Amplifier chip - Severe oscillation on one channel

[Vincent]

I had a TA8258HQ amplifier chip laying around, along with the speakers the chip drove from day one inside a projection TV. The circuit I built has been on a breadboard forever (not the greatest idea, distortion levels go through the roof), and today I decided to transfer it on a much better suited protoboard.

I pretty much built the test circuit from the datasheet, with the addition of a couple of 470 ohm resistors to attenuate the gain a little bit, and a power on LED. Everything worked fine the first times, although I did notice the heatsink went unusually hot. Then one of the channels started to trip at power up, with a much lower output as a result and severe clipping at lower frequencies if I force the volume up.

This was the perfect opportunity to grab the oscilloscope I saved from scrapping and investigate a little bit.

At first I took a look at the supply. I see some of the switching from the transistor, altough I think both channels would oscillate if this was the problem. Unsurprinsly adding an inductor improved the ripple, but didn't fix the problem.

Then I measured the signals at the chip's outputs. One channel has around 0.2V oscillation, nothing alarming to me. But the other channel.... DAYUM! almost 2V p-p at around 400kHz  . The waveform is a severely distorted triangle.



There's gotta be something going on. Components in the feedback loop checked OK. I suspect a loose or broken connection somewhere, but any feedback on this will be appreciated. Maybe it's a really stupid thing (it usually is xD).

EDIT: Never mind, I knew it was something really trivial. Turns out the wire closing the feedback loop broke, no wonder the gain went through the roof! 

[mikerj]

These integrated bridge amplifiers can be very sensitive to PCB layout, the grounding in particular.  If your pre-ground and power ground are daisy-chained on your protoboard that would probably be enough to cause oscillation.  As a first attempt try attaching your PSU 0v connection to the other end of the daisy chain, or preferably re-wire the board so that the grounds are star pointed (even better, use a PCB with a chunky ground plane).

[Vincent]

Both the power supply and the input signal generate their own noise. Although now that I found the issue and took care of it, the chip works just fine, and I'm not even sure I can hear the remaining distortion at high power. But even then, it would top at less than 30W with the 32V 940mA power supply I'm using. I could push it to 40W if I wanted to 

Looking back at the datasheet, the connection that broke was a capacitor tied to the feedback loop, with the addition of a 470 ohm resistor I added to attenuate the gain as suggested. I assume the capacitor is there to filter this high frequency noise that would otherwise cause oscillation as you mentioned.

It doesn't seem to mind having the pre and power grounds connected together. But I did connect the speakers and power grounds the closest to the power supply. That's what common sense tells me to do LOL.