The PAM8403 is a dual channel amplifier, but I only need one channel. I want to utilise both channels as one, in order to get a single channel with higher output power.
I have been reading through the PAM8403 datasheet, but can't find any hints on whether this was recommended or not.
here's the response I got after emailing Diodes:
We would advise NOT to connect the two outputs of PAM8403 in parallel. We recommend two ways to get higher power :
a ) increasing the supply voltage, from 5V to 9V for example---PAM8320 working at 4.5V to 15V is a good candidate for this solution.
b) reduce the loading impedance, from 4Ohm to 3Ohm for example---PAM8406 can drive as low as 2.5Ohm/Ch that make PAM8406 is a good choice.
Conclusion:
You cannot bridge the PAM8403 (as you can't bridge two pairs of differential outputs), and Diodes have recommended against using the two outputs in parallel.
the outputs are 250KHz pulse modulated square waves, if these signals were put through inductors on each pin, the +R and +L could be commoned and ditto the -R and -L as long as the left and right inputs are also commoned. I have not tried this yet. Ferrite beads may be enough.At the worst it will overheat or blow up.
Is it worth it I ask? Even if you could easily and efficiently bridge the two outputs you only (at best) get double the power - that's a 3dB increase in power and ditto for sound pressure level - equivalent to about a 30% increase in loudness. So, is it worth it?
You could use two 1:1 transformers - primary_one connects to +OUT_R and -OUT_R and primary_two connects in the same way to the left channel. Then you wire the two secondaries (anti-phase) in series and connect the ends to the speaker. You'll get twice the voltage out but you'll need to use an 8 ohm speaker to prevent overloading the chip. This means power doubles.
You could argue that the transformers are small and could be useful components in the EMC filter (needed for class D) but I don't think it's worth it.
