your own

Making your own inverter revolves around the transformer as this is the main component in the circuit and is the centre of focus in this article as it is quite a complicated component to understand. 
Before we start on the circuit and its operation, there are 4 "secrets" behind how a transformer works. 
The first "secret" is well-known. When a rising or falling voltage is applied to one winding, an exact replica is produced on any other winding. This is the basis of "transformer action". The amplitude of the voltage coming out of the transformer will depend on the turns ratio. i.e. the number of turns on the primary and the number of turns on the secondary. 
The second "secret" is the output waveform can be exactly the same as the input waveform or "reverse." By simply reversing the connections of your detecting device (such as a CRO) , it will appear that the output voltage is falling when the input is rising and vice versa.  
The third "secret" is very important. We know that a transformer only passes an alternating waveform. In other words, a DC voltage connected to a transformer (such as from a battery) will not be transferred to the other winding.  But the extension of this must also be noted. Take for example, the first half of a waveform. This is the same as applying a DC voltage to the winding. As the voltage is rising, the secondary is producing an increasing waveform, but when the input voltage rises to its maximum, the transformer does not see a CHANGE IN VOLTAGE and the voltage on the secondary winding immediately stops. At this point we have maximum voltage applied to the primary and no output from the secondary. In other words, we are delivering maximum current into the transformer and getting nothing out the secondary. 
The fourth "secret" is also very important. When the supply is turned off, the voltage produced in each of the windings can be considerably higher than the applied voltage. This is due to the collapsing magnetic flux cutting the turns of the  windings. This voltage can be 10x, 100x or even 1,000 times greater than the applied voltage. This is called "flyback" voltage and always occurs when power is removed. 
These are the four secrets behind how the inverter circuit works. 

The circuit has two modes of operation. When the double-pole switch is in the down position, the electroluminescent panel (or EL wire) is ON continuously. When the switch is in the up-position, the EL material blinks. The centre-position is off. 
You will notice the transistor is not taken out of circuit in the off-state. The collector and emitter terminals are still connected to the supply rails. The switch merely removes the bias from the base and this causes the circuit to draw almost no current. The only current is  through the transistor and is called leakage current. With modern transistors this is much less than a microamp and will not reduce the life of the battery.  
The circuit operates at approx 500Hz and this is determined by the inductance of the transformer and the load placed on the circuit by the size of the electroluminescence panel.  The circuit is shown below:

When it is first switched ON, the base bias resistor (4k7) charges the 10u electrolytic to approx 0.65v,  and the transistor conducts. The 4k7 (and the 10u) is not required for continued oscillation and we will show how the cycles continues to repeat in a moment. The first part consists of the cycle we will call the beginning cycle. This is shown in the  animation below: 

Mouseover for the first cycle

When the switch is turned on, a small amount of current is delivered to the base via the 4k7 resistor. This causes current to flow in the feedback winding to turn the transistor on harder. This current is due to the flux in the core and at the moment this flux is called EXPANDING FLUX.  The transistor is turned on more and more until the core is saturated. At this point the feedback winding immediately ceases to produce a current and the transistor is turned off. The current through the primary winding  stops flowing and the magnetic flux stops being produced. 
This causes the flux in the core to collapse and produce a reverse voltage in all the windings that is opposite to the previous polarity. This voltage is also much higher than the previous voltage and the voltage in the secondary winding (also called the tertiary winding or over-winding) is high enough to energise the electroluminescent particles in a panel and cause them to fluoresce. 
This circuit is not a push-pull arrangement and the EL only fluoresces during the time-period when the transistor is turned off. During this time the transformer produces the high-voltage flyback pulse.

After the first cycle, to understand how the circuit continues to oscillate we can reduce it to two components - the transistor and transformer.  One point to note: this circuit will not self-start. It needs a voltage on the base of approx 0.65v to start the circuit operating and that's the purpose of the base-bias resistor and capacitor (electrolytic). They allow the base to rise to 0.65v to start the circuit operating. But once it has started, it will keep oscillating.

