wireless doorbell
TRANSCRIPT
Wireless Doorbell
An educational project using a $10.00 doorbell
No kits are available for this project
9-9-2006
You cannot beat the Chinese. The project described in this article is a ready-made item from a $2.00 shop! And it is made in China!It is a wireless doorbell with a cost of about $10.00. It is so cheap, you could not build it from components for $10.00!So, we have to change our attitude to building projects and use items such as this to learn how things work and modify them to suit our own requirements. This doorbell uses quality components. It's not rubbish. The circuit is quite unbelievable. You can not obtain some of the components individually and the effectiveness is magic. We have so much to learn!
Transistor Type:
Collector current:
Max frequency
8050 NPN 1.5A 100MHz
8550 PNP 1.5A 200MHz
9014 NPN 150mA 150MHz
9018 NPN 50mA 700MHz
The first thing you will notice is the clever circuitry. Some of the design goes against everything we have learnt in electronics. That's why we have to study other people's designs and realise "the more you know, the more you realise you don't know."The oscillator circuit is very interesting, but first we will look at the RF oscillator. The doorbell operates on the 303MHz band and the 30 metre range (100ft) is obtained without the use of an antenna! The circuit is actually radiating from the printed track of the tank circuit. The Tank Circuit is a single-turn coil and a small capacitor (5p & 4p in parallel).
In this project we show how to add a small antenna to the circuit to get double the range - plus two other improvements to increase the range.Some of the improvements will load the circuit and alter the frequency at which it operates. Others can be done without any effect on the circuit.Fortunately, the transmitting stage is what we call "tight" and is not affected by surrounding "stray capacitance." Normally, this stray capacitance is a persons hand or body, touching or coming near the transmitting (output) stage and altering the frequency. The circuit has been kept near the power rails by the use of a choke in the positive rail. The positive rail is then reflected to the negative rail via the battery. This feature helps us when we want to add an antenna. A 7cm length of tinned copper wire is connected to the collector of the transistor and bent around the board so that everything can be put back into the case. When the project was tested inside the authors house, the range was increased to double. When the transmitter was taken outside, the range was over 60 metres (200ft) and the full range could not be tested as the sound from the doorbell was too faint to be heard!
We can learn so much from a product that is already on the market. Firstly, you know it works, it is reliable and you know what can be built for $10.00.Secondly, you know what type of components can be purchased cheaply and what to expect from them. In this case the transmitting transistor has the highest gain - so they have taken a special effort to get a good quality transistor. Now, let's look at the transmitter circuit:
THE TRANSMITTER CIRCUITThe transmitter circuit is made up of two building blocks - the 303MHz RF oscillator and the 32kHz crystal controlled oscillator.
The 303MHz oscillator consists of a self-oscillating circuit made up of the coil on the PC board and a 9p (9 puff) capacitor (actually 4p and 5p in parallel). The circuit starts-up by the transistor producing noise. This rising-and-falling signal on the collector is passed to the parallel tuned circuit (the tank circuit) and the base sees a very smooth sinewave at a frequency of 303MHz. This sinewave is then amplified by the transistor and this is how the 303MHz frequency is generated. Now we come to the purpose of the 15microhenry choke on the tank circuit. When the circuit oscillates, it takes a larger and small amount of current. This current passes through the choke and the turns produce a back-emf or back voltage that fights against the flow (change) in current. The end effect is a voltage created at the point where the choke is connected to the track-work on the board. This effectively allows the track work to produce a waveform and since the frequency of this wave is very high, a percentage of the energy is radiated into the air as electromagnetic energy. The choke allows the track-work to effectively rise and fall while providing a very low resistance path for the flow of current during certain parts of the cycle.
The second building block is the crystal oscillator.This is made up of a two-stage DC coupled amplifier with feedback via the 2n2 and crystal. If the crystal is removed, the oscillator is seen as producing very narrow spikes with a frequency determined by the 2n2, as shown in the diagram below:
When the crystal is added, the frequency increases (because the effect of the 2n2 and crystal in series creates a lower capacitance than 2n2) and as it rises, the amplitude of the feedback signal increases until it reaches a maximum at the resonant frequency of the crystal. The crystal exhibits the lowest impedance (the highest capacitance) at the resonant frequency. This is how the circuit stabilizes at the frequency of the crystal. When the device is turned on, the 150k on the base of the second transistor turns the transistor on. The third transistor has 0.65v on the collector and the base also receives very close to 0.65v, via the 220k resistor. The third transistor is not fully turned on and it produces a small amount of noise. This noise is passed to the second transistor and appears on the collector. The collector passes this noise to the base of the third transistor and the noise very rapidly increases to a maximum. It comes to a point where the waveform above is generated and the reason why the spikes are so narrow is easy to explain. When the middle transistor changes from an OFF state to an ON state, the capacitor will be partially charged and the voltage on the end connected to the base of the third transistor will drop about 6v and put a negative voltage on the base of the third transistor. This will keep it off and the middle transistor will be kept ON via the 150k base-resistor. The capacitor will gradually charge in the opposite direction via the 220k and 150k and when the base of the third transistor sees about 0.6v, it begins to turn ON. This causes the middle transistor to turn OFF and the collector voltage rises. This causes the capacitor to charge and create a current-flow in the base of the third transistor. Both transistors are now turned ON and the capacitor charges very quickly via the 12k and base-emitter junction of the third transistor. This creates the very narrow high-period in the waveform.
