# Audio Amplifier Help

Discussion in 'Homework Help' started by zero_coke, Nov 9, 2011.

1. ### zero_coke Thread Starter Active Member

Apr 22, 2009
294
1
Hi!

So we have to build an audio amplifier that takes 100mV peak signal with Vcc = +5 V and have to build a power amplifier.

So what I did was build an opamp to get my gain of about +8V because it doesn't reach 10V due to non-ideality of the opamp I think.

And my question is, this is an audio amplifier so what should the output power be? I mean, when they say "build an audio amplifier", do they mean try to get the highest power at the output??

And one more thing, audio signals are 10 Hz - 20KHz and I heard from one my classmates that this design we make is supposed to work for any input frequency between 10 Hz and 20KHz. But whenever I change my input frequency to some other value than 1 KHz, the output power changes dramatically. Is this supposed to happen?? Also, at 1 KHz, I'm getting about 1.25 Watts at output. What is the theoretical maximum output power I can get having a max Vcc of +5V and input signal of 100mV??

2. ### #12 Expert

Nov 30, 2010
16,023
6,537
A drawing would help.
5V/.1V = 50
5V/.2V(p-p) = 25
No 8's in those numbers.
The resistance of the speaker matters.
and yes, an audio amplifier should have the same voltage output at all frequencies.
That's why a drawing is needed...to find your frequency error.

Dec 26, 2010
2,147
298
It also matters whether you only have a 5V Vcc and ground, or also a -5V line (Vee ?). If you are really getting 8V (peak-peak?) output, I would assume you have more than a single 5V, or else you are using a bridged amplifier. Beware however of simple AC simulations, which usually take no account of voltage limitations. These can give ridiculously big output levels based on a small-signal gain analysis. Use transient simulations to estimate maximum levels.

The maximum output level from an amplifier is normally determined by the maximum available voltages and currents, together with the load impedance. Typically the input level is not a limiting factor, as nowadays gain is a relatively cheap commodity. 100mV is actually a rather high level, compared to the measly millivolt or so coming from a microphone, nevertheless routinely amplified to give enough power to fill an auditorium.

Typically frequency response issues in an audio amplifier may result in some loss of gain at quite low frequencies, say below 100Hz, or in the higher kHz frequencies. A response flat within a few dB from tens of Hz to tens of kHz is indeed usual, unless a filtered response is required. Any dramatic variation of gain with frequency in the middle of the band suggests that there is a significant problem.

4. ### zero_coke Thread Starter Active Member

Apr 22, 2009
294
1
Ok thanks guys, I'll try to be more clear now:

We have available to us a +/-5V DC source, a few opamps, and regular transistors (2N222A) of both npn and pnp type, and power transistors MJE172 npn and MJE182 pnp. We can use passive components like resistors and capacitors, etc.

So what we did to make this audio amplifier is take the provided 100mV input voltage and at the first stage we tried to get max voltage gain using an opamp and the resistor ratio of Rf / Ri to get this.

Since Vcc was +/-5V, our theoretical maximum gain should be ~5V peak or 10V peak-to-peak right? Well we're getting now is 4V peak or 8V peak-to-peak and now at the power amp stage we just have a simple voltage buffer and then a Class AB power amplifier. My questions are:

When my input is 100mV @ 1KHz, my output power using a wattmeter shows 1.25W. But whenever I change this frequency between 10 Hz - 20KHz (the audio range), the power at the output changes. Is this supposed to happen? I did an AC analysis in multisim and it showed a straight line from 10 Hz to 20KHz but my power at the output would change whenever I changed the frequency so I'm confused...

What could be the thoeretical max output of an audio amp having +/-5V Vcc and an input of 100mV between 10 Hz - 20KHz frequencies??

5. ### Wendy Moderator

Mar 24, 2008
20,735
2,499
Just a question, do they dictate a Class AM amp? Could you use a Class D (PWM digital) instead? Just curious.

YOu will need good high quality op amps, else you will run into slew rate (which looks like frequency response but isn't) and frequency response.

Power is dictated by power supply and speaker impedance. A 10V power supply (what you described) could come with 10VDC/2.828 = 3.5V RMS (sine wave assumed) waveform, max. If this is fed into a 4Ω speaker then your power would be just under 1 watt, max. To increase the output power you need to up the power supply voltage or lower the speaker impedance.

