weekend. By the way, thank you very much! Because in my region, there is power cut on daily basis and inverter is used for providing backup, I'm planning to detect when mains goes off and comes back, same on the inverter side so as to know the backup time, inverter fault, logging these time info. etc.

Interesting slant on this design you have there, Bob. You seem almost on the point of trying this out so I'm going to hand over the baton to you to follow through and see whether your theory pans out in practice. I look forward to hearing the results in due course!

If the LED lights up when the switch is off, it means that that wire is still "hot" but the switch is preventing current flow. Typically in a house, the wires are live with a branch to the switch. The neutral wire will never give a reading as it is the equivalent of the Earth wire (N and E are connected together either just outside your house or by the electricity company on the way to your house).

I don't get how this works, John. How do you connect it up and what is the relevance of the LDR? I'm intrigued, do tell!

@@RalphBacon I only needed to detect the presence of the AC, and log when it went off and came back on. I didn't care about measuring the voltage. So I used a small neon indicator lamp such as: www.ebay.com/itm/20pcs-Neon-Bulb-Indicator-Lamp-Red-Neon-5X12mm-With-Resistor-Neon-Light-Lampada/252468642466?epid=850361273&hash=item3ac84dc2a2:g:BS0AAOSwLN5WlmQO:rk:1:pf:0 with resistor across the AC line. These are the same kind of tiny lamps found in orange light up switches. If AC (220v or 110v) is present, the neon will light. I use the Light Dependent Resistor LDR to detect if the neon is lit or not. The LDR is easily read with a Microcontroller pin using a pull-up or down resistor. Use heat shrink to couple the Neon with the LDR in a short piece of drinking straw. Essentially, this is a DIY Optocoupler, with an AC side, and a DC side. Providing complete isolation. Also, you can use a photo-transistor or photo-diode in place of the LDR. I've found this solution works nicely for only about 25 cents.

Ah ha! So basically it's what I demoed in video #111, just a bit more DIY. Cool, see we're all thinking along the same lines.

Yes, I'd had similar thoughts about using screened wire up to the point where you actually want to detect it, so spurious signals were not being detected along the way. But all bets are off during thunderstorms, who knows what happens then (we sometimes get power cuts then anyway, so I'd know that the lamps were off!) I'll try and do some experiments once I've built it onto stripboard. Thanks for posting.

You could check several times to see if the signal persists longer than a lightning bolt.

Yes, indeed, it would be sensible to sample the inputs (over one second maybe) and make sure you get three identical values before acting.

Ralph S Bacon decadies ago I went my feet a bit in trying to detect thunder/lightning and that wasn't easy. I suspect this circuit may be insensitive to lightning strikes, unless it's really close. Nevertheless, the proof of the pudding is in the eating

Didn't you do a vid entitled "useless projects" or something like that !...……….any how that involved a Tesla Coil...….Well now ya have ya lightning...……test away!

It's certainly easy enough to build, Tony (on a breadboard first, perhaps) and you'll find a use. It might flash if you put it near your mobile cell phone - it should light up when you receive a call. You can let us all know whether that happens!

You are correct inasmuch that this detector only detects the _presence_ of mains, not the flow of current. As you suggest you need either a current sensor, either something non-invasive like this: www.ebay.co.uk/itm/1Pc-SCT-013-000-YHDC-30A-50A-100A-Retractable-Current-Transformer-SCT0130-YM6K/313337299840 or something where you have to make a box with a socket etc: www.aliexpress.com/item/32890540897.html The problem with the first method is that the tumble dryer/washing machine etc can sometimes just sit there having a think, and during that time very little current is consumed - making your sensor suggest that the power cycle has finished. Even the second method has that problem but is slightly more accurate (but not much) because your current sensing device must be rated appropriately (so for a tumble dryer or washer, more than 13A, probably 20A is the next value). How does such a device detect just 100mA current? Poorly is the answer. What I did is have a configurable delay, so that I would wait for 5 minutes after a "no current" measurement and then sound a "washing finished" alarm if it didn't start up again (and reset the 5 minute delay). Tumble dryer was easier to detect than the washing machine.

Oh. I've just checked my video and the "white board" drawings there do seem to include the 4k7 resistor - so please tell me where the error is so I can fix it (or at least make a note so others don't get mis-led). That said, the 4k7 resistor should just ensure the capacitor doesn't remain charged (and hence give a permanently "on" signal). But I'm glad you got it working at last!

