LM48511: Should I use filters on the output of a filter-less Class D amp?

Hi, Randy,

Based on the details of the LM48511 datasheet, it should be able to support the test conditions you mentioned. I can actually handle 5.4W with a good THD+N level of 1% and a 4ohms load (VDD = 5V). And 6.7W with a THD+N level of 10%.

Due to the mechanical conditions of the speakers, the Class-D switching frequency should not be a problem. Actually, the mechanical characteristics will omit the high switching frequency, leaving only the audio content.

Have you taken some measurements of the output and see if the output power and THD+N are accordingly to the datasheet. You would need to use an RC filter for Class-D output measurements.

https://www.ti.com/lit/an/sloa107/sloa107.pdf

Please let me know if you have additional questions or comments.

Best regards,
Luis Fernando Rodríguez S.

Hello Luis,

Thank you for you response.

Ended up building the recommended filter circuit to measure the output.  Once I did this and started my measurements, the amp died.  Our TI rep. sent a second eval board.  That too died during testing.

Before I start having our repair/surface mount techs. replace the IC on the eval boards, are there any known issues that could be causing this?  i.e. signal in max., can't change resistor gains while powered up, the has an issue with having both a 4ohm speaker and the above RC filter circuit in place?

Modification on the eval board are,

Installed low pass filters before the CIN +/- caps.

Change CIN +/- caps to .1ufd,  R5/7 to 10k 1%, and R6/8 to 29.8K (to increase gain and adjust filter to a lower freq.)

On the first board I installed a 330ufd tant. cap. on Vdd.  On the second board, this was omitted.

Boost, Amp, EN, and FB = High.

Equipment used:

DSO with single ended input.  Measuring across the RC circuit OUT +/- , with the RC filter ground tied to DSO's earth ground.  Unit Under Test does not have earth ground.

Any info would be greatly appreciated.

Thanks again,

Randy

Hi, Randy,

Just to confirm, is the EVM also damaged when no loads are connected or when you use higher loads? The settings you used seem to be correct, so I'm checking if there could be an overcurrent event or something that could damage the device.

Best regards,
Luis Fernando Rodríguez S.

Hi Luis,

Thanks  for getting back to me.

The Eval boards were damaged with a 4ohm speaker connected in parallel with the recommended RC filter (sloa107.pdf).  I was attempting to achieve max volume before distortion when failure occurred.

Thanks again,

Randy

Hi, Randy,

Thank you for your response. Are the eval boards damages when the RC filter is not connected? I mean, if you only use the 4-ohms load, are the units damaged?

Best regards,
Luis Fernando Rodríguez S.

Hi Luis,

Thanks for your prompt response.

I had one of our surfacemount repair techs. replace the IC on one of the EVAL boards. So far with the new IC in place and the RC filter not installed, just the 4ohm speaker, I have yet to damage it. (Knock on wood.)

Thanks again,

Randy

Hi, Randy,

Do you mean that the RC filter is not installed and the EVM is working properly with the 4 ohms speaker?

Best regards,
Luis Fernando Rodríguez S.

Hi Luis,

That is correct.  The RC filter is not installed, just the 4ohm speaker is connected to the amp.

Thanks,

Randy

Hi, Randy,

Thank you for your response.

Can we close this e2e thread? Or do you have any other question or comment about this case?

Best regards,
Luis Fernando Rodríguez S.

Hi Luis,

Might as well close it.

From what I saw with the RC filter failure, it is not recommended to put a filter on the output of the amp (before the speaker).

Regards,

Randy

Hi, Randy,

Understood. We will perform some testing from our side in order to see how related is the RC filter with the failure. Thank you for your feedback.

Best regards,
Luis Fernando Rodríguez S.