From the animation above we can see the feedback winding turns on the transistor until the core is saturated. The voltage (and current) to the base ceases and the transistor turns off. The magnetic flux collapses and create a voltage of opposite polarity in each of the windings.  This voltage is higher than the supply voltage and energises the panel into emitting light. The collapsing magnetic flux is converted into electrical energy and reduces the high negative pulse towards zero and then the voltage on the winding becomes positive again. When the voltage is +0.65v, the transistor begins to turn on to start the next cycle.  
There is one other way to see how the circuit operates. The animation below shows how the forward and reverse voltage on the feedback winding turns the transistor on and off.  
The secondary winding actually produces a negative voltage to the EL panel but since the panel is not polarity sensitive, the polarity of the voltage does not matter. Also, the voltage on the primary winding reverses direction during the operation of the circuit but this too, does not have any effect on the oscillations. 
To keep the animation simple, we have only shown how the feedback voltage changes. 

One point to note is the voltage on the 10u capacitor rises to 0.65v at the beginning of the first cycle but when the feedback winding takes over, it is charged very slightly in the reverse direction during the action of the feedback winding delivering its voltage to the base and discharged via the base-bias resistor during the second half of the cycle. Thus we don't see any overall charge on the capacitor. But it is important to know the capacitor charges and discharges.
It works like this: 
The 10u and 100u electrolytics are actually around the wrong way in the circuit as they get charged in the opposite direction to the markings on the diagram. I don't think the circuit designer took the effort to check the actual operation of the circuit with a CRO - otherwise he would have placed the electro the other way around.  
Electrolytic can be charged slightly in the reverse direction without damage and this is what happens in this circuit.  
To understand how the 10u gets charged, connect the negative end of the electrolytic to the positive rail. This will turn the transistor on (due to the electrolytic being uncharged). In the process of delivering energy to the base, the electrolytic will charge and in this case it will get charged in reverse. If you can't see it getting charged, remove the transistor and place the negative to the top rail and the positive to the bottom rail. Now, it gets charged in reverse!
This is what happens. During each cycle is only gets charged (negatively) a very small amount and when the feedback winding turns the transistor off, the electro gets changed in the forward direction by the base-bias resistor and the voltage across it is removed. This knowledge is needed to understand the flash mode. 

The power switch has two positions. The UP-position produces the FLASH or Blink mode and the lower position produces the ON mode.  You will notice the circuit in the flash mode is almost identical to the ON mode. In fact the only extra component is the 1k5 resistor. This resistor allows the 100u electrolytic to charge during the ON part of the cycle, as you will see.
The cycle starts in the normal way by the base resistor (15k) turning the transistor on slightly. The operation has been described above, but this time there are some very slight differences in the feedback section. The base bias resistor is a higher value so that it does not fully discharge the electrolytic during each part of the cycle and the 1k5 resistor allows the electrolytic to charge a small amount during each cycle. This is all designed to create the long on and off times for the flash. 
During each cycle only a very small increase in voltage appears across the electrolytic and this voltage opposes the positive voltage produced by the feedback winding. The amplitude of the feedback winding in flash-mode is 15v, and eventually the voltage across the electrolytic prevents the feedback winding turning the transistor ON. At this point the circuit stops operating.  The negative voltage across the electrolytic is removed by the 15k base-bias resistor and it begins to charge in the forward direction until 0.65v is across it. At this point the circuit begins to turn on again and the flash cycle repeats.

The circuit looks simple enough to use almost any type of NPN transistor. The problem is the circuit is more complex than first meets the eye. The instantaneous current when the transistor is turned on is very high and some small-signal transistors will not deliver the required current. Ideally you should be able to compare a substitute with the original to see how much difference is produced. 
If you want to try substitutes and check the pin-outs of other types, go to our discussion on transistor pin-outs.

Our circuit has been taken from the 3v inverter shown above. It will drive up to 1 metre of Space Light in "ON" or "Flash" mode. 

Although the transformer looks like 4 simple windings, a lot of experimentation has gone into the number of turns for each winding. The turns on the feedback winding for the continuously ON mode actually determine the amount of current delivered to the base and thus how hard the transistor turns ON. In our case the amplitude of this voltage is 7v. Obviously the base does not need 7v, but this is the amplitude needed to deliver the required current to turn the transistor on. The reason is the winding is only capable of delivering a very small current and this is due to how hard the transformer is being turned on, and when "X" number of turns are provided, the circuit works.  
The same reasoning applies to the flash mode and the amplitude was measured at 17v. These have all been arrived at through experimentation, and are also dependent on the values of the surrounding components. 

