Polarized vs. Non-polarized caps

[swb1]

Hi,

When do I know when to use a polarized capacitor vs. a non-polarized capacitor?? Is there a rule of thumb??

Thanks,

swb1

 

[itsmoked]

I agree that is odd. I have used them in P too lots of times.

My bad?

 

[OperaHouse]

We had a piece of equipment that had to go in a mine. It had to be modified with resistors placed in series with each of the caps as a current limiter so it was incapable of creating a discharge that would cause an explosion. Probably that tant was in backwards or at least marked backwards. I found a diode once in production test with the band printed on the wrong side. Perfect diode, just marked backwards. Stuff happens!

 

[zeitghost]

Tantalum capacitors are notorious for going short circuit.

They really don't like reverse voltage either.

Some surface mount tantalum caps have a fuse built in to remedy this.

I had a tant bead on a pc motherboard short out once... a very unpleasant smell indeed.

I read somewhere that the usual arrangement of a 100nF cap per chip with a 4.7uF electrolytic every now and again is intended to damp the ringing that might occur on the power rails with the high Q of the ceramic caps.

Substituting a tant with a much lower esr won't have the same effect.

 

[Higgler]

Regarding tantalums in parallel, I recall some verbage such as they needed a small series resistance ? 1/2 ohm or so cuz at first turn on of voltage (maybe first dc power up) they can cause a problem. I'm certain I remember that, although maybe I remembered the wrong info. Our power supply tantalums definitely blew up during First Article test.

kch

 

[itsmoked]

Higgler you must lay off the juice while reading verbage... Stick with verbiage.

OperaHouse; That is a classic "intrinsic safety" technique.

 

[zeitghost]

Just found a "green" led that lights up orange... looks quite surreal.

 

[OperaHouse]

Twenty five years ago that was my first introduction to "intrinsic safety." After designing IS devices and bringing them through agency approval, I can tell you there was nothing safe about what we did back then!

 

[felixc]

I remember reading about some families of tantalums that have restrictions on dIdT and that they do require some inrush current limiting in power supply applications. In many cases the presence of a fuse or a PTC takes care of it. But I can't relate it to the addition of resistors when you parallel tantalums.

 

[Warpspeed]

I have never seen a series resistor with a tantalum, and cannot see the advantage of fitting one. It all seems rather pointless if you are trying to bypass high frequency noise with a low shunt impedance.

Most power supplies start up reasonably softly, at least slow enough for there not to be an inrush severe enough to be a problem. I cannot see a one time inrush being destructive either. Very high frequency high amplitude ripple current may cook a tantalum, mainly because they are so small, but that should become very obvious at the initial prototype testing stage, and a resistor is not going to help.

All very strange.

 

[zeitghost]

So why do the manufactures go to the trouble of putting fuses in surface mount tants?

 

[itsmoked]

Reverse bias => draws huge currents => causes rapid disassembly => technician eye damage => attracts lawyer vermin?

Hence: put in fuse.

 

[zeitghost]

The other thing about tants is this:

In 1976 I bought an immensely expensive Yamaha CT7000 tuner.

In the audio path it had the then newish tantalum bead capacitors.

It's built like the proverbial tank with a 7 section variable capacitor in the rf section.

Sadly someone installed the tants the wrong way round: after about an hour or so I used to get a deep sort of rumbling under the audio.

Eventually took it apart & changed the tants for something a bit more audio friendly, if a lot larger: a couple of polycarbonate caps.

 

[felixc]

But the fused tantalums have a rather poor ESR. Duh! All that remains is that they last longer than electrolytics.

There even were 3-leaded tantalum drops. To make sure that there was no way to insert them backwards. I have seen holes bigger than a quarter in a multilayer PCB that had a reversed tantalum. We've had to install smoke detectors and halon (that was long ago) systems to make sure that the burn-in chambers weren't scrapping whole lots because of a single tantalum.

Tantalums in high-end audio? Even if connected correctly, their leakage currents may generate hiss depending on where they are in the path. Nice decision zeitghost.

 

[zeitghost]

Yes. I was somewhat surprised too.

It cost £450 in 1976... which is knocking on for £4000 today.

I must have been mad.

 

[felixc]

I have a vintage Hafler preamp which had tantalums too. Heresy, but it sounds great!