EEVsmoke #2

[EEVblog]

[rs20]

Awesome stuff!!!

I'd really like to try replicating (and videoing) the captured-smoke failure. Seems like a worthy exception to the "natural failures only" rule 

[Vgkid]

Some major self destruction, on this one.

[SL4P]

Gives you an idea of how good coffee is for us! 
They say cockroaches will survive a nuclear blast - but not an active coffee machine!

[Halcyon]

Gives you an idea of how good coffee is for us! 
They say cockroaches will survive a nuclear blast - but not an active coffee machine!

It looked more like a cricket to me.

[blackbird]

About the last picture of the escaped smoke (the Jaycar UPS), I noticed the most right component looks still OK, at least it has not exploded or burnt.

Could it be that because the most right one has a sil-pad underneath it survived, or better, because the other three does not had the sil-pads made them go 'booom'?

[G7PSK]

The fourth mosfet in the Jaycar UPS looked liked it had got rather hot as well. I once had a 400KVA gen set that had the insulation on the stator removed by ants, but unfortunately I do not have any photo's of it as it was quite a few years ago now.   

[lowimpedance]

The captured magic smoke is amazing, worth framing  .
Also an interesting picture of the 'dark side of the moon' ......... well maybe not that dark! 

[EPTech]

Hi there,

Nice episode. I remember now that this is the actual meaning of debugging, LMAF. 

I especially liked the power module that blew up "discretely", capturing the act in the gel, very nice. This would be difficult to recreate actually.

To dave: I would use your friends "joule bomb", not sure what the technical term it for that demonic device, the one you used on the multimeters. The trick would be to determine just the correct amount of energy not to blow up the IGBT completely but still have the effect of the smoke caputered in the gel. The setup itself is quiete simple. Hook up source and drain to the "joule bomb" and put an isolated control signal over gate-source to make it conduct, like a 9V battery. If the battery blows up with the rest, that is not a big loss.

[max_torque]

Here's the DC link contactor that died alongside that IGBT from me:



Current monitor on DC link opened it automatically when IGBT shorted, but contactor had 850A flowing through it at the time!  Which it did break, but resulted in significant arc damage to the contactor contacts and cracked the porcelain of the housing.


Love the fact the contactor looks soo sad in the picture!   

[vinicius.jlantunes]

That magic smoke captured picture is mesmerizing... the video was good in its entirety but that picture alone would have made it worthy no doubt.

[Len]

About the last picture of the escaped smoke (the Jaycar UPS), I noticed the most right component looks still OK, at least it has not exploded or burnt.

Could it be that because the most right one has a sil-pad underneath it survived, or better, because the other three does not had the sil-pads made them go 'booom'?

More likely, the one with the sil-pad failed previously and was replaced, that's why it's mounted differently, and then the other 3 failed.

[dansan]

I can't be the only one who saw in that first example a Chinese emperor holding a Super Bowl trophy.

[retrolefty]

Cochrouch my ass, that's a grasshopper. 

[electrophiliate]

The captured smoke in the IGBT gel is impressive.

[EPTech]

Here's the DC link contactor that died alongside that IGBT from me:



Current monitor on DC link opened it automatically when IGBT shorted, but contactor had 850A flowing through it at the time!  Which it did break, but resulted in significant arc damage to the contactor contacts and cracked the porcelain of the housing.


Love the fact the contactor looks soo sad in the picture!   

Hi there. I have not encountered a contactor like that in the drives I repaired, with the porcelain base I mean. Was it and older drive or maybe a high voltage one?

[SgtTech]

Excellent stuff Dave, l like the magic smoke stuff. Maybe need to take another look at the multimeters you blew up! or maybe blow up a few more with a slow mo camera?

[EEVblog]

80k views, 5k thumbs, and a normal 5% thumbs down ratio.
So the segment seems very popular.
Although it could still be anomaly because of the awesome captured smoke photo.

[Brumby]

You never know, this series might find a cult following.

But aside from that, I am beginning to warm to the content - and it's not because one of my photos was in there (but that didn't hurt  ).

What is beginning to appeal to me is, I think, the dramatic nature of the failures - and the chances of seeing those one-in-a-million shots ... like the smoke captured in the gel.  That was just awesome.


I'll be interested to see how this theme tracks over time.

[Brumby]

Anyone have some good examples of a MOV failure?

I had one, but I cleaned it up and repaired it.  A dead short across 240V makes for an interesting afternoon.

