I built a LED driver with PT4115 IC. The board lighted LEDs up but sparks appeared around SW, VIN & GND pins after 10 seconds and the phenomenon didn't damage the IC. I recorded it by my mobile phone.

Spark around pins

Below is the schematic and the PCB board: schematic PCB

I removed R2, D2 and R3 so the DIM pin is floating. It should help us to concerntrate to this problem.

I checked every pin of the IC and didn't find any short or something wrong. I thought that something conductive around these pins and start to conduct after 10 seconds. It is very strange and confused me.

Any good suggestion?

  • 1
    \$\begingroup\$ You should not be seeing sparks here. Can you give a schematic of your circuit? \$\endgroup\$ – Hearth Apr 29 '17 at 15:27
  • 1
    \$\begingroup\$ The datasheet also cautions against using a silicon diode for D3 (called D1 or D in the datasheet), because poor reverse recovery time and high voltage drop can cause high voltages to appear on the SW pin. Though it looks like the spark in the video is between CSN and GND. \$\endgroup\$ – Hearth Apr 29 '17 at 15:55
  • 1
    \$\begingroup\$ Also, did the sparks occur during open circuit operation, or when the LED was connected? \$\endgroup\$ – mkeith Apr 29 '17 at 16:18
  • 1
    \$\begingroup\$ Can you post a bigger picture of the board. ANd are you sure that diode is not fried. \$\endgroup\$ – Trevor_G Apr 29 '17 at 16:24
  • 2
    \$\begingroup\$ @BruceHsu So you used plumbing solder? That's a real problem - the acid is not only highly conductive but it's also corrosive. Try to clean it very well. \$\endgroup\$ – Spehro Pefhany Apr 29 '17 at 18:26

Assuming the diode is still good, the layout of the board is really bad.

That current sense line loops around the chip, crosses and runs parallel with the drive line. That's bad. The inductor is also a long way from the chip and Vin also crosses v-out..... .

The PT4115 needs to be turned round 180 degrees with that part layout. But in truth the whole layout needs redone.

More like this...

enter image description here

See also "Layout considerations" on page 14 of the datasheet.

Vital parts are:

Minimize ground noise by connecting high-current ground returns, the input bypass-capacitor ground lead, and the output-filter ground lead to a single point (star ground configuration).

The SW pin of the device is a fast switching node, so PCB tracks should be kept as short as possible.


It is particularly important to mount the coil and the input decoupling capacitor as close to the device pins as possible to minimize parasitic resistance and inductance, which will degrade efficiency.

  • \$\begingroup\$ BTW C1 is missing and should be located between the diode and the PT4115. \$\endgroup\$ – Trevor_G May 1 '17 at 17:29

Making an answer out of my comment:

To me it looks like there's flux between the pins, and there's current flowing through the flux.

It also seems to me that I see a wisp of smoke drifting over that chip when the sparking stops. That indicates that the problem is the flux.

Use some alcohol to remove the flux, then let it get good and dry, then try it again.

Given that cleaning the area around the chip fixed the sparking, it seems to have been correct.

It would still be a good idea to follow the suggestions in the other answers to improve the layout and to check the type and function of the diode.


The sparks are inductor's way to keep the current on if all low voltage drop routes are off. Your diode D3 is bidirectionally non-conductive or disconnected.


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.