If you use a full H bridge in linear mode you could double that.

6. ### zero_coke Thread Starter Active Member

Apr 22, 2009
294
1
Hi Bill, no they did not specify such things...they just told us to make an audio amplifier which has a peak input of 100mV and I guess we're free to use Class A,B, or AB (these are the ones we learned in class so far so I'm assuming one of these 3 is fine).

As for the OpAmps, we're using the UA741CN ones...

Here's how my schematic looks so far but I'm not getting the theoretical Vcc/R_load = 5 V / 4 Ohms = 1.25 A

<pictures removed>

AC Analysis: (Note: IR3 and V3 are voltage and current at 4 Ohm speaker load)

<pictures removed>

I'm getting like 1.15 W...is this ok? How can I reduce the distortion?? Why are my waves square-like? I think its because its clipping due to max gain of Vcc of opamp but when I reduce the ratio of the resistors for the opamp, I get less power at the output and more sine-wave like waves. My opamp only gives me a sine-wave like output when the output of it is about 4 Volts, so I can't get 5V even though theoretically its supposed to clip at 5V, not 4V.

Last edited: Nov 14, 2011
7. ### R!f@@ AAC Fanatic!

Apr 2, 2009
8,514
714
U forget a lot of tings.

First the 741 has a bad slew rate. low power out. No good driving power stages.

Second is when driving complementary output stage, you need to bias the transistors to avoid cross over distortion.

Bias requires a lot of components and it needs to be temperature compensated to avoid changes with temperature rise.

Go here and start Reading articles

8. ### Audioguru New Member

Dec 20, 2007
9,411
895
I guess you did not look at a datasheet for an old 741 opamp.
It is designed for a plus and minus 15V supply, not plus and minus 5V. But some 741 opamps work (poorly) with a plus and minus 5V supply.

With a plus and minus 5V power supply, the minimum output swing of a 741 opamp is zero when its load current is 5mA because its output transistors are darlington (a total of four emitter-followers for pulling up and for pushing down). You are lucky that your 741 opamp has any output.

Your output transistors are also emitter-followers and have a total voltage loss of about 2V.

If you use a more modern opamp like an MC34071, its minimum supply is plus and minus 1.5V so it is fine with your supply.
Then its minimum output swing with your supply is plus 3.7V and minus 4.7V.

Your output is severely clipped because the opamp drives its output as high as it can which is not much. Turn down the input or reduce the gain until the output is not clipped.

Your 'scope does not show how many volts per division. 2V? Then your amplifier's output is a max of 6V peak-to-peak which is 2.12V RMS and the power is 2.12V squared/4 ohms= 1.1W when it is not clipping or 2.2W when clipping very badly.

File size:
25.5 KB
Views:
21
9. ### zero_coke Thread Starter Active Member

Apr 22, 2009
294
1
Thanks AudioGuru, I changed the OpAmps and here's the result. I made four different configurations of my amplifier, so I need your help guys to determine which is the best...I"ll show you the AC analysis of each below:

Each of these designs have Vcc = +/-5 V and input is 100mV peak.

Design #1:

Design #2:

Design #3:

Design #4:

And this is the output for Design #1:

From the looks of these graphs, I think I should go with Design #1 right? I mean, the signal doesn't die out at higher frequencies even doh we're only concerned with audible range of 10 Hz - 20KHz and all designs are linear in this region, but I don't know, how can I pick the best one?? All their power outputs are almost the same, but the distortion is 0.3% for Design #1 and 0% for the rest 3 designs. How can this be?? I did use negative feedback wires from output at load all the way back to opamp voltage gain stage could this be why?

10. ### Audioguru New Member

Dec 20, 2007
9,411
895

You don't show a Design #1 and don't show its results.
The results for Design #2 and Design #2a look the same.

Your numbers do not appear to be correct:
1) Your 'scope shows an output of 10V p-p but that is impossible with a 10V supply. The typical output is 5.4V p-p maximum which produces only 0.91W into 4 ohms, not 2.8W.

2) If the output is an impossible 10V p-p then the output into 4 ohms is 3.1W into 4 ohms.

11. ### zero_coke Thread Starter Active Member

Apr 22, 2009
294
1
Hi AudioGuru, I can't really show my schematic because as much as I want to there are other classmates lurking these forums and I just don't want them to take my design and copy/paste it into multisim.