When the push button is pressed, the circuit produces a 303MHz carrier with a 32kHz tone. The receiver detects the 32kHz and turns on a SOUND chip.
THE RECEIVER CIRCUITThe receiver circuit consists of a number of stages and we will go through each one separately.
The first stage is actually a 303MHz oscillator that is operating all the time. It produces a clean 303MHz signal and this frequency is too high to be detected or processed by the 4069 chip, as the chip will only operate to about 1MHz.The theory behind using this type of stage is quite simple. It is easier to "upset" or modify a stage that is already oscillating, rather than get a non-oscillating stage to begin oscillating. There are all sorts of electromagnetic radiation at 300MHz and the 2-turn coil picks up all this noise. The 303MHz oscillator is firstly set into operation by the noise produced by the transistor and this is passed to the tank circuit made up of the 2-turn coil and 2p capacitor as a parallel tuned circuit. The transistor keeps amplifying this until it gets to a stabilized point where the collector is producing "hash" (junk) of about 300mV. When the transmitter is activated, the receiver circuit will detect a signal as small as a few micro-volts and the 32kHz signal will be included with all the other noise. There is a little bit more behind this receiving stage, than first meets the eye. The stage is actually a transmitter, but we will still call it the receiver circuit. Yes, it is a very weak transmitter and it fills the surrounding with a clean 303MHz signal. When the 303MHz signal from the transmitter enters this space, the signals interfere with one another and the receiver takes more and less current as it tries to maintain the signal strength. When the 32kHz signal is present, the receiver takes a varying current that corresponds to the 32kHz signal and this is how the receiver circuit produces the waveform to correspond to the 32KHz. A low-impedance path to the 0v rail is provided for the emitter by using a 82uH choke and a 2n2 capacitor across a 560R resistor. This low impedance path is needed so that the transistor has a high gain. The circuit is put into very delicate oscillation by using a 1k5 from the positive rail. It operates from 3v and the current taken by this stage is less than 1mA.
THE BRASS TUNING-SLUGThe coil in the tank circuit is tuned via a brass screw (or slug or core). This is alters the frequency at which the circuit operates, when it is turned via a small screwdriver. At 303MHz, you cannot use a ferrite core as it will completely absorb the magnetic radiation produced by the coil and prevent the circuit operating. Another choice is air. If you use an air cored coil, you can use a trimmer capacitor to adjust the frequency. A cheaper approach is to use a brass core. Brass has a permeability very close to air (µ0 is the permeability of free space) and it has very little effect on concentrating the magnetic lines of flux or moving them apart. Materials that cause the lines of flux to move farther apart, resulting in a decrease in magnetic flux density compared with a vacuum, are called diamagnetic. Materials that concentrate magnetic flux by a factor of more than one but less than or equal to ten are called paramagnetic; materials that concentrate the flux by a factor of more than ten are called ferromagnetic. However, at very high frequencies, such as 303MHz, the magnetic flux causes eddy currents to flow in the brass and this decreases the available flux so that inserting a brass core causes the frequency to drop. This is how the receiver is tuned to exactly the same frequency as the transmitter. The most critical capacitor in the receiver is the 2p. This sets the frequency. The 4p is merely a feedback capacitor. The 2n2 and 1n are called "pass" capacitors and allow high frequency signals to pass through them. The 1n actually keeps the positive and negative rails rigid while the 2n2 prevents the emitter moving up and down when amplifying the 303MHz signal.The signal on the collector passes to the first input of the chip via a 5k6 resistor and 1n (pass capacitor). A lot of the high frequency component is removed with the 1n capacitor connected to the 0v rail.
The first inverter has a 1M connected between the output and input to set it to mid rail so it becomes a high-gain amplifier. The second and third inverters also amplify the signal and on pin 6 we have a signal greater than 0.6v containing a lot of noise and an identifiable 32kHz waveform. The 32kHz crystal only allows the 32kHz signal to pass and the base of the transistor sees a very clean signal. Any other frequencies will not appear on the base of the transistor. The 32kHz is further amplified with two more stages and appears at pin 10 of the chip.It is then passed to a diode pump that charges a 47u electrolytic.Normally, this electrolytic is uncharged and pin 8 is HIGH. The PNP transistor is not turned on and the sound chip is silent. But when the electrolytic charges, the transistor turns on and the sound chip operates.