@@RalphBacon On second look, yes, there is a resistor in parallel with the capacitor. Thanks for clarifying! Again, great video!

strange, when i put in to resistor, it doesn't work. The values starts at 5V (I used analog pins to check the voltage), when main is off. When I turned on the main, the value goes to 0 and then back to high values (4.xx).....

Yes, coil the sensor wire around the cable carrying the current you are interested in, it should work...

If we're talking about the circuit at 6:00 then we are only running from 3v3 in the first place so it's just the inherent resistance in the circuit and the fact that only a few milliamps will flow because we have a 1N4148 diode in series with T3 too (forward voltage drop of about 0.7 - 1v [max]) and then via T4 (another 0.7v drop). This brings the voltage for the LED down to below 2V3 which is safe enough to ensure the max current of 20mA is not going to be exceeded. If you notice how dimly the LED lights you will see that my hypothesis is correct! Even at 5V VCC it will still be fine!

@@RalphBacon Have just tried your circuit, never once has a circuit worked, and worked perfectly the first time of asking. was looking at all sorts of solutions an this is by far the simplest. I will be looking into a 2 darlington configuration 10k gain at@ 1mA - little more expensive but same TO-92 package and hopefully mitigate T1 and T2. will let you know. Thanks again for the vid Ralph

No. but at least one of wires must be connected to mains voltage for it to be detected. Note, however, that if you have a light switch, the "line" (aka live) wire will be hot up to the switch. After the switch it will be off but usually the two wires to a light switch run in the same outer sheath so would be difficult to differentiate.

OK! The "orange thing" is a capacitor, connecting the antenna to the base of the first transistor (T1). It is a filter to reduce interference. It's probably about 0.1μF but you may need to experiment with the value a bit (say, from 0.01μF [10nF] to 0.1μF [100nF]). See what works best. There's another capacitor between the base of T4 and GND to eliminate jitter (ripple) and is "fed" by the diode from T3. The voltage at the base of T4 should be steady (800mV+ or 100mV). In other words the final OUT is either HIGH or LOW, not jumping about. Does the LED in your circuit remain steady ON when the mains has been detected? This will indicate whether your circuit is working properly (if not, then no wonder the Arduino detection does not work either). If you have a scope then it will be easy to see the outputs at each of the transistors emitters but if not check the above in logical sequence. Good luck!

Running off mains means that noise, as you suspect, will probably make its way into the circuit. Either seriously decouple the supply to the detector (low value resistors, in series, with caps to ground at each junction. Say 10Ω resistors and 100nF-ish caps) or just run off battery when you need to detect the presence of voltage. If you have a scope check what the voltage line looks like.

@@RalphBacon Thanks Ralph. I suspected as much but was not sure how to clean up the supply to the sensor. I'll will try your suggestion although in the meantime I bought the mains optoisolator from your other video - (sensibly encased in a project box with a transparent lid before the h&s zealots start kicking off) and that is very reliable!

Well, Steve, the idea is not to give up on this channel, the idea is to make the channel pay its way. So whilst Patreon _might_ be a way forward (see my response to John Crawford here) I'd much prefer it if my subscriber base increased, viewing figures increased and RU-vid promoted my channel more than it does. But, that said, it's a competitive marketplace so I will have to work harder and smarter. Thanks for your support, lovely to hear from you.

I imagine this would detect all mains-level voltages (eg 110 volts in the USA) but the lower the voltage the less the field around the wire (it must be AC). I don't know what the lowest voltage is that it will detect, I've never tried that.

Sorry, Stefano, I wouldn't like to comment on the suitability of your project, although detecting 24VAC should be tested first, as this was designed for 240VAC (or thereabouts).

It's *great* you put it together, Ed, and must be quickest anyone has ever followed one of my circuits, but I'm not quite sure what you mean about not much difference whether on the live wire. On a standard mains cable (eg one running to a heater, lounge light or anything that has two or three core cable with a plug on the end) it will detect the mains "as is". The problem is with the dual-core wire (plus earth) running to light switches in or behind your walls - you have to ensure the sensor wire is only looped around the switched wire, as per the demo, otherwise it will _always_ detect the presence of electricity. Perhaps I didn't quite understand what you mean, care to just clarify for me?

