by Harry Lythall - SM0VPO


Now that you have seen the basic transistor biasing methods, now let us try to 'jumble up' a few circuit blocks and produce a couple of practical circuits that we can use in reality. As always, you should use these circuits as a start before playing with yourself.

Cascading Stages

If we begin by taking two basic stages given in the previous article - 1 to make up a simple transistor power amplifier, we can see that the individual stages are quite simple. Power is VOLTAGE x CURRENT. So if we want a power amplifier then take a voltage amplifier and couple it to a current amplifier:

In order to couple signals between the stages without destroying the DC biasing conditions we need to add capacitors to 'block' the DC component from the previous stage.

C1 prevents any resistive or DC component in the input from changing the DC operating conditions of TR1. Typical value is 1uf.

C2 prevents any DC from upsetting the bias to TR2 and TR3, that bias is provided R3 and R4. Typical value is 1uf.

C3 is needed to pass the audio signal from TR2 and TR3 to the output, which in this case could be a 16-Ohm speaker. Typical value is 1000uf.

This circuit will work fine and give quite good results. It is, however, a little wastefull with components. In Biasing transistors - 1 we saw that the collector of TR1 should be set to 50% of the supply voltage by adjusting the value of R1. We also saw that D1 and D2 should be set around 50% of the supply voltage (+/- 0.7v); in other words, they should be the same voltage. So why not allow R1 and TR1 to set the operating conditions for TR2 and TR3? In this way we can save two resistors and a capacitor.

Now that looks a lot simpler. The most important point that we want to have 50% of the supply voltage at the junction of the emitters of TR2 and TR3. So I have taken the liberty of returning R1 to that point. The function of D1 and D2 is exactly the same as before, but this time they are driven by the current through TR1 and R2.


The output waveform of this amplifier would tend to 'flat-top'. As the output rises towards a peak, then the voltage across R2 would fall. This would reduce the current capability through R2 that feeds TR2. In other words, TR2 will be starved of drive current, just when it needs it the most, resulting in an output waveform like this at high volumes:

There is a remedy for this. It is called 'BOOTSTRAPPING'. If our +ve supply voltage were, say, 12v, then the voltage across R2 would be a little less than 6v. Wouldn't it be nice if we could raise the voltage at the top of R2 if the output waveform needs to rise? We can do it, too!

All I have done is added C2 and R3. R3 will typically be 10% of the value of R2, and C3 is typically 100uf. C2 top will be anchored at almaost 12v via the low resistoir value R3. The bottom of C2 will become 6v. So there will be 6v across C2. When the output waveform needs to rise to 12v, then the top of C2 will be the output 12v PLUS the 6v charge in C2 = 18v. R2 will now have 18v at the top of it at the peak of large signals. The output waveform from this amplifier will therefore look something like this:

Problem solved. This circuit is quite practical and will work exactly as drawn. Typical component values would be:

R1820KAdjust for 6v at TR2/TR3 emitter
R22K2Increase if TR2/TR3 get hot
R3220R10% of R2
C11ufAffects low-frequency response
C2220ufAffects high-volume distortion
C32200ufAlso affects low-frequency response
TR1BC547Any old NPN (2N2222 etc) Adjust R1
TR2BC547Any old NPN (2N2222 etc)
TR3BC557Any old NPN (2N3904? etc)
D11N4148Any old SILICON diode
D21N4148Any old SILICON diode

AF Oscillator

It has long been said "If you want an oscillator, build an amplifier." That was spoken in jest since many constructors amplifiers oscillate and their oscillators just sit there and stubornly refuse to oscillate; they amplify. There is far more truth in that statement than you would at first think. It was Edwin Armstrong, who in 1912, developed the 'amplifier with an infinite gain'. He had the idea that by sending a signal through an amplifer, then passing the amplifed output signal back through the amplifier again, and again, then the amplifier would have an infinite gain. He did not fail, but developed the oscillator, at about the same time as Dr. Lee DeForest. Let us do the same, but I will not bother with valves; too much playing around. Take a simple amplifier from Transistor biasing - 1:

Now let us connect the output back into the input.

It looks good, but it doesn't work! "If you want an oscillator ...". The output of the amplifier has the oposite phase as the input signal. As the input voltage rises, then the output will fall. In other words, the amplifier output cancels out the input, which is the oposite of what we want. We need to change the phase of the signal. C1/R1 form at time-constant that will shift the phase by only 60 degrees (at a certain frequency), but we want 180 degrees. Ok, then let us try again, but with three R1a and three C1s in series:

Ok, that did the trick, now it oscillates at audio frequencies. The 10n capacitors determine the frequency, which in this case is about 1KHz. As long as the losses in the feedback capacitors and resistors is less than the gain of the amplifier then it will oscillate. You can simply scale the capacitors for other frequencies, it is a lot easier than playing around with formulas.

RF Oscillator

For radio frequencies we need to feed some signal back to the input, but instead relying an a phase-shift being right at only one frequency, we will use a tuned circuit. Begin with a basic RF amplifier again. L1/C3 determine the frequency it will amplify.

Now rip out the input terminals, add a coupling winding to L1 to make it into a transformer (T1). Feed T1 secondary back to the input via C1. Now it will oscillate and T1/C3 decide the frequency.

The primary of T1 and C3 you can choose yourself using THE formula. The secondary of T1 should be about 10% to 20% of the number of turns as are used on the primary. This circuit works first-time every-time if you have the winding the right way round. Incidentally, if you replace C1 with a crystal then you have a crystal oscillator. It will not oscillate unless T1/C1 is resonant at the crystal frequency, or an odd multiple (odd harmonics).


You should now have all the information you need to examine the other circuits on my homepages and begin to understand them all. If you want to go through that exercise, then I suggest you start off with the building blocks chapters.

Very best regards from Harry - SM0VPO - Lunda - Sweden.

Return to INFO page