Ceralink capacitors

Yup, they're electrets. Were featured in the Google Little Box inverter winner a couple years back.

Downside (besides price): don't push them towards Tc or it goes way. Makes reflow soldering kind of impossible.

Presumably you'd be able to re-polarize them in situ, but I'm not sure how to go about doing that (it should just be applying voltage bias, then "freeze" it in by cycling above and below Tc; not sure how much leakage you need to supply in the process, though).

Tim

The data sheet pretty much says that these things are useless.

kes

ow

you

Temperature tolerance? Or the mad price? I only got 1/3 way through the dat a sheet.

I reckon it would be easyish to make a cheap imitation of these things. Use a couple of nasty Y5U ceramic caps back to back with the centre tap connec ted to a big voltage via a high R. As the applied signal approaches the bia s voltage, capacitance goes up to rated value.

NT

Don't expose to AC, sunlight, human fingers, liquids, reduced oxygen, heat, humidity, whatever. Shelf life is 6 months.

They do have impressive specs for CV and AC current, except that we're warned against applying AC.

ter

Makes

e how

ge you

data sheet.

Use a couple of nasty Y5U ceramic caps back to back with the centre tap con nected to a big voltage via a high R. As the applied signal approaches the bias voltage, capacitance goes up to rated value.

Back to back Y5Us could do far better, albeit requiring a connection to the Vcc rail.

NT

On Friday, July 21, 2017 at 7:16:36 PM UTC-7, John Larkin wrote: ...

Some of those are the usual requirements for components that would undergo reflow - the constraint on oxygen partial pressure must be above 1 mbar is a bit unusual.

AC current does not require alternating voltage - they are designed to be used under DC bias conditions with high ripple current.

They are not for resonator use - with the large change in capacitance with voltage that would probably create a lot of distortion anyway.

kevin

Useless indeed. The price alone ran me off. Sorry for the post, guys.

Well, it's interesting. PLZT ceramics are really weird, electrically and optically. But they seem to be kind of erratic.

I've lost a lot of money around optical PLZT applications. Like on this one:

Well, no:

What's the value of (small capacitor ) (large capacitor)?

That's what's special about those thingys, and why they're interesting for inverter applications (assuming you have an application where the space, cost and usage precautions are all suitable).

Tim

Wow, I would have thought this defies some kind of physical law, but I guess not. Glad I don't need these in my designs, they ARE pricey.

Oh, wait, they have (Gasp, horror) LEAD in them! Ohhh, Nooooo!

Jon

That would work if using the right values, but let me improve on it with a Y5U in series with one that does't vary much with applied V.

. Vcc . | . R . | . Cin -----||---+

Great! Pack it into the same volume as the Y5U, and I'll pay you $20 each for the finished component. ;-)

Tim

Do they have any real-world customers?

NT

kes

ow

you

They for "....compact solution for the snubber and DC links of fast-switchi ng converters based on SiC and GaN." I wonder which part of the topology is the "DC link?" The voltage ratings are pretty high. These are for state of the art switchers, high frequency, compact- and not for the commodity swit cher market. The price probably makes sense.

hing converters based on SiC and GaN." I wonder which part of the topology is the "DC link?" The voltage ratings are pretty high. These are for state of the art switchers, high frequency, compact- and not for the commodity sw itcher market. The price probably makes sense.

I'm guessing they're something like Y5U with a built in electret so relativ ely compact. What app makes $10 caps worthwhile just to reduce cap size? Sp ace missions?

NT

DC link, main DC rail bypass. So, 300 to 700V for industrial inverters.

Tim

Very good. I see now some really esoteric very high efficiency topologies made possible by the GaN. Off line w/o input rectification for one.