I understand your point Ralph, but, I noticed that no matter what (single) wire I coiled it around, as soon as the (electric) circuit is broken, the sensor picks up on that. But it's the continent, things are different here I guess ;-) Anyway, it works. That is the most important thing :-)

Ah ha! The plot thickens. I'm guessing from what you've said that the wiring layout and switching arrangement may be different on the continent (from the UK) in some way. But as you say, it works so that's the most important thing. Thanks for clarifying this.

@@RalphBacon (dunno if you still care, 5 years later, but...) could this just be a case of it needing to be passing a certain amount of current before the detector will register? That's my interpretation of what Ed reported, though I've not (yet) tried it myself, and I'm far from an expert, so I could easily be wrong. That's just how I'm interpreting it. Would be interesting to know! Would also be interesting to know if one could (without a much more complicated circuit) make at least some crude level of detection of different current flows -- not necessarily precision, but just having an ability to distinguish between, say, 10W and 1kW or whatever.

It detects the presence of voltage in the wire, even if no current is flowing (eg it is switched off). Good luck!

@@RalphBacon I made up 2 on a bit of veroboard to sense my Boiler Termostat demand and the Oil Sonenoid demand. The Heating controller demand I've got sensing from the battery charging circuitry. BTW I've also got 3 x 18B20s sensing the cabinet, water out and water return temperatures. A bit of Arduino code for an 8266 and I've now got all of this being reported to my Home Assistant through an MQTT broker. So thanks for this video which enabled me to complete my ambitions safely. I did use some different resistor values - 47k for T2, 2k2 instead of the LED, and 100k with the 10mf smoothing bit. One of my circuits is using 2N3904 and the other is using 2N2222. The latter is slightly more sensitive. In fact both were over sensitive until I made sure the mains cables being detected were properly connected to Neutral. The cheap JYETech home built oscilloscope really came in useful. Thanks again.

Update - I made up a circuit and put it in my boiler with the antenna on the wire coming from the internal thermostat and it is just too sensitive giving false on readings. When the power goes off it momentarily goes off too then over a few seconds the led indicator comes back on and says on. I've even tried putting the circuit board right next to the mains wire inside the boiler and keeping the antenna really short. Any clues on what I can do to stop this?

I have two devices in my workshop that have to act differently when the mains power supply is interrupted; both have battery backups, from a small Li-Ion battery of about 1000mAH. Anything from 500mAH to 2000mAH is small, slim and works great. The way I do it is to run the ESP32 board (which has an inbuilt battery/charger) from a USB-style wall wart; I also connect the incoming 5v (eg from the USB socket) via a suitable 3K/2K resistor divider to one of the GPIO pins. So far, so easy. When the power is interrupted, and the wall wart stops supplying power, the ESP32 keeps on running uninterrupted on its battery power. But it knows the power has gone off because the GPIO pin (monitoring the incoming 5v) reports zero volts. It then, for example, sends a WhatsApp message or email or whatever to alert me, or just does some other processing as required. It only requires the battery to run for a minute or so after power interruption, although I've discovered the ESP32 can run off my Li-Ion battery runs for 15 hours+ before it is depleted. If you put the ESP32 to Deep Sleep and wake up every 15 minutes to check the power state, it can run for a couple of days (at least, I've never depleted the battery!). My ESP32 board with built-in battery socket and charging circuit is a TTGO T8 v 1.7.1 but there are many others out there that have a similar setup, some quite a bit smaller than the one I mentioned. Just throwing this out there for you to consider too!

Let's hope someone reading this takes you up on your offer, Antoine.

I didn't have that issue AFAIK but I can only think that the 5v (switched) PSUs are introducing noise into the system that basically makes the circuit insensitive to anything. Try filtering the 5v via a very low resistor eg 100ohm followed by low value cap eg 100nF to 10uF.

I guess you could do that, John, but I'd like to see the advantage of an opamp over the circuit I proposed. KISS and all that! And it gives viewers a chance to actually use some discrete components here. Woo hoo!

@@RalphBacon I use op amps in these cases because they are an integrated solution to a high gain amp. Kinda like your power amp video I just watched. They are cheap and ubiquitous. I use them anywhere I don't need CURRENT amplification. Of course you can get high current op amps as well. I love your videos and am just throwing out alternatives. Back in the 80 I got an op amp circuit book and studied it well. You can use them for low / high / band pass filters, very linear amplifiers etc. Because of the high gain they bring to the table.

