Power supply problem / PWM

[Morcego]

Hi All,
I have a problem using (what I considered) a simple circuit.
I use a microcontroller that generates a PWM signal that feeds a gate of a N-channel Logic level MOSFET. When I switch the LOAD on and off (without PWM) the 5V (that powers the micro) is perfectly stable. When I use PWM to switch the load the power changes as you can see in the attached picture. The micro of course goes crazy from time to time.

-The load is a LED board powered at 12V 25 groups of 5 leds with current limiting resistors.

- The 12 V are provided by a bench power supply.
- The 5V are regulated on board using a 7805.
- No gate resistors
- I have all decoupling caps in VCC /GND
- Changing PWM frequency as no effect.

Can some one help please?

Thanks in advance.

 

[MacGyverS2000]

Your switcher needs time to react to changing current requests... when you turn on a ton of current-gulping LEDs, the switcher is not able to give you immediate current. You need to either consider multiple phases or give that switcher some serious Farads for cushioning the blow your giving it.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[rodrigof]

Try powering groups of leds at a time, how much power rated is the power supply? what is the PWM freq? are using PWM to just dim the panels?

let us have a look at the schematic

regards

rodrigo

 

[itsmoked]

Either your bench supply can't hack the step load function you're hitting it with, or your leads are too long between the PS and your board.

I have had both problems occasionally.

Put a large >1000uF cap right next to your board in parallel to to your supply lead and lets see what transpires.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[itsmoked]

Man I hate when I respond with a short answer to a zero-response-so-far question and multiple responses have already occurred..

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[Morcego]

I did what Rodrigo suggested and the problem was less evident...

The large capacitor at the end of the power supply leads (at the input terminal of the 7805) solve the problem. Also I have increased the value of the capacitor at the ouput. Now I have a clean VCC at any duty cycle.

Learning every time...

Thanks for your help guys!

 

[GonzaloEE]

You might put an anti-parallel 1N400x diode between the IN and OUT pins, in order to bypass reverse surge currents from your decoupling caps, if the input pin were suddenly shorted.
I killed a couple of 78xx's this way.

Good luck!

 

[IRstuff]

One problem that you might also have is with Miller capacitance coupling from the MOSFET to the micro. This might cause the micro to try and sink more current than is specified for the output, and the current surge through the ground inductance of the micro might also cause upsets in the micro.

TTFN

FAQ731-376