Bike alarm

This is a big page - please give it time to load. All photos are thumbnails - click on them for a larger view.

Keeping your machine secure when out and about is always a worry. I use the dutch lock, or in very high risk areas, a heavy duty shackle, but these only protect the bike, not the stuff attached to it. For a long time I had thought about building some kind of paging alarm, where I could carry a small device with me that would bleep when the bike was disturbed - I envisaged some kind of radio link. After some rascal made a concerted, but unsuccessful effort to nick my front lights whilst I was quaffing beer in the local, the idle daydreams took on a new urgency. As I saw the problem, there were several solutions.

These ideas rattled around. I bought some 418Mhz modules to experiment with - the range and reliability were terrible. The FM idea was a non-starter too; I'd fiddled around with those as a youth, and the stability was poor, and the resultant output would stick out like the radio equivalent of a foghorn. I became aware of 446 PMR - personal mobile radio - some time before. It's a totally legal system, marketed for the leisure and outdoors markets, consisting of hand-held 2-way radios operating on the 446Mhz waveband. Each unit has a fixed antenna, and a quoted range in the region of 2km. The units generally have 8 base channels, and further 38 pseudo-channels are facilitated by CTSS, taking the effective channel count to 304. To find out more about PMR 446 radio from a real enthusiast(!), check out Delboy's PMR page.

It was while rooting around at my annual visit to the Midland Amateur Radio Society's annual rally, at Drayton Manor Park in Tamworth, Staffordshire, that I found what I was looking for. Radio Rallies are huge electronic jumble sales, where everyone from recognised traders to car booters get together and do business. Wandering through a marquee, I spotted a stall selling PMR sets - some were shop soiled, some were repaired guarantee returns, some were boxed as new. Spotting my opportunity, I haggled with the trader and bought a pair of Binatone RoamAround 100's for, I think, £15. I know it wasn't much. The seller even demonstrated that they worked for me.


Drayton Manor Radio Rally: no prizes for spotting anoraks... if you can't find a bargain here, you haven't got a pulse.	It's amazing the misfits you come across at these events. Clive Roberts admires Terry Basford's bulging sack, while Roger Mosedale amuses himself.	The Binatone RoamAround 100: cheap at twice the price... note the orange 'call' button top right.

When I got home, I started to have a play with my acquisitions. Having no manual, and being unable to find one on the net, I played around. I found that the radios could be set up to monitor their channel for activity, and would beep if none were found, whilst at the same time transmitting periodic bursts of signal on that channel. This has the effect that if a pair are set up the same, they beep when out of range of each other - I could use that feature to detect if the alarm had been disabled, by for instance, having the power source disconnected. I also noted the presence of a 'call' button, which, when pressed, broadcast a few seconds of high-pitched sound. That would be my alarm signal, all I had to do was to interface my vibration detection device to the call button, and I was sorted.

The idea I had was to use a mercury vibration switch - unlike a tilt switch, which operates on moving the switch from one angle to another, the vibration switch I had in mind made contact briefly whenever it was subject to a certain amount of vibration. Now the pulse this gave would not be long enough to be useful without a bit more circuitry. I needed to extend the signal it gave, in order to cause an alarm signal to be transmitted to the other radio. To do this, I used a timing chip - the NE555N timer. This can be used to produce a single output pulse from a much shorter input pulse. The 555, in turn, would operate a relay, which could be wired to mimic the action of the call button.

Taking the unit apart, I discovered that the 'call' button was actually bridging 2 lines to a common ground. It was going to be hard to solder wires onto the small connections available. Leaving those concerns aside, I designed the following circuit...

(Clicking on the diagram will open a larger, more legible version in a separate window)

I notice now that the 10k resistor between trigger and 6V is rendered completely redundant by the sense LED - It's the kind of thing you only notice afterwards...

The supply would come from the 6 volt lighting battery on the back of the bike, and this presented a problem - the radio takes 3 'AA' cells meaning it runs on 4.5 volts. Since I wanted a self contained solution to the problem, I didn't want to have to keep taking the alarm apart to change batteries. The solution was to use a couple of diodes to drop the 6V line for the radio. This would approximate to 4.5V, or even a bit over, which wouldn't hurt too much. The timer was designed to extend the vibration switch output to about 11 seconds - I had already found that pressing 'call' broadcast for 5 seconds after the button was released, so a total tone output of around 16 seconds seemed about right. Within the circuit I incorporated indicator LED's for arm (on), sense (switch activity) and transmit (for the signal to the radio) as well as a handy loop for the luggage - more on that later. I ordered the parts, and when they came, commenced construction.


The RoamAround 100 runs from 3 'AA' Batteries sitting in a compartment in the rear of the device	Taking the top off, one can see that the buttons operate by shorting conductive rubber pads over fingers on the main PCB	Removing the circuit board assembly from the rear housing revealed that the batteries were connected by springs from the rear housing	The call button shorts two lines on the PCB to ground (highlighted in the larger picture)

The rear of the case. The battery compartment is facing downwards, the main connections being made by the spring contacts on the back.	All the components ready for assembly. I used stripboard for the timer circuit, far easier than making a custom PCB. Note the polycarbonate housing the alarm was to be housed in	The vibration switch. One must be careful not to get it too hot when soldering. The leads on the device were rather heavy gauge and hard to solder	Assembly and testing of the timer took about an hour. For the time being I was running it off a mains DC power supply. I bodged the LED's in order to check the circuit status

Assembly of the timer circuit was routine stuff. I used stripboard for least hassle, and took care to allow enough space for external jumpers. Once the board was ready, I tested it using a 6V supply from an old mains adaptor, and it gave the results I wanted. The next task was to get the connections I needed out of the radio. I had ordered a housing of a reasonable size so it would fit on the water bottle mountings on the down tube of the bike frame, yet be strong enough and thin enough not to get knocked off during normal riding. This housing would accommodate the assembled radio in the base, with the rest of the electronics in the lid, the box being held together with four screws. The antenna of the radio would need to be left uncovered (the protective cover was removable from inside the disassembled radio) and bent upwards, but otherwise it would fit just fine.