@@RalphBacon an op amp will be an 8 pin dip with 2 amplifiers. Input and feedback resistors, A diode and a capacitor. Much simpler circuit. KISS.😉. Love your stuff tho.

Are you picking up extraneous signals from your hands or other wiring nearby (eg other wireless equipment)? You can use a shielded box for the components, make sure the power is ripple free (use a battery for testing that) and shielded cable for the sensor up to the point where you want to sense power. Different transistors may have different gain (Hfe) that could affect the result - talking of which what results are you getting, John?

You're acting as a ginormous aerial, Peter, and you will have plenty of electromagnetic interference (ever touched the input to an audio amplifier and jumped out of your skin with the loud buzz that ensued?). You could use screened cable, of the type used for microphones or other audio inputs. The outer, coaxial screen goes to ground and the inner wire is your sensor that you can then extend (without the screened outer) for the actual detector wire (short-ish). Finally, try a high value resistor in the sensor wire, 100K to 10MΩ to reduce the sensitivity. And/or a high-ish value from the base of the first transistor to ground. Your transistor setup is probably a higher gain than mine was (just random luck) so is picking up more mains hum floating about in the ether. Good luck!

Hmm. Does the red LED also flicker wildly or is it steady? The 10µF should make the Arduino see a steady voltage not a rapidly moving one - you are using pin A0, I assume? If you put a multimeter on the base of the last transistor (T4) is it a steady voltage?

​@@RalphBaconthanks for the quick reply....the red LED is study and and I see .5V of Steady voltage at T4 Base. first I used digital pins to see the HIGH/LOW but it was continuously changing between HIGH & LOW then I switched to A0 to see what's happening. it is going like below, and I have changed T4's many and also tried using 2N2222 but no luck.. what is happening here? is there any way I can fix this.. I though this is a great hack to use in my project but not working only for me... 933 1023 1023 131 1016 1023 2 592 1020 1023

@@santh028 l have the same issue. Did you find a fix for this?

Whoops, forgot the rectifier/filter, which you still need.

BTW, why might you use an IC instead of discrete transistors? These are either dual-op amp (or dual Schmidt triggers, such as the SN74LVC2G1) or quad op amp (or even 6 Schmidt triggers, such at the 4016). Lots of inputs may be detected with a very compact circuit. Use care connecting multiple antennas to avoid crosstalk.

An interesting alternative that you suggest there, Richard; I especially like the idea of multiple inputs. Care to draw out a quick (but complete) circuit diagram and I'll mention it in a future video? Thanks for posting this anyway!

Hi Ralph, I decided that I shouldn't rely upon a 70 year's old 40 year-old memory (the last time I thought about this topic), so I prototyped my concept. It works, but not as a direct substitute for the transistor version. The comparator output isn't a digital signal. Well, it is, but it is either high (power detected), or 50/60 Hz pulses - I'm in the US, so 60 Hz is the name of the game for me. Thus, my Arduino code detects either a pulse train, or its absence. Why? Well, the "antenna" is just that, an antenna, and it picks up any stray powerline signal(s) in the general area, not just the signal in question. I used a thin coax cable, with the shield grounded only at the comparator, and allow the center conductor "antenna" to extend just a few inches beyond the at the far end, then to be wrapped around the target switched mains feed. The pulses are generated because stray signals, even with a shielded cable, exceed the comparator threshold when there is no local electrical field. You can see this by touching the oscilloscope probe tip with nothing else connected -- you are the antenna. However, when the power switch is energized, the stay signals are reduced by the effect of detected electrical field AND that the actual amplitude of this new induced signal has lower peak-to-peak excursions than that of stray signals. It actually feels like there is a little black magic involved. Counting pulses also acts a type of digital filter. If there are fewer than 25 in a period of one second (30 here), then something went haywire; ignore, reset the count and try again next second interval. This should avoid false positives, or in this case "negatives". I'd also monitor the analog value using a separate AD input of the Arduino (need more hardware on a Pi!), if there were no counts during a period of one second. If the analog value is not "high" but is low, something else went southwest. It is interesting to see the input signal on an oscilloscope and to come to a sort of understanding of what is going on. It isn't really intuitive, at least to me. Anyway, I can email you a video showing my results and a schematic, if you want. Your transistor design might be best, mounting the detection circuitry right next to the source, so that the antenna is very short -- as short as possible, so that stray signals are minimized. The output signal could be connected to a central Arduino via something like Ethernet cable, so that the circuit power could be furnished over the cable. My "design" would facilitate connection to the source via the coax (distance limits probably apply, TBD), while the circuitry could be near the Arduino itself. BTW, the components I used are a single op-amp (I used a BiMOS CA3040 - the modern part is CA3140 -single OA and CA3240 -dual OA. Because I had a 3040 in my parts drawer) and 3 resistors. Two resistors set the comparator trigger level (about 5 mV, I think) and the third terminates the coax at the op-amp. One might add a fourth resistor as positive feedback for the op-amp, if there are noise problems in practice. Pretty simple, but only investigation would tell if there is anything to be gained here. Just to be clear, I could not get this circuit to work with unshielded wire. Dick (www.hardandsoftware.net)

