Adding Mains Adapter to Doorbell

I have an Arlec DC371 "wireless door chime." It works fine and is very handy as I can clip the chime to my belt and wander out into the garden.

The downside is that the chime part is powered by two AA batteries which do not last very long at all. The manual suggests that a set of batteries should last for a year, while I find that a month is more accurate.

Therefore, I wanted to modify the chime so that it could be powered from a mains supply most of the time, but be disconnected and automagically switch to its internal batteries while I carried it around.

This circuit uses a switched jack to disconnect the ground for the internal battery pack when the plug is inserted.

The PSU used for the project was an old Nokia phone charger which provided 6V at up to 400mA.

To drop the 6V supplied by the PSU down to the 3.2V could be done by designing a circuit around a PSU controller such as a TL499 or some kind of variable voltage regulator like the LM317. But since I had a drawerfull of signal diodes and only needed a small amount of current, I decided to daisychain a bunch of diodes.

The datasheet for the diodes showed they could handle 100mA constant, with transient peaks of 500mA which seemed like plenty for running what is basically a very low power radio receiver with occasional audio output.

Silicon diodes should drop about 0.7V, so four of them would equate to a drop of 2.8V which should be good enough for the purpose. I would draw a circuit schematic, but I'm having problems with ExpressSCH at the moment, so hopefully you'll be able to follow the design from the description.

AA batteries have a nominal output voltage of 1.5V.

The Nokia charger which I used had a nominal output voltage of 5.7V.

Silicon diodes have a nominal forward voltage of 0.7 volts.

In the first picture, I am measuring the voltage at the output of the battery pack for the chime, and in the second I am measuring the voltage at the output of the phone charger.

We have six volts, we need three point two volts. If we put four silicon diodes in series then their forward voltages will add up to two point eight volts, leaving everything hunky-dory.

I cobbled together a breadboard circuit with a small load resistor and four diodes and the drop across the diodes is two point six volts, which is good enough.

By lucky chance, there was space in the chime case right next to the battery compartment.

Since the switched jack socket disconnects the outer (sleeve) voltage, and since that is generally the ground, we disconnect the ground from the battery pack where it reaches the circuit board. We then route that through the shunt terminal on the socket. The ground from the socket sleeve connector and the positive voltage from the tip connector are connected to the circuit board.

When the plug is not inserted into the socket, then the ground from the batteries is connected to the sleeve of the socket and the current flows from the batteries to power the chime.

When the plug is inserted into the socket, then it disconnects the shunt terminal and the battery ground is no longer connected to the circuit board and the battery is isolated from the circuit.

I drilled a hole in the case with an 8mm bit which was just big enough to allow the barrel of the socket to be fitted through. The three terminals were connected, with heat-shrink tubing being used to cover all the exposed conductors.

After this was done, the case was re-assembled and the batteries fitter. To test that nothing had gone wrong, the chime was activated with the batteries in place and it worked fine. Then the plug (not yet connected to anything) was inserted into the socket and the chime was tried again when nothing happened. This showed that the batteries were being isolated when the plug was inserted.

To keep the phone charger functioning as a charger, I retained the end of the charger cable which had been clipped off and soldered it in to give a Y-cable which can operate either a phone or a doorbell.

I wired up the doorbell plug with the tip positive and covered all the terminals with heat shrink, also running some heat-shrink over the end of the plug strain relief (not pictured here, but see later) to give a bit more physical strength.

I found a scrap of VeroBoard to hold the diodes. I picked a piece which was a bit smaller than the widest heat-shrink which I had so that I could cover everything when it was finished.

The diodes were soldered in place and then the wires were attached. The Nokia plug was connected at the input end, and the ground for the chime was connected there as well. The positive wire was connected at the end of the chain of diodes.

Once this was tested and working I covered the sharp (soldered) side of the board with a scrap of heatshrink, just resting in place. I added a couple of cable-ties to the wires at each end of the board to provide a physical grip and then put heatshrink over the whole lot.

Finally, I added cable ties over the outside to close off the heatshrink where it hadn't shrunk enough and to grip onto the cable-ties which were underneath it.

This worked great, and I don't have to worry about the batteries going flat and not hearing trick-or-treaters at the door.

When I drilled the hole for the socket, I was so busy avoiding the PCB that I forgot to make sure that its position would leave room for the terminals. There was not room for the terminals. Fortunately, the plastic forming the battery compartment was pretty thick, so once that was pared away with a chisel the socket fitted and there was still enough room for the batteries.

The other problem was that the ground wire from the battery compartment was only just long enough to reach the socket. That was soldered in place without checking, and it was sheer blind luck that it was _just_ long enough to reach if it had been an eighth of an inch shorter, I would have had to splice in some more wire, which would have been ugly. Always check your cable runs before you terminate!

This is going to save me a fortune in AA batteries, and the same technique can be used in a lot of other places too. Dropping voltages with diodes is really ugly, but it's easier than using a regulator and its good enough for very low current applications.