Holy c(r)apacitor dielectric, Batman!

See "Capacitors_X7R_Y5U.zip" on the Device Models & Subcircuits page of my website... modeling temperature _and_ voltage effects with these capacitor types. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| STV, Queen Creek, AZ 85142    Skype: skypeanalog |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 

             I'm looking for work... see my website.

As you found out, they're not good. One thing to watch out for with X5Rs is that they're only rated to 85C. I only use X7Rs and C0Gs because the rest are crap.

Murata has curves for every capacitor. The difference between caps isn't obvious, some being much worse than others that are very similar otherwise. I look up each capacitor I use to make sure it's what I think it is.

Yes, or use a physically larger cap. Larger caps tend to have a shallower DC/dV curve but, again, it varies depending on the particular cap.

Yes.

Yep. Designers should know these things. ;-)

I'm well aware. I think I was one of the first to point out the problem decades ago, and still have a reel I marked for eventual use as varactors :-).

But the quest to make ever-smaller SMD caps has made even dielectrics I previously measured and considered "safe" now "unsafe", at least in this example.

Cheers, James Arthur

;-)

The tolerance is pretty poor, though.

The amazing thing, as you noted above, is that the performance of similar caps is so variable.

I'm searching vendors now, wasting time looking for an honest bypass cap.

So far TDK is the only vendor I've found that's up-front about capacitance loss under bias--curves plotted right there in their data sheets.

MuRata, Yaego, AVX, Kemet, Taiyo Yuden, Samsung all get failing marks--the info's not in their short data sheet, and not easily discoverable on their websites.

I consider these ratings fraudulent. A 1uF/35V-rated cap should measure 1uF @ 35V.

Cheers, James Arthur

I posted about the same effect recently, and north of a hundred (redacted) individuals went pantybunch over my penmanship. The thread may hit 200 posts, 95% or so nonsense. Grumble indeed.

Since those caps can stand many times their rated voltage - 10x at least - and since they lose most of their capacitance well below rated voltage, how do the makers decide what the voltage rating should be?

Note that the rated voltage on the part that you link to is

1V (35Vdc)

with application (up to 50V)

I think it's technically a 1 volt cap.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

Power parametric amplifiers and frequency multipliers and variable attenuators.

I'd like to make an oscillator with a ceramic cap as the active element, but I suspect it's not possible.

This is interesting:

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

Here's a TDK X7R 1206 2.2uF / 25V--

It's down 50% at 25VDC.

This would be a lot more manageable if there were easily-accessed real data sheets for each part, but there mostly aren't. Digikey mostly gives links to 168-page full-line catalogs, but *the catalogs don't have dC/dV info!*

I spent hours on this today, turned shy after nearly getting bit trusting my fellow humans' specs. Very annoying. And still not much the wiser--each cap is a crapshoot.

My current task has some 'hot' caps with 300-400V on them, and I'd like more than 20% of rated capacitance, please. Need it, actually.

Cheers, James Arthur

The best capacitors in the world are on-chip silicon dioxide. People sell 16, 18, 24 bit A/D converters that depend on the stability and low DA of SiO2 caps. Big CPU and FPGA chips now have substantial on-chip bypass capacitors, which means that don't need heroic off-chip bypassing.

I wish someone would find a way to make multilayer SiO2 caps using semiconductor lithography.

There are some MIS caps for sale, mostly single-layer, low values.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

With some (or actually many) other manufacturers this sort of data is hard to cajole out of them. Maybe because it doesn't look good. So I usually prefer Murata caps when it comes to situations where capacitance retention at higher voltages is critical.

Why not? All it takes is the application of a DC voltage across the capacitor to vary the frequency.

A nice use of an otherwise nasty side effect. Or "How to turn unwanted slag into gold". We did that in another domain on Monday, making bread from the trub (yucky residue) from beer brewing. We could not stop eating.

--
Regards, Joerg 

http://www.analogconsultants.com/

No, I meant a circuit that takes in DC and makes AC. A negative real impedance, like a tunnel diode, can do that, but a negative reactance, like a Z5U capacitor, apparently can't.

An AC pumped nonlinear-cap parametric oscillator should work. Some old

20 Hz telephone ringer supplies were magamp oscillators pumped at 60 Hz, so some capacitive analog should be possible.

If DC to AC were possible with a nonlinear cap, somebody would have done it by now.

Beer and bread have a lot in common. The Tassajara Bread Book suggests that, when we bake, we pause for a moment and give thanks to the yeasts that are about to die for us.

The other wisdom in that book is that, if you are going to use sugar, don't hold back.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

Ok, if you meant a passive-only parts oscillator that isn't likely going to work with such a capacitor.

The amazing thing is that you sprinkle a small packet of yeast into the wort, it makes beer and as a side effect it also makes another 3-4x as many yeast cells as went in. I re-used part of the trub with the newborn yeast in there for a Stout but in order not to over-pitch I had to do something else with the remaining 60-70% of harvested stuff.

Someone on a brew NG had an example where a British brewery used the same yeast strain since 1957. In the hobby world it is considered best not to go past 5-6 generations because of the chance of increased wildness and contamination.

And if you didn't hold back enough make sure there are cleaning utensils for when the dough raises above the pot and runs down the sides.

--
Regards, Joerg 

http://www.analogconsultants.com/

don't they all use the same yeast to get their specific taste, ever since Emil Christian Hansen at Carlsberg in the 1870's figured out how to isolate a pure yeast?

-Lasse

I don't know about that. The main yeast manufacturers at least per what I can buy are now in France, the UK and the US.

It doesn't have to be pure. What I pitched into the Stout had lots of residue from a Cream Ale with it. That direction works. The other one, darker beer's yeast cake towards lighter beers, not so much.

--
Regards, Joerg 

http://www.analogconsultants.com/

when some one like Carlsberg is brewing in ginomous batches and it all have to taste consistently they don't take any chances.

I've seen a documentary on Heineken, they had their specific Heineken yeast in a cryocooler and grew batches of it in the lab

I believe that for some historic reason you can go get yeast for free at the Carlsberg brewery

-Lasse

The really big breweries with a full-blown food chemistry department will probably all do that. Not quite my taste, I prefer craft beer from very small breweries and I go to great lengths to get it. Like a rather muddy 24 mile mountain bike ride yesterday to visit this brewery operated by just two guys:

Busted a rear suspension shock mount on the way back :-(

You wouldn't catch me drinking it :-)

That can be practical but only for people living near Copenhagen because they will likely provide their yeast in liquid form and not freeze-dried.

--
Regards, Joerg 

http://www.analogconsultants.com/

How quickly do those critters react? They show in Fig 3 drive with PWM at 100Hz (!)

piglet

Hi-K ceramic properties are bad enough already, without introducing time-domain effects!

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

Some San Francisco sourdough bread biology has been around since the Gold Rush. We know a member of the Boudin family whose great-grandmother rescued their starter breed as the flames of the

1906 earthquake were closing in.

Sourdough is a pH-mediated equilibrium between a unique lactobaccilus and a yeast.

You don't like the kind of beer that I like, but the Guinness Blonde is made in the USA with Irish yeasts. We love it.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com