Wireless Doorbell

An educational project using a $10.00 doorbell

No kits are available for this project

9-9-2006
13-1-2019   -  5k6 added to circuit


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!

Here is the corrected circuit:

5k6 needed to allow the voltage on the second transistor to rise above 0.65v

 

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 CIRCUIT
The 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 CIRCUIT
The 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-SLUG
The 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 (
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 FURTHER
One 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 ANTENNA
The 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.

THE 82uH INDUCTOR IN THE RECEIVER
I covered the amazing improvement of adding an inductor in the emitter circuit of a transistor in Spy Circuit Page 1.
I am not going to re-write the discussion but merely copy it onto this project and you can work out the relevance.
It's the inductor in the emitter. This can be in a transmitting circuit or receiving circuit.
Let's look at what is does and how it works.
The following circuit is a common-base COLPITTS circuit and oscillates at 303MHz.
The components that set this frequency are: the length of the track on the PC board (shown in thick black) and the value of the capacitor across this track.
The component that makes this a "common-base" circuit is the capacitor on the base.
The 1n has a capacitive reactance of less than ONE OHM at 303MHz!   
This means the base will not move up and down during the time when the stage is oscillating.

The circuit turns ON via the 22k resistor and this is because it puts a voltage on the base that is 0.7v higher than the emitter.  The current delivered by the 22k can now turn the transistor ON a certain amount - BUT it is not allowed to turn the transistor ON fully. If it did, the other components in the circuit (that turn the transistor ON - by an increased amount) would not be needed and there would be no components in the circuit to turn it OFF. 
So, we have the first part of the cycle where the 22k starts to turn the transistor ON.
Before we go any further, we have to mention the second way a transistor can be turned ON.
If we hold the base of a transistor rigid and "pull" the emitter lead towards the 0v rail, the "gap" between the base and emitter will increase and although this increase may be only a few millivolts, it will cause the transistor to turn on MORE.
Now, there is one more thing we have to cover.
When the voltage on the collector falls, due to the transistor turning ON, the energy in the 3p3 will be passed to the emitter to reduce the voltage on this lead.
But there is more to this. When the transistor turns ON, the relatively high voltage across the 3p3 will be converted to a much lower voltage to push the emitter lead a few millivolts lower but the current it will deliver will be considerable.
The emitter lead is considered to be a very low impedance lead and it requires a considerable current to move it lower.  
Let me explain the capacitor phenomenon again:
You already know about this phenomenon, but have forgotten.
If you take a large electrolytic and charge it on a 12v battery with a resistor in series, the capacitor will take a long time to charge and the current-flow will be quite small.
Now take the electrolytic and short the 2 leads on the shaft of a screwdriver. You will see a bright flash and the screwdriver will be "burnt."
That's because the electrolytic is capable of receiving a low current over a long period of time and releasing a high current for a short period of time.
The 3p3 is doing the same thing.
The low impedance of the "emitter circuit" is due to two things. It is due to the junction inside the transistor and the 100 ohm resistor.
If we were able to remove the 100 ohm resistor, the impedance of the "emitter circuit" would increase - but obviously the circuit would not work.
If the impedance of this part of the circuit is increased, the 3p3 will be able to turn the transistor ON more and thus the overall amplitude of the circuit will increase.
That's exactly what the inductor does. It effectively moves the 100 ohm resistor away from the emitter lead so the 3p3 sees the "emitter circuit" as a higher impedance.
How does it do this?
Firstly you have to understand how an inductor works when placed in this type of circuit.

The circuit above is called an LC HIGH PASS FILTER.
It only allows high frequency signals to pass through the filter.
The signal enters from the left and the left lead of the first capacitor rises and falls and passes this action to the right plate.
The incoming signal is not reduced by the capacitor if the signal has a high frequency.
This means the incoming signal appears on the top of the inductor with very little attenuation.
When the signal rises, a small amount of current enters the inductor and produces magnetic flux. This flux is called EXPANDING FLUX and it cuts all the other turns of the "coil" and produces a voltage in these turns that is of opposite polarity to the incoming voltage. Say the incoming voltage is 10v.  The inductor will produce a voltage of 9.999v and this means only 1 millivolt is available to deliver a current into the inductor.  
The end result is this:  only a very small current will flow into the inductor.
Exactly the same thing will happen when the voltage is falling.
The inductor will have very little "grip" on the voltage and it will allow the voltage to "pass over" the inductor without reducing its amplitude.
The "out" capacitor will pass the amplitude without reducing it and so the incoming waveform will pass though this circuit without being effected. 
If the waveform has a low frequency, the first capacitor will attenuate it and the inductor will reduce it and the "out capacitor" will reduce it.  The output will be almost zero.  
Now, how does the inductor in the 303MHz circuit work?
It allows the signal from 3p3 to enter the emitter lead and have very effect on attenuating the amplitude.
If the inductor were just a short length of wire, it would not have the same effect and a lot of energy from the 3p3 would be passed to the 100 ohm resistor.
I know the two circuits are not identical, but the basics are still the same.
The addition of an inductor DOUBLED the range of another transmitter circuit we were testing in the laboratory and when you realise the extra power needed to do this (about 8 times), its effectiveness is very impressive.



22-10-2014
13-1-2019