Advertisement

Electronics and Electrical Engineering Design Forum

Where you can find electronics and electrical engineering forum questions & answers.
no image Sunday, January 14, 2018 by huddslad huddslad

Biasing emitter follower power amplifier

Hi I have been reading the latest practical everyday electronics (epemag) on the lateset power amplifier the SC200 which was deveoped by the australian magazine siliconchip who gave permission for epemag to use the same design article. I posted a couple of questions which but havent had a reply back.

Can anyone else help with the questions below.

 I have also posted the circuit diagram.


1.) does it matter which version of BC546 is used, ie BC546B, BC546C and also can BC556B be used for the SC200 power amplifier.


2.) i have a query about the 4.4mV drop each emitor resistor Re. I read elsewhere specifically in Randy slone power amplifier construction manual and douglas self power amplifier handbook that the optimal V drop for low distortion with emitter follower output stages need to have a much higher voltage drop across both emiter resistor ie Randy slone quotes a value of 47mV and Douglas self quotes 50mV. With Re restors of 0.1 ohms this means the current per transistor pair should be 0.047 / (0.1 x 2) = 0.235 Amps.  So total current for both sets of ouput transistors = 2 x 0.235 = 0.47 Amps. These values do seem ridiculously high to me but the fact that 2 different Authors have quoted very similar Re Voltage drops for emiter follower outputs circuits raises the question why this design has only 4.4mV.Am i getting confused misunderstood about emiter follower output stages and the optimal biasing required ????.


Regards

Heath

Comments

  • by  David Ashton (edited)
    Hello again Heath.


    1. it would not matter much which version of the BC 546 is used, if you can get the higher gain ones BC546B or C I would do so.  The gain of the amp is set by negative feedback (the resistors on Q2 Base) rather than the transistor gains, but selecting the higher gain types means they have to work that little bit less.  It would be an interesting exercise to try the circuit with low gain BC transistors (A types) and with high gain (C types) and see if there is any difference.  You might see a slight increase in distortion but not much else I think.  The same goes for the PNP BC556s as for the NPN BC546s.

    2. 4.4mV is quite a low drop for these resistors.  They are there because the Base-Emitter voltage of the main output transistors Q13-16 varies a bit with temperature, and if you put a fixed voltage on a base-emitter junction, the collector current will increase with temperature.  In extreme cases this can lead to "thermal runaway" resulting in the destruction of your output transistors.  So you put a small resistance in the Emitter lead, and if the current in the transistor increases, the voltage across the resistor will increase too and tend to reduce the actual Base-Emitter voltage (so it is a stabilising effect.

    All other things being equal, 4.4 mV would be a bit low for this.  However, in this case the biasing transistor Q10 is acting as the load for Q9 which (with Q6) is configured as a constant current source.  So if Q10 tries to pass more current due to thermal variations, the biasing voltage will actually reduce.  You will see that it is thermally coupled to driver transistors Q11 and Q12 which will enhance this effect.  You're right in that if you biased to get 50 mV on these resistors, your current would be large.However, if the designer wanted that much he would have used 0.47 Ohms or 1 Ohm resistors instead, to get a larger voltage for around the same current.  But this will take power away from the output when the amp is working hard.  So this is actually pretty clever design.

    Bear in mind too that the 4.4 mV is at quiescent conditions, ie no signal.  When the amp is working hard you will see considerably higher voltages across these resistors.  The reason for the quiescent current is to have a smoother transition between one side of the output stage working and then the other side  during the "zero crossing" between positive and negative output.  This is why this type of amplifier is called type AB - it's not totally linear like a type A, but it's not pure class B (where positive and negative drive are completely separate).  In a small area in between the amp is working in class A and this considerably reduces what is called "Crossover Distortion".  If you want to see what this looks like, get an oscilloscope on the output and reduce the voltage on those emitter resistors to as low as it will go.  you will see a bit of discontinuity between the + and - half cycles on the output.  You don't want this in a good amp.

    It would be nice if Alan Winstanley could get the designer to comment as well (almost certainly he will know more than me about these things :-) but I think you can rest assured that the designer has done the sums and that the emitter resistors are correctly dimensioned for stable performance.  Silicon Chip (where the design was first published) is a very respected Australian electronics mag and the design would have gone through a lot of vetting before making it to the magazine.


Add Comment

You must log-in to comment.