ceramic capacitor fire

[pucho812]

outside of a crack or some other compromise to the cap itself, what would cause a ceramic to catch fire and burn./release the smoke?

 

[Bo Deadly]

I'm not speaking from experience but logic would suggest that if a capacitor fails short, such as because it was damaged either physically or from excessive heat or age or some other process, with enough power, it will cause the shorted material to be burned up. The type of material will dictate what happens next. Either it gets burned away and the short is "fixed" or nearby material starts to melt increasing the shorting contact at which point it might run away and maybe actually catch fire.

 

[JohnRoberts]

obviously heat... heat is caused by resistance and lots of current, like a short across a PS rail.

Some capacitor shorts could be caused by a high voltage punch through the insulation layer. Some types of capacitors designed for use across mains have specific self healing properties, to prevent possible fires.

JR

 

[pucho812]

I'm not speaking from experience but logic would suggest that if a capacitor fails short, such as because it was damaged either physically or from excessive heat or age or some other process, with enough power, it will cause the shorted material to be burned up. The type of material will dictate what happens next. Either it gets burned away and the short is "fixed" or nearby material starts to melt increasing the shorting contact at which point it might run away and maybe actually catch fire.

well, it's like this. unit(audio rack mount, I am leaving names out on purpose) is new from the factory. I assume the design is tried and true as the company is selling them. my buddy picked one up and it didn't pass audio. He had me look at it where I saw the smoke leave a ceramic capacitor that had that amber/orange glow we have seen before in other components. I have never seen a capacitor burn. I have seen many a capacitor failure over the years but never actually saw a capacitor burn up before. The unit has other oddities but the burning cap, was something to behold.

 

[JohnRoberts]

You've seen the rare incandescent ceramic capacitor....

JR

 

[musipol]

Was it an MLCC or ceramic disc? I would guess the former.

 

[MrDude]

I remember that there are some Marshall guitar amps that have some sort of ceramic bypass on the anode of the output tubes.
That capacitor is only rated to 500V and is known to fail and burn up, in some cases burning a hole into the PCB.

 

[Newmarket]

Certain it is a ceramic cap' and not tantalum ?

 

[gyraf]

I've had a whole series of catastrophic faults on MLCC capacitors. Some 20-30 short-fails between 2019 and 2021 - in two different batches of 10uF/50V MLCC's. Not a single fail at initial tests, fail rates peaked ca. 5 months in -takes time for cracks to build momentum

Micro-cracks are the main reason - on leaded parts it's VERY hard to avoid mechanical stressing when mounting.

I've come to the conclusion that for MLCC, I can't get anything over 1uF if I want to ensure stability

/Jakob E.

 

[musipol]

Wow, I knew they were somewhat fragile, but not to that extent. Good info.

 

[pucho812]

Certain it is a ceramic cap' and not tantalum ?

Yes. 100% positive.

 

[pucho812]

Was it an MLCC or ceramic disc? I would guess the former.

Mlcc

 

[Matt Syson]

Over the millions of 0u1 50 Volt ceramic caps (MLCC) I have been nominally 'responsible' for over the years I have seen them fail, seemingly at random. 50 volt rated across a 17 volt DC rail that suddenly 'let go' after a few months. Each time one has failed 'catastrophically' it is usually only One (or perhaps 2) out of a mixing desk with around 1000 of them, when you consider all the channel strips combined.
As they are used with about 30 in parallel in a channel strip there is no rhyme or reason as to why any would be stressed more than any other. So I suppose a microscopic manufacturing weakness.
They may have been a batch issue as I remember (30 years ago) more failures being reported within a few months of each other. Sometimes they just went 'leaky' and as they were used for switch debouncing with large value resistors around them simply stopped the switch operation being recognised.

 

[templemark]

Voltage

 

[ccaudle]

...I saw the smoke leave a ceramic capacitor that had that amber/orange glow we have seen before in other components. I have never seen a capacitor burn.

In my experience almost certainly mechanical damage if a ceramic capacitor gets a short. A reliability engineer once showed me thermal camera video of a ceramic cap getting so hot it melted the solder and fell off the PCB. The cause was microscopic cracks caused by mechanical stress during assembly, and multiple thermal cycles caused the cracks to propagate.

 

[ThierryAudio]

I've had a whole series of catastrophic faults on MLCC capacitors. Some 20-30 short-fails between 2019 and 2021 - in two different batches of 10uF/50V MLCC's. Not a single fail at initial tests, fail rates peaked ca. 5 months in -takes time for cracks to build momentum

Micro-cracks are the main reason - on leaded parts it's VERY hard to avoid mechanical stressing when mounting.

I've come to the conclusion that for MLCC, I can't get anything over 1uF if I want to ensure stability

/Jakob E.

Hi Jacob,

Z5U, Y5V, or X5R, X7R dielectric, and what brand, please?

Because we use a lot of X7R and have never had any issues.


Have a Nice Day

Thierry

 

[JohnRoberts]

Hi Jacob,

Z5U, Y5V, or X5R, X7R dielectric, and what brand, please?

Because we use a lot of X7R and have never had any issues.


Have a Nice Day

Thierry

Is that a question.? Most conventional ceramic caps are mature technology and reliable. Chose the dielectric that fits your application.

JR