A very informative experiment, indeed, Dick. I'm thinking that if the input signal is sufficient to trigger the comparator without the real signal being present, then maybe either a series resistor may help, or a potential divider or a 1M resistor across the sensor wire and ground (think: x10 oscilloscope probe), with or without a 20pF capacitor in parallel. Just throwing ideas around. Your idea of shielded cable has been discussed in other comments; I'm surprised it didn't offer better shielding. Perhaps your circuit is just too sensitive as it stands (or you have rubbish shielded cable, gasp). I might try this later this week (I'm actually working on Monday with a 7am start!) The idea of counting pulses, or detecting a certain analog value was the very thing I was trying to avoid, actually. High or low is what is going to be the most reliable in this case, but you never know. I'll put my circuit through some further experiments regarding the above so I can compare. Thanks for doing this, I hope you had fun even if you did end up waving a dried seaweed root over your oscilloscope at one stage! Black magic indeed. Stand by for more experiment details but don't hold your breath.

I've powered the Nano from a standard 5V USB hub, David, but I'm powering the transistor circuit from the 3v3 pin. Check the voltage at the OUT point - it should be the same as VCC (3V3) with no mains being detected. It will drop to near zero when the mains is detected. If you have a scope you can check your DC supply. Try putting a 100Ω - 500Ω resistor in the VCC line to the circuit with a 100nF cap either side. Also add in a 47µ-100µF after the resistor (ie transistor side). That should clear any ripple.

Unfortunately not, Robert, as this will really only detect mains (240v) level AC voltage. With a speaker output you could use an optical isolator, perhaps.

Yes it will interfere, Nitin. This cannot differentiate between different live wires, only detect the presence. So if one wire is already live, switching on another will probably have no effect (it _might_ make the LED a tad brighter). I don't know what you mean by the 230v to 5v modules being "better or more reliable" than my circuit! In what way? You can run my circuit off 3 AA batteries, about as safe as you can get too. You'll have to clarify that question, Nitin.

@@RalphBacon Lets take another scenario, you have 3 wires, you want to detect if one wire is live or not by winding up a wire of your circuit and 2 other wire are also nearby. Now, if any wire from those 2 other wires goes live, even if the wire we are interested to detect doesn't go live the interference from that wire from those 2 other wires may be picked up by and we can have a false positive. Now, about the 230V to 5V modules, those modules have mini transformer along with other components all encapsulated in a plastic case with only 4 wires coming out, 2 for main inputs and 2 for 5V DC output. You can get such module for PCB for a rating of 200mA. If we use those after our relay or switch, that we are interested to detect if it is on or not or mains is available or not, we can use those module which will turned on when our mains/switch/relay is on, we will get a output of 5VDC. It does not interfere with other wires, and it can power another circuit if so need be.

The 5v SMPS units you refer to (which can deliver more than 200mA BTW, see my video #105 (ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-xodLuR6C8N4.html ) would require actual physical connection to do what you describe - the opposite to the thrust of this video! But yes, it would work. Another way to differentiate wires, if you have access to the wires, is to wrap the "antenna" part of this circuit around/along the wire you expect to go hot; that way others will not really affect it too much but should that wire suddenly carry 240V it will be very obvious.

@@RalphBacon Yes, I'm aware of the challenges. It is a interesting problem when it comes to reliable detection and being economical at the same time. 5V SMPS has it's own problems that it need to be protected by fuse, once it is fixed inside a panel, not easy to replace and then some constant power loss, though smaller but still some power loss of those module/units. I'm building a DIY 3-phase smart meter for my home, I can get currents reading reliably now if they are above 0.2A using a 70A Hall sensor (may replace it with lower current rating). So mains/live wire can also be detected by hall sensor and they also have minimum to no interference, they are cheap and constant loss of power is also less than smps modules. Of course, we have to clamp or pass through wire we are interested in. Perhaps, we can modify your circuit where it coils around wire. We can have more winding when wrapping it on the wire we want to detect, and we can insulate those winding from outside so it can minimize the magnetic flux from surrounding wire. Investigation of such kind of tape/material which can provide of insulation is needed.

