I'm working on a basic application with a microcontroller controlling several components including a:

  • 12V DC pump (Darlington array)
  • 12V PWM fan (Darlington array)
  • humidity sensor (I2C)
  • and LED controllers (SPI)

Online circuit can be found here.

I've setup and tested each component individually to ensure they work given my circuit. From there, I transitioned from breadboard stage to a more intermediate stage using a protoboard along with various wires soldered in the right place.

After placing everything into a small housing, I recently went to connect everything only to end up frying the microcontroller and the UBEC. Since all the components were previously working, I'm not certain exactly what caused it except for perhaps a short caused by my mess of wires on the protoboard.

enter image description here

I know general questions like this are generally not within guidelines, but I'm just looking for insight into whether I'm completely off the mark somewhere.


Per the comments I added:

  • diode to protect the PWM pin on the fan (although I'm not certain I actually need this - I figured the circuit built into the fan would already have this type of protection; unfortunately their datasheets don't share the circuit)
  • fuse on the +5V & +12V rail
  • power description on fan & pump
  • added link to schematic

enter image description here

  • \$\begingroup\$ You didn't mention what kind of power do you need for the pump and the motor. \$\endgroup\$ – Ale..chenski Nov 17 '16 at 4:34
  • \$\begingroup\$ Also, the connection of your fan is highly suspicious. When the fan ground is off, 12V can go into PWM pin trough fan circuit frying your MCU. \$\endgroup\$ – Ale..chenski Nov 17 '16 at 4:44
  • 1
    \$\begingroup\$ If you fried the converter, I bet you've got a pretty short-cirtcuit somewhere. It would be a nice idea to add a fuse after it. And maybe also one in the main supply rail. Also, as @Ali Chen comments, the fan connection seems dangerous. Maybe you could use some optocoupler there. \$\endgroup\$ – Adam Calvet Bohl Nov 17 '16 at 5:22
  • \$\begingroup\$ After the smoke clears, check the wiring, old Chinese proverb. When making proto boards, I always double check under magnifying lamp/microscope how the solder side looks like. Tiny whisker is enough to fry things. Try to keep those connections neat, "mess of wires" does not sound good. If you can afford it, current limiting lab supply is your friend. I use one at work all the time. Set it to 100mA and you're unlikely to blow anything before you have the chance to switch it back off again. Add fan etc high power components after. \$\endgroup\$ – Barleyman Nov 17 '16 at 18:32

The first thing that jumps at me is the common ground connection of the buck converter. There is no need to connect the negative side of the 12V supply to the negative side of the 5V supply. There should be no common connection between the two supplies!

Eddit: Since you are using the 12v supply for the fan, you do need to connect the grounds.
After looking closer to the schematic, I noticed that the 12v supply is a Wallmart 12v, 5A. These cheap supplies are notorious for not using isolation transformers. The result is that when you use any test device that is grounded, there is a high probability that the 120vAC will find its way into your circuit and sap it! Test the 12v supply (by itself) by using an AC voltmeter and see if there is 120vAC on either the + or - connection?

  • 2
    \$\begingroup\$ I thought it was common practice to tie the GND for multiple DC voltage sources together? forum.arduino.cc/index.php?topic=13452.0 \$\endgroup\$ – Constantino Nov 18 '16 at 22:59
  • \$\begingroup\$ Yes it is. I did not notice you are using the 12v supply for the fan. In this case, you do need to connect the grounds. However, this comes with a risk, which I detail in my edit. \$\endgroup\$ – Guill Nov 23 '16 at 9:18

You mentioned using a small housing, ensure that the compact space isn't causing any components to bend and touch each other when the unit is assembled and closed. Look for any exposed leads that may need a bit of heat-shrink tubing around them to keep them isolated from other nearby components.

I'd recommend using a multi-meter in the continuity test mode and validate all the wiring to ensure the circuit matches your schematic exactly. Also check for continuity between things that shouldn't be connected, like a power rail and ground.


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.