So my design #1 consists of one opamp to with R2/R1 = 50 gain so I get 100mV * 50 = 5V peak gain theoretically, but I get something like 4.98 or something anyway which is fine.

Then, I have a buffer connected from this opamp output to a power bjt configuration (class B) and the output is what I showed before to be. It shows 2.790 W and I don't know if the meter is wrong or if my circuit schematic is wrong...

Last edited: Nov 12, 2011
12. ### Audioguru New Member

Dec 20, 2007
9,411
895
Why don't you understand that with your plus and minus 5V supply, the opamp clips at 0V to 3.8V peak and the output emitter-followers reduce its output level about 0.8V each?? Then the max output is only 3V peak (maybe less) but your scope wrongly shows 5V peak.

You must reduce the gain (or reduce the input level) so that the output is not clipping, or increase the power supply voltages.

We have no idea what is your secret since you don't show the schematic of your "buffer" and your "overall negative feedback".

3V peak is 2.1V RMS then the output power is (2.1V squared/4 ohms)= 1.1W, not 2.79W.

13. ### Georacer Moderator

Nov 25, 2009
5,142
1,266
I would like to note that it is understandable that some people don't want to show their schematics over the internet for their reasons.

It is also unreasonable for those people to expect to get full help when they don't disclose all of the information related to their project.

Choose the path you will take, but you can't have it both ways.

14. ### R!f@@ AAC Fanatic!

Apr 2, 2009
8,514
714
what a predicament...?
Use PM.

15. ### zero_coke Thread Starter Active Member

Apr 22, 2009
294
1
Yes I'm sorry guys I already placed a suggestion for this forum to be able to allow certain people to "see" the attachments. I'm not a big fan of not posting schematics but like I said there's other students most likely lurking these forums and I know they'll simply copy/paste they're not here for help or anything. I'll PM you guys my schematic. Thanks and I do understand I can't get full help without showing the schematics so I'm trying to explain the problem as much as I can.

16. ### Audioguru New Member

Dec 20, 2007
9,411
895
When I went to school I didn't have forums, instead I had a helpful teacher and an excellent text book.

This school kid does not have a text book and has no teacher.

17. ### bountyhunter Well-Known Member

Sep 7, 2009
2,498
507
I don't think I could have created a more eloquent reason as to why such submissions should not be allowed by students.

And why students should be required to build things to understand them.

personally, I never figured out a way to hook a scope probe to a simulation...... but what do I know.

I love the smell of solder flux in the morning..... it smells like ELECTRONICS.

Last edited: Nov 12, 2011
18. ### bountyhunter Well-Known Member

Sep 7, 2009
2,498
507
We had terrible textbooks but pretty good teachers (for the most part).

19. ### Audioguru New Member

Dec 20, 2007
9,411
895
I received a PM from the school kid. It has a link with many letters, numbers and punctuation to a weird website that hosts the schematics with a password.
The site said it no longer supports my fairly new Internet Explorer and recommended Firefox but again mine is too old for it. I upgraded my Firefox then it said the web address doesn't exist.

20. ### zero_coke Thread Starter Active Member

Apr 22, 2009
294
1
I'm sorry but that's the only website that I know that allows files to be uploaded and shared with a password and I didn't know that it didn't support older versions of Firefox since I'm using the latest and did not have a problem while uploading and viewing the files myself.

As for the school kid part, we all need help sometimes. I mean, the textbooks, sure, the teachers, sure but they're not enough. This day and age if you went to school I'm sure you'd see for yourself the difference of professors that teach now vs. the professors you had then. And the textbooks, well, they pretty much remained the same with just minor changes in format.

You know what, here's the designs posted here. I really don't care if someone steals it anymore. The professors rarely have time in lecture to explain all the nitty-gritty details and I cannot find a satisfying answer online without having to read hundreds of lines of articles/journals. I just want an answer to my question which is why forums were purposely meant for this and I really don't want to sound ignorant but I really don't have time to go do all the research. I have four concurrent projects, 8 assignments, 3 lab reports all due next week and I hope you understand. I'm not saying you have to answer, and I contribute as much as I can to this forum as well. I just need help, that's all.

Last edited: Nov 14, 2011