Hi this circuit (see attached image). Where Im switching some leds about 3v drop on each led. So the current is about 35-40ma. The power I'm using is a power switching wall adaptor 9v with 150ma max. Now I'm swthcing the leds at 117hz, and varying the duty cycle from 0 to 50% and I see oscillations/ripple match to 117hz when probing the 9v supply around 400mv. My main concern is that this 9v goes to a linear regualtor where i get my 5v, when probing the 5v for power my microcontroller is stable not ripple on the otuput and the regultor does not get hot or anything, but will this ripples/oscillations of 400mv be bad on the long run, and create problems with my stable 5v? Second, how can this be fix, I can only increase the frequency up to 200hz but its the same, I tried placing a 100nf capacitor but still the same. This oscillations seem to happend becuase of the draw of current on and off, I could lower the leds current to minimize the effect but that will dim my leds. I can not change my power supply, I'm stuck this this one? And ideas thank you for the help.

 

Forgot to mention that I'm using a LM-341-TO252 for my linear regualtor 9v to 5v.

 

Maybe you should try using an LM7805 linear regulator, specifically designed to have a 5V output? The LM341 has a variable output with a minimum of 5V (so you are nudging the limit) and 500mA whilst the LM7805 works up to 1.5A. You don't mention the processor you are using, is it powered by 5V or does it convert the 5V to 3.3V on board? And I'm not sure what you mean by "the long run" - if it works, I'd suggest it should go on working. I expect increasing the PWM frequency would be a good idea and I wonder, if you include a parallel resistive circuit to mimic the LED current, turned on when the LEDs are off, would this reduce your ripple?

 

Yes, the regulator that I have is a fix 5v output version. I have tried the lm7805, but still the same. But the total current consumption is only less them 40ma. What I mean by long run is will somehow at some point my linear regualtor will fail and start to pass through and affect my steady not ripple noise 5v output, right now even thought 9v have these 400 to 500mv oscillations the 5v is stable.

 

It is temperature that kills the regulator not ripple voltage.
The spec for PSRR ( Power supply rejection ratio) tells you how much ripple it is able to remove without passing it through to the 5V supply.
Generally, the variable ones (LM317, for instance) are better than the 7805, probably just because the 7805 is an old design.