Author Options:

How to reduce signal noise in a circuit (for Amplifier)? Answered

I am building a sound reactive RGB light. Using a LM386 and a 16f88's AdC. I got it to work great with just 1 RGB LED on a breadboard, powered from a 7805 I then built it on perfboard, watched the amp's output on my O-Scope, worked fine.

Used this schematic for LM386

Once I hooked up 3 Constant current regulators with 3x 3 watt LEDs(red-green-blue) through 3x MOSFETS, the Amp's output signal becomes blurry/chaotic and the light malfunctions. The regulators negative output must be connected to logic GND, which i know to be where the interference is coming from.

How can I reduce the noise coming out of the Amp? Even if I gotta use different type of Amp. I don't know to much about them.


For the schematic, I am using only the LM386 parts, with a 10uf capacitor across its VDD & VCC instead of a 100nF. Trial and error led me to find a 10uf worked better.

As for running the Current regulators through FETS, I have the light(with the 3x 3 watts/3 regulators) running on an auto color-fade mode, and it works fine.

The current regulators ground have to be connected to logic ground to be able to drive the FETs.

Current Regulator positive ==> LED+ out LED- ==> into FET drain ==>FET source is common ground for Logic & 3 current regulators negative

Erm....Constant current regulators driving an LED that is in turn being modulated by a FET? That seems like a big no-no to me. That in itself could cause a great deal of problems. Maybe not at very slow modulation speeds, but at audio frequencies that constant current regulators may not be able to cope with the forced modulation of the drive current and are spiking to try to drive the current back to "normal"

What part of that schematic are you useing? All of it? Can you post your schematic? You could pm a copy to me if you don't want to post it. I'm not sure why the regulators have to be connected to logic GND, I'm hoping to see that from you schem. The regulators may be injecting garbage back into the system and you probably already thought of that. I've been working with LM386's lately and one thing I've found is that if run low on voltage or current they whack out. I spend several hours trouble shooting a small amp I built. 500 hz. sine wave in and on the scope the trace was erratic. I happened to hook it to my new variable power supply and turned it up just a little and all of a sudden it cleaned up.