Hi,
I am using the TAS5630B in the DKD package for a stereo BTL application. I have the rail at +51V. The PCB layout is not exactly as the EVM module but we tried to closely match it within the constraints of the mechanical package. The output filters are using 10uH 7A Ice Component shielded inductors and 0.47uF/100V film capacitors. Everything else is the same as the EVM board.
The Problem I am having is when the device approaches a clipped output, there appears to be some sort of distortion (oscillation?) prior to clipping. One interesting point is that when I power up my board and apply the signal the output seems almost normal at first but then degrades quickly. If I remove the signal, wait a few seconds and re apply it, again this behavior is observed. Seems that when I let the PCB idle with no signal for a few minutes and apply the input signal intitally the output provides almost normal clipping but then quickly degrades.
If I allow the signal to dwell at the level where the distortion appears for too long, the TAS5630B will shut down or be damaged.
This is the second PCB layout. The first used the PHD package...which would not even get close to the 240W (4 ohm load) spec. The second PCB layout with the DKD package almost achieves the stated 300W at 10%THD+N but has this problem.
I have moved the output snubbers to the layer that has the ground plane in an attempt to reduce the inductance through the vias. I have also directly connected my +51V supply directly to the traces running to the TAS5630B PVDD. These changes seemed to make a small improvement but still the problem persists. I have changed the input to the +12V regulator from my low voltage (+15V) source to the +51V source with no real improvements. I did notice that when the distortion occurs the current pulled from the +12V increases from 75mA to over 100mA.
Below are two pictures of the output waveform at clipping levels. The first is the immediate application of the signal, the second is after after a few seconds of running time. Notice the degradation at clipping....this is my problem. Any Ideas?
