[sdiy] Using non-linear passives

[Mattias Rickardsson]

On 5 April 2017 at 20:38, Tom Wiltshire <tom at electricdruid.net> wrote:
> On 5 Apr 2017, at 19:11, Mattias Rickardsson <mr at analogue.org> wrote:
>> On 5 April 2017 at 20:03, Tom Wiltshire <tom at electricdruid.net> wrote:
>>> It looks to me there's something we're not understanding (or
>>> misunderstanding) here. There's no way a capacitor that is spec'd to 35V
>>> loses 90% of its capacitance by 35V. That simply wouldn't be *useful*.
>>
>> Yes, they are still useful for many applications - if you know the
>> voltage dependency and your DC voltage, and and choose your capacitors
>> accordingly. :-)
>
> Ok, in that case, I guess my question would be why call it 4.7uF/35V if it isn't?! :)

All specifications for all components are not guaranteed under all
conditions. For active components we are used to take all kinds of
strange values into account when we interpret the figures in the
datasheet, and we even feel proud when we learn how to deal with
component data in our constructions and not trust everything we read.
But for a passive little component like a capacitor we just can't
accept that it's not ideal! :-)

> I notice all the parts talked about are in the uF range, and these are very small caps we're looking at. Would this problem decrease for smaller values where you're not trying to squeeze so much capacitance into such a small package? 220nF? 47nF? 100p?

Look closely on all the colours in the first graph - not just all the
lines, but also the different parts of the lines:
http://www.edn.com/design/analog/4402049/2/Temperature-and-voltage-variation-of-ceramic-capacitors--or-why-your-4-7--F-capacitor-becomes-a-0-33--F-capacitor
There you can compare different sizes of caps with the same value &
voltage rating, and the graphs clearly indicate that bigger is better.

> Thanks for pointing it out though. I wouldn't have imagined such a spec was remotely realistic.

I agree, it's unexpected and almost shocking! :-)

/mr