The TI TPS26631 datasheet states: “CIN is a local bypass capacitor used to suppress input noise. TI recommends a minimum value of 1 µF for C(IN) to limit the rate of change during the peak test”

So I went to the Murata website to look for one and found the GRM188D72A105KE01# In its datasheet, there is this diagram showing that at 48V (my voltage) the capacitance drops by 84 per cent, falling from 1uF to approximately 0.161uF



What I’m asking is:
- when a manufacturer recommends a capacitor with a capacitance of x Farad, do they mean that I have to find one which has that capacitance at my operating voltage? ..so in this case I cannot use GRM188D72A105KE01#
- or do they mean that it’s enough just to choose one with that capacitance and that’s that, regardless of the derating curve I’ve attached? ..so in this case I can use GRM188D72A105KE01#

Do you think the 865081740001 would be the best option for me, then? we-online.com/en/components/products/WCAP-ASLI#865081740001

 

TI recommends a minimum value of 1 µF for C(IN)

Since 1 uF is the minimum required value, you don't do anything wrong by selecting a capacitor with a higher value, e. g 4.7 uF. You can also select a capacitor with a higher nominal voltage (e. g. 200 V) this reduces the capacitance derating at 48 V, too.

 

avx.com Here is a paper. Every company has a different version of this graph. Each type of cap has a different graph.

 

Good engineering practice would dictate that you meet the minimum capacitance requirements at the maximum expected operating voltage.

As others have already pointed out, the best approach is to be quite conservative with the voltage/capacitance ratings, to ensure that you meet those minimum requirements.

PD; also consider temperature related effects!!!

 

I designed a power supply for a Texas company. It needed to run down to -40C and it did.
A year later they started using the product in Canada and it turns out it would not start-up below -40C. It would run just not start up. So, I had to make a change to get it to start to -50C.

 

Consider that the engineer that designd an amplifier produced that design for the purpose of it being duplicated. THAT in turn means using standard components with specified values printed on them. The actual values may differ some depending on the voltage present.
And many folks evaluate an amplifier by how it sounds, rather than anything else. It all was based on the sound at performance sound levels. At least on the amplifiers that I took care of.
Switch-mode power supplies are a different situation. The demands on the capacitors are not the same as a "linear mode" amplifier. Not even close. So the designer needs to be aware of the different demands and understand the actual performance of all of the reactive components in a switcher supply. The design of a GOOD switch mode power supply is among the most challenging design tasks around.

 

In this case it seems that there must be serious consideration of the application. so quite probably using the capacitor that does not meet the specification at the working voltage is the wrong choice. GOOD switcher supplies ARE a challenge! Using a part that satisfies the design intent in the actual working condition is the correct choice. That claim should stand up in a design review of qualified folks.