I want to plug in a microphone into a line level audio input of my computer and for that I need an amplifier, so I decided to use the popular one from the schematic below. I tested and it work very well.

But now I would like to power it from a USB line with 5V instead of 9V. I tried to mess around with the resistor values and add stages, but I fail miserable, the waveform either doesn't get as much amplification or it hits the rails and get deformed.

All I have at my disposal are resistors, caps, PNP/NPN and some LM358 op-amp(works with single power supply with output swing from -V to +V - 1.5V, so from 5V I only get ~ 3.2V swing and I need a bit more than 3.4 V).

So how can I make this work with 5V but achieve the same amplification ?
(at 9V, a 10mV p-p signal in the schematic below is amplified to 1.7V p-p ;at 5V it drops to 950mV p-p )

 

After a lot of jiggling around in the simulator I got something that will work with the 5V but there will be a bit of distortion. Time to test it irl

 

Why don't you ditch the transistor and get a GOOD audio op amp. You will have inaudible distortion from the amplifier as long as you don't overdrive it, and you can have variable gain. As drawn, the gain is adjustable from approximately 10 to 100.
Jameco has the amplifiers for $2.75 in onesies. You might find a better deal elsewhere. They also have the dual version (OPA2134).

 

I added the transistor because the swing of the op amp was not large enough and signal was severely deformed close to maximum output voltages.
OPA134 seems to have also a limited output swing, specifically (V–)+1.2 (V+)–1.5, which is even worse than LM358 (swings from GND to V+ - 1.5), plus I couldn't find anywhere in the datasheet any specification about single rail operation.
Anyway, it doesn't matter because I'll not be able to find this op amp at my local electronics store. I could probably find it on Farnell, but I was rather looking for a quick cheap solution.
I tested my solution and it seems to be acceptable quality, nothing fancy, but because of the high amplification, the microphone is very sensitive, which is what I was looking for since the beginning. During the test, there was a bit of static, but I was using some long wires that probably picked up a lot of noise, and hopefully it will go away once the circuit is going to be nice and compact. I should probably switch to ceramic caps as well instead of the electrolytic ones for decoupling, like you did in your schematics.

Thanks for the input

 

I added the transistor because the swing of the op amp was not large enough and signal was severely deformed close to maximum output voltages.
OPA134 seems to have also a limited output swing, specifically (V–)+1.2 (V+)–1.5, which is even worse than LM358 (swings from GND to V+ - 1.5), plus I couldn't find anywhere in the datasheet any specification about single rail operation.
Anyway, it doesn't matter because I'll not be able to find this op amp at my local electronics store. I could probably find it on Farnell, but I was rather looking for a quick cheap solution.
I tested my solution and it seems to be acceptable quality, nothing fancy, but because of the high amplification, the microphone is very sensitive, which is what I was looking for since the beginning. During the test, there was a bit of static, but I was using some long wires that probably picked up a lot of noise, and hopefully it will go away once the circuit is going to be nice and compact. I should probably switch to ceramic caps as well instead of the electrolytic ones for decoupling, like you did in your schematics.

Thanks for the input

Look more closely at the datasheet. With a 10k load, the output swings from 0.5V to 3.8V, or 3.3V p-p. Line level for consumer audio is about 900mV p-p. If that is inadequate, then maybe you should have been more specific about your requirements.
LM358 has crossover distortion, but then if you don't mind that nasty even harmonic distortion that your transistor amp exhibits, then the crossover distortion probably won't bother you either.

 

Hmm, I see. Frankly, I don't know what the input impedance of my sound card is or what voltage level it accepts, but the test was ok. All the instrumentation I have is a lousy DVM .
During test the sound quality seemed to be fine, I couldn't notice any distortion in the voice band, except for the really low frequencies at which the microphone might be more sensitive and after amplification the output was probably ramming into those line level voltages. But for normal speech it sounded really good for me .
The only thing that I'm worried about is the noise, which was pretty loud, but as I said I used very long wires and that might have been the reason. I was also able to hear noise coming from the USB line (there where certain squeaky noises during different PC activity) even though I used a 470uF filter capacitor on the USB line.
Does LM358 or BC547 have such a bad reputation regarding noise? Are electrolytic capacitors noticeable more noisy than ceramic ones?
Any advice is appreciated.
Thanks again for the help. Respect

 

Hmm, I see. Frankly, I don't know what the input impedance of my sound card is or what voltage level it accepts, but the test was ok. All the instrumentation I have is a lousy DVM .
During test the sound quality seemed to be fine, I couldn't notice any distortion in the voice band, except for the really low frequencies at which the microphone might be more sensitive and after amplification the output was probably ramming into those line level voltages. But for normal speech it sounded really good for me .
The only thing that I'm worried about is the noise, which was pretty loud, but as I said I used very long wires and that might have been the reason. I was also able to hear noise coming from the USB line (there where certain squeaky noises during different PC activity) even though I used a 470uF filter capacitor on the USB line.
Does LM358 or BC547 have such a bad reputation regarding noise? Are electrolytic capacitors noticeable more noisy than ceramic ones?
Any advice is appreciated.
Thanks again for the help. Respect

If you are not listening to music, or if you have a tin ear (like me), then you probably won't hear distortion.
Op amps all have good power supply rejection. Your transistor amplifier has NO power supply rejection. A 470uF cap helps a little. Adding 100Ω between the USB input and the 470uF will help a little.
Your biggest problem, I think, will be the fact that you have ignored the loading of the sound card. I found a site that says this:

most sound cards have an input impedance of 600 to 2,000 ohms

This will severely limit your peak-to-peak output swing of the transistor amplifier.
If you don't care about crossover distortion, you can replace the transistor with the second half of the LM358. Even with the limitation on its output swing, you will probably get at least 3V p-p before severe clipping occurs.