Voltage Regulator not working 1

[Elenappli]

Hi, I have just gotten my first pcb printed and back to me to start testing. I am having trouble figuring out what why my voltage regulator isn't working, it should be outputing 5V. I am using the LT3663

http://www.linear.com/product/LT3663

I have attached my schematic. One change is that the C7 is really 1uF and not 22uF(There was a mistake in that footprint so I had to size down the cap.) If that is what causing it not to work? If so could anyone suggest a common way fix the board by hand so I don't have to order a new one. Thanks in advance.

 

[IRstuff]

Is this for school?

TTFN
faq731-376

 

[Elenappli]

No... I guess I am just wondering what the effect change C7 value had on the output of the regulator. From reading the data sheet it said that a cap larger than 10uF was better to achieve low output ripple, but I am not sure. Also since this is my first pcb I was wondering if there are any general tips for debugging it and any hacks to fix them so I can test the rest of my board with proper power. I am sorry if this sound like a hw question. I have checked for shorts between my power and gnd. I have check my cap and res values. I am just trying to understand the effect the C7 has on the output of the reg. Thanks

 

[IRstuff]

Well, since you've not even described how it "isn't working" there's not much to go on. The output capacitor is an integral part of the filtering mechanism, so as a guess, your circuit is 22 times worse than it's supposed to be. As an engineer, that shouldn't really be that hard to make that leap, given that the datasheet provides a functional block diagram of the regulator.

TTFN
faq731-376

 

[DRWeig]

Yes, Elenappli -- just exactly what does "isn't working" mean? Does that mean you see an output voltage other than 5? Or is it totally dead? Fluctuating with load? Did you put a socpe on the output (please upload pic if so)?

We can't read minds through the internet (but I'm guilty at times of trying).

Good on ya,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[itsmoked]

To get around your cap problem you just stack more caps on top and solder them just like 'they' were sitting on board pads. Instead they are sitting on top of their neighbor.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Compositepro]

Don't they still sell alligator clips?

 

[Comcokid]

Given that this is a 1.5 MHz switcher, power/ground planes will be critical. Trace inductance to your capacitors is critical. One rule-of-thumb is that a trace has 1uH per mm. Now, this rule doesn't take into account trace width, but lets suppose that your 1uF capacitor is ideal (has no inductance in itself), and that it is located 5mm from your IC. Then, the trace inductance from the +5 and ground is 10mm total giving 10nH of inductance. The resonance frequency for 1uF and 10nH is 1.59 MHz! So, you effectively have no capacitance! Now, with a via or two, and the (maybe) 1nH lead inductance of a SMT capacitor, the resonant frequency could be well below the 1.5 MHz switching frequeny, so your capacitor may actually act like an inductor! Alligator clips will not help.

Not to mention that when you pick one of these type of MHz range switchers your 6.8uH inductor has to be rated for the frequency. Now, if that inductor uses a powered-metal core rather than a ferrite (and it will have to be a ferrite rated for the frequency because not all ferrites go that high) then your inductor may be acting like no more than a few turns of wire. Given skin effect for the frequency, it may be acting like no more than a wire-wound resistor.

Now, if you do have good ground and power planes, and you selected your components with prior experience in switch mode at MHz range, then I would check that your pin 9 (ground pad under the part) was properly reflowed to your PCB - a common assembly-type issue.

These MHz range switch-mode parts are neat. But unless you take into account the component and layout limitations they will cause you to prematurely age. Unless you need the super-small size that comes from having a very small inductor, I suggest you choose a part that switches at a lower frequency.

 

[Comcokid]

I meant to say 1nH/mm for the rule-of-thumb.