# How to connect this amplifier to my symmetrical power supply?

I am relatively new to electronics.

I have this power supply. It is symmetrical because I need negative voltages to other parts of the circuit. The power supply input J3 receives 18 VDC. The 7806 and 7906 are working perfectly and the voltages measured there are 6.0V and -5.98V.

On another part of the circuit, I have this amplifier, using the IC TEA2025B. Because I am using a symmetric power supply that is derived from a VDC power supply, I am not sure what I should connect to V+, virtual GND or V-.

I have marked all points I don't know in blue, from 1 to 15 and at the end of this question I will show how I connected that.

First notice R1, R2, R5 and R6. I am using that to add a DC voltage to the audio, to shift it to the middle of the rail.

This is how I have connected it:

1. to virtual GND
2. to virtual GND
3. to +6V
4. to virtual GND
5. to virtual GND
6. to +6V
7. to -6V
8. to -6V
9. to -6V
10. to -6V
11. to -6V
12. to -6V
13. to -6V
14. to -6V
15. to virtual GND

But I am hearing an "helicopter sound", something like a feedback and as I increase the volume the pitch gets higher...

I have tried to connect these points differently but none works correctly.

How should I connect these 15 points? I think the problem comes from these 15 points incorrectly connected.

Unfortunately I do not have an oscilloscope, just a multimeter.

• Use your soundcard as a scope – PlasmaHH Apr 10 '18 at 6:24

The TEA2025 isn't designed for use with a symmetric power supply.

The easiest solution would be to only use the +6 V and virtual ground for the TEA2025 circuit, so completely ignore the -6 V. Then only point 6 (VCC of the TEA2025) will be on +6 V, all other points at virtual ground.

However you would then be loading your symmetric supply in an asymmetric way. This is because the TEA2025 will take supply current from the +6 V but the return current will have to flow through the LM358 opamp and that opamp cannot handle much current! So this solution is likely to give issues.

A solution for the asymetric loading would be to only use +6 V and -6 V lines, so completely ignoring the virtual ground for the TEA2025 part of the circuit. Again then only point 6 (VCC of the TEA2025) will be on +6 V, all other points at -6 V.

This could introduce an issue if you use the output signal of the other circuit, which has its ground reference at the virtual ground (6 V above the TEA2025's ground reference), and feed that to the TEA2025. However we can solve that with AC coupling capacitors, in fact, we can use C4 and C16 for this. Then connect point 2 to the virtual ground (so that the other circuit sees a load to virtual ground), then C4 and C16 will take care of the 6 V DC voltage shift.

I do not see why you would connect point 3 to +6 V, that will introduce DC at the volume control resulting in a grinding sound when you change volume. Only connect the VCC pin of the TEA2025 to the positive supply!

In summary:

• all points (except 2 and 6) should connect to the ground reference of the TEA2025 circuit, that ground reference is the - 6 V of the symmetrical supply so: connect to -6 V.

• point 2 should also connect to -6 V unless you're using an output signal from the other circuit you have which is using the virtual ground. Then it might be better to connect point 2 to that virtual ground.

• point 6 is the VCC of the TEA2025, it must be connected to the + 6 V.

• thanks but my brain is hurting at this point. Can you add to the end of your answer the numbers 1 to 15 and how you would connect it? Is it: 1, 2 and 15 = GND, 3 and 6 = +6V, 7 to 14 = -6V? ... I noticed that you talk that I should not connect 3 to +6V but you don't mention what should I do with 3 – Duck Apr 10 '18 at 6:38
• There's no need for your brain to hurt, just read carefully. You only need to think about points 2 and 6. 6 is easy: + 6V. 2 should connect to the local ground of the circuit which provides a signal to the TEA2025. If there is no such circuit, connect 2 to the -6 V. All other points: -6 V. – Bimpelrekkie Apr 10 '18 at 6:42
• I don’t see any problem with DC coupling this chip on bipolar supplies but the LM358 need complementary emitter followers inside the feedback loop. So consider +/-5V from PC – Tony Stewart Sunnyskyguy EE75 Apr 10 '18 at 6:44
• @TonyStewartEEsince1975 - care to translate that visually? Some link? – Duck Apr 10 '18 at 6:45
• What Tony means is that there are ways to increase the current handling capability of the LM358 by using extra transistors. I would not go that way as it adds complexity and I do not think it is needed. – Bimpelrekkie Apr 10 '18 at 6:49