0
\$\begingroup\$

Short version: Extremely loud sound a few seconds after power on (400 V AC, 3 phase) in high humidity, no smell: is it arcing? Is it possible that a component exploded without smell?

Long version: Today, we did some testing on a 15 kW power supply (EA-PSE 9750-60 : main page, manual datasheet), trying to find the (small) leakage between the output and earth (isolation between output and earth is >20 MΩ in dry weather, but can go down during high humidity (below 2 MΩ in some cases, causing our monitoring system to prevent power on / cutting power)).

Most testing was performed with the supply not powered, at 75% humidity (supply is rated for 80%).

Once we fixed the leakage problem (between a fan and the chassis, by adding an isolated padding), we kept testing for some time with power off (isolation was always >5 MΩ, which is a big improvement for such high humidity).

When we powered on the supply, it worked fine for a few seconds (with isolation decreasing to 3.4 MΩ), before making a very strong noise (like an explosion, my ears are still ringing one hour after). My colleague removed power within a few seconds. No fuse, circuit breaker nor ground fault circuit breaker triggered.

So my first hypothesis was that some component exploded, but there was no burnt smell, nor have I found any component with visible damage so far.

NB: although I haven't seen any on the circuits afterwards, there might have been some condensation (or a drop falling from the "roof" of the humid cabin (we used a water boiler to increase humidity) we used for testing).

Can components explode without smelling? (So far, in by experience, when a component pops open, it smells quite strongly, but maybe there are exceptions?)

Or might it have been an electric arc? Is this coherent with the loud explosion without smell? If it is, is there any physical sign to look for?

Voltages involved:

  • AC mains, 3 phase : 400 V. The supply is split in 3 modules, each receiving 2 phases (i.e. 400 V AC) whose outputs are connected together.
  • DC output stage: 320 V DC (isolated from AC and earth): not powered on during the incident
  • Auxiliary voltages: up to 24 V DC
\$\endgroup\$
8
  • 1
    \$\begingroup\$ A loud explosion isn't typically associated with an arc, at least at sub kV voltages. (It's more of a zap, snap, crackle, fizz or pop.) There would also typically be visual evidence of the arc if it were big enough to cause that much noise. Very odd situation as it's usually obvious what exploded. \$\endgroup\$
    – John D
    Commented May 14 at 16:53
  • 2
    \$\begingroup\$ Electric arcs generate ozone, which you may smell if you are in the right place at the right time. Once it dissipates though, there is no lingering odor. \$\endgroup\$
    – gbarry
    Commented May 14 at 17:47
  • \$\begingroup\$ @JohnD : from the sounds I found on the internet, I would say probably a snap (pond5.com/fr/sound-effects/item/…), maybe a pop (pond5.com/fr/sound-effects/item/205185140-electrical-spark-pop). Is there something one can conclude from the sound? \$\endgroup\$
    – Sandro
    Commented May 14 at 19:35
  • \$\begingroup\$ @gbarry : the cover from the supply was removed, and I was less than 1m apart, in a rather closed enclosure (about 15m³ separated from the rest of the room by plastic sheets). I haven't smelled any ozone smell. After maybe a minute (once sure power was off), I smelled from close, without smelling anything either \$\endgroup\$
    – Sandro
    Commented May 14 at 19:38
  • 1
    \$\begingroup\$ explosion sound? it's a spark not an arc. \$\endgroup\$
    – Jasen Слава Україні
    Commented May 15 at 3:54

1 Answer 1

Reset to default
2
\$\begingroup\$

It's possible that your humidification setup caused condensation between two points that vaporized when power was applied. You'll see that the humidity spec states "non-condensing" and using a water boiler to increase steam is definitely creating a condensing environment; try an ultrasonic humidifier.

If you were using a megger to test resistances, then it's also possible that you applied an overvoltage somewhere and damaged insulation.

\$\endgroup\$
3
  • \$\begingroup\$ So current flowing through water, causing the water to evaporate very fast causing the sound? If this is the "failure", would it cause any visible sign or damage? (There might of course be indirect damage if the current path went through some low voltage component) \$\endgroup\$
    – Sandro
    Commented May 14 at 17:38
  • \$\begingroup\$ For the monitoring, we use a continuous line tester, that applies +/-50V between earth and output (isolation is rated >1kV, so it shouldn't be a problem, and never was so far) \$\endgroup\$
    – Sandro
    Commented May 14 at 17:40
  • 1
    \$\begingroup\$ I would look for faint splatter marks on the board near exposed HV conductors and give special attention to any vertical areas where water could have dripped down the board or flat areas where it could have pooled. \$\endgroup\$
    – vir
    Commented May 14 at 17:42

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