Other Parts Discussed in Thread: TPS51120
Hello, I have this audio amplifier installed on a PCBA and have the /SHDN and GAIN pins floating for always enabled and lowest gain setting. However, when I apply 5V to the the circuit, I do not see the boost converter running, nor for a 100mV to 500mV peak-to-peak single-ended audio input do I see any output. The IN+ is AC coupled via 1uF capacitor as is the -IN input to SGND via a 1uF capacitor. I have series 49.9 ohm resistors to each input from the AC coupling capacitors and a 100nF differential filter capacitor with 10nF capacitors to GND for high frequency suppression. The OUT+ and OUT- outputs are DC coupled to the TDK PiezoListen speaker with 10 ohm resistor in series with OUT+ and 22uH inductor in series with OUT- and a 1kohm 4.7nF series snubber on each output to GND. All of this follows either the data sheet circuits or those from the TI evaluation board for the most part. The boost circuit consists of a 10uF input capacitor with 1uF at the VDD pin, 1uF at the VBST pin a 470pF at the VAMP pin, the boost inductor is 4.7uH rated for 1.9A and a NRVBSS13HE Schottky boost diode. The top side copper and ground layer copper is almost identical to that of the suggested layout.
Given this part is a BGA, and the /SHDN and GAIN pin pads are not intended to be connected and inside the array, I cannot verify that they are in fact floating without an X-ray test, nor is easy to tell if the outer ring of balls are all soldered properly. I want to confirm that floating the /SHDN and GAIN pins is valid in all configurations?
I am also wondering just how sensitive this part is to voltage transients? The data sheet gives 6V absolution maximum, but if there was a transient above 6V, how high must this be (or at what energy) for an immediate failure--say gate rupture or at what avalanche current or energy? (Kind of a loaded question, I know.) The 5V rail is generated by a TI TPS51120 in current mode using a 82mohm current sense resistor with an intended 1A maximum load, 150uF of ceramic output capacitance, and a compensation network of 10pF in parallel with series 6.7kohm 2.7nF. This rail seems stable and well controlled during power with very low ripple (<30mV).