After the experimentation is finished and the circuit works perfectly, we find the primary will have about 5 turns and the secondary about 1,200 turns. The feedback windings will have about 11 and 26 turns. The ratio of the number of turns on the primary to the number of turns on the secondary is called the TURNS RATIO. This can be any number such as  2:1, 5:1, 20:1, 100:1, 1:20, 1:50, 1:100, or any value. When the first number is larger than the second, the transformer is a STEP-DOWN transformer. This means the output voltage will be lower than the input voltage. For a turns ratio of 2:1, the primary winding will have twice the number of turns of the secondary. For a turns ratio of 20:1, this is read as: 20 turns on the primary for every turn on the secondary. 
If the second number in the turns-ratio is larger than the first, such as 1:2, 1:5, 1:20, the transformer is a STEP-UP device. For a step-up transformer, the output voltage is HIGHER than the input voltage. In other words the secondary voltage is higher than the primary voltage.  For a turns ratio of 1:2, the secondary voltage will be double. For a turns-ratio of 1:20, the output voltage will be 20 times the primary voltage.
But this theory only applies when a smooth sinewave is delivered to a transformer. In our case the transistor turns off and the transformer operates in flyback mode. The output voltage of the circuit above was measured at 800v p-p when no load was connected to the output. Thus we cannot apply the turns-ratio formula, however we can state the transformer is a step-up device. 


If a transformer has 5 turns on the primary and the primary voltage is 3v, the turns per volt = 5/3  = 1.7. Again, this theory can only be used when the transformer is receiving a sinewave voltage as this turns-per-volt value does not reflect in any of the voltages you will get on the other windings as the "fly-back" voltage will over-ride your calculations.  

The core is the material in the centre of the winding. It can be AIR or a MAGNETIC MATERIAL.  Cores made of a magnetic material can be made from iron laminations or a solid ferrite material. 
An air core has a value of 1. Metal cores can have a value as high as 1,000. If a ferrite core has a value of 1000, it is 1000 times better at collecting, absorbing, retaining, delivering and returning magnetic flux than air. 
Air has a fixed value of 1 and does not alter. The value of a magnetic material core varies according to the frequency of operation. As the frequency increases, the "value" of the core decreases. It can fall to "1" and even below 1. This is why high frequency transformers often have an air core. A ferrite core may only improve the performance of the transformer by a factor of 2 or 5 or 10 and if it is as low as 2, an air core may be quite suitable. 
If a 10 turn coil on a ferrite core (with a value of 2) has an inductance of 10uH, it will have an inductance of 5uH if the ferrite core is removed. Alternatively the coil will require 20 turns on an air core - or the diameter of the coil will have to be increased to achieve the same inductance. 
Thus a coil wound on a magnetic core is physically smaller than an air coil - that's the main advantage of a magnetic material. 

The size of a transformer depends on the amount of energy required to be transferred. A certain amount of energy is transferred during each cycle and if the frequency is increased, the energy transfer will also increase. 
A typical example is a power transformer. At 50 or 60 Hertz (cycles per second - this is the normal frequency of the "mains" - the power entering a property) a 100 watt power transformer may weigh about 2kgm. At 50kHz, it may only weigh 200gm. 
The normal operating frequency for electroluminescent displays is about 500Hz. This is one of the factors that governs the size of the transformer. 
Electroluminescent material (displays, sheets, Space Light etc) will operate at frequencies as low as 50 - 60Hz but the light output is not at a maximum. In addition, the colour of the emission will change slightly from the expected colour (more deviation in some cases). Some displays are sold with a "power-plug" transformer. This "power plug" may be one of three different versions:
1. It may consist of capacitors connecting the EL material to the mains. No isolating transformer is present and the panel operates at "mains" frequency (50 - 60 Hz). 
2. It may consist of an isolating transformer. The display operates at mains frequency and the output is approx 120v AC.
3. It may consist of an inverter to convert the frequency from 50Hz to 500Hz (500 - 750Hz, depending on the load). This is the best (and most expensive) supply for EL as it operates the material at the optimum  frequency


1: Name the component that starts the inverter circuit into operation.
2: Name the three windings of the transformer:
3: How does electrical energy get from the primary winding of the transformer to the secondary winding?
4: How is the high voltage produced?
5: Is the output of the transformer AC or DC?
6: How does the transistor turn off?
7: Draw the oscillator circuit:
Click HERE for circuit

Go to: Page 5