No need for VC tuning diodes, we have VC variable ceramic capacitors.
Use any large-value ceramic capacitor. Measure the uF while applying a DC bias, and watch what happens.
I tried this on a 22uF 10V cap, in 0805 package. At 2.6V, capacitance decreased from 21uF to 10uF! At the rated 10V, capacitance was 2.55uF.
WTF
The suppliers have apparently been concealing the fact that, for very small ceramics with large values, the capacitance value is not stable, and across the voltage rating, can decrease by a factor of at least ten! (I only measured one example. I bet there are others which behave even worse.)
Try finding this fact in a typical spec sheet. Your 10V capacitor might have an advertised 5% tolerance, but only if you never actually apply 10VDC to it.
What!?! Bill, surely you knew about this, it's certainly not been a secret. The reason you don't see it laboriously spelled out on MLCC capacitor datasheets, is because they specify ceramic type, X7R, X5V, Y5V, etc., on the datasheet. Then they separately publish data and capacitance-vs-voltage curves for their ceramic types.
As Win says this has been known about this for a while. 'Your mission Mr Phelps(Beaty), should you choose to accept it, is to turn this cap into a voltage tunable filter.' or maybe a parametric amplifier ?
Nonlinear ceramic caps have been used in high-voltage NLTL shock lines, to speed up the rising edge of kilovolt pulses.
They are not very useful as tuning devices because the caps with large voltage coefficients have low Qs and horrible temperature coefficients. But then diode varicaps have bad TCs too.
Somebody (Murata?) makes a 3-terminal voltage-variable cap, for RF tuning, that is ceramic based.
One could make a parametric amp with nonlinear caps, for fun, but it probably wouldn't be useful.
formatting link
formatting link
formatting link
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
The problems are the tolerance and tempco. I have a reel of 22 nF Y5V caps that are pretty good varactors. I've been meaning to try making a paramp out of some, to see how good a LF noise figure I could get. That would be a super amusing product. ;)
Cheers
Phil Hobbs
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510
http://electrooptical.net
https://hobbs-eo.com
Vitramon brand comes to mind, and I forget which others. Meaningless curves -- other manufacturers make values all over the place (e.g., a 1uF
16V 0805 that performs better at rated voltage than a 1uF 50V 0805 does at the same voltage), so I have no reasonable expectation that their products are any different.
Do they only make parts that in fact conform to their suggested curves? I have no idea without testing. The market has been spoiled by everyone else. This is why we can't have nice things.
At least some are finally bucking the trend. Samsung parts -- when their freaking datasheets don't actually 404 -- about 30% of the time have a characteristic sheet that shows C(V). Taiyo Yuden is doing this too, IIRC (but in the regular datasheets, that don't 404). Murata, Kemet* and TDK have characteristics on their web-accessible database.
*When the parts show up there. I think most of Kemet's variations fail (you can't find a flex-term part, but presumably the base PN is listed??).
Ferrite beads are where it's still dire. Murata infrequently has Z(I) curves. Only Laird's catalog has complete coverage. I don't bother even checking anyone else.
Careful with the generalization -- some C0G are available nowadays that might be considered "large-value". Not that they would necessarily be your first choice, given the price.
It's not characteristic of the large values, but of the dielectric type, namely, 2.
Your very example proves it's not "at least"... "Up to ten" would be more honest. I've seen a Taiyo Yuden "50V" that was surely constructed like a
6.3V part, and had >10x reduction at rated voltage. They aren't common, but they are definitely out there.
X7R is as good as you can find, and is defined as R = +/- 15% over temperature. (P is 10). You can get "5%" X7Rs but the tolerance is only applicable at room temperature. It's not clear to me if they include aging, probably not (which will blow that tolerance as well after a few years/decades).
Note that aging is set from when the part was last annealed, i.e., heated above Curie temperature. This is typically soldering, so, from the date of final assembly. Aging proceeds roughly as ln(age) so the first few percent are gone in weeks, but consuming the full tolerance can take decades (at least for the common 10% parts).
No no no. Not the known variability associated with any particular material. Read again. This is about capacitance instability for the smallest smt sizes. (Actually a cool topic ca. 2012. If not already in AOA text, it should be.)
