Audio noise filter advice

[codlink]

I have a MP3 Trigger V2 http://www.sparkfun.com/products/11029 (Manual)
With a Mono Amp: http://www.sparkfun.com/products/11044 (Datasheet)

But I get alot of static in the speaker. 

Right now I am powering the Trigger with the 5v supply from an Arduino.
And it's wired on a breadboard to a .5W 8Ohm Speaker.

I do not know much about audio so I thought I would ask here on advice for a filter to cut down the static.  If you need more info please ask as I do not know what info is needed.

Thanks!

[Magician]

There is a note on amplifier web page:

Note: The amplifier's class-D design outputs a 250Khz PWM-like signal that is restored to an analog voltage in the speaker's coil. This is what makes the amplifier so efficient, but because of the switching frequency, you should keep the amplifier as close to the speaker as possible to minimize possible interference.

Can you test just amplifier, applying audio (music) from other sources, mp3 player, PC  or radio?

[Grumpy_Mike]

It is not the audio you want to filter but the power supply. You need lots of decoupling on it.

[codlink]

I have read quite a bit about decoupling the power supply from the audio.  Would it be beneficial to use a separate power supply for the amp?

If you could give me some advice on how to decouple my circuit (maybe a starting point for experimentation) would be most appreciative.

[Grumpy_Mike]

Yes a seprate supply would help, remember to join the grounds.

To start with have you amp powered through a 10R resistor with 100uF caps on each side.

[MarkT]

Would it be beneficial to use a separate power supply for the amp?

If you could give me some advice on how to decouple my circuit (maybe a starting point for experimentation) would be most appreciative.

Best not to use a switch-mode supply if possible.

Decoupling:  something like 2,200uF electrolytic, 10uF electrolytic and 0.1uF ceramic (some of these will already be on the board).

One glaring issue is feeding a 48kHz sampled audio signal into a 250kHz class-D amplifier, lots of heterodyne whistles to be expected - the class D clock should ideally be derived from the sample clock.  I suspect a multipole low-pass filter is needed on the audio signal to eliminate the sample clock.

Also note the amp's input is differential so you must pass both audio signal and ground wires between modules for the audio as well as common supply ground.

[codlink]

I have the amp on a different supply (switched) for testing, but will use simple batteries when the project is complete.  Just using a different supply has taken out most of the static I was experiencing.  But I still get a little static.  Right now I am using low power amps and such to get things working and to learn.  I will be using a bigger amp and speakers to get more sound in the future.

I am seeing that there are multiple ways to de-couple.  Could anyone be willing to make a schematic that I would need in my situation?  

The help is much appreciated.

Also, I am not seeing any heterodyne whistles.  The grounds are connected expect for the switching power supply I am using for the amp.  

Another thing I am noticing is alot of bass.  The audio is just a female's voice.  But the sound seems muffled, but "muffled" seems like a harsh word for it.

[Grumpy_Mike]

The grounds are connected expect for the switching power supply I am using for the amp.  

All the grounds need to be common.

Could anyone be willing to make a schematic that I would need in my situation?  

Only if you post the schematic of what you have can we suggest where to put the decoupling.

[codlink]

Attached is an image of the setup from fritzing.  It does not show the ground wire for the audio.  Just the + signal wire.  But there is an audio ground wire running from the MP3 Trigger to the OpAmp than out to a speaker.

Thanks for the help.

[Grumpy_Mike]

OK, for a start get the ground wires right. All the grounds from the various modules must meet at one point. This is called star grounding. Do the same for the + wires. Then put a large cap between these two points. In the +ve lead to the amplifier put a 10R resistor and another large cap across the supply close to the amplifier.

[MarkT]

Ah, normally true, but the amp has differential inputs so you want to run two wires from the mp3-trigger audio jack direct to the amp board's inputs.  That way no power supply currents flow on the audio ground lead.

[Grumpy_Mike]

but the amp has differential inputs

Makes no difference to the supply decoupling though.

[MarkT]

Indeed.  To be clear those decoupling capacitors go across the supply (from +5V to ground), electrolytics the correct way round and rated at 6.3V or more, and close to the boards, supply wiring should be short if possible.  The boards should all have 0.1uF ceramic or similar already of course so don't worry about that.  Mains hum is the hardest to get rid of as its a low frequency - but the differential audio signalling should really prevent this being an issue.  Make sure the 5V supply is powerful enough to drive the amp at full load (probably needs upto 0.6A peak).

[codlink]

I have implemented your suggestions and didn't hear a noticeable difference.  Attached is an updated pic. 

The "muffled" sound maybe just the tiny speaker I'm using..