LM4946: Having issue with harmonics.

Ravi Shrestha

Part Number: LM4946

If I give 1 kHz sinusoidal signal from function generator to the input (MONO_IN+) of the amplifier (MONO_IN- to GND), and analyze the sound on the handsfree speaker using spectrum analyzer, I can see 1 kHz, 2 kHz, 3 kHz and 4 kHz . While 3KHz is understandable as sinusoidal signal form function generator are not always perfect and will produce odd harmonics, I can't figure out how 2 kHz and 4 kHz is appearing.

I used the same function generator, power supply and speaker on a different amplifier circuit, and this problem goes away. This points out it is definitely a problem with the amplifier (LM4946), circuit surrounding it, or the PCB routing.

Can anyone steer me to the right direction please? I'm suing the same circuit as advised in the datasheet of the amplifier, which I have laid out in my own PCB. Development board from TI is obsolete, so I can't conclude the PCB routing as a reason easily.

over 5 years ago

0 kai klaas69 over 5 years ago

Guru 140200 points

Hi Ravi,

I would repeat the distortion measurement with 8R and 32R resistive dummy loads. This would rule out the influence of loudspeaker on the measurement.

Kai

0 Luis Fernando Rodriguez S. over 5 years ago

TI__Guru* 84915 points

Hi, Ravi,

Welcome to E2E and thank you for your interest in our products!

Our team is taking a look at this. We will reply as soon as possible.

Best regards,
Luis Fernando Rodríguez S.

0 Luis Fernando Rodriguez S. over 5 years ago in reply to Luis Fernando Rodriguez S.

TI__Guru* 84915 points

Ravi,

Do you get the same results if you connect the function generator to the line inputs (Lin or Rin)? And when the unused analog inputs are connected to a 0.22uF - 1uF capacitor to GND?

Best regards,
Luis Fernando Rodríguez S.

0 Luis Fernando Rodriguez S. over 5 years ago in reply to Luis Fernando Rodriguez S.

TI__Guru* 84915 points

Hi, Ravi,

Excuse me, do you have any update on this?

Did you get the same results if your use the Audio input RIN and LIN? Do you get the same harmonics in the input lines? I mean, have you tried to measure the input signal and verify if the harmonics are generated even before the amplifier section?

Please let us know if you have more details about this. This would help us to have a better approach to this issue.

Best regards,
Luis Fernando Rodríguez S.

0 Luis Fernando Rodriguez S. over 5 years ago in reply to Luis Fernando Rodriguez S.

TI__Guru* 84915 points

Hi, Ravi,

I will close this E2E thread for now. But feel free to provide additional feedback in this post in case you have any information. Also, you can create a new E2E thread if you have a different issue.

Thank you.

Best regards,
Luis Fernando Rodríguez S.

0 Ravi Shrestha over 5 years ago in reply to kai klaas69

Prodigy 60 points

Hi Kai

The distortion is still present with the 8R and 32R resistors. When there is no load, the harmonics disappears.

Ravi

0 Ravi Shrestha over 5 years ago in reply to Luis Fernando Rodriguez S.

Prodigy 60 points

Replying to "Do you get the same results if you connect the function generator to the line inputs (Lin or Rin)? And when the unused analog inputs are connected to a 0.22uF - 1uF capacitor to GND?"

Hi Luis

Yes, I get the same result when I use the line inputs (Lin and Rin) and ground the unused analog inputs through 1uF cap. I also get same result when I use the headphone output (Lout and Lout) with both configuration (Capacitor-less and Single Ended). I tried all possible combination of input and output, but the result is same.

Thanks
Luis

0 Ravi Shrestha over 5 years ago in reply to Luis Fernando Rodriguez S.

Prodigy 60 points

Hi Luis

Harmonics is only present in the output, never on the input. I have verified this on all possible combination of input (MONO, LIN, RIN) and output(MONO, ROUT, LOUT).

Regards
Ravi

0 kai klaas69 over 5 years ago in reply to Ravi Shrestha

Guru 140200 points

Hi Ravi,

is your supply voltage clean and stable?

Can you show us a schematic and your layout?

Kai

0 Ravi Shrestha over 5 years ago in reply to kai klaas69

Prodigy 60 points

Yes, the power supply is clean and stable. I have removed the voltage regulator in the circuit and applied power from a voltage source directly as well. Also tried out with all the components removed beside the speaker circuit and microprocessor.

Ravi

0 kai klaas69 over 5 years ago in reply to Ravi Shrestha

Guru 140200 points

Hi Ravi,

I see two issues:

1. You should not split the ground plane. If you do so, you have to filter every line you route across the gap. But this would make the circuit very complex and expensive. Splitting the ground plane is only helpful in very limited applications. Here, I would not split the ground plane but would use a solid ground plane.

2. You must heavily reinforce your GND wiring arround the LM4946!!!!!!! Use a ground plane or at least large ground fills also on the top layer of PCB (in the "red" plane) and connect all the ground pins of LM4946 to this ground plane (by using heat traps for better soldering, of course). Use as many vias as possible to connect the "red" ground plane to the "blue" ground plane.

Kai

Because of the Thanksgiving holiday in the U.S., TI E2E™ design support forum responses may be delayed from November 25 through December 2. Thank you for your patience.

Audio

Audio forum

LM4946: Having issue with harmonics.