Your 22uF ceramic becomes a 3uF ceramic at the rated working voltage, BUT ONLY FOR EXTREMELY TINY SMT VERSIONS, somewhat regardless of ceramic type. This is about the *vast* change for *overly small* smt capacitors having *large* values. See:
Temperature and Voltage Variation of Ceramic Capacitors, or Why Your 4.7?F Capacitor Becomes a 0.33?F Capacitor
formatting link
"Introduction: I Was Surprised "A few years ago, after more than 25 years of working with these things, I learned something new about ceramic capacitors. I was working on an LED light-bulb driver and the time constant of an RC circuit in my project simply did not seem to be right. I immediately assumed that there was an incorrect component value installed on the board, so I measured the two resistors making up a voltage-divider. They were just fine. I desoldered the capa- citor from the board and measured it. It, too, was fine. Just to be sure, I got new resistors and capacitor, then measured and installed them. I fired up the circuit, checked that the basic operation was proper, and went to see if my RC time-constant problem was resolved. It was not. NOT ALL X&Rs ARE CREATED EQUAL"
Here's their graph for various 4.7uF smt sizes:
formatting link
My quick test with a BK LCR meter isolated by 1000uF electrolytic: Murata GRM21BR61A226ME51L, 22uF 10V 0805 X5R
21uF at 0V 10uF at 2.6V 6.8uF at 5V 2.55uF at 10V
In other words, repeat the same test with X5R, but in 0805 package
0.1uF 25V. Nothing unexpected.
Or, try the 22uF value NOT in 0603 or 0805 but in a 1210 package or larger. Works fine, no eight hundred percent decrease.
Again, this isn't about ceramic type. This is about stuffing a
47uF cap into an 0603 package, then not warning users that the tiny capacitor will NEVER give you 47uF, except at nearly zero voltage.
This is only old news, if "old" means five years ago. Then again, maybe it was heavily discussed here in 2013-2014?
Why Your 4.7?F Capacitor Becomes a 0.33?F Capacitor
formatting link
The huge effect appears for extremely small package size and low voltage versions, while keeping ceramic type the same. Well known since five years ago. I only found out yesterday.
--
((((((((((((((((((((((( ( ( (o) ) ) )))))))))))))))))))))))
William J. Beaty Research Engineer
beaty a chem washington edu UW Chem Dept, Bagley Hall RM74
billb a eskimo com Box 351700, Seattle, WA 98195-1700
ph X3-6195 http://staff.washington.edu/wbeaty/
OK, now I'm learning something. But first let me say that I am often putting various SMT caps on the SMT fixture for my HP 4192A LCR meter, and certainly have seen the severe 10:1 reduction at rated voltage, even in large package sizes. But I've been unaware of what must be a specsmanship game, where apparently manufacturers are playing with the voltage ratings for some of their parts. I've often wondered exactly how they come up with MLCC ratings, because it doesn't always seem to be leakage / breakdown related. Instead for many parts it appeared to be give by the 10:1 voltage change, but now you're saying that's not so. OK, thanks, I'll take a look at your links and go educate myself.
I only find one, and from five years ago. This is about uF variation versus package size, only when using large-value ceramic filter-caps stuffed into tiny SMT. They go way, way way outside their spec tolerance when running at their normal WVDC spec. See the maxim appnote 5527 link
We can't use 22uF dk part 490-10511-1-ND, it turns into 6.8uF when run at 5VDC! Instead of 0805 We'll try a 1210 package, like DK part 1276-3395-1-ND?
--
((((((((((((((((((((((( ( ( (o) ) ) )))))))))))))))))))))))
William J. Beaty Research Engineer
beaty a chem washington edu UW Chem Dept, Bagley Hall RM74
billb a eskimo com Box 351700, Seattle, WA 98195-1700
ph X3-6195 http://staff.washington.edu/wbeaty/
uF variation versus package size, only when using large-value ceramic filter-caps stuffed into tiny SMT. They go way, way way outside their spec tolerance when running at their normal WVDC spec. See the maxim appnote 5527 link
We can't use 22uF dk part 490-10511-1-ND, it turns into 6.8uF when run at 5VDC! Instead of 0805 We'll try a 1210 package,
formatting link
???
The AC voltage response is interesting, too. Initial permea\\ittivity is about half the average value. Very standard behavior for ferro{electric|magnetic} materials, but not often mentioned outside of general material properties.
But yeah, 22uF at 5V is more a 1210 thing, think I've seen 1206s that are reasonable too.
I've, on rare occasion, even seen 1206s that are better than 1210s of similar value and rating. Only way to do, check the characteristic. Tedious as hell; such is engineering...
Bill that was very interesting, thanks. (I would only use cog/npo ceramics in a filter.) But it does go to show that ~energy storage goes as the volume. (hmm, not sure that's true. Well at least that's ~true for film caps.
So Phil, my para-amp comment was a lark. I've got some idea how to do a tunable filter with V-caps. (I need two of them in series.) But what would an opamp V-cap para-amp look like? ... You might be able to sell that just for the fun of the idea. Heck I would like to build one even if I can't sell it. (I found lotsa stuff behind paywalls
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.