GOING FURTHEROne of the main reasons for presenting this project is to show how to get the best range from a transmitter. Normally you need very expensive equipment to help you, but a very clever alternative is to use our method. All you have to do is place the receiver about 15 metres from the transmitter and in a very poor reception area. The aim is to get the receiver to be at the extreme end of the range so that if you move the transmitter away by as little as a metre, the receiver will not detect the signal. Now the receiver is a very sensitive RF indicator. By moving the transmitter to different places, the receiver should not detect the signal. You are now ready to add an antenna to the transmitter and determine its effectiveness. We have already mentioned the transmitter circuit is classified as "tight" and adding an antenna should not shift the frequency.
ADDING AN ANTENNAThe 7cm antenna is added to the point where the choke touches the PC coil on the board.
Adding the 7cm antenna
This point is highly active but it does not interfere with the operation of the circuit, when the antenna is connected. Bend the wire around the edge of the board and about 0.5cm above it and test the transmitter. You should get an increase in the range, up to double. If the range does not increase appreciably, the receiver may not be tuned exactly to the transmitter. To do this, adjust the brass screw in the tank circuit of the receiver. Before making any adjustment, make sure you know the original position by noting the alignment of the slot. Only turn the screw about 15° in either direction and take a "field test." If the range reduces, you know the screw has been turned in the wrong direction.
INCREASING THE OUTPUT Another thing that will increase the output of the transmitter involves rewiring the LED. At the moment the LED is in series with the positive rail of the transmitter. This effectively drops about 1.7v and the circuit only gets 7.3v from a fresh battery. By connecting the LED between the positive and 0v rail, with a 470R load resistor, the transmitting section will see 9v and this will increase the range. One more thing that will increase the output is to add a 220R across the 220R emitter resistor of the transmitter transistor. This increased the range in our prototype, but the frequency-adjusting screw on the receiver had to be turned about 15° clockwise to compensate for the slight change in frequency.
PARTS LISTWireless Doorbell
$10.00 from $2.00 shops
1 - 220R resistor1 - 470R1 - 7cm tinned copper wire
If you want to build this project using your own components, but do not have the 32kHz crystals, here is a modification that does not need them:
The only components you will have to know how to make are the coils. The size, shape and wire diameter are important as the frequency is very high and the coil must be identical to those in the doorbell products we bought for this article.
ding-dong Turning a Heath / Zenith Wireless Doorbell into a Remote Control Relay
Synopsis : With the addition of one capacitor, one relay, and a small piece of wire we can convert an inexpensive wireless doorbell into a remotely controlled relay useful for a wide range of applications. These include remote PC starting, lighting control, Halloween effects control, or virtually anything that can be controlled by a pair of relay contacts.
une 27, 2006 : Motivation and First Steps
This project has two beginnings. First, some time ago I was looking for a cheap remote control that would turn on my driveway floodlights from the car. Yea, an IR motion detector would work, and I even tried that for a while. The problem is that we have coyotes, javalina, and other nocturnal creatures that wander around here and I didn't want the lights to turn on automatically and scare them off. Further, I support the folks at Dark Sky. I started watching for old garage door remotes but none ever crossed my path. That was about a year ago. Then a few days ago a guy named Brett posted a question to the Make Forums saying he wanted to remotely turn his PC on and off. Toward that end he had bought this wireless doorbell and wondered how to run the receiver off of the PC power supply instead of batteries. As of this writing that conversation is ongoing, but it reminded me of my old flood light project. So, since we needed to reverse engineer the receiver to solve Betts' issue, and it seemed like a perfect fit for my lights, I confirmed the part number and headed off to Home Depot.