[EEVblog]

What is beginning to appeal to me is, I think, the dramatic nature of the failures - and the chances of seeing those one-in-a-million shots ... like the smoke captured in the gel.  That was just awesome.
I'll be interested to see how this theme tracks over time.

Yes, will be interesting.
100% reliant upon viewer submission though, and to be honest there were a lot of submissions that really weren't worth showing IMO

[rrinker]

 That captured smoke is going to be very hard to top. I'm with Lowimpedence - that's suitable for framing as art.

[Brumby]

That captured smoke is going to be very hard to top.

It is.

[StuUK]

80k views, 5k thumbs, and a normal 5% thumbs down ratio.
So the segment seems very popular.
Although it could still be anomaly because of the awesome captured smoke photo.

And the keywords in the title I suspect.....

[amirm]

Anyone have some good examples of a MOV failure?

Here you go:

[bktemp]

Anyone have some good examples of a MOV failure?

Here you go:

What is this device that needs so many MOVs? And what is the black thing with the toriodal core on top of it?

Whenever metal vaporises, it leaves a horrible mess.
That's why you need HRC fuses for multimeters, because cheap glass fuses can explode while trying to interrupt a full mains short.

[Brumby]

Ah!  Found the pics I took of my MOV drama:

The direct view....


From above (with enhancement)


The PCB after cleaning...


The casualties:  (Yes, everything in the yellow frame is was the MOV):

[Brumby]

Dave - I'll send the Hi-Res to you ... in case you are interested.

[amirm]

What is this device that needs so many MOVs? And what is the black thing with the toriodal core on top of it?

It is a service entrance/whole house SPD (surge protection device).  I assumed they paralleled the MOVs to get higher energy absorption capacity.

That's why you need HRC fuses for multimeters, because cheap glass fuses can explode while trying to interrupt a full mains short.

That's right.  Many of these SPDs unfortunately have no fuse so when the MOV goes, it goes really bad.

[Fungus]

Ah!  Found the pics I took of my MOV drama:

Now you know why some people wrap them with heatshrink.... 

[Mike Warren]

From quite a few years ago. The magic smoke had escaped, but it wasn't particularly dramatic. Looks like the mouse could smell fried lizard.

[guido]

There's also magic liquid. One cap turned ballistic and tried to lift off.



Got it working again, it's an off the shelf powersupply in a Kode time interval analyser.

[strawberry]

From some old TV 80's

[iwasz]

Mains went through USB port. Pictured below is an ESD protection IC.

[Nobody2]

From quite a few years ago. The magic smoke had escaped, but it wasn't particularly dramatic. Looks like the mouse could smell fried lizard.

And what's the mouse doing there? Or where both the mouse and the lizard fried? How did they get in there in the first place?!?

[EPTech]

Hi there,

I was repairing a VFD and got another power IGBT module with a blown up brake chopper, massive gate failure. The TVS over the gate and emitter took the over voltage like a boss, shorted out and the mostfet driver seems to have survived thanks to the series resistor to the gate. Although not as impressive as OP picture, The magic smoke got caught in the act. Additionally you can see the bonding wires melted off the emitter, just below the black spot. The intense heat also boiled the gel around the area.

[Mike Warren]

From quite a few years ago. The magic smoke had escaped, but it wasn't particularly dramatic. Looks like the mouse could smell fried lizard.

And what's the mouse doing there? Or where both the mouse and the lizard fried? How did they get in there in the first place?!?

Sorry, I missed this. Both animals had been killed by touching the 240V terminals. My guess is that the mouse smelled fried lizard and went to get some to eat.

[madsbarnkob]

80k views, 5k thumbs, and a normal 5% thumbs down ratio.
So the segment seems very popular.
Although it could still be anomaly because of the awesome captured smoke photo.

I am happy that you like my photo and contribute it some success, it was however sent in by someone that has seen it on my website and did not include proper credits for it.

Please link back to my original article where the photo is from: http://kaizerpowerelectronics.dk/technology/teardown-schneider-electric-frequency-inverter-altivar71/

On your question on how the layers of smoke settled in the gel has happened, it is my thesis that the long and flat flower like black paths in the goo is actually the gasses from the  exploded IGBT die, the gas expands out into the goo that blows a bubble and collapses as soon as the pressure is gone. Leaving a black trail of burned silicon, metal and goo.

[EPTech]

Hi all,

This one reminded me of this topic. Almost a piece of art. Magic smoke flowers. 😁

Greetings and happy tinkering.