Trying to solder flying leads to the 'call' button fingers proved too difficult. The common was easy, I just bared the coating off an area of copper in the large area beside the contact, but the tiny semi-circular pads just bridged over when touched with a soldering iron. Vowing not to be defeated, I resolved to find another way - I studied the tracks on the PCB to see if there was another, easier soldering point. Both disappeared under the display. Looking closely, I noticed that all other functions were performed with one line, so why two on the call? I picked up the other set, turned it on and fiddled with the 'call' button. I found that if it was pressed to the left, it mimicked the 'mode' button, and on the right, it mimicked the 'PTT' (Press to talk) button. I pressed those two buttons together, and the radio called.

I went back to the other radio, and found more suitable solder points near the 'mode' and 'PTT' contact areas. Soldering was still tricky, but with some swearing, I finally got the wires right. Next, I bonded the wires in place with 5 minute epoxy, to ensure they didn't snap off. I soldered two power leads onto the battery contact area on the rear of the PCB, and and after burning a hole in the corner of the battery compartment with the soldering iron, carefully routed the cables round the circuit boards and out of the case. I reassembled the radio, and put some batteries in just to check it still worked - luckily it did. shorting the three signal wires together produced the call signal required. I connected the signal lines and power lines to the timer PCB and the whole thing worked like a dream.


The signal wires soldered onto the main PCB. Note the the epoxy to stop them tearing off during reassembly of the case	The reassembled radio, assembled to check the fit of the signal lines. Later, power connections were added. The brass insert on the left was for a belt clip - this was later used to mount the radio in the box	Testing the unit running from the timer PCB, prior to final reconstruction. Study the careful routing of the red power lines to avoid fouling the case	The radio reassembled for the final time. Note the antenna cover has been discarded for fitting into the alarm housing

All that now remained was to fit the electronics and radio into the box. I drilled all the holes, fitted the LED's, key switch (to arm the unit - i.e. switch it on) and fit the terminals for the luggage loop. The idea for this came to me after a visit to Dixon's. I noticed they had goods on display wired to an alarm by passing a thin cable through their handles... thus, were they to be taken, the wire would be broken and the alarm would go off. I incorporated a similar idea into the circuit for the alarm - with the camera on the bike, I could run a thin piece of wire through the case latches and through pannier clips and suchlike, so should they be opened, the wire would be severed, setting off the alarm. I used a couple of spring terminals for this, and they sit with a short loop of wire between them when not in use.

The device has worked flawlessly for eight months now. The following photos show the alarm as it is now, after many soakings (it's fully sealed using silicone) and a few thousand miles of vibration. The only problems I had were that the key switch I used originally was awful, and failed. I replaced this for a heavy-duty one which was bigger, requiring me to space it off the box face with a 20mm plastic conduit locknut. The other RoamAround started to exhibit weird behaviour 3 weeks later (lucky it wasn't the one I built into the alarm!) and I purchased a smaller, much-improved Binatone MR500 from the same trader at Elvaston Castle radio rally a week later for £10. It's behaviour has since been impeccable.

The range of the unit is excellent - the signal travels well through pub walls, and is good enough to leave whilst shopping with the bike locked up outside the local supermarket. It has served me well in catching a few miscreants tampering with the bike, the chance to silently creep up on them in the act being a jot to the soul, and also serves as a good talking point over beer...


The alarm on the bike. It runs from the battery that runs the lights - should anyone disconnect it, the channel monitor on the handheld receiver soon beeps to alert me of a problem	Contrary to appearance, the device fits well and doesn't interfere with the pedal action at all. It's only a bit wider than the bottle and holder it replaced. The terminals are for the luggage loop	The connections to the lighting battery are getting a bit messy. It now serves the front lights, the alarm and a modified rear light - but still runs 5 days before recharging.	The radio is mounted by the belt-clip screw, and stays firm. The timer is mounted on the underside of the enclosure lid. All of the base openings are sealed with bathroom sealant to keep moisture out.

How the timer PCB fits in with the controls. The wiring is a little messy, but works just fine. The small black transistor device soldered onto the 'arm' LED is a flasher IC, making the arm light flash when turned on	The vibration sensor is in the middle of the PCB, which is attached to the box lid by double-sided foam tape. This softens the sensor against nuisance tripping	The alarm box is screwed to the frame by stainless steel countersunk screws, into threaded bushes designed to hold a water-bottle cage. The screw holding the radio in is highlighted in the larger image	The Binatone MR500 - running of four 'AAA' rechargeable batteries, it's small and loud enough to tuck into a coat pocket, even if it does look like a baby monitor...

I plan to build a version 2 of the alarm sometime this year. The aim is to build it smaller - I have seen that RS components stock a Eurowave PMR set made by Entel (RS stock no. 389-4094), measuring just 80x50x22mm, and weighing in at just 120g. This is half the length of the RoamAround 100, a third thinner and a quarter narrower. I envisage the replacement having the timer built circuitboardless into the battery compartment of the Eurowave, and the main arm-disarm being performed by a magnet operating reed switches in the case with a magnet instead of a key. A better solution would be used for the luggage loop terminals, and I'd like to fit the unit on the seat post instead.

All that, however, is for the future. I realise nothing will stop the determined thief from stealing a bike, but the alarm, cheap and fun to build, helps deter the casual ne'er-do-well, and gives peace of mind while I shop or enjoy a pint.