I didn't see that behaviour at all, Nick, I must admit. I've just checked the video and I see I was powering the circuit from the Arduino itself which could be quite noisy - but maybe not as noisy as a SMPS. As well as smoothing caps, try putting in a 100nF cap too (in parallel) as that defeats high frequency noise on the power lines (and why we put such caps right next to the power pins of ICs and MCUs). Nope, I don't understand why a 10μF cap would not work for you. The upside of all this is that you researched and modified the circuit to get it to work for you, so kudos indeed!

I had a similar experience detecting 110v ac using 3.3vdc as the circuit supply. The 10uF cap didn't have enough energy storage to keep T4 on for the entire 60Hz cycle.

Most of the power here is drawn by the LED but even that might not be much, depends on the LED. Some consume only 2mA. Too much for a battery though, if continuously driven. You could use an ESP8266 that sent out a quick MQTT message every X seconds/minutes and then goes back to sleep. Others might have more experience of this so we will wait to see what they say.

Ok, maybe i Try it without The LEd. Thought a ESP32 has a Deep sleep Funktion, it should only wake up if The Front door Bell rings and Send mqtt. Should work Fine because The esp does Not use much power in Deep sleep.

Yes, the trick is to use a µController that allows deep sleep and only wakes up every X minutes. A doorbell is, perhaps, a little more time-critical (you don't want to be notified 5 seconds after someone has pressed the bell) but for mains detection, it doesn't matter too much. I'm making a note even as I type this to look into sleep modes for the ESP8266 / ESP32.

Ralph, I noticed that you have a capacitor between the antenna and the transistor. What's the value of this capacitor?

It should work even without the cap, but if you use one use something like a 10nF. Even a 1nF might work OK.

Same issue. Did anyone find a fix for this?

It will detect the hot (live) wire only as that is the only one carrying voltage. Whether a bulb is plugged in or not, I'm pretty sure, as that wire is generating an electromagnetic field around itself.

@@RalphBacon with 9014 and 4.7uF electrolitic cap it doesn't work

It's probably mains-borne interference (mains hum) making its way up the wire. Try running it off a battery of suitable voltage (eg 3 x AA) it will be good then.

@@RalphBacon It's no good mate as I intended to use the detection circuit as a sensor in an Attiny project so I can't rely on battery due. But your answer does explain why my other experiment failed. I tried creating a linear power supply as I assume the fault is i the high frequency switching in the switch mode power supply. I created it but it still gave false positive. I'll probably try something else like creating a switching circuit for it, sth to temporarily power the detection circuit for a while to charge the capacitor before measuring. Then open the circuit and let it run on the left-over juice from the capacitor for accurate measure. Thanks for the reply, mate! It's heartwarming to still got help from a video so long ago

It might work with a single package Darlington, I haven't tried, it will depend on the gain of the Darlington pair. How close it can be depends on that gain - try and see. Transistors are cheap and fast and hardware is sometimes better than software but it's your choice, try it and see!

I got my circuit down to this: electronics.stackexchange.com/questions/328897/using-240v-mains-on-off-as-input-to-microcontroller/429820#429820

Ah, I guess you must have missed my video #111 Mains Optoisolator which does pretty much what you describe here. The module I use there is available in single detectors or in a three-unit module like the one in the video. Regarding the cutting of wires to control them, that is absolutely true, but not the focus of this video! Thanks for posting.

Hmm, I'm not sure we cater generally for hardware failure conditions, do we? This is already a backup of my Sonoff/RasPi automation hub, so I get independent verification that things are doing what the computer is saying it is doing. So if this fails (seems a bit unlikely) I guess I will be "blind" for a bit until I realise it is not working. And, yes, there are many ways to do this and this is just one way, but as you say it's pretty basic and therefore simple. Thanks for posting K3n7, you raised an interesting point.

Good point. If you turn on the lights and this doesnt register, then you'd know about the fault, or that your light bulb has failed.

