0
\$\begingroup\$

I am designing a board and the problem I am facing on the board hardware is that the core power supply and ground has very low impedance of around 0.3 Ohm. This is leading to a conclusion of two nets as if these are shorted together. Although the current drawn for that particular assembly for core power supply is 9A the impedance would be low(may be around 4-5 Ohm).But is 0.2 Ohm a correct value. We have done the x-ray analysis of BGA, and no solder bridging is observed. Can anyone suggest what could be the issue?

\$\endgroup\$
6
  • 3
    \$\begingroup\$ You could also just have a damaged component that is causing the short. A way I have "solved" this is by connecting the system to a beefy current source and just pushing a few amps through it. Start low and you might be able to measure/observe what gets hot. \$\endgroup\$
    – Joren Vaes
    Mar 29 '18 at 14:49
  • \$\begingroup\$ Have you measured resistance between the two nets on a blank board, no components installed? Could you have a diode in backwards, or another active component with incorrect pattern? \$\endgroup\$
    – AlmostDone
    Mar 29 '18 at 14:50
  • \$\begingroup\$ First test a blank PCB. PCB's should be factory tested to prevent copper filaments. Also check for 'assembly' errors e.g. chips mounted backwards. Ideally rent/borrow a micro ohmmeter to find the location. Last resort is to 'blow-out' the short. (But my experience is that you may need e.g. 50Amps for that). \$\endgroup\$
    – Oldfart
    Mar 29 '18 at 14:58
  • 3
    \$\begingroup\$ I have had to resort to "blowing out a short" that was mistakenly designed into a board artwork. In my case it was a short between to GND planes that were meant to be isolated through a single point ferrite bead connection. We knew there was a short because we could measure it but no amount of work with a micro volt meter was able to isolate the short due to the multi-copper plane nature of the board design. In the end I had to grind into the planes and solder on 10AWG wires. The wires were then connected to a fully charged 12V car battery. Many Amps of current located the short in seconds! \$\endgroup\$ Mar 29 '18 at 15:15
  • \$\begingroup\$ @MichaelKaras - I've done something similar with a few 10,000uF caps charged up to 3.5V to blow out a whisker short between the power & GND in an assembled PCB... \$\endgroup\$
    – brhans
    Mar 29 '18 at 18:58