Introduction
I needed to upgrade a spare room at home to convert it into an office. The ceiling has some low-voltage halogen lights fitted which had gradually failed over the years (more to do with the heat damaging the connections, rather than the bulbs failing).
I wished to upgrade them to LED lights, and replace the low-voltage power supplies too (replacing with fancy MeanWell supplies). For reasons I won’t get into, I wished to trace the wiring rather than replace all of it, but it was very confusing figuring out what was going on at the other side of the ceiling! I only had access to the ends of the wires, poking out of the holes where the old lighting fixtures used to be. A mains cable locator is useless here because the cables (they are low voltage) are not powered up.
I wanted a device that could clip onto a wire so that I could almost ‘buzz-out’ the wire at the far end. The device needed to be quick to assemble. I could have just used an LED and battery, but I figured surely there must be a slightly better way. After some thinking and hacking, this project was realized. I’m sure it’s not original, there are probably far better commercial versions out there.
Please note that the tool described in this project must not be relied upon, it won’t meet the standards of any commercial device. It should only be constructed and used at your own risk, for locating wires that are safe and where there is no risk of harm if the wrong wire is located or if no wire is located.
Anyway, this is my attempt at a crude radio-based wire locating tool – see the 40-second video for a demo.
How is it Used? How does it Work?
The project described here consists of a box containing a small transmitter circuit and a 9V battery. The only connection from it is a single alligator clip. The clip is attached to one end of the wire to be located.
Next, a pocket radio receiver is tuned to a certain frequency (540 kHz on the AM broadcast band) and the radio is moved around near the far end wires. Once you’re close to the correct wire, you’ll hear a tone. The wire has now been identified!
Internally the WirePulser contains an RF oscillator which is gated by a low frequency (350 Hz) astable multivibrator. Together this creates an AM modulated output which can be heard as a 350 Hz tone on a radio receiver placed close to the wire under test.
Is it Legal?
The power output is very low; a pocket radio cannot pick up the transmission unless it is within a few feet. Beyond this distance, any commercial radio station received on the AM band will be more powerful. I believe it could be legal (common household equipment can emanate more RF power than this project), but please check the legislation in the country it is to be used in. The circuit as it stands cannot be tweaked for significantly higher output. It would require additional circuitry for that.
Circuit Details
The two building blocks can be seen in the circuit diagram (click to enlarge). The right portion forms the RF oscillator (it is a topology known as a Colpitts oscillator), and the left portion contains the ICM7555 astable multivibrator to gate the oscillator.
The oscillator circuit was mainly designed by trial-and-error; there isn’t (as far as I’m aware!) any easy way to know for sure how such an oscillator will perform, other than trying different values. Some things are known, like the approximate frequency, and that a 1:10 ratio for C1 and C2 is a reasonable start. I played with values in a simulator until it sort-of worked (often it can hang a simulator! Also, sometimes one of the components needs a ‘kick’ to start off the oscillation; I did that by setting an ‘initial condition’ for one of the components, in my case I specified that component C3 had 0.1 V across it at the start of the simulation). When I tried the circuit for real it didn’t work the first time, because I used different ferrite cores that were not designed for 540 kHz region operation, and the inductance was not what I’d (mistakenly) expected. After I swapped out for FT37-43 ferrite cores, all was well. Type 43 cores produce a very flat, stable inductance through to 540 kHz and beyond to at least a couple of MHz.
There are three wound components; two identical inductors L1 and L2, and one auto-transformer T1. All three are built using FT37-43 ferrite cores which are low-cost (about $0.30 each) but unfortunately not available from any major large distributors. However there are plenty of smaller suppliers for these globally; they are often used in ham radio projects, so ham radio suppliers will stock them. Personally, I would avoid Aliexpress for the ferrite cores, because visually it is impossible to tell if you’ve been supplied the correct core or a different one. It would be difficult to troubleshoot if you were not sure if the core was correct or not. A trusted supplier should be used.
A short length (no more than a meter) of 0.2 mm and 0.3 mm enameled wire is used in total. The wire thickness does not need to be precise, and as long as the wire has that approximate thickness, that is good enough. Wire removed from a toy motor or a small transformer could be used.
The number of turns needed is described in the diagram above.
To wind the inductors L1 and L2, arrange 17 turns such that they cover the entire ferrite, with a few millimeters of a gap at the ends.
Incidentally, I used a BF256C JFET to construct the circuit since that was what I had at home, however, a BF256BBF256B JFET is much more easily available from distributors. Unfortunately, I didn’t have a BF256BBF256B JFET to test with, but it should work.
The oscillator building block was constructed first (everything between the points labeled A and B in the circuit diagram shown earlier) to test it in isolation. To simplify the construction, some surface-mount parts were soldered to the underside of the proto-board, but through-hole components can be used if preferred.
The point labeled B was temporarily connected to 0V to enable the oscillator, and the signal at point A was measured with an oscilloscope (using a X10 10 Mohm probe). If no oscilloscope is available, a radio can be used instead! Just tune around 540 kHz, place it very close to the circuit and listen out for a position on the radio dial where the background noise fades into silence. You’re now listening to the result of the AM carrier transmission. If you can’t hear it, the circuit needs to be debugged. Once it works, the rest of the circuit can be constructed.
If the frequency needs to be adjusted slightly, then the winding on inductor L1 can be squeezed or stretched a bit. It will allow for a slight frequency alteration. The ‘scope trace above shows the oscillation at point A, it happened to fall on 540 kHz without squeezing/stretching the windings. Incidentally, if you can use C0G/NP0 capacitors for C1 and C2, then the frequency should be quite stable.
Now disconnect point B from 0V and continue to build the ICM7555 portion of the circuit. The ICM7555 is used to generate a square wave which will gate the oscillator. If you use a 555 chip instead of the ICM7555 then some tweaks may be needed, I didn’t test with a 555.
The tone can be adjusted by altering the value of resistor R5 (smaller value results in a higher tone).
Finally, the circuit to the right of point A can be constructed. It is an attempt to better match the random wire to be located, to the oscillator output.
To wind the 48-turn auto-transformer, I found it easier to do by folding the length of wire in half, and winding 24 turns starting from the center of the wire, and then winding 24 turns using the other half of the wire. A tap needs to be brought out while winding it, 8 turns from one end of the transformer. The total number of turns is not critical, it can be more or less than 48, however, the tap should be at 8 turns.
After powering up, the output connection J1 was examined with an oscilloscope (again with a X10 probe).
Trying it out
The project was briefly tested with a short length of wire, and then I tried it with a 20-meter length of Ethernet cable.
Initial tests were good, so then I tried it on the lighting wiring in the room, by clipping the alligator clip to one of the low voltage wires (with the power disconnected). I was able to successfully identify that the wiring was split up into two banks of lighting using the WirePulser tool! The earlier video shows it in action.
Summary
A very basic wire locating tool was constructed, based on a weak RF signal transmission principle. Although it works for basic DIY purposes, it was later discovered that commercial tools exist for this purpose too. They are likely more accurate and have more features.
The design here could be improved further. One easy improvement would be to add a second ICM7555 chip to produce a more interesting tone sound (perhaps with a gap, or two alternating tones or a sweep, because a single constant tone might get annoying!
Thanks for reading.
Top Comments