Hi everyone,

First of all, this is a great forum - I've picked up a lot of handy bits of info just sitting in the background. But I'm having trouble with this idea:

What I'm trying to make is a non-contact voltage detector. It simply just detects the presence of voltage or not - no need for level indication or anything special like that. Working with 230v in Aus (50Hz). Fluke and heaps of other manufacturers make them in a pencil sort of thing, but I'm looking to integrate this into my project, and this seems the safest way to do it.

I have thought of other ways to do it, such as
- a plug in 230v coil relay that closes signal contacts (but I can't always get access to the end of the cable, and if its hard-wired to an appliance then I can't do anything with that anyway)
- inductive current sensor (but i need to detect voltage, not current because there may not always be a load drawing current)

So far I have tried a floating input into a PIC (16F628A), a hall effect sensor into the PIC, and a (dodgy) inductive coil sort of thing. These have SORT OF worked, but nothing reliable enough. This will be used for AN INDICATION only - so I'm looking for something fairly reliable, nothing that you'd bet your life on though.

Has anyone worked on something like this, or got some good information, because from what I have seen, this 'voltage-detection' seems somewhat elusive.

Thanks for your time,
Tom.

 

I can visualize an op-amp circuit wired as a precision rectifier. That will give a DC output to run an LED. The input has a variable resistor to adjust the sensitivity and a wire to a flat piece of metal. We all know that power mains hum is all around us, so dialing that background hum down with a variable resistor would leave the circuit sensitive enough to perk up when the flat plate sensor is held against the insulation of a live wire.

 

Thanks Bychon,

Analogue electronics is definitely not my strong point, so I have alot of reading to do!

I've been reading this site http://sound.westhost.com/appnotes/an001.htm and I'm thinking the Half Wave Precision Rectifier would probably get me A signal to use that the PIC with interpret. But they don't specify the op-amp, so if its not too hard, could you please suggest an op-amp that I could start with?
Thanks.

 

Just any piece of drek op-amp will do this if you give it both a positive and a negative power supply. Common numbers are 741, TL081, OP-07.

You will need a diode in series with the input to force a threshold voltage that the hum has to exceede to get into the amp. Try 1n4148. Make the 10k resistor 100k, and variable. But why use a microcontroller to interpret "on" and "off"? Just put a resistor and an LED on the output of the op-amp. If you need a microcontroller to interpret, "the light is on" and "the light is off" you can add that later.

 

Just any piece of drek op-amp will do this if you give it both a positive and a negative power supply. Common numbers are 741, TL081, OP-07.

My circuit is based on 7805 5v fixed reg from a 9V battery, do I NEED a minus 5v rail as well?

 

If you find an op-amp that will respond to signals at barely above zero volts, you don't need a negative supply. You could also make a "fake" center voltage with 2 resistors and use that as ground for your signal sensing.

 

You could also make a "fake" center voltage with 2 resistors and use that as ground for your signal sensing.

Sort of a resistor divider network/ladder thing? I think that might be a better option (if it works well).

 

...2 resistors...the only possible circuits you can make with them are series or parallel. Which circuit will have a voltage besides the voltages you started with?

It's time to get your solding iron warmed up.

 

The input to a 'volt stick' style probe is typically a FET with an (almost) floating gate.
The wire from the gate lead is in the insulated tip of the probe.

The fet needs a resistor to stop the gate floating and building up a DC charge, but this should be something like 10M to 100MOhms to the fet's source (which is ground).

I'd then try a fairly high value drain resistor to V+ and a simple charge pump rectifier from that. (ie. cap to middle of two series diodes, anode of one grounded, cathode of the other is output).

 

Righto, I'll start trying some things with the stuff I've got here already. Thanks Bychon.

Another good idea thanks rjenkins, I'll look into them too. Thanks again.

 

You can sense the electric filed from a conductor. if its an AC voltage, then you merely need an op amp with a capacitor between the output and the inverting input, and a low-passed DC feed back path to stabilize the operating point. A small capacitive plate acts as an atenna to chance the electric field to an AC curretn on the invering input, which is then changed to a volatage that is linearly proprotional to the sense voltage (providing the opamp can keep up with it).


If it is a DC voltage, then use the same circuit but also include a rotating, grounded, conductive disk with a hole cut into it. The disk is a shield to prevent the capacitive pickup from "seeing the Dc voltage being sensed, except when the hole comes around. This arrangement, called a "field mill" chops the DC into pulsating DC, which will register on the output of the opamp.

This web page is difficult to read because of its busy bacround, but the principle is explained well.

 

This web page is difficult to read because of its busy bacround, but the principle is explained well

Which page?

 

So far I'm getting pretty good results with a FET with a floating gate and a comparator to turn it into a digital signal.
Thanks for all your suggestions.
Tom.