Considerations on selecting parts for a Class AB Audio Amp

Thread Starter

logans-electronics

Joined Sep 1, 2009
36
Any recommendations or specifications I should review/consider/search for when selecting PNP NPN bipolar transistors for the output stage?

I am ready to purchase some and get started...just having trouble knowing where to start.

I am building my power supply now using +/ 20 VAC rails and I have a lot of MPSA06 and MPSA56 transistor laying around for drivers and pre-amp related circuits.

I have been looking at the 2N6488, 2N6491... or ....TIP41C TIP42C.

Are there any particular specs that I should look for to possibly make biasing any easier?...or anything easier?

Thanks!:)
 

studiot

Joined Nov 9, 2007
4,998
It used to be that every electronics engineer (and lots of others) would brew up an audio amp in his backroom/garage so I can understand the appeal of getting going.

Most of these attempts resulted in rehashing / blown output transistors / excess distortion / and many other delights but it was all good character building stuff.

I don't really think it matters that much which pair of transistors you pick. The TIP series were the first to have a halfways decent Ft (3M) so their gain would not disappear at the top end of the audio spectrum. They have the added advantage of being readily and cheaply available in bulk from the far east.

You will definitely need to buy quite a few pairs.

I say this because, whilst biasing itself is pretty easy, the problem comes with matching.
Most people think in terms of gain matching but that is not the issue. The thing about class AB is that you give each transistor a small on bias in an attempt to eliminate (reduce) crossover distortion. The forward bias is provided by a common circuit so the individual biasing is not available for each transistor.This presents a problem if Vbe is not the same for both transistors. It is this that needs to be matched, remembering that they are inherently different for PNP and NPN transistors and even vary quite widely for transistors of the same type number.

So you will want to test select pairs.
 

hobbyist

Joined Aug 10, 2008
892
I am ready to purchase some and get started...just having trouble knowing where to start.


Are there any particular specs that I should look for to possibly make biasing any easier?...or anything easier?

Thanks!:)
Hi.

How are you going to aproach this design, do you have a specific load and power output to the load in mind, and do you have a specific input voltage in mind?

I'm asking because you said your having trouble knowing where to start.

If you want I could share the procedure I took in designing my power amp (for my CD rom, making it into a CD player)

I feel like I've been hogging up everybodies threads lately so let me know if you want me to post my design tips..
 

bountyhunter

Joined Sep 7, 2009
2,512
It used to be that every electronics engineer (and lots of others) would brew up an audio amp in his backroom/garage so I can understand the appeal of getting going.

Most of these attempts resulted in rehashing / blown output transistors / excess distortion / and many other delights but it was all good character building stuff.
Well now you're just making me feel nostalgic.

The tricky part is building in some circuits that do "turn on delay" from the output to speaker so you don't blow them with the turn on thump. Many amps use relays for this, which can also act as over current protection if they use current sensing.
 
Last edited:

bountyhunter

Joined Sep 7, 2009
2,512
Most people think in terms of gain matching but that is not the issue. The thing about class AB is that you give each transistor a small on bias in an attempt to eliminate (reduce) crossover distortion. The forward bias is provided by a common circuit so the individual biasing is not available for each transistor.This presents a problem if Vbe is not the same for both transistors. It is this that needs to be matched, remembering that they are inherently different for PNP and NPN transistors and even vary quite widely for transistors of the same type number.

So you will want to test select pairs.
Set up a test circuit to run some current through the B-E junction and measure the voltage. Write it on the transistor with a sharpie. When you get done with all the transistors, match up the ones that are closest into pairs.
 

bountyhunter

Joined Sep 7, 2009
2,512
Any recommendations or specifications I should review/consider/search for when selecting PNP NPN bipolar transistors for the output stage?

I am ready to purchase some and get started...just having trouble knowing where to start.

I am building my power supply now using +/ 20 VAC rails and I have a lot of MPSA06 and MPSA56 transistor laying around for drivers and pre-amp related circuits.

I have been looking at the 2N6488, 2N6491... or ....TIP41C TIP42C.

Are there any particular specs that I should look for to possibly make biasing any easier?...or anything easier?

Thanks!:)
The 2N5881 and 2N5885 (and their PNP complements) are really good audio power transistors. Gain bandwidth is about 4 MHz.
 
