6
\$\begingroup\$

I’ve been working on a project making a lab bench power supply. I designed the schematic, the layout, got the PCB made by OSH Park and ordered all of the components separately from digikey and have been assembling the components by hand.

The general block diagram goes AC to DC converter using a full bridge rectifier with an isolating transformer, which goes into a shunt FET, leading into a synchronous buck converter (a TI IC), which goes into a high current LDO, which finishes at the output. The shunt FET is driven by an op amp and acts as the current limiter. This whole setup is meant as a max 25V and max 5A adjustable power supply, where the buck-LDO relieves the need of a big heat sink but gives relatively noise free power. Most of the PCB uses SMD components.

When I was putting components down for my AC to DC converter, I noticed some arcing that occurred on the undersides of some of the capacitors. The arcing was shorting the power rail and was clearly visible and could be smelt. Desoldering the capacitor and cleaning the ash residue that was left would fix the problem. The voltage on that line was a max of 35V depending on the load, and I though these components were big enough to not cause an arc (package sizes ranging from 0402 to 1210).

I initially thought this might be because my soldering iron was too high of temperature and when soldering the components with flux, there would be some solder splashed around the components and that could cause a short. However I reduced the temperature and have been cleaning my board religiously and the same problem persists as I’ve been finishing the buck converter.

Another note is I literally saw a black line spreading between ground and the power rail when I was debugging the arcing problem for the AC to DC converter. It would take around 5-10 seconds for the line to connect ground and power and when it did it would start arcing. The black line looked like spreading lightning.

Any ideas on how to avoid this?

EDIT: Here are is the schematic of the AC to DC and buck converter respectively:

enter image description here

enter image description here

The arcing also occurred when I plugged the thing in, not while soldering. This has only happened on the power-GND caps so far.

\$\endgroup\$
  • \$\begingroup\$ Without a schematic and with power off I imagine the only aource of an arc is from a faulty iron. Be careful on isolating parts from grid voltages from stray leakage . This includes surface contaminants (creapage) and stray cap. Leaka at frequency of current. Air gap slots of 2.5mm are used for this reason. \$\endgroup\$ – Sunnyskyguy EE75 Jan 2 '18 at 7:40
  • 4
    \$\begingroup\$ There is no way that 24Vrms is literally "arcing". You have a different problem with conductive contaminants on your board. You're going to have to be a lot more specific about how you're assembling it and the materials you're using before we'll be able to help you. \$\endgroup\$ – Dave Tweed Jan 2 '18 at 13:28
  • 2
    \$\begingroup\$ "This happened when the board was powered."... So, you are soldering the components to the board while it is powered? \$\endgroup\$ – Tyler Jan 2 '18 at 13:37
  • 2
    \$\begingroup\$ Yeah as the others said, add pictures of the damages. This is not arcing. \$\endgroup\$ – Vladimir Cravero Jan 2 '18 at 13:42
  • 1
    \$\begingroup\$ Thanks for the responses, after more reading I think I am using the wrong flux and I'm having containment creepage. This is my first hobby project so I'll probably have other dumb mistakes, thanks for helping out! \$\endgroup\$ – Michael E Jan 3 '18 at 9:17
3
\$\begingroup\$

-Edit. Woops, forgot references...Information shared by Bob Willis during a PCB Webinar hosted by Vision Engineering Eliminate Printed Circuit Board Problems & Failures

-Edit. And for the Slide-Show

Sounds like contamination and voltage tracking; as a function of the contamination and voltage potential, dendrites are forming. See attached.enter image description here

\$\endgroup\$
  • 2
    \$\begingroup\$ Yup, this was definitely the problem, thanks a lot! \$\endgroup\$ – Michael E Feb 8 '18 at 18:14
  • \$\begingroup\$ Why doesn't the dendrite burn up from the current flowing through it and cease to be a problem? (Apologies if stupid question.) \$\endgroup\$ – Sean May 28 at 1:31
  • \$\begingroup\$ Vaporized conductive material condensing again on something nearby might not help.... \$\endgroup\$ – rackandboneman May 28 at 23:04
2
\$\begingroup\$

If you are using the wrong kind of flux (eg. acid plumbing flux) this can happen.

Use only electronic grade solder and flux materials. Solder should be 60/40 or 63/37 Sn/Pb if lead-based and the flux should be something like RMA rosin.

\$\endgroup\$

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.