(RANDOM NOISE: I think I'm going to start shopping at Handy Andy's Hardware, or any other small hardware stores I can find. It's not that I have a problem with the prices or products at Home Depot, it's their "self-check-out" system. And it's not so much the systems as it's the fact that the Home Depot closest to me very often has NO human checkers!! They make due with one clueless young girl to monitor the machines. Now if everyone shopping there was as smart as you or I, there would be no problems. I, however, always seem to get there the same time as the grandpa trying to figure out how to run the machine, the clueless goon trying to take 16 sheets of plywood through, and the young girl that managed to pick out the only 45 PVC fitting in the whole store that have unreadable barcodes. And then there are those security anti-shoplifting tags buried inside half of the things there. You know, the sensors at the door can see one of those things from 8 feet away. So why the hell can't the AutoSelfCheckersFromHell machines tell that you just put one in your bag?!?! Yea, all the alarms went off when I tried to leave the store with my doorbell. )
I finally got home with the doorbell and the batteries. First things first; let's make sure it works. This is the first step because it's really really difficult to return something like this to the store as defective AFTER you've ripped it's guts out! Batteries installed, the doorbell works as it should. Granted, it sounds pretty crappy, but then I'm more of a 'solenoids striking huge brass chimes' doorbell kind of guy. This thing is simple enough that my Mom could install it. Now for a range
check. The package says that the doorbell will work with the transmitter up to 100 feet away. When I installed the batteries I could see that the antenna on the transmitter is a small 0.75 inch wire looped at one end of the case, and on the receiver side, there is a small disc capacitor with one leg soldered into the board, and the other hot-glued to the component-side of the PCB for the antenna (upper left corner of the receiver PCB photo to the left). While it worked great sitting on my desk with the transmitter and receiver separated about 12 inches, I was a little suspicious of the 100 foot claim. For the open air test I set the receiver on the patio table on my back deck. Since my lot is about 330 feet front to back there was plenty of space. Well, reliable operation stopped at about 69 feet. From 70 to about 80 feet I could get it to work changing the orientation of the transmitter (I had been holding it as if it were screwed to a wall). Then I placed the receiver in my kitchen and went outside my front door, a distance of about 30 feet. The unit worked fine. Then I started walking away from the house and the chime continued to work until I was about 50 feet away (from the transmitter, not the house). Again, by changing the orientation of the transmitter I was able to get another 10 feet out of it. Interestingly, the instructions for the doorbell state that installing the transmitter or receiver near metal studs, on a metal door, or near a concrete floor can reduce the transmitters range. By my guess this rules out about three quarters of all the homes build in the past ten years. So the range, while not what was specified on the package, seems perfectly acceptable for a doorbell application. This, however, is a hardware hacker project and I imagine that that limited range could become an issue. A couple of things come to mind. First (and simplest), I have to believe that if you reworked the antennas inside the units you could achieve a bit more range. The antenna in the receiver is a short piece of wire, roughly vertical, and bent into a "C". The transmitter antenna is a small loop situated horizontally. I -think- that if you rearranged the receiver antenna so that is was straight (not a "C") and completely vertical, and then mounted the transmitter sideways so it's antenna was vertical that you would see increased range. The second option would be to eliminate the two internal antennas entirely and figure out what a 1/4-wave vertical whip would be. This would CERTAINLY increase the range. I may play with that later, but the current quest is to make this thing drive a relay instead of a speaker, and see if we can power it from a PC power supply. So tomorrow it will be time to deal with the guts.
June 28, 2006 : Electro-Vivisection
Last night I made a pot of coffee, stuck my copy of Colossus into the DVD, ( "Colossus" is in my top-10 movies for hackers and geeks. Besides, Susan Clark is kind of sexy in a 1970's, geek-chick sort of way) and set out reverse engineering the receiver. I have a method of reverse engineering circuit boards involving digital pictures, PhotoShop, overlays, and CorelDraw that is probably suitable for an article all it's own. But for now suffice it to say that by the end of the movie I had a schematic, a decent cafine buzz, and I really had to get to the bathroom. If you click on the circuit image at the left you'll open a new page with the entire schematic for the doorbell. I've said many times that I might be sixteen kinds of geek, but RF-geek isn't one of them. Perhaps one of my radio-gifted readers will contribute a decent discussion of what's going on in the left two-thirds of the schematic (and maybe even tackle the antenna question). Luckily for me in the immediate project the RF stuff really doesn't matter. Looking at the schematic we can see that the bulk of the parts count is for the RF section. The transistor in the upper right is the audio amplifier, and the COB is the brains of the thing.
(RANDOM NOISE: "COB" stands for Chip-On-Board. For a regular integrated circuit (IC), a tiny piece of silicon is bonded to a lead frame and then encapsulated in plastic. The lead frame ultimately becomes the ICs legs, and are soldered to traces on a circuit board. For COB, the tiny
piece of silicon is bonded right to the circuit board and traces, and then a blob of epoxy is dropped over it to protect the die and die bonds. This saves time, space, and money for the manufacturer and you'll see this a LOT in consumer stuff made in China. )
The address header allows the user to select one of 128 distinct addresses for the reciever and its transmitter. This opens up the potential for some interesting remote control hacks but for now here's what we need to play with:
What we have is a numbingly simple audio amplifier. It kind of surprised me that there's no current limiting through the speaker at all but then it occurred to me that it might not be an 8-ohm speaker. A quick check with the meter and sure enough, the speaker IS 8-ohms, so it's Ohms Law time!. Current = Volts / Ohms, so 3 volts / 8 ohms = 0.375 amps, or 375ma. Power = amps X volts, so 0.375 amps times 3 volts = 1.125 watts. Now, granted, the calculations are for DC current and the speaker is seeing some kind of pulsed or modulated DC, driving that tiny speaker with a watt seems a little close to the limit. Since they have been selling these things for a number of years we'll presume it's not a design flaw, and just chalk it up to 'art'. Next we need to know what sort of signals are going through this circuit. Using an oscilloscope I'll make measurements at the three labeled points of the circuit. But for now all I have in front of me is a meter so let's check out the power.
