High value MLCC capacitors vs electrolytic or tantalum

[Richard Lewis]

Seems that high value MLCC capacitors (around 100uF) with a decent dielectric (X5R) can be obtained for around $0.50 in small quantities. Other than possible issues with them making noise due to piezo effects are there any reason not to use them in place of electrolytic or tantalum for power line filtering in a digital circuit? I have asked a similar question before in the RC2014 group but in a more general sense. I guess my question would be: "If I require a 100uF capacitor for the purpose of power line filtering for a RAM/ROM board I'm working on. Will I be OK or should I use an electrolytic instead?"

[Phillip Stevens]

Richard Lewis wrote:

Other than possible issues with them making noise due to piezo effects are there any reason not to use them in place of electrolytic or tantalum for power line filtering in a digital circuit?

I've used 3x for filtering +5V and +12V on my YAZ180.

I've heard they can spontaneously explode and ignite a fire when they get older and their insulation weakens.

I also know if you solder them reversed, by accident or stupidity, they will make a crater in your PCB when you power up and they go bang.

Yep. Did that.

Phillip

[Richard Lewis]

When I was into vacuum tube audio many years ago I accidentally put +300VDC on the wrong terminal of an electrolytic. Was a pretty spectacular mess I had to clean up. 

Maybe I'm misguided but I was thinking 

[Sergey Kiselev]

MLCC are not polarized, so they are not going to explode because of the wrong polarity.
I've used 10 uF MLCC throughout my projects, and I haven't really had any issues with them.
One disadvantage though, is that their capacitance decreases when voltage is applied to them... The higher the voltage - less capacitance. Getting capacitors rated for higher voltage doesn't help. Instead get higher capacitance or more capacitors.

With that being said, aluminum polymer capacitors are more reliable than electrolytic capacitors, have low ESR (useful for power filtering), and not that expensive.
I'd consider using them for anything 10 uF and more.

[Mark T]

I seem to remember there were concerns about ceramic decoupling capacitors back in the 80s. There was a low but not zero probability of a ceramic capacitor failing short circuit. The concern was due to the high current rating of power supplies, not uncommon to have 5v @ >5A, so if a decoupling capacitor went short circuit It could burn quite hot and may not blow the fuse. One of the safety requirements we had was to ensure the psu had foldback current limit and did not act as a 5A constant current source. This was back in the time before smd was common, so the failure of the decoupling caps was probably not due to cracking from flexing of the pcb. It was also possible to get capacitors that included a fuse to protect the circuit if the cap went short circuit but they were a bit expensive and not common. I always wondered if this might have been one of the factors driving the use of local regulators on systems like S100, as the local regulators would limit current at 1-1.5 A.

Mark

[Tom Storey]

They do still fail short spectacularly. Dave from eevblog did a video on a small power supply that had a ceramic cap soldered in between the output (screw) terminals, which had failed due to mechanical stress, likely from when the output leads were attached.

I think it's fair to say you need to put some thought into where you place ceramic caps, avoiding areas of the board that are likely to experience flex (e.g. around screw holes or terminals), or orienting them appropriately to minimise stress from what flex there will be.

On Thu, 4 Jul 2019, 07:40 Mark T, <mark...@gmail.com> wrote:

I seem to remember there were concerns about ceramic decoupling capacitors back in the 80s. There was a low but not zero probability of a ceramic capacitor failing short circuit. The concern was due to the high current rating of power supplies, not uncommon to have 5v @ >5A, so if a decoupling capacitor went short circuit It could burn quite hot and may not blow the fuse. One of the safety requirements we had was to ensure the psu had foldback current limit and did not act as a 5A constant current source. This was back in the time before smd was common, so the failure of the decoupling caps was probably not due to cracking from flexing of the pcb. It was also possible to get capacitors that included a fuse to protect the circuit if the cap went short circuit but they were a bit expensive and not common. I always wondered if this might have been one of the factors driving the use of local regulators on systems like S100, as the local regulators would limit current at 1-1.5 A.

Mark

--
You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/4fa8ccbd-93af-45e4-9908-67c7c8a9c6b0%40googlegroups.com.