I’m surprised by how well these circuits work. They’re not doing anything clever like heterodyning high frequencies down into the audio range. This circuit will only resolve signals up to 20kHz which excludes many signals of interest such as Wi-Fi and Bluetooth up in the Gigahertz range.

Obviously there will be a lot of mains hum at 50Hz or 60Hz which I would have expected to swamp other signals but surprisingly there are plenty of interesting sounds from probing things such as routers, mobile phones and PCs.

It’s a very straightforward circuit which I haven’t modified much except tweaking some component values.

R6 and R7 are high value resistors in a potential divider creating a virtual ground at half the supply voltage. Electrolytics C6 and C7 smooth the power supply. The other electrolytic capacitors are simply DC blockers/high-pass filters.

I added a 100n decoupler to the virtual ground to stabilise it which is probably not really necessary as clean audio is not important in a circuit whose whole purpose is noise.

The inductors, which are simply sealed coils of wire, act as pickup coils. The circuit detects the magnetic field component of any EMF radiation so the more turns in a coil the greater the induced voltage and therefore the greater the sensitivity. The inductance of a coil also increases with the number of turns so the higher the inductance the better. Commonly available inductances seem to max out at 100mH so that’s what I used.

The op-amps are wired in a standard configuration with the non-inverting inputs tied to the virtual ground and feedback resistors from the outputs back to the inverting inputs. The gain in this arrangement is simply the feedback resistance divided by the input resistance.

Initially I tried fixed resistors for a fixed gain but some signal sources were uncomfortably loud so I added variable resistors in series with the fixed resistors to give me variable gains. Some of the circuits I’d seen on the web used variable resistors of 500k which are uncommon and seemed unnecessarily high to me so I used common 47k ones. The gain is set by the ratio of two resistances so absolute values aren’t important.

A dual-gang variable resistor would have been more elegant than independent left and right volume controls but dual-gang PCB-mounted presets are not common.

Designs I found on the web are stereo which didn’t seem really necessary. But dual op-amps are cheap enough (and seemingly no more expensive the single op-amps) so there was no real point in not using one. The only additional cost is a second set of passive resistors and capacitors and a second inductor (about 10p/13c on AliExpress). Stereo jack sockets are more available than mono ones which makes sense as you can wire a stereo socket as mono anyway. Both coils will pick up basically the same signal but you do get a panning effect as the detector is waved in front of a signal source.

The LED is just a power indicator but adds a little visual interest.

Two sets of components leads to a nice symmetrical PCB layout (right down to symmetrical vias and traces on the top layer) which I find attractive.

The 9V PP3 battery fits neatly underneath the PCB. This does introduce the possibility of shorting out components on the back of the PCB so I used a self-adhesive Velcro pad on the PCB as an insulator.

Yellow silk screen on black resist is a combination I haven’t tried before and is satisfyingly garish.

Sample recording from a mobile phone.

Many thanks to PCBWay for sponsoring the boards for this project. Fast, efficient service with high-quality boards and affordable shipping options.