Powerplay
Brett wants to run this off of his PC power supply after all so we need to know how much power (current) the receiver draws. A long time ago I made a silly little widget to make this sort of measurement a lot easier. It is a little piece of double-sided printed circuit board material about the size of a dime. I soldered a piece of test lead wire to each side and soldered female banana jacks to the wires' ends. To measure the current draw of a battery powered device (like our doorbell) you simply slide the little PCB in between two of the batteries or between the end of one battery and the spring. You then plug your current meter into the banana jacks and away you go.
(RANDOM NOISE: By the way, this is a handy way to add remote control or timers to battery operated devices that you do not want to physically modify. Using the same little PCB between the batteries, use a relay to short the wires together to turn the device on or off. For example, turn off the smoke detector while you're actively burning dinner, turn a boom-box into an alarm clock, turn on the lantern when you open the shed doors, etc.)
With the receiver sitting idle my meter indicated a current draw of 0.22 milliamps. For comparison consider that a regular old red LED draws 20.0 milliamps, or about 100 times as much current. 0.22 milliamps is virtually nothing, but this is with the receiver sitting idle. What happens when someone pushes the bell? Well, something interesting happens: When the button is pressed the receiver current jumps to a little under 10 milliamps for about 200 to 300 milliseconds, then when it starts sounding the current jumps to about 110 milliamps. The current fades back to it's idle level with the diminishing tone of the chime. This is just a guess, but I think that probably the module that makes up the brains of the receiver can power-down the 'made-a-sound-like-a-chime' part of itself, leaving only the RF section powered to conserve battery life. It would be interesting to know how much of the power is used by the module, and how much is consumed by driving the speaker. Easy; we just remove the speaker from the circuit and measure the current again. Repeating the measurement with one wire to the speaker unsoldered gives us the same 0.22 milliamps idle current. When we push the button now, however, the current jumps to about 8 milliamps and just stays there until it's done chiming. So, bottom line (and to use conveniently rounded numbers), the receiver draws 10ma, and about 100ma is driven through the speaker. Now, if you're really paying attention, you will have noticed that this 100ma through the speaker doesn't jive with the ohm's law figures above. Well, this is the difference between steady DC calculations and pulsed or modulated DC. The figures above were a guess, and having made the current readings we have real numbers so reworking the equations gives is an average of 0.300 watts, or 300 milliwatts going through the speaker. This is a MUCH more reasonable number. All of this doesn't necessarily mean that we can just connect a 3-volt relay where the speaker is. Again, the signal going to the speaker isn't really DC, it's some sort of pulsed, or modulated DC shaped like the audio of a bell ringing. So we need to know a couple of things; what does the waveform across the speaker look like, and is there enough constant DC in it to drive a relay. We should also know how much current can be driven through that transistor. A quick check of the C8050 data sheet tells us that the transistor can handle up to 1.5 Amps so power handling isn't a problem. Since we want to drive a relay, we'll have to find one rated to work at 3 volts. Running off of batteries and with a current capacity of better than an Amp through the transistor we really don't have to worry about the current draw of the relay. This could, however, be an issue running it inside a PC. When the PC is on there is plenty of power available to run the receiver so turning the PC off should be easy. But when the PC is turned off the only power available is the feeble amount available to let the power switch operate, and possibly to keep the ethernet circuitry operating (for "Wake On Lan"). Another participant in the Make Forum, "Lars-Phobe", offered that the current AXT power supply spec calls for at least 10ma for the stand-by 5 volt power, but that 720ma is recommended and that 1 to 2 amps is not uncommon. Hrmph. That's a hell of a range, anywhere from 10ma to 2 amps! On the one hand I'm thinking that most manufacturers (I was in that field for 15 years) take a recommended spec as the actual spec, so by Lars' number we have about a 700ma budget. But some of that power needs to be used by the motherboard so if we can make the receiver work at 200ma or less we should be fine. On the other hand the current spec is for current computers, and we hackers / Makers have tons of old PCs lying around and I haven't a clue what the 1990 power supply spec was. Since I have a half dozen operating PCs here, and some are pretty darned "vintage," I'll be doing some real-world tests. My homework: Scope the points on the schematic and see what the signals look like. Read the specs on the power supplies in my PCs. And see if I have any 3V relays in my junk pile....errrr.... Piece Parts Inventory.
June 29, 2006 : Wave Riding
Well, I did find some relays buried in my inventory, I didn't read all the PC power supplies,
and I did scope the receiver board. I decided that the power could wait. Let's just get a relay working on battery power, and then we'll deal with PC power.
