muRata GC3-series dielectric has capacitance superior to X7R under d.c. bias

We were just brainstorming a product that would have three switched timing ramp capacitors, the largest being 1 uF. I was assuming that no ceramic would work, what with tempco and DA. So I was figuring on using a big thru-hole film cap (not! a surface mount part).

In another project, I'm using a lot of 4.7 uF 50 volt caps in parallel at 48 volts, and figuring they might be about 0.4 uF there.

It's about time we had some better dielectrics.

COG is very good for DA. Your array would be a little sizeable though.

The film caps are excellent if you can stand the ESL. But sometimes, you can't.

Yikes! That's particularly awful.

I'm using a 2.2uF 450V MLCC. It's about 10 or 15% of 2.2uF where I'm using it.

That sure would be nice.

I chatted with a physics-type working on super-caps at the university a few years back. They were just a few years away from gigacaps that would change everything. And they probably still are.

Funny how that works.

Cheers, James

This thru-hole ceramic is getting sort of interesting:

If you look at the d.c. bias graph, X7T retains half its capacitance at rated voltage, versus ~18% for X7R.

The Z(f) graph suggests ESL

Energy storage needs volume! Even 1210 caps have better c:v curves than 1206's.

In the graph I posted, energy is going up as CV, not C*V^2. Well, at least it's going up.

I may as well just use a lot of 1206's, which we have in stock. The pick-and-place will shoot them onto the board, and we'll have less inductance.

We have! People just want to be cheap - myself included.

NT

Yep. I'm using 2220's, nominally 2.2uF/450V (pick one). They're down to

I just went through all of Digikey's offerings yet again and didn't find anything better.

I'm using ceramic for the fast cache and TDK film caps for the bulk store. Works well. (I need to hold .1%.)

Cheers, James