Electronics > Projects, Designs, and Technical Stuff

Problems Building a Class AB Amplifier

(1/3) > >>

ojl1997:
Hi there!

I'm very new to electronics but I have been trying to build a transistor based amplifier to amplify a signal.

I've attached an image of my circuit [attachimg=1], which I designed in LTSpice.
It has a power rail of 55V and an input signal of 0.64 V peak-peak, giving a gain of around 50 [attachimg=2].

I've built this circuit on a circuit board, however I can't seem to get it working with a 55V power rail. Whenever I increase the voltage above 35V the current draw becomes too large (it starts running away to upwards of 800 mA!), which is a much larger current draw than I'm expecting based on the models (500mA). [attachimg=3].

I'm using high power transistors, NPN: TIP 3055 and PNP: TIP 2955, which have a max base current of 7A and a max collector current of 15A. They get very hot when I turn the circuit on (I am using heat sinks) and once the current starts to run away above 35V the power supply shuts the circuit down (due to inbuilt safety feature).

I can't understand why this circuit isn't working at the full 55V power rail and why it is drawing such a large current? Is anyone able to explain this to me and suggest a way I could get the circuit working?

If you have any other ways I could get an amplifier working with a gain of over 50 with an input voltage of 0.64V that would also be appreciated!

Many thanks!
Olivia

RoGeorge:
Welcome to EEVblog   :)

- The bias current through TR1 and TR2 is set by the voltage drop on D1 and D2.
- The voltage drop on D1 + D2 increases with the current passing through them.
- Bigger V2, bigger I through diodes, bigger voltage on the diodes, bigger current through the power transistors.

One way to decrease the current through D1 D2 would be to increase the value for R3.  Another method would be to adjust R1 and R2.

In practice there are a few other ways for a better control of the bias current through the power transistors.

Another thing is, the schematic used here is good for didactic purposes, but in practice you may want to add some short-circuit protection for the final transistors, fine tuning for the bias current, maybe some filter against self-oscillations, too.

P.S.  When posting LTspice questions it is a good practice to attach the .asc file, so others can run the same circuit and modify it or probe other nodes of interest.

BrianHG:
Also add 2 small value series resistors in series at the emitter outputs and take your output from the center of those.  With the emitters shorted, thermal run-away can go haywire without some small breathing room leading to blowing your driver transistors, or blowing your power supply by consuming too much current.

I would used to use 0.2 ohm in some of my audio amps in the past and I had some heavier biasing as I did like to run my amps nice and toasty.

Zero999:
In addition to the above comments: D1 & D2 should be thermally coupled to TR1 & TR2, so their forward voltages also decrease, as it heats up.

Don't expect brilliant performance, from such a simple circuit. The output will be very distorted. The gain of TR3 is dependent on its collector current, which varies, as the voltage across R3 changes. Consider using a lower gain and adding a pre-amplifier.

David Hess:
I would replace R3 with a current source so that the bias current does not change with supply voltage, and add feedback to control the gain.

In a practical circuit, D1 and D2 would be replaced with an adjustable Vbe multiplier composed of a single transistor and a pair of resistors, one being adjustable.