The first measurement, as marked "A" on the diagram above, is the output of the COB module to the resistor which then drives the audio amp transistor. When the receiver first detects the signal, the output rises to about 0.60 volts. about 300ms later the chime starts. The chime signal is a modulated squarewave that starts of oscillating between 0.60 volts and about 1.5 volts. As the ringing fades, the amplitude of the signal drops until it is again floating at about 0.60 volts, and then about 200ms later the output drops to zero. I hadn't thought about it before hooking it up to the scope, but it actually cycles twice. Once for the "ding", and once for the "dong". Since the ding is a higher frequency than the dong, there is a bit of overlap in the middle of the complete cycle. The fact that during this overlap the output is actually to different frequencies of squarewaves, my scope had a heck of a time syncing to the signal.
The second waveform ("B") is from the other side of that resistor, and displays pretty much what we would expect. The signal is the same as before, but the resistor is limiting the current, and therefore dropping the peak voltage to about 0.80 volts. Since this signal is applied directly to the base of the transistor, it controls how much current flows THROUGH the transistor. That is, controls how much current flows from the +3Vdc of the battery, through the speaker, and then through the transistor to the negative pole of the battery. (understanding that is going to become pretty important in a minute. So, for example, if we lowered the value of the resistor, less voltage would be dropped across it, and our 0.80 peaks would move higher, say, to 1.20 volts. This would have the
effect of making more current flow through the transistor, and thus more current flow through the speaker. The end result? The sound from the speaker will be louder.
Next we look at the line between the transistor and the speaker ("C"). It's really the only one to check because the other side of the speaker is the +3Vdc of the battery, and the other side of the transistor is the negative pole of the battery. When the chime isn't chiming, the transistor is turned off, and there is no current flowing through the speaker. That means that with the chime off we see the full +3 volts on both sides of the speaker. Thus, this waveform floats at the battery voltage. When the button is pressed, we can see about a 0.20 volt drop on this line as the voltage applied to the base of the transistor rises to 0.60 volts (not enough to turn it fully on).
Now consider things from the point of view of the speaker. When the chime is off there is +3Vdc on BOTH of it's leads so it "sees" 0 volts. When the chime first starts but has not start ringing yet, the speaker still has +3Vdc on one side, but the other side has dropped to +2.80 Vdc due to that little voltage applied to the base of the transistor. With 3 volts on one side, and 2.8 volts on the other side, the speaker "sees" 0.20 volts across its terminals. Now, when the chime starts sounding the speaker has 3 volts on one terminal, and the other terminal switches between 2.80 volts and about 1.80 volts. This means that the speaker "sees" a signal that oscillates between 0.20 volts and about 1.20 volts.
What this all boils down to is that at its maximum, there is only 1 volt across the speaker. That's plenty to make sound come out of a speaker. It is not enough, however, to operate a relay. I guess that if you could find a relay out there with a 1 volt coil it would work. But in all my years I've never seen one. Now I can hear a bunch of you yelling "What about a solid state relay???". Well, one of the big selling points of solid state relays is their complete electrical isolation between the control input and the output, and that is done using an opto-isolator which is nothing more than an LED aimed at a photo-transistor. 1 volt still isn't enough to drive the LED with any confidence. What we need to do is to get that 1 volts expanded to as near 3 volts as we can.
And I already told you how to do it...
Up there ...
... where I was talking about how reducing the value of that resistor would increase the volume
at the speaker? Remember?
Exactly! As we reduce the value of the resistor the transistor "opens" more allowing more current to flow, and thus more voltage is dropped across the speaker. So what value of resistor do we want to put in there? Well, we're looking for the maximum current through the transistor so we want as high a signal to the transistors base as possible. Since any resistance in there is going to limit the current it's pretty obvious that to get the maximum current we won't replace the resistor, we'll eliminate it completely. Besides, it's a lot easier to short out the resistor that's on the board than it is to unsolder and remove it, and then solder a replacement onto the board. Now this isn't completely without risk. We have no way of knowing how much current we can pull out of that COB pin. I spent a LOT of time and there are simply -NO- datasheets or other information available on the part. So, as it does with a lot of hacking, you come to a point where you just have to cross your fingers and dive it. Heck, the absolute WORST case is that I'll have to go brave Home Depot again and buy another $14 doorbell.
BINGO!
As we said, with the resistor out of the circuit, the signal shown in the very first waveform above is now applied directly to the base of the transistor, and the transistor is now toggling between fully open and fully closed. You'll notice that the signal doesn't actually go all the way to zero. This is because even when the transistor is fully turned on, it still drops a little voltage across it. In this case, about 0.18 volts. So now our speaker "sees" 2.80 volts and then about 0.2 volts. This span gives us a swing of 2.6 volts which is certainly enough to drive an LED, and is usually enough to pull in a 3 volt relay. But, we still have a problem.
