Originally Posted by bountyhunter

Good point. When I used 2003 for a bridge, I did put a trimmer in it to zero out the DC offset. Schematic attached.

Your offset adjustment is in a positive only direction. What if you need a negative offset adjustment, swap the amp ICs?

Did you look at my previous attachment?

It makes a huge difference how the return lines to the audio channels are connected.

Those aren't bridged.

All right my last schematic contained an error shorting the +V to 0V.

Here's the corrected version:

It's not bridged, my previous incorrect schematic wasn't bridged, it's connected as two separate channels as is the corrected version.

The problem with the second schematic in the picture is that the long input return leads both go back to the power supply, rather than to the IC's 0V pin.

Do you understand both of the above statements?

Cramped is good.

Is each channel on a different board with the returns going back to either the battery or just one board? That could be your problem.

Originally Posted by audioguru

Your offset adjustment is in a positive only direction. What if you need a negative offset adjustment, swap the amp ICs?

No, if I remember correctly (and it's been 20 years) the DC outputs both sit at about 1/2Vsupply. I recall the TDA2003 has an internal resistive divider which always sets the output OP point to half supply. The trimmer is centered at that voltage and can swing above and below it. That means it can adjust the offset in both directions around the nominal value to get zero Volts across the speaker..

The "single ended" design centers the operating point at 1/2Vsupply, has a cap couple to the speaker whose neg side is connected to ground which means the max signal swing that the speaker can ever see is 1/2Vsupply. The bridge setup allows one output to swing up and the other to swing down at the same time, so the peak signal swing is Vsupply. However, I believe the DC bias (no signal) set point for both outputs on the bridge design is 1/2Vsupply. They can not both sit at ground because then the "negative going" side has no swing room.

Originally Posted by GooseElectro

I will have to post my wired circuit I am making a stereo amp not a briged amp ! Bothe circuits work fine on there own. Somthing to do with connecting them.

IMHO, the problem will disappear if you do this:

1) use a "single point" power ground for the point where both load (speaker) negative lines come back and tie to the power supply ground.

2) Run a wide and short trace straight from the single point power ground to EACH of the ground pins of the 2003's and DO NOT TIE OTHER COMPONENT GROUND LEADS TO THOSE TRACES ALONG THE WAY. This gives you a separate power ground lead to each IC's ground pin from the single point ground.

3) Tie the ground leads for the other circuit components together ant then connect them to the ground pin of each 2003 (do it on each channel) right at the IC ground pin.

4) The R-C compensation on the output (R4 - C3) should tie directly to the ground pin AT THE IC.

He's building it on Veroboard so I don't think that's an option.

This circuit really needs to be built on a real PCB.

You've got the inputs and outputs connected at one end of the PCB so it's no surprise it's giving you trouble.

Connect the input, output and return as close as you can to the respective IC and the problem should go away.