5
\$\begingroup\$

We are working on a product that contains an LED module consists of 66 LEDs (33s x 2p - 100VDC load voltage).

We'll use aluminum PCB due to the thermal requirements but the customer requirement claims that the insulation voltage for the aluminum PCB should be at least 1kV.

The PCB company manufactured the PCBs for us according to the requirement above then we tested it in our lab. Here's the test procedure:

  • Short the supply terminals of the LED module.
  • Apply increasing AC voltage gradually between the back side (i.e. aluminum) and the front side (the shorted terminals) until the generator stops (The generator beeps then stops automatically under a faulty condition -e.g. short- so we can determine the insulation voltage).

The results are quite interesting:

  • The insulation voltage is lower than 100Vac for an unused, empty PCB.
  • The insulation voltage is around 1.2kV for a final LED module (i.e. LEDs placed on the same PCB).

The image below shows an empty PCB (top) and final LED module (bottom):

enter image description here

Is this normal? Where can this difference come from?

\$\endgroup\$
2
  • \$\begingroup\$ Comments are not for extended discussion; this conversation has been moved to chat. \$\endgroup\$
    – clabacchio
    Commented Jul 27, 2018 at 6:49
  • \$\begingroup\$ @OP also, feel free to insert in the question any important detail risen up from the discussion in the comments \$\endgroup\$
    – clabacchio
    Commented Jul 27, 2018 at 6:50

1 Answer 1

Reset to default
2
\$\begingroup\$

The difference is most likely in the method of test. Looking at the wires in the two pictures with most everything being the same, its probably the crap between the wires.

I'll bet that if you took only the top board and connected the wires, then did the insulation test, and repeated the test several times, the answers would vary even with no LED's connected.

Either keep the wires consistent between tests, or make sure they are connected in the same way as possible. Clean off the residue between the contacts with flux remover.

\$\endgroup\$
5
  • \$\begingroup\$ The difference is most likely in the method of test. Looking at the wires in the two pictures with most everything being the same, its probably the crap between the wires. I don't think so because the cables have the same physical properties (num of conductors, cross-sectional area etc) except the thickness of the outer isolation. \$\endgroup\$
    – Rohat Kılıç
    Commented Jul 26, 2018 at 19:52
  • \$\begingroup\$ I'll bet that if you took only the top board and connected the wires, then did the insulation test, and repeated the test several times, the answers would vary even with no LED's connected. Today we've got an interesting result: During the tests, one of the alligator clips was attached to the cables (i.e. shorted) and the other one was touching to the aluminum base. We got different insulation strength voltages (still lower than 1kV) for different points of touch on the aluminum base. ... \$\endgroup\$
    – Rohat Kılıç
    Commented Jul 26, 2018 at 20:02
  • \$\begingroup\$ ... Then we put the bare board onto an aluminum panel, attached the latter alligator clip to that panel and repeated the test and wow! We got around 1.5kV of insulation strength voltage! \$\endgroup\$
    – Rohat Kılıç
    Commented Jul 26, 2018 at 20:07
  • \$\begingroup\$ Glad you found the problem, that's a weird one. \$\endgroup\$
    – Voltage Spike ♦
    Commented Jul 26, 2018 at 20:16
  • 2
    \$\begingroup\$ With added resistance by only touching (or laying on) poles, higher insulation- values are unsurprising. You need to connect the probes to the aluminium directly, bypassing the oxide layer. Solder the probe-leads to the board's sides in every trial. \$\endgroup\$
    – loonquawl
    Commented Jul 27, 2018 at 5:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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