# TDA2050 feedback capacitor placement, and why do I need to bias the input?

Firstly here is the datasheet.

I see that cap C4 is placed before R4 for a single supply voltage. However, for the dual supply, it is placed afterward (in between R4 and ground).

Questions:

1. What will happen if I also place C4 (for a single supply setup) between R4 and ground, which is different from what is shown in the figure above?
2. Why do I need to bias pin 1 at the midpoint of the supply for a single supply?

1. Nothing will happen. You can swap around the C4 and R4 and the circuit is exactly the same. Think of C4 and R4 as single AC impedance. If nothing connects between C4 and R4, there is no difference.

2. It does not work without midpoint bias. You need to bias to mid-supply so a signal such as a sine wave can have positive and negative excursions around the midpoint. If you don't, your sine wave can't go above supply or below ground, the input and output must be within the supplies.

• Be careful of C4's polarity...whether you put it ahead or behind R4, it's positive terminal always points toward TDA2050 "-" input. And C4's negative end always points toward ground. Commented Sep 5 at 17:09
• What is the point of C4 as well? Commented Sep 5 at 17:16
• @JoeyB to avoid amplifying the opamp's own offset by the gain factor you designed for. At DC, the amplifier degenerates into a buffer, so the offset appears at the output with a gain of 1. Commented Sep 5 at 17:17
• @Designalog oh, so at DC the cap is "seen" as a open circuit therefore R5 basically now forms the buffer and allows the mid point bias through. I see very clever.... And at AC the cap C4 is seen as a short thus forming the amplifier gain... Commented Sep 5 at 17:21
• @JoeyB you got it. Commented Sep 5 at 17:22

The order of components in series rarely matters, you could also swap the positions of the resistor and capacitor in the Zobel network (R4/C7 in the split supply version) and it shouldn't make any difference. The only reason they are shown differently in the two schematics is that whoever drew them wasn't consistent.

In both circuits the input is biased at the midpoint of the supply voltage, it's just that in the dual supply version midpoint is ground while in the single supply version you have to derive a midpoint with a voltage divider. Take for example a +5 V/-5 V supply versus a +10 V supply, both are 10 V, and for both the midpoint is 5 V above the negative rail.