In this single supply configuration that amplifies an audio signal, why are the two .1uf caps there? I understand the 10uf cap is to remove the DC bias from the output, but why the other two?

Thanks

 

The upper .1uF cap is there to isolate the DC level of the input signal.
The lower .1uF cap keeps the junction of R3/R4 quiet. Current flowing through resistors causes electrical noise; caps to ground filters out that noise.

 

Good morning, Wookie. You seem faster than I am today.
Easy question, slow to log in.

 

The upper .1uF cap is there to isolate the DC level of the input signal.
The lower .1uF cap keeps the junction of R3/R4 quiet. Current flowing through resistors causes electrical noise; caps to ground filters out that noise.

Could you reword "isolate the DC level of the input signal"? The signal coming in would not have any DC offset since it's intended for a speaker/driver load, right?

 

The 0.1 uf input capacitor would allow the average signal voltage to be at +Vcc/2. In other words, removing the DC reference to ground.

The 0.1 uf input capacitor in combination with the 1k resistor roll off frequencies at 1.5 kHz and below at -20 db/decade.

Also:
The 10 uf capacitor on the output will allow the output to be connected to a grounded load without drawing DC current.

 

R3 and R4 create a DC level on the + input. The amplifier is going to duplicate that DC on its - input. The top, left, 0.1 stops that DC from escaping into whatever is bringing the AC signal, and the 10 uf on the right stops that DC from running through the speaker.

 

Ok, sounds good. One last question: why are the resistors 47k? why not 100k or 1M to minimize power loss? Does the voltage drop become less precise as you get higher resistances?

 

Yes. Depends on the op-amp. An LM741 has much different input impedance than a TL082

 

K, thanks a bunch!

 

Resistor RMS thermal voltage noise is proportional the square-root of the resistor value so higher resistor values can cause excess circuit noise.

 

R3 and R4 create a DC level on the + input. The amplifier is going to duplicate that DC on its - input. The top, left, 0.1 stops that DC from escaping into whatever is bringing the AC signal, and the 10 uf on the right stops that DC from running through the speaker.

The only type of speaker this would drive would be a piezoelectric speaker (or maybe earbuds). Another amplifier is needed to drive coil speakers.

 

Last questions, I promise.

Why is the value .1uF chosen for the smoothing capacitor and not a higher capacitance? Is there a rigorous way of calculating the appropriate cap size?

Also why is the DC blocking cap on the input .1uF? I understand the 10uF on the output because that won't block out important lower audio frequencies. But won't those lower frequencies already be blocked at the .1uF input cap?

 

The size of the filter cap on the + input isn't critical I think. It's just there to filter noise on that input & .1uF gives a cutoff of 33Hz. It could easily be some other value.

The input cap really should be larger for an audio amp. You're right that the low frequencies will be filtered by it. DickCappels gave the high pass cutoff frequency.

 

The way the circuit is drawn, it has an inverting gain of 100. Without the left side cap, *any* small difference in voltage between whatever is coming in the input and Vcc/2 (including the opamp's internal input offset voltage and errors caused by its input bias current) will be amplified by 100. Small voltage differences quickly can saturate the output of the opamp. The left side cap has an infinite impedance at DC (assuming no significant leakage current), so at DC the gain of the circuit is almost perfectly 0, greatly reducing errors at the output.

The input cap and the 1K input resistor form a high-pass filter. In this case the cutoff frequency is 1.6 kHz, too high for quality audio but maybe intentional if this is a specialty amplifier intended for high frequencies only rather than a full bandwidth preamp. A 10 uF capacitor on the input would reduce the corner frequency by 100 (10/.1) to 16 Hz.

ak