Ah no, it doesn't matter whether the light bulb has failed or not, as this detects the _presence_ of electricity, not the _flow_ of electricity. Yes, if I turn the light switch on and nothing happens the bulb may have blown but the phone app (or whatever I end up using) will say that the lights are on regardless. That's what I like about this circuit!

If you have a live wire going to a lamp but the lamp is switched off then the neutral is not carrying any current. Switch the lamp on and the neutral also becomes "hot" and this circuit should detect it. Have a go, let me know!

@@RalphBacon Thanks for replying. I wish it were that easy. I have switched the lamp on and still the same issue. I made a video to show you what's happening with it: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-k1MrWLVsumo.html

Your video described it perfectly. You are supplying the lamp with both Line and Neutral. Once the current hits the neutral wire it cannot be detected (otherwise it too would be "hot"). You won't get a shock by touching the neutral wire (but don't try it) only the line "hot" wire. Neutral is, in fact, connected to earth either at your premises or in the street somewhere. Make sense?

Hmm, that's quite a deviation from my circuit design. I use silicon NPN transistors, you're using germanium PNP transistors! I shall to say to you what I say to all my viewers who can't get a circuit or sketch working: follow my example *exactly* before customising it. You know my version works, so there's your answer. Good luck (not that you will need it, believe me!)

@@RalphBacon ah ok thx I'll work on it. As you can tell I'm not very good at the electronics part. I'm ok with cobbling together Arduino and sensors but lack the experience with discrete electronics. Thank you for answering so quick. I'll have to find those transistor you used.

@@RalphBacon i'm looking at the datasheet for the 2n2904 and it says its a PNP transistor. ?? did you mean 2N3904 ?

My bad. The circuit does say 2N2903 (or any similar *NPN* transistor) so it should indeed say 2N3904. I was probably getting carried away with the 2N22222 which I use all the time. Cheap as chips too. Sorry about that I shall make a note in the video description.

@@RalphBacon I'm still working on getting it right. I tried with the bc547 but still couldn't get it going. The led lights, but having trouble at the out signal stage. Anyway is a good learning experience for me. I haven't really done much discrete electronics before. Will order all the components you used and try again until I get it right. Thanks for your videos. BTW hope your eyes get better. My friend had the same thing happen to him both eyes within a year of each other.

It's the alternating current that makes this work, Jorge, so it won't work with DC, I'm afraid.

If you touch the base pin of the BC547 does it light up the LED? If not, perhaps it does not have the required high gain for this circuit? Build the circuit starting with just one transistor and work up in stages to see when it stops working.

@@RalphBacon Hi, thanks so much for your answer. My LED is on most of the time, seems to catch up noise, I have ordered the 2n2904 so I plan to wait. The gain math is one or three levels over my skills. I will use the circuit for my pond pump and and well pump and send cycles to my Home Assistant.

That means, Flemming, that you would be wiring up a USB supply which doesn't qualify as "non-contact". And, quite frankly, after my video #90, I don't fancy wiring up any more cheap, Far Eastern wall warts which have this tendency to burst into flames! Especially as this (in my case) is supposed to be "behind the scenes", that is, buried in the wall cavity, or similar, of my (wooden) workshop. Additionally, as several posters here have mentioned, it should be possible to have several of these mini-sensors connected to a single Arduino, possibly via screened cable. That's not to say your method would not work, (of course, it would) but it just doesn't seem elegant enough for my workshop automation project. So you didn't miss anything obvious, but it just doesn't fit into my plans at this time. But you raised a very good point so thanks for posting.

I don't buy 220V connected stuff outside Europe - nobody should. A USB PSU from an old phone, ignored for it's ability to only deliver 500mA, would do the trick safely. However, I do see your point. And yes, I'm probably more pragmatic than elegant ;-)

From what I can recall, Jim, and I've just looked at the schematic, a normal electrolytic capacitor of 1μF - 10μF with a voltage rating of about 10v (you can go higher but it will get bigger) will suffice.

@@RalphBacon thank you

I'm not sure a mad tranny from Down Under is going to help detect mains voltages but any Quad Transistor Array chip (such as a TPQ3904) should work, assuming the gain is sufficient (around 200 - 400 per transistor). Google might be your best friend here, dammit, Janet!

thanks, i don't even know how to search for these stuff :)

Yup, that was pointed out already (will you all STOP trying to earn your *EagleEyes* awards please?) but I'm hoping the context and transistor symbol points to NPN-type transistors. Thanks for posting JH, good to hear from you.

