2
\$\begingroup\$

Using TI's "power designer", I've taken their 33630A and designed it for a 5v output. My actual circuit has a slightly different value for the divider to raise the voltage up to about 5.2v, otherwise it matches this design:

TI circuit designer

..and here's my design. I've reviewed it a few times.

my circuit

My circuit, and the PCB from it, are on easyeda. After assembling the circuit, my output is right around 1 volt and the chip is overheating. I can't figure out why! I've checked for short circuits, and I've done in-circuit tests to ensure the passives are approximately correct.

As a wanna-be-EE, I don't know where to go next. I have a low-end 'scope and a high-end meter. I'm not using this attached to anything (e.g., it isn't actually connected to a Raspberry Pi), and I'm supplying it with 12-16v. I also removed LED1 so I could rule that part of the circuit out.

\$\endgroup\$
  • 2
    \$\begingroup\$ Unrelated to anything, that "pokitmeter" thing looks like a handy thing for hobbyists dipping their toes into EE. Thanks for inadvertently bringing it to my attention. \$\endgroup\$ – Hearth Jan 22 '19 at 17:41
2
\$\begingroup\$

This turned out to be the answer for this specific problem: Double-check that the pad really is soldered down. All kinds of Bad Things could happen without that analog reference being where it belongs.

Going out one layer; when you're working with a prototype you can't assume that a problem is either a design issue (anywhere, schematic or PCB in this case) or an assembly issue. You have to check both. With assembly issues, you have to check everything -- it's easy to get a soldered connection wrong and have it look right. After doing it for a while you come to be an expert at noticing the subtle hints, but it doesn't come quickly. Any soldered connection that's hidden (like a BGA, or those thermal pads) is especially difficult. It's not uncommon for circuit board houses to x-ray boards to verify that all balls on a part are soldered, and that complete coverage between chip pad and PCB has been attained.

| improve this answer | |
\$\endgroup\$
4
\$\begingroup\$

You appear to be using a nice beefy inductor, so you've avoided that common noob mistake.

But layout is important for these switchers and yours doesn't much resemble the LMR33630EVM - particularly the trace between the IC and inductor, and inductor & output filter caps (which should all be short & fat).
Similarly, your input caps are too "far away" from the IC - in particular look at the path between pin-1 and the ground pins on those caps - it goes all the way around the edge of your board. Lots of vias would probably help a bit here, but rearranging things to bring them physically closer together with more direct routing would help more.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Thanks. I originally had one that was very close to the spec sheet's pattern but was part of a larger circuit. This was my simplified board to see if I could get it working in isolation. While I wouldn't expect it to be perfect, I wouldn't expect it to outright fail like this. \$\endgroup\$ – tedder42 Jan 22 '19 at 17:26
  • 1
    \$\begingroup\$ I agree with this, the OP should look at the switch node on a scope to see what's going on. Also could get closer to 5.2V output with 150K and 35.7K divider resistors. (Can find optimal divider resistors and output tolerance range at syncbuck.com ) \$\endgroup\$ – John D Jan 22 '19 at 17:26
  • 1
    \$\begingroup\$ Also double-check that the pad really is soldered down. All kinds of Bad Things could happen without that analog reference being where it belongs. \$\endgroup\$ – TimWescott Jan 22 '19 at 17:33
  • \$\begingroup\$ thanks all. I'll look at the SW pin and also make up another PCB in the next 24 hours; AGND could very well be the culprit, as I've desoldered and resoldered the 33630 chip on this PCB. If AGND was floating would that explain the chip overheating too? \$\endgroup\$ – tedder42 Jan 22 '19 at 19:05
  • 1
    \$\begingroup\$ the problem was definitely AGND. Now that my second try is working, I see the switch node going from 0-2V and the inductor is bouncing about 0-8v. So I definitely need to rearrange, but ultimately @TimWescott is correct (please post an answer..). \$\endgroup\$ – tedder42 Jan 23 '19 at 5:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.