It's not DC.
The power is being applied to the speaker (or our LED or relay) in pulses. In the speaker this just makes sound. In an LED, the light actually turns on and off as fast as the signal, and this is going to cause some heartburn in a solid state relay. And for our conventional relay, you have to average the power across time and in this
case the average doesn't add up to enough to actuate the relay. Actually, when I tested this to get the waveform you could actually hear the chime sound coming from the relay as it struggled to close it's contacts. It was very quiet, but I could hear it.
What we need is something to apply power to the relay during those gaps to bring the average up. Something that will save some of the power when it's there, and then give it back when it's not. That is what a capacitor does. If we place a capacitor between the speaker-transistor leg of the circuit and ground, it will (in a manner of speaking) save up some power when the signal is there, and let it go when the signal is gone.
(Okay, okay, it's not actually applying power, it's keeping that leg of the circuit from getting back to +3 volts as quickly as it might. Some of you can congratulate yourself on picking a nit. The rest, if you didn't understand this, don't worry about it)
Now I would like to be able to tell you that I applied my extreme mathematical prowess to some complex formulas that I know by heart to calculate the value of the capacitor to insert into the circuit. But that would be a lie, and life is too short to spend any of it in my attic looking for my old high school electronics textbooks.
2.2 uF didn't work.10uF didn't work.Neither did 22uF or 50uF.Then I stuck a 100uF cap in the circuit ....
There ya' go. It's not the prettiest or cleanest waveform on Earth, but it will do. With the 100uF capacitor in the circuit pressing the doorbell button resulted in a firm, authoritative, 'CLICK' of the relay pulling in. Well, actually, it was "CLICK" "CLICK". The relays closed twice for every press of the button.
I have to confess that I had to stop and think for a minute before I figured it out....
"Ding" ... and then ... "Dong".
Oh jeeze! Brett want's this thing to turn on his PC, and I had no clue how computers responded to double-clicks on their power buttons. What's more, this will make my flood lights into a kind of Disco strobe light. I started trying to think of ways to bridge the 300ms gap between clicks and then remembered the instruction sheet for the doorbell.
The doorbell will actually play different sounds for front and back doors (if you buy a second, optional transmitter). The sound for a back door is a single "Ding". The very last jumper position on those headers in the transmitter tells it whether it's at a front door or a back door. I added a jumper to that position and the problem was solved. Pressing the button once clicked the relay once. Done! I've done all of the above testing using jumper leads and alligator clips. Let's clean that up and make the mods to the receiver.
July 5, 2006 : Hackin'
Okay, the 4th of July weekend is over. One of the kids got bit by a dog that was scared by the fireworks, plenty of sunburns to go around, I ate too much, drank too little, and though it was a grand weekend I'm glad to be back at the bench instead of cleaning up the pool and the backyard like I should.
To accomplish the relay mod we need to do three things ...
Short out the resistor at the audio amplifier (or replace it with a shunt). Add a 100uF capacitor to the circuit. Remove the speaker and replace it with a relay.
Finding the relay is probably going to be the biggest challenge. The one I have is from a lot I bought on eBay about a year ago. I still see them from time to time, and a number of surplus electronics on-line stores have something that will work. You're looking for a relay with a 3 Volt DC coil. Mine has a coil resistance of about 12 ohms, however anything in that ballpark should work. Since the receiver runs on 3 volts, just about ANY 100uF electrolytic capacitor will work. You might want to tend toward the lower voltages (ie. 10v, 16v) because they are physically smaller. Up to this point the whole thing has been running with a bunch of alligator clip test leads connecting the receiver, the batteries, the capacitor, and the relay. The receiver has been sitting powered since I left it about a week ago and a quick check with a meter shows the batteries still have a full charge. I didn't expect anything else, but it's nice to confirm that my changes to the circuitry didn't start killing the batteries. I'm just going to list the steps of the hack here as opposed to writing them all out. You can
follow along with the pictures.
Remove the receiver board, unsoldering the wires to the battery terminals. This just makes it easier to work with the board. Solder a wire shorting out resistor R11. This is easiest if you solder a straight piece of wire to one end, then use a fine tipped needle-nosed pliers to bend it into shape, and then solder the other end. Now solder one lead of the capacitor to the end of the same resistor closest to the middle of the board. I found a short piece of scrap wire and stripped about 2 inches of it. Not for the wire, but the insulation. You can slide the insulation over the leads of the capacitor before soldering it into place. Solder the other lead of the capacitor to a ground point on the circuit board. I found that I could slide lead into the same hole and along side the black wire that goes to the battery terminal. Now we need to assemble the relay. I found that originally there was not enough room for the relay inside the case of the receiver, but then figured out that if I removed the clear case of the relay it fit just fine. Also, I like my projects to have a finished appearance so I dig out a 1/8" mono earphone jack I had salvaged out of something ages ago. This jack will carry the relay contacts to the outside. Solder the earphone wires to the relay contacts, and two pieces of wire to the coil leads of the relay. Unsolder and remove the two wires on the receiver board that used to go to the speaker, and replace them with the wires that come from the coil of the relay. Drill a hole in the side of the case to accept the earphone jack. Install the jack, and hot-glue the relay into place in the bottom of the receiver case. Reconnect the power wires from the receiver board to the battery terminals. I replaced the wires in mine with longer ones just to make my life a little easier. Reinstall the receiver board in the case, snap the cover on, and your done!
