Tantalum Capacitor Derating?

[ACharnley]

Hi,

What is the consensus on derating tantalums as typically applied to aluminium electrolytics?

In my case I need to extract the most performance from a couple of 220uF 10V tantalums and wondering how close to 10V I can drive them.

(I'm aware from physical experience that tantalums catch fire on over-voltage)

Andrew

 

[Nigel Goodwin]

Hi,

What is the consensus on derating tantalums as typically applied to aluminium electrolytics?

In my case I need to extract the most performance from a couple of 220uF 10V tantalums and wondering how close to 10V I can drive them.

(I'm aware from physical experience that tantalums catch fire on over-voltage)

Andrew

As I've said here many times, in my previous life as a TV Engineer, various manufacturers introduced tantalums at various times, EVERY single time was a disaster, with mass failings. These were always 25V tantalums on 12V regulated rails.

As far as I'm aware the recommendation is that you always specify twice the maximum expected voltage for tantalums, so less than 5V for your 10V rated ones.

However, I'm somewhat vague about 'maximum performance'?, what do you mean by that?.

Personally, I don't see how manufacturers can claim a component is 10V, then state you can't use it above 5V? - in my book that's a 5V component.

 

[ACharnley]

I don't see anything in datasheets about derating by 50%. My understanding is aluminium electrolytics have a much lower life expectancy when run close to the rating while tantalums don't (citation needed!). Indeed tantalum's are supposed to self-repair to an extent.

I've ran the 10V caps in question at 25V by mistake and they caught fire in quite an explosive manor. I'm really asking for a risk analysis for running 10V caps at 10V (or 9.1V which is my assumed safe alternative).

I'm after the maximum capacitance/voltage in a space critical design where every bit over a 5V level is a win, hence Tantalums work for this particular application.

 

[tvtech]

It's not about the voltage. Tantalum is not able to provide the rated microfarad smoothing for constant duty. They heat up and destroy themselves quickly if used in the WRONG application. Military radios I worked on in the eighties were full of them. I cannot remember ever having replaced one. They were never designed to replace Electrolytic caps.

Tantalum hates ripple currents. Not made for that.

When they are used in applications they were designed for they give zero problems.

 

[ACharnley]

Ripple is 0.2A max but will average at no more than 0.12A.

0.1 ohm ESR & 0.2A = I^2R = 0.002w dissipation. Can't see them heating up. The application will never rise above ambient temperature so say 50c worst case.

Interestingly the data sheet has a surge voltage that is 13v for 10v rating.

 

[tvtech]

Ok... you've all the answers. Why are you even posting for help?

 

[ACharnley]

Double check if I have it right...

 

[narkeleptk]

looking at the snippet of datasheet you provided it looks to me the 10VDC model is made for systems with 5V or under in mind. It will probably work for what you want but I would expect higher failure rate and shorter life span.