Last edited:

Thread Starter

logans-electronics

Joined Sep 1, 2009
36
Hi.

How are you going to aproach this design, do you have a specific load and power output to the load in mind, and do you have a specific input voltage in mind?

I'm asking because you said your having trouble knowing where to start.

If you want I could share the procedure I took in designing my power amp (for my CD rom, making it into a CD player)

I feel like I've been hogging up everybodies threads lately so let me know if you want me to post my design tips..
Thanks for offering your assistance, I would be very happy if you would share your sample circuits. I am not looking to directly copy it just understanding how it works, and what I need to do to make it run.
Any formulas you have would help as well!

Thanks!:).....Logan
 

Thread Starter

logans-electronics

Joined Sep 1, 2009
36
Well now you're just making me feel nostalgic.

The tricky part is building in some circuits that do "turn on delay" from the output to speaker so you don't blow them with the turn on thump. Many amps use relays for this, which can also act as over current protection if they use current sensing.
Thanks for the heads up....I will use a 8 ohms dummy load and measure the time delay (add a second or 2) and use a relay ran by a 555 once it is powered up.
 

hobbyist

Joined Aug 10, 2008
892
Thanks for offering your assistance, I would be very happy if you would share your sample circuits. I am not looking to directly copy it just understanding how it works, and what I need to do to make it run.
Any formulas you have would help as well!

Thanks!:).....Logan
I'm writing all this with an assumption your very very new to this, if not than this may just be stuff you already know.

I'm sure my circuit has a lot of flaws, so no schematic here, BUT, I think my procedure might be of some help.

First I took the signal source (CD rom) and tried to get a close estimate of PK. voltages with the volume controle at min and max value. either multimeter or osciloscope. recorded that data.

Then I calculated how much power output I wanted at the spkr.

Then calculated the amount of volt. gain needed.

Then breadboarded a simple class B amp, using complimentary pair. Knowing the supply voltage and the calculated RMS output I set up a dummy load 7.5 ohm power res.

Then I applied a signal from the audio generator at the max output voltage I wanted across the speaker.

Because a power amp. needs to amp. power not voltage, so whatever RMS volt. I needed at the output is the same voltage I put at the input.

Using my osciloscope, I watched for clipping and crossover distortion., When I got cross over distortion before the input voltage reached max. value, I then lowered the two base resistors (both res, were same value) than cont. increase the input signal.

When input signal was at it's max. and crossover sistortion was min. to none, then I knew what base res. I needed for that setup, at that supply voltage.

Once that was done than I moved to the preamp, I calculated the amount of Av. needed to bring the CD rom input signal to the Vout value at the power stage.

Then calculating roughly the Zin at the power stage I biased a common emitter amp with a RC value as low as possible with respect to the Zin of power amp.

Then designed it to have the proper gain needed to amplify the input signal to the desired output value.

So just remember that signal voltage amplification happens in the preamp stage, and the output of the preamp stage, should be as close to as possible the voltage output at the power stage.

The preamp cannot drive a lot of current into your speaker, But it can produce the same voltage amplitude that your power amp will apply to the spkr.

So just remember it is the preamp that amplifies that original small mV. signal, but cannot supply the right amount of current into a low res. load like your spkr.

So that SAME voltage output of your preamp can drive a high impedance load (small current) which is your input to your power amp.

SO NOW, that same voltage from your preamp, (that was current limited), passes that signal at a 1:1 ratio to the power amp, which can take that same voltage output and in itself has the capacity to supply all the current needed to a load that is so low that would otherwise bog down the preamp, so power amp, is a go between with the extremely low res. of a load to a high restance output from the preamp.

Power amp,In a sense maybe it's like a impedance matching circuit.

One more thing do not expect to get the max values calculated there is always going to be some variations, due to Vbe drops and ect... but you work the values until you get a satisfactory working circuit close to what you desire.


Hope I explained it well enough to get some idea about power amps.

See told you my posts are long.
 
Last edited:

Audioguru

Joined Dec 20, 2007
11,248
I disagree. All power amps have plenty of gain (20 times to 100 times) so that a line-level input can drive them to clipping and they have plenty of negative feedback to reduce distortion to almost zero.

A preamp for a modern amplifier has a low gain (nobody uses a microphone at home nor an old phono anymore) and has tone controls.
 
Top