We have now test using the TPA3126EVM and we can measure the same high THD+N hump we measure on our design. If we change the class_d switching mode from hybrid to BD the THD+N profile changes and improves slightly.
Hybrid mode is the setup stated in the datasheet and the THD+N figures stated for 6KHz are much better. Our setup is 26V PVCC and 4Ohm load so we have compared against Figure6, page9.
Also see TPA3116EVM THD+N test which is much better and inline with the datasheet. Because the TPA3116EVM measures ok we believe our test setup is ok.
For reference our AP measurement setup:
- Differential input. No GND on input terminals.
- Balanced float output config. from AP.
- Chassis GND from AP to PGND on EVM.
- 4ohm load.
- Differential output from 4ohm load contacts into AP input.
- 22-22kHz BW.
- LE03 PSU 26V without cap-board. Sweep limited to below 25W output power.
- AM pins GND as per EVM default, same as LE03 mainboard config.
From our testing, because the THD+N hump is measured on the TPA3126EVM, we believe the problem may be on the amplifier IC on with our design implementation?
Best Regards
Rob
