How to blow (up) fuses

[Phil1977]

Hi,

I wanted to test an quite old high speed camera that is capable of doing 1200fps. So I thought let´s do something dangerous and load up a 350uF/1500V foil capacitor to 750V and short it with some glass fuses.

I switched it with a quite beefy contactor. The whole circuit has a resistance of around 0.3Ohm, so the short circuit current will be somewhere below 2.5kA. That´s a short circuit current that a low-impedance mains socket is also able to deliver, so the test parameters are not absolute unrealistic.

First I connected a 3A / 250V 5*20mm fast glass fuse. Surprisingly it succeeded to interrupt the current flow! The foil cap just discharged from 750V to 730V, the fuse was black but okay:





I wanted some fireworks, so I took a 20A/250V 5*20mm slow glass fuse. There the result clearly shows why this fuse can not get a higher CAT-rating:






It exploded with quite a bang, and the arc that is visible in the second last image discharged the cap down to 29V. It clearly was not able to interrupt the current flow.

PS: Please don't do these experiments if you have just a slight concern about your safety. Such a capacitor is roughly equivalent to a medical defibrillator - it can easily interfere with your heartbeat.

[tooki]

Very nice! Great demonstration of why we really do need $10 fuses in our multimeters!



Maybe the lead inductance was enough to slow current flow just enough for the fast-blow fuse to react in time?


The maximum current depends on the cap itself, too.

At work, my old boss built a doodad that uses two 2kV, 800µF electrolytics in parallel. Because the goal is huge currents, they were selected accordingly to have very low ESR, and each is the size of a 2 liter soda bottle.

When tested at 1.5kV, the discharge pulse peaked at 14kA. 

(And that's discharged through a little coil the size of the cap of a 2 liter soda bottle. They're potted in epoxy and cooled in liquid nitrogen, but still are basically single-use. )

[pcprogrammer]

The person of this youtube channel knows how to blow a fuse. 

[johansen]

The maximum current depends on the cap itself, too.

At work, my old boss built a doodad that uses two 2kV, 800µF electrolytics in parallel. Because the goal is huge currents, they were selected accordingly to have very low ESR, and each is the size of a 2 liter soda bottle.

When tested at 1.5kV, the discharge pulse peaked at 14kA. 

(And that's discharged through a little coil the size of the cap of a 2 liter soda bottle. They're potted in epoxy and cooled in liquid nitrogen, but still are basically single-use. )

i bought 40 8uF 4Kvdc caps for 5$ each back in 2010. wish i had bought 200 of them but i ended up buying the remaining 40 in stock.

they have a 50khz internal resonant frequency and so its practical for me to discharge them through a coil and get a 20Khz frequency. which is sort of still too slow for coin shrinking.

2560 joules at 8kv (or 4, if wired for 4kv)

I was able to bend a quarter but not shrink it. when i get around to it i'm going to make some coils and pot ceramic around them and then fire it in a kiln, and then press fit the ceramic tube into a copper pipe, to reflect all that good magnetic energy and also contain the coil.

i once wrapped the coil in 1/4" of fiberglass packing tape, and it blew it all up.

[tszaboo]

You can also blow up 3-5mm LEDs, with the right amount of voltage and current it might shoot off it's lens.
The worst mistake I saw was with a 10KA DC fuse, that someone blow.

[tooki]

The worst mistake I saw was with a 10KA DC fuse, that someone blow.

Go on...

[Phil1977]

Maybe the lead inductance was enough to slow current flow just enough for the fast-blow fuse to react in time?


The maximum current depends on the cap itself, too.

At work, my old boss built a doodad that uses two 2kV, 800µF electrolytics in parallel. Because the goal is huge currents, they were selected accordingly to have very low ESR, and each is the size of a 2 liter soda bottle.

When tested at 1.5kV, the discharge pulse peaked at 14kA. 

(And that's discharged through a little coil the size of the cap of a 2 liter soda bottle. They're potted in epoxy and cooled in liquid nitrogen, but still are basically single-use. )

The cap I´m using is specified for 25kA surge current and 1.9mOhm ESR. It´s a non-polar foil cap 8.5cm in diameter, 24cm long and 1.5kg (Electronicon E50.N23-354N50). I already did some short circuit tests with it and so far it´s capacity is stable - but I chickened out at 1kV so far, the energy and available power of this thing (>10MW peak power!!!) are nothing but scary.

[tszaboo]

The worst mistake I saw was with a 10KA DC fuse, that someone blow.

Go on...

It was with a lead acid battery, single cell about 20KG heavy.
Batteries do that. We were also making testers that went up to 4000A.

[schmitt trigger]

These hi-speed photos give one an appreciation on fault currents.
It would have been enlightening to actually see a video, but video cameras with those frame rates are devilishly expensive.

There is a YT channel of a pair of guys, The Slow Mo Guys, who specialize in very high frame rate videography and who love chemical explosions and bullets.
I ignore whether they have the knowledge to perform an electrical “explosion”.

[Phil1977]

In fact I took the pictures out of a video, but decided to just publish the stills because even in the 40x-slowmotion (1200fps record speed / 30fps view speed) everything happens more or less instantaneously. It´s quite obvious that the explosion of the 20A-fuse together with the discharge of the cap took something between 1 and 2ms.

Recording it with a Chronos or Phantom (as the SlowMoGuys) would be great fun, but since I forgot to click order on bitcoins 20 years ago (yeah - my fault!) I can't do this right now. If someone in the area of Munich has a high speed cam of that type I´m readily available to blow something up.

And though it´s an ABB advertisement, there is at least one interesting high speed video of a circuit breaker available on Youtube:

[Gyro]

That's nowhere near a clean an arc clearance as I expected. It seems to hang around the arc chute plates for quite a long time (relatively). Interesting.

[Sensorcat]

In fact I took the pictures out of a video, but decided to just publish the stills because even in the 40x-slowmotion (1200fps record speed / 30fps view speed) everything happens more or less instantaneously. It´s quite obvious that the explosion of the 20A-fuse together with the discharge of the cap took something between 1 and 2ms.

Recording it with a Chronos or Phantom (as the SlowMoGuys) would be great fun, but since I forgot to click order on bitcoins 20 years ago (yeah - my fault!) I can't do this right now. If someone in the area of Munich has a high speed cam of that type I´m readily available to blow something up.

It can be embarrassing if one presents experiments, and then the spectators tell what they would have done, but anyway (sorry), here is my proposal, if you want to continue with blowing fuses:

Measuring the voltage of the capacitor with a high-voltage probe and almost any oscilloscope should be affordable to get more insight into the dynamics of the blow (compared to the camera with 1200fps), because this measurement would not suffer from overloading/saturating something (like the pixel capacitors in the camera at least possibly could). At least not if the derating of the probe is taken into account. While the SlowMoGuys are in their own league, another channel of said scope could be used with a fast photo diode or photo transistor, to get a really high-speed affordable cam with one pixel! 'Cam' could be tested with a photo flash, such that nothing must be blown to find useful parameters for sensor and circuit.

I'd like to see that, but just realized that your goal was to test your camera, not to blow fuses...

[Phil1977]

Don't worry, I think it´s nice to see other peoples improvement ideas.

As you said, this was only to do something interesting with the camera, and the exploding fuse didn't disappoint there 

If I´d want to build a test stand for blowing fuses, I´d go for your ideas. I´d even enhance them a bit:

- A fast hall sensor near the bus bar to measure the slope of the current
- Two filtered photodiodes: One in the spectral range of glowing metal (yellow) the other one in blue or even in UV to discriminate the burning arc
- A remote control with a remote discharge function so that the setup can be put into a safe state from a distance.