Ceramic capacitors

Shhhhht! Don't let out all the secret tricks :-)

Did you verify that the 2V and 36V supplies were effective shorts to ground for AC?

Inneresting results -- makes me want to go check Digi-Key to see if that mix is available for lower capacitance ranges, too.

There were one film cap + 100u electrolytic to ground + whatever is in lab supply

Yup:

Many younger engineers consider Y5V and Z5U junk. Little do they know ...

That's another point. I have an application (an AC tweak on an improved laser noise canceller)that is crying out for a MVAM109 varactor (500 pF), except that they don't make them anymore... I'm looking at alternatives, e.g. a 100 pF cap with an inverting op amp driving the other end, with gain set by a dpot. Unfortunately, getting good noise performance requires low resistance feedback, which in an inverting configuration requires a BF862 buffer in front, which needs AC coupling,... about a dozen parts and 30 mA of supply current when it's done. (Blech.)

Jameco has some 2200 pF Y5Vs made by "Prosperity Dielectric Co", but that isn't too confidence inspiring from either a quality or supply perspective.

I need it to be a pretty good adjustable bypass cap (one end grounded), nice and capacitive from about 10 kHz to 10 MHz, ballpark 100-500 pF. Keeping the excess phase shift low at high frequencies is the key to this.

Any suggestions?

Thanks

Phil Hobbs

Yes, check out these:

Datasheet:

But they only come in through-hole.

Well, I'm not a young engineer any longer and tend to consider Y5V and Z5U as junk. I don't even regard this as "secret trick". Sure you can build a VCO this way, but then you also have to add a thermostat, which will turn the secret trick into an expensive trick. Not to mention that the thermostat won't cure the thermal hysteresis problem.

Klaus

It's the same with varicaps, FETs and whatever else you want to use in this function. A good designer designs out those dependencies or makes the circuitry live with them :-))

It is not trivial to build a variable resonant circuit in the kHz range where the varying element is supposed to cost no more than five cents, other than with Z5U and such. I've also done super-cheap spectrum spreaders that way for EMI mitigation. There, a temperature-dependency doesn't matter at all.

Of course, in Phil's case he needs to watch out where he puts the capacitor. It should not be in a path where there would be large signal amplitudes across it, else the whole thing will become uncomfortably non-linear.

It's going across a current mirror, so the maximum swing should be less than 50 mV even if everything is going nuts.

Cheers

Phil Hobbs

Then just take a look at Z5U and Y5V for some other cap. They won't be grossly different but AVX doesn't have the curves in the HV cap datasheets. This would give you a first estimate of how bad it would be at 50mV. Should be pretty miniscule.

P.S.: Digikey carries over two million of this Y5V capacitor in stock,

That many engineers can't be wrong :-)

Think parametric amplifier!

I've seen kilovolt shock lines that use ceramic caps as the nonlinear elements.

John

Back in the "old daze" when varactors were new and expensive, some fiddled with high voltage diodes and some seemed to be rather useful. It seems that every maker uses different diffusion schemes, so that (making this up) a 1N4007 from one maker may work but a 1N4007 from all others does not work. Perhaps 200V diodes may do better than 1000V diodes.

=20

this.

Some 30 years ago i was testing low tempco chip caps that had a huge voltage coefficient. I wish i could remember what they were made of.