Does anyone have experience/familiarity with the TDA Multiwatt 11 family of audio amps, specifically, the TDA7296A, 14W/ch stereo audio amp?

[DCook.1255]

 I've been able to get the TDA7269A to work as a standalone audio amp, but it won't work on a system board (or as expected).  The data sheet offers 3 different ways to configure for single-supply power and I've been thru them all several times, trying to disprove insanity:

1. 1K R's shorted to gnd - amp goes into shutdown. This works for my standalone amp.
2. 1K R's tied to gnd thru 5.6nF cap's - only get x2 gain.
3. 1K R's tied to Pin 9(Ref) - Lower half of the output waveform is lopped off.

I have no idea what to try next. Thoughts?

[Muhammed Güler]

If the signal is clipped in this connection, you can try to apply half of the supply voltage with high ohm resistors on pin 9.

With the current flowing through r4-r5 and r6-r7, the capacitor c2 should charge and carry the signal up, but this seems to be insufficient in your scenario.

[DCook.1255]

Hi Muhammed,

Thanks for the suggestion. I did try biasing (Vin/2) on Pin9 (Ref) using 2 magnitudes of values:

1. 18k R's - This did not seem to impact/change the signal. The bottom half of the waveform is still loped off.
2. 2.2k R's - This significantly suppressed the output signal, and did not change the waveform shape.

I'm am very confident the IC is good. I tested it in my standalone amp board before I installed it into this system board and it works fine. And I'm getting the same behavior out of this known amp as the previous one installed. So it has something to do with the way this Amp is installed in this system board.

Power input is from an External Supply @15V with a common gnd plane for all circuitry.

If you have any other suggestions I should look at, pls let me know.

Thanks,

Dave

[DCook.1255]

Ok, I'm not sure I've found the root cause, but I have found a solution thanks to Muhammed. Rebiasing Pin 9 did not work, however treating the Amp like an OpAmp does work. Block the DC input and rebiasing the input signals on each channel separately. The new DC bias point is not Vcc/2. I used a pot to find the point at which clipping begins equally top and bottom. A better approach may be to use a single fix 1% resistor on one side and a decade box on the other an dial in 1% values. Otherwise, if you have to convert voltage levels to resistor values and you may or may not hit the make. You may find that the new DC bias point(s) is(are) not the same for the 2 channels.