Ceramic capacitors for higher voltage switchers

Well, take a guess at power dissipation and find the damping factor (X7R is what, 1.5% or something?). If you're running hot, low power is probably a good idea, like 1/8W or something. And use lots of copper (ha, like you ever have the space).

So stack them horizontally instead? :-)

Or if you're cramped for horizontal space, maybe you could make a capacitor daughterboard so they're stacked horizontally, on a vertical board...

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms

Just solder the caps in so they are on edge (vertical); one may be able to use 4 in parallel this way and use the same PCB board space.

I wish they were physically a tad larger.

The client won't like that :-)

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Can't do that, it must remain machine-solderable.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Use hermetically sealed military versions then.

The difference in performance is worth the difference in price.

Pretty much very labor intensive assembly work. Not very good for any automated process, so not good for anything that has intentions of large volume production.

For short run, hand assembly though, no problem.

Vertical daughter cards are the way we did it in the HV lab.

50 Watts is a lot of loading for a ceramic cap bank to manage the ripple of.

I know, and I do that. But it's often not easy to find some for higher voltages with low enough ESR _and_ being available without huge leadtimes.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Haven't used these particular caps but I have run some X7R 250V surfamce mount 2220 caps at 1A p-p ripple current at 10Khz. No noticable heating at all.

A also run a 500V 2225 .47uF caps on the output of a 150V rail that can put out 75W (300Khz ripple) with no noticable heating.

You really have to hog some current on the ceramics to heat them up.

That said I did have some x7r .1uF cap with teh yellow epoxy turn red when a

5A audio amp IC started oscillating into it. It was in an RC damper on the output. Scratched my head for a while trying to figure out why that cap got hot. I had thought before that ceramic caps were nearly ideal. :)

I can't use these large SMT sizes in this case because of thermal stress issues, cracks, and all that.

That's one reason why I asked. I've managed to turn capacitor ceramic into bubbly green glass. But that was several MHz and several amps through just one. The usual, had only one, did a quick experiment, "Come on, hang on another 10 secs" ... *PHOOMP*

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Rock | Joerg | Hard Place ;-)

The designs that get chucked over onto my desk are nearly always like that :-)

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

You have been reported to the 'Society For The Prevention Of Cruelty To Components'.

--
Greed is the root of all eBay.

You can get capacitor characteristics from TDK.

The resistance, reactance and voltage/temperature dependencies are illustrated graphically.

A 0.1uF 100V X7R part in 1206/3216 body size, 3216X7R2A104M, exhibits an internal series resistance of 30mOhms, over the range of common power conversion frequencies.

RL

I bet pick-and-place can do it..by using a small dot of adhesive to hold a cap that is placed, another dot nearby for nest cap, etc (all dots could be placed first and then the caps). Solder paste mask can be same as if one cap laid down in conventional form. "Never" say "never".

Oppenheimer wannabe?

-- "Electricity is of two kinds, positive and negative. The difference is, I presume, that one comes a little more expensive, but is more durable; the other is a cheaper thing, but the moths get into it." (Stephen Leacock)

Sounds trivial but 30mohms times 5A squared is .75W on a 1206 cap. You now have made an LED. :)

At 100KHz, that would be 80Vac across a 0.1uF capacitor - a little more ripple than is anticipatd here.

The internal series resistance of most LEDs is some order of magnitude higher.

Degrees of triviality are possibly more relevant electronic units of measure than most appreciate. Especially in this news group.

RL

Fred Abse Inscribed thus:

:-) I've had ceramic caps get hot enough to melt the solder and fall off the PCB.

--
Best Regards:
                     Baron.

Sure, everything can somehow be done but once you deviate from industry standard practice -> $$$.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.