Have any of you used this DE-SW050 switching power regulator before? Did you require any additional filters or circuitry?

I picked up a DE-SW050 switching regulator thinking it would be a fairly easy and cost effective power supply for my project that involves an Atmel AtTiny84 microprocessor. As I understand the data sheets, it has capacitors on both input and output and should be able to run without any additional diodes or capacitors. I put it in my breadboard and the attiny ran funny. The program would mostly work (pot values seemed to read slightly differently than my bench power supply), but the chip would become unresponsive after a few second and require a reset. Thinking maybe I was pulling too much power driving these relays (through a transistor), I put my amp meter inline and power consumption never got above 100 mA. And in fact with the amp meter inline the chip behaved more normally (amp meter acting as a filter?). Although after my test run, the chip stopped responding and was burned out, so I had to toss it and flash a new chip. Fortunately they are really cheap and I have a bunch to play with.

So I'm missing something vital here. Do I need some more filtering on the output power?

Thanks so much for any advice!

 

Add a 22μF cap between power and ground on the breadboard, add another 0.1μF cap (Ceramic, preferably, though metalized film type will work) between Vcc and Vdd on the uC as close to the uC pins as possible, with as little wire between the cap and the power lines. See sigline on bypass caps.

See if that helps.

 

I will try that. As I stated, the regulator has caps already on both input and output. Part of the package (I can see them under the shrink-wrap). I will try adding external regulators though. I'm interested in anyone who's used these switching regulators specifically from dimension engineering.

 

Put a scope on the power rail, AC coupled, and see what it looks like.

 

Yes I did have that thought but forgot about it. I will do that. What do you mean by "AC coupled?"

 

It should be an option on the scope. It blocks the DC bias of 5V, allowing you to see only the AC Ripple that's "riding" on the DC voltage at a higher scale (instead of scaling for 5V)

Though if you don't have an AC Coupling option, expanding to low volts/div and lowing the trace so the top is visible, you'll see the exact noise as well.

Another option is to put the probe through a capacitor, though that will sometimes cause impedance issues, hence the internal compensated capacitor coupling in "AC Coupling" mode.