loading

Noise in Amp from LED Strip & servo

Hi,

I am working on a project in which TL072 is amplifying an audio signal which is going to speakers & arduino analog input. Amp is powered by a wall wart with 12V 2A output. The same 12V is also regulated to 5V and used to power an AVR. The AVR is being used to PWM, using a power transistor(TIP120), an LED strip taking 1A from the same 12V supply. There 

1) Now the rapid switching of current for LED strip is messing up the amplifier output. It's working fine with LED Strip off but when using PWM, there is a high frequency noise in the amplifier output. I have a 0.1uF and 10uF caps near power inputs for both AVR and amplifier. This is probably due to insufficient decoupling. Can someone please help me make it better?

2) There is another independent source of noise I'm finding which I'm unable to explain. The audio is going into Arduino which is controlling a servo as well. As soon as the servo is connected, there is a constant loud noise in amplifier output. This remains even after powering the arduino and servo using seperate USB supply(only ground connected). I placed audio setup and servo data wire far apart since they may be interfering due to PWM used for servo, but it wasn't of any use.

Since in both cases, amplifier is producing noise, I think my amplifier diagram might be at fault. If I remove the amplifier, the audio stream reaches arduino without any noise . Anybody got any ideas how to remove these 2 noises?

Attached images: amplifier schematic, graph of audio input in Arduino analog pin.

-Antzy

Picture of Noise in Amp from LED Strip & servo
amplifier.bmp
sort by: active | newest | oldest
1-10 of 20Next »
gmoon3 years ago
Here's something to try...it's a bit of a long shot, if you're certain the noise isn't RF:

Add a resistor on the non-inverting input of the opamp--something between 33K and 100K. Coupled with the innate input capacitance of the opamp, it should form a low-pass filter. See these in tube amps all the time--it's called a "grid stopper," and it helps prevent high frequency oscillation...and picking up radio stations.

With the high input impedance inherent in a non-inverting amp it shouldn't attenuate your signal very much...
Antzy Carmasaic (author)  gmoon3 years ago
That might work but won't passing the input signal through a resistor decrease the volume of output as well?
It shouldn't effect the output much--a resistor in series with the input doesn't change the signal voltage (as a voltage divider would). It may limit current somewhat, but if the input impedance of the non-inverting opamp is high (and it should be very high, possibly 10M or even more) then very little current is required for the opamp's input to function well. 50K or 60K of in-line resistance here probably wouldn't even be noticeable.
It sounds looks like you didn’t isolate your circuits.

From your 12 volt supply have a diode go to each circuit and after the diode a large capacitor.

The diode keeps the capacitor charge from back bleeding so you don’t get a loading signal through your power supply.

If you don’t isolate your circuits one circuit creates a loading effect on the other.

So from your 12 volt supply you have two diodes separating the 12 volt supply to two large capacitors then one goes to the 5 volt regulator and the other to the amp.
The diodes didn't reduce the noise. But at the moment, I'm only working with servo and amplifier. They may help when I add LED strip later on.
what is the frenqucy of the noise
It "sounds" like 100Hz. I don't have an oscilloscope so I matched the noise to various frequencies created by a frequency generator app and 100Hz sounded just about right.
It sounds like your amp may have turned into an oscillator, Flip C4.
I uses a modified version of amanda's amplifier Instructable: http://instructables.com/id/Arduino-Audio-Input/
Will try reversing it. But I think the noise is now at a bearable level. I will put up the whole circuit on the weekend on a breadboard. Thanks for your help!
Put a lot more heavy caps near the motor supply connections, bypass them with 100n ceramics.

Run heavy current lines separately back to a common ground "star point"
1-10 of 20Next »