How to eliminate audio noise in my LED clock?

This is basically custom ATMega644p based talking Alarm clock I'm working on. I made PCB via OSH Park, added Adafruit's wave shield and Sure's 32x16 bicolor LED matrix for display. I'm now trying to amplify audio form WaveShield, so built a breadboard TDA2822 based amp (schematic straight from datasheet). It's getting a lot of (digital?) noise, especially when I'm scrolling text on the screen as you can see in this video:

Ignore Temperature, I don't have sensor connected yet :slight_smile:
This is kind of a show stopper for me, since it's supposed to be alarm clock, and I don't want it humming and making weird noises all night near sleeping person's ear. Since I'm somewhat new to electronics, can someone make a suggestion what I can do to eliminate idle humming and reduce digital noise? Perhaps I need some kind of gate or filter? I have no idea... Maybe there are inexpensive ICs that can do that?

Another idea I had is to turn off amp when no audio is being sent (via transistor connected to digital pin on ATMega chip perhaps?).

Without seeing your schematic it’s going to be hard for anyone to help you.

A gut feeling would be lack of or insufficient decoupling though.

tack:
Without seeing your schematic it’s going to be hard for anyone to help you.

A gut feeling would be lack of or insufficient decoupling though.

Sure, here it is.

Gotta clean up the wiring to start. Its a little antenna farm now, lots of radiating sources.
Also, try putting some caps across pins 5/6 on WS2, see if that quiets the power supply noise some.

CrossRoads:
Gotta clean up the wiring to start. Its a little antenna farm now, lots of radiating sources.

Not sure what you mean? :slight_smile: If I remove wiring it won't work...

CrossRoads:
Also, try putting some caps across pins 5/6 on WS2, see if that quiets the power supply noise some.

Cool! I'll give it a try, thanks!

I mean start making it into a neat assembly. The prototype works, yes? Now neaten up the wiring so there is not such a jumble of wires.
When current flows and signals change state, they emit energy, which is being picked up & amplified by wave shield.

CrossRoads:
I mean start making it into a neat assembly. The prototype works, yes? Now neaten up the wiring so there is not such a jumble of wires.
When current flows and signals change state, they emit energy, which is being picked up & amplified by wave shield.

Thanks! I made some progress btw. Looks like main issue was breadboard wiring. I rebuilt amp using LM386 chip and soldered everything to a perfboard. Now noise is almost gone, it sounds very acceptable.
But I still want a way to turn off/on power to the amp on demand from ATMega chip. Anyone done something like this? I presume NPN transistor should accomplish task, but I never worked with transistors...

It is hard to say about switching power as your schematic doesn’t show any amplifier.
If anything, a PNP transistor or P-channel MOSFET could be used as a switch in the power line to a LM386.

google ferrite beads.

CrossRoads:
It is hard to say about switching power as your schematic doesn’t show any amplifier.
If anything, a PNP transistor or P-channel MOSFET could be used as a switch in the power line to a LM386.

I didn’t have it there at a time,just added. Will it work like described with NPN transistor?

dhenry:
google ferrite beads.

Good idea about using ferrite! Thanks a lot!