Power Amplifier Project

[Epicdemise]

Hey guys, new to the forum and I am not sure where to start on my project. I have uploaded the requirements and I can do the power supply but I am not sure how to/where to start on the power amplifier. Thanks for the help!

 

[ronsimpson]

So you need 10watts into 8 ohms rms. So you need about 14w peak. Now you can find the peak voltage needed on the output. Now find peak to peak voltage. Next add another 5 to 10volts to get the supply voltage.

 

[Epicdemise]

Alright so I found the Vp to be ~10.5V and the Vpp to be ~21V. Which from your advice the supply voltage would be ~30v. I understand basic opamps and mosfets/transistors but I am not sure how I would setup the tone controls and etc. In my mind I would need a high pass and low pass filter (from previous classes) but I can only use an opamp driver so I would have to use transistors for the rest. Which is mainly where I am stuck at.

 

[ronsimpson]

Search for audio amplifiers on the internet. Watch out some of these are really bad.
You could also look at application notes from places like TI.COM There are some really nice application notes. Old national semiconductor application notes may still be available.

 

[KeepItSimpleStupid]

If your single supply limited too, your amp probably should be AC coupled. So, think the topology that has a capacitor bypassed resistor and a capacitor coupled speaker and a transistor in the output stage.

 

[Epicdemise]

Okay, so after doing some research and etc. It seems to me that I will have an inverting op-amp with a Vin of 100mV and a Vout of -11V, so a gain of -110V/V. Which would leave me with a current of 1uA, therefore I would need to use a version of the class AB amplifier but modified with a darlington pair to increase the current to the desired amperage to get the power output of >10Wrms. Is this correct?

 

[KeepItSimpleStupid]

Where does the 1 uA of current come from?

Unless your using a reallly wierd OP Amp, I think you would have between 1 and 20 mA available for drive.

I "don;t think" you need class AB. With a single supply, that may be tough.

There is always bridging where you drive one amp in phase and another out of phase, but usually what that does is double the voltage. It's usually used in car amps so the amp thinks it's working with a +-12 V supply.

And I do want to re-iterate some of the above posts, When you have a spec as 10 W, 8 ohms, you do have to figure out both I and V, such that you can meet that spec.

Note the "Average power" spec. Don't take it as RMS. See here: https://www.learnabout-electronics.org/ac_theory/ac_waves02.php

 

[Epicdemise]

So from the attached picture in the original post, if I use the line level input of 100mV into 100k Ohms I have a current of 1uA. I think need to amplify the current to around 1A (could be wrong) by using a class AB amp but with multiple transistors.

I believe I can use a dual supply.

 

[ronsimpson]

OK draw a schematic.

 

[Epicdemise]

It works but there is a bit of crossover distortion and I probably need to lower the power supply voltage due to the op amp I am using. I am not really sure what I can do to remove more of the distortion and if I lower the rails the way the schematic is the output waveform clips.

 

[KeepItSimpleStupid]

R2 is in the wrong place. Move the side at the output of the OP amp to the speaker or if you need more voltage gain than your OP amp can provide, divide the output so it's in range and feed that back. That feedback network can tailor the high frequency response.
and you need a filter at the input so it starts at 20 Hz and won;t amplify DC.

Component tolerences, like different beta's should give you a slight AB advantage. Make the beta 40 higher for one transistor. Crossover should go away.

 

[ronsimpson]

Also D2 is in the wrong place.

 

[Epicdemise]

I am not entirely sure what you mean by dividing the output and etc. We were supposed to cover feedback today but professor was sick. Here is the updated schematic and waveforms are also provided. The orange is the output voltage at the op amp and the blue is the output voltage across the speaker.

 

[ronsimpson]

Attach schematic.

 

[Epicdemise]

Sorry was editing

 

[ronsimpson]

R12 should be on the emitter. Small thing.
Where is vpin9? I think vspkr is the output voltage.
Are you sure the supply is correct on the op-amp?

 

[Epicdemise]

vpin9 is between the two diode and yes vspkr is the output voltage. Changing the two supplys on the opamp does not change anything in the output waveforms and same goes for moving R12 to the emitter. It is clipping and I think that is because the opamp only supports 110V gain when I have a 130V gain. I don't really know how to fix it.

 

[ronsimpson]

Now I see. The output of the op-amp goes from - supply to + supply but the emitters (speaker) only goes +/-5 volts. This is because the current gain in the transistors is too low. R4,5 provides the current into the base(s) of the output transistors.
two options.
1) make more current pass through R4,5. Change to 220 ohms.
2) use a better transistor. (darlington) Or use total of 4 transistors to get more current gain.

You really should go look at amplifiers on the internet to see how this is done. Don't invent this from nothing. Learn from others.

 

[KeepItSimpleStupid]

Move R12. Think symmetry.

Remember the diodes should be mounted in thermal contact with the respective transistors.

A potential issue you have now is that the feedback exceeds the maximum input of the OP amp.

The bandwidth isn't to spec.
1. you have to reduce the bandwidth of the amplifier
2. You have to set the lower frequency.

you have to re-do the initial assumptions for Ave power, not rms power.

 

[audioguru]

Have you calculated currents?
The typical output current from an MC33078 is about 18mA. Then it can drive the 1k resistors that bias the output transistors, but it can't drive a lower resistance.

But can the 1k resistors turn on the output transistors?
If the peak current in the transistors is 0.8A each and their typical current gain is 100 then their base current is 8mA peak.
When the 1k base resistors have a current of 8mA peak then they have a voltage across each one of 8V.
Then the bases have a voltage of 18V - 8V= 10V peak each. Then the output swing is about 8.8V which is an output power of only 4.8W.

Your problem is that the opamp is not turning on the transistors, instead it is turning them off. The resistors are turning on the transistors but they do not provide enough current for a higher output power. If the resistor values are reduced then their currents will be too high for the opamp to drive.

To fix it replace the transistors with darlington transistors (and bias the darlingtons with 4 diodes instead of 2 diodes). The higher current gain of the darlingtons allow a much higher output power.