Using Many Ceramic Chip Capacitors Series vs Parallel

Maybe you could use a variable inductor instead? You know, control winding, variable permeability, etc.

Or a switched array of trim caps, & use a smaller varactor for the residual.

(Of course there's always the dual varactor, make a VCO with the spare, & monitor the drift...)

Best, James

We're planning a credit-card-sized delay generator, so we haven't the space or power for saturable magnetics. TCs might be interesting there, too.

We have used the Maxim Flecap to do this.

At powerup time, we center the varicap then step the flecap to the closest frequency. That reduces the required varicap range by 32:1.

I guess you could do this with a tiny dip switch, too.

Or start with a magical super-accurate, low-drift, low TC, high-Q resonator that doesn't need a lot of pull range.

John

Hey thanks for posting that.

If I could repost my first post, it would be like this:

"If only given ceramic chip capacitors*, what would be the best series/parallel arrangement for Ct= 50uF (200V rating) with a 1.2Arms ripple current."

• I'm using lots of ceramic chips due to a >11 lifetime spec and I have a very tight space. I've allowed a shocking .00 budget for the capacitors.

I wish I could use a single electrolytic but I don't have the space. It has to be less than 9mm in height. I can use popular 8mm diameter throughhole electrolytics. I can get 50uF by paralleling about a dozen, but I have to parallel more to set a per-cap ripple current for long life.

I think I got enough hints to figure it all out anyways..

D from BC myrealaddress(at)comic(dot)com British Columbia Canada

Are electrolytics used in military equipment?

D from BC myrealaddress(at)comic(dot)com British Columbia Canada

It seems to me a lot of devices in parallel is like a test to find the weak sister. You take a part that it 99 percent reliable, parallel 50, and you get 61 percent. Of course, your caps will be better than 99% reliable, but still, I'm not sure I like paralleling 50 devices.

I'm not sure. When we have to test our stuff to -20 or sometimes -40C, we can't use regular aluminums. MnO2 dry tantalums are OK at low temps, but we don't like them as bypasses on power lines, because they tend to explode.

The military choice is silver-cased wet-slug tantalums, good down to around -55C.

John

"4) Taiyo-Yuden JMK105F474ZV-F, 470nF/6.3V, Y5V, 0402 0v 440nF (520nF on first measurement, 440nF after voltage had been applied then removed.) 1v 238nF 2v 134nF 3v 82nF 4v 64nF 6.3v 40nF"

Are these numbers correct? This doesn't even look like a capacitor? If I calculate the charge (Q=3DCV) it looks like no more charge went on as you raised the voltage.

But maybe I'm misunderstanding what you measured.

George Herold

I measured #4 yesterday. Yes, that's what they do.

Small-cased packages have greater capacitance-loss with applied voltage, so #4 (in an 0402) is (deliberately) a ...

... worst-case scenario.

Grins, James Arthur

Is it the small package or the material. (Well, not that there is much distinction.) I've never used the Y5V ceramic. I'll have to order some and make my own measurements. It's not that I don't believe you, but until you've done it yourself...

Hmm, Now I'm wondering how I'll measure it. Clearly I can't measure a long RC discharge time. I can wiggle one end through a resistor and find the frequency at which the 'wiggle' voltage is down by 1/2. Then add a DC bias.

Do you realize your numbers suggest that once the cap gets 240 nano Coulombs of charge on it, it just can't hold any more. (Sorry for repeating myself, but this just looks weird, more like a battery.)

Scratching head, George Herold

Smaller package implies higher-K material, ergo higher dC/dV.

pg. 39-41:

Cheers, James Arthur

the capacitance is the slope of the Q/V graph so it's not quite that bad.

.)

.)

on

Thanks for the links! More than ever I hope I never have to use these. It's a bit of a misnomer to call these 440 nF/ 6.3 Volt caps since they are down to 10% the stated value at the maximum voltage.

Barium Titinate is weird stuff.

George Herold

Yup thanks Jason, I realized this on the drive in to work this morning.

George Herold

I knew Z5U/Y5V were bad, but I'd forgotten just how bad. It's almost a crime to call them 'capacitors'. Maybe 'piezotemporo-voltemp transducers' instead.

Excellent for PC green job applications though--these guys

claim a 500-mile capacitor-based electric car:

One could imagine the following fine print: Note: 500 mile estimate based on average driving habits and capacitance rating with zero volts applied. Your mileage may vary.

Cheers, James Arthur

Yes, mostly wet and solid tantalums. Wet aluminum in ground, mobile, and seagoing equipment with large power appetites; much less in airborne. Missile, space, and satellite applications eliminate wet types pretty much entirely, using batteries instead.

Further musing: a series string of very nonlinear ceramic caps, driven by AC, could be dynamically unstable. One cap may "run away".

John

Yep. That could be interesting, failure-wise.

Cheers, James Arthur