Archived Forum

http://www.harizanov.com and look for Detecting AC.

Excellent, thanks. I really like the idea of just sticking a coil of wire on a resistor... very clever

I think you need to look a little closer it's a small neon and an LDR, built one myself, works well.

There's two solutions on the the harizanov.com site - one is an LDR and neon (which involves tapping into the wiring), the other one is a coil and resistor into the ADC. The coil is less developed but could work.
I'd be interested in seeing if there are any refined non-contact voltage detectors available. It would be handy for sensing things like valve position in a heating system without touching the wiring.

Another option is to use an airflow switch, but you would have to be able to install in the duct. Current monitoring better if you can make it work.

As I wired the loft myself, I'm not afraid to break into the wire if I have to, but a non-contact method would be preferred, so there's no issue with curious fingers getting fried. 

This topic seems to confirm the concept... I'll report back once I've given it a try.

Hi Schism,

I was in the same situation than you and I decide to solve it with an easy, safety and cheaper solution :

Buy a power supply for mobile telephone. means 220V input (use in parallel of the fan input)  and USB output will give you a stabilized 5V  (directly understandable by the emonTx digital input, when the fan will be supply)

the circuit is already protect against all possible damage, reverse polarity, thermal dissipation or transient , so really safety, and cost 5€.

good luck.

Thanks, that's also a great idea. It would be easy to wire in a single socket (although it would need to be clearly labelled in case anyone later tried to plug a hoover or kettle and blew the lighting circuit!)

OK, I just put a dozen turns of wire around a lamp cable, into the emonTH analogue pin, and used a very weak pull-down (2x10M resistors to ground).

Easily detects when the mains line is energised:

ON :  507
ON :  507
ON :  508
ON :  509
ON :  509
ON :  510
ON :  509
OFF:  28
OFF:  28
OFF:  29
OFF:  26
OFF:  28
OFF:  29
ON :  510
ON :  509
ON :  509
ON :  506
OFF:  29
OFF:  28
OFF:  30
OFF:  29
OFF:  34

I missed the coil/resistor method so humble apologies but I'm glad you corrected me as I have a situation that could use that solution, thanks.

The only unfortunate thing is that it's using the only analog pin out on my emonTH, which I think will interfere with D18B20 operations.

I need to experiment and see if a longer coil of wire will increase the 'on' value enough for IRQ0...

 I've almost managed to get this working on IRQ0, despite the maximum reading I get on the analogue pin being ~500. Unfortunately it misses a transition every few times, which is ultimately no good.

I've tried different lengths / coils of wire, but I don't understand the physical principle. Is there a way to adjust the threshold on the digital pin? (Google not encouraging)... or any other ideas?

How about re-purposing something like this to drive a digital input rather than lighting the led?

http://www.ebay.co.uk/itm/Led-Electric-Voltage-Tester-Detector-Sensor-Pen-Alert-Stick-Volt-Sensor-90-1000V-/141133522580?pt=UK_BOI_Electrical_Test_Measurement_Equipment_ET&hash=item20dc36a694

Schism: Is this http://harizanov.com/2013/08/non-contact-ac-detection/ what you don't understand?  It appears to be relying on stray capacitance between the live conductor and the pickup wire. The 1 MΩ resistor is actually reducing the sensitivity, but it's probably necessary to lower the probability of false detection.

To make it more sensitive, the easiest way is to put the "aerial" in close proximity to the live conductor, and put as much length as possible parallel to it, so as to increase the capacitance.

That's the one, Robert. Thanks for the advice - I've got 20MΩ as a pull down, about as weak as it gets!

I'll try a longer cable and tape it on parallel to the mains and let you know the result.

thanks,