Just a few random notes...
At a few points through assembling this thing I stopped and reconnected everything with the test leads just to make sure it still worked. This may seem silly to some, but doing so will help you avoid "finish line failure". That is, getting everything assembled, putting in the batteries, snapping on the cover, and having it not work. Then you have to backtrack and figure out at which step you made your mistake.
I thought about adding a power switch, but since the receiver will run for about a year on a pair of "C" cell batteries I saw no real need. There is room, however, under the circuit board alongside the earphone jack to accommodate a micro-toggle switch if you want one.
You could also add another jack for the antenna if you like. A little panel-mount BNC, "F", or banana jack should also fit under the circuit board. If I have trouble with range I'll go ahead and add that later. But for now, 50 or 60 feet of range should do fine.
It might not be a bad idea to add a diode in parallel with the relay to shunt any back-current from the field of the relay coil collapsing when power is removed from it. This would protect the transistor from a big reverse voltage spike. However, I didn't see anything nasty on my scope without it, and I left it out to save current for the relay.
Another addition would be an LED indicator that lights when the relay closes. I would be concerned that the LED would draw too much of the available current and the relay would no longer close. If you want it, give it a try. I have to leave *SOMETHING* for you folks to figure out.
8-) Besides, the closed receiver case makes a really good resonator, and you can definitely hear the relay clicking.
I wired the earphone jack to the Normally Open (NO) contacts of the relay, so pressing the button will turn something on. You could, alternatively, wire up the Normally Closed (NC) contacts so that pushing the button would turn something off. Better still, instead of a mono earphone jack you could use a stereo headphone jack, bringing both the NO and NC terminals out of the receiver making the device even more universal. What would you want to turn off remotely? Well, the stereo in our kid's car when it get's within 100 feet of our house is the first thing that comes to mind. (grin)
Next up ... we'll get this thing running off of a PC power supply.
October 5, 2006 : I'm back
Well it's been a crazy few months. In short, I started a new business. What it is doesn't matter to you folks, but any of you who have gone through the ordeal know what a huge task it is. Things are settling back down now so it's back to "work". :-)
Earlier this week I got out the doorbell and transmitter. It's been sitting since my last entry for this project and I was pleased to find that the batteries are still good and the relay still works fine. To start to address the power supply question, I removed the batteries and with test leads connected the battery contacts to a well regulated bench power supply.
As you might expect, the doorbell worked just fine with the power supply set at 3 volts, the same as the batteries. I slowly increased the voltage, a half a volt at a time, testing at each voltage. I went as far as 6 volts and it still worked. I didn't want to go any higher than that and chance blowing the thing up. Besides, we're looking to run it at 5 volts inside a computer, and that is a VERY well regulated power supply. So, we're golden.
Then, just for giggles, I started turning the power supply down. The receiver worked all the way down to 1.5 volts, but relay stopped pulling in at about 2 volts. That's a good thing to know if you plan on using this mod as a battery powered receiver. While I'm sure that in it's original state the doorbell would actually sound down to 1.5 volts, our relay will only work down to 2. That means that while you could expect the two "C" cells to power the doorbell for more than a year, with our relay mod the battery life will be greatly shortened.
I have some down-time early next week, and I have an idle PC sitting in my office, so I'll get this thing installed and then update the project.
If you want to be notified when this page gets updates, send an email to: dingdong(at)hackersbench(dot)com
If I were to rank the top evil things in the universe, the devil would be second. Spam email would be first. You can absolutely be assured that your email address will be used for nothing else.
( By the way, third in the list is Microsoft) 8-)
October 6, 2006 : Email Feedback
Dave B. wrote ...
Nice hack, and a very clear instructable! One thing you mentioned was a kick-back suppression diode. A lot of relays come with the diode built-in, but if you want to add one, place it across the relay coil, not across the 100 uF cap, as you suggested. Wire it with the cathode (+) terminal at the +3V, the anode at the collector of the 2SC8050. Dave
Oops. I must have been typing faster than I was thinking that day. I've changed the 'capacitor' to 'relay' in the above text. Thanks Dave! Both for pointing out the typo, and the compliments. John