I'll have a read, thanks Rupert.

You can try reducing the sensitivity by adding a resistor between the base of the collector and ground. Try 100K first.

See the video description, all details that I have are in there!

@@RalphBacon oh! I misunderstood, I thought you built that "little device". Now I get it. thanks!!

They use split current coils, of the type linked to by whitefields5595 (as per my demo) and they generate a voltage proportional to the current flow; I don't need this, just the presence of mains is enough for me, and this circuit works whether current is flowing or not. (I, too, have a current clamp meter and it works pretty well but just not what I need here.) But if you need to measure current flow that is certainly one way to do it.

I must say that I learned something here. First I apologise for thinking you might not know of current clamp meters. My problem was that I could only picture a "no contact" sensor as being inductively coupled. It didn't occur to me that capacitive coupling would work. Of course your setup couldn't have been inductively coupled because only one end of the antenna (aerial) was connected to the sensing circuit. I've been out of the business too long.

No apology of any kind required, captxrox, I didn't take any offence, we're all friends here. I'm just glad this circuit seems to provide a way forward for my "backup" verification for my workshop automation project!

Heavens be praised, I've been set free! A huge burden has been lifted off my shoulders. Thanks for that whitefields5595, nice to hear from you, hope you didn't mind me quoting you!

Ralph, for homework I'd like you to compare and contrast the hardware (charge pump) and software methods for determining mains presence at your final stage. Some of the baying mob out there are going to get solder anxiety so it would be useful to see the pros and cons of "doing it in software" vs hardware. Also you are still going to find it a tight fit in places so the 4 transistor DIL chip would be good so you could make it long on thin for tight spaces ..... This would then be a good project taking first principles of non-contact, then HW vs SW rationalisation and finally miniturisation/ component reduction / packaging. Maybe use the ATTiny or similar?

See the responses from Richard Grier - I think that qualifies as getting my homework done by someone else!

Ralph, homework cannot be outsourced. Richard is proposing a different method of INPUT detection using an op amp. Your homework is to compare and contrast conditioning the OUTPUT signal using the binary nature of your charge pump vs (quasi-analog) pulse train counting in software. There is always a decision to be made between hardware vs software that would add a richness to your projects .....

Hang on, BaronVonBiffo, let me count, 1..2..3..4..5..er, hang one where did I get to, 4...5...5...8, no ...7... er... Oh WTH let's call it 10, we're all friends here. What? Oh alright, it's eleven then. 🤓 In my defence, I had to "count" whilst I was filming the video so it wasn't a bad estimate, hey? Heh heh, thanks for the correction. And yes, it's quite a nice circuit, cheap AND safe!

Oh dear, did I say 2N2904, they were, in fact, a *2N3904* or any other general purpose *NPN* transistors, such as 2N2222A or BC327. The link I put in the video description is for a selection of NPN and PNP transistors, very cheap and useful. And I have now painstakingly resorted them all back into their correct locations, took about 15 minutes. Now I'm _very_ careful when opening the lid. Thanks for correcting this. I once did a course on valves (or tubes, as our American cousins might call them). They seemed even more dangerous than all the mains projects I've ever tackled! And hot! And big! Yet nowadays I yearn for a retro mini-tube amplifier, with exposed tubes, they just seem so cute! But far too expensive. Patreon? Hmm. I've read their new terms and conditions. They have increased their overall markup percentage for small amounts so that my original plan (if I was going to do it at all) of a $1 monthly donation would lose 5c to fees but also about 10c to payment processing fees. Plus a fee when money is transferred to my bank account. Not good at all for small, affordable donations. So to make it even vaguely viable I'd have to ask for $1.50 or even $2 which suddenly subscribers would notice more. I'm not sure about this at all until I have twice the subscriber base I have now. But never say never, as James Bond might say! Thanks for the encouragement, John, we shall see how things pan out over the next 6 months or so.

That is good about the 2N3904 as I have some and 2N2222 in my selection box.

Let me know how you get on, John!

Just quoting your post, Martin: "Darlington input must be ... for the circuit to work. very very unlikely!" But surely the video proves it does work? I don't fully understand your objections or concerns, care to elucidate?

Um, you may be right. Is it important? Did I say something indicating otherwise (oops)?

Ralph S Bacon No no you did’nt. Being Indian I surmised from the spelling of the name.