I was just wondering... when should I use polarized caps? What's the benifit here over a regular one? I'm setting up a voltage regulator... a LM317.... and I'm putting a cap at the input voltage and a cap across the output voltage... can I use non-polarized caps here? Does it matter?
The polarity is necessary in the capacitor design. All but small value capacitors will be polarized, its part of the process. If the capacitor you are using is not polarized you need not worry about it usually. A filter capacitor in your output will be polarized. You can obtain non-polarized capacitors in larger values but these are special use and are not necessary in your case.
Thanks for the reply, so... basically it's something more to do with the way the capacitor is designed than it is something that's useful in any particular kind of circuit? I can see that in AC circuits it could be bad to have a polarized cap... but in DC circuits, does it not matter whether a cap is polarized or not... they can be used interchangeably... it just happens that maybe certain values of caps have a smaller footprint if they're polarized...
Polarized capacitors are Aluminum or Tantalum Electrolytics - they typically have much greater capacitance than a non-electrolytic type of a similar size, so are used where large capacitance is required.
Peter Bennett, VE7CEI
peterbb4 (at) interchange.ubc.ca
Ohhh... "A filter capacitor in your output will be polarized." why can't it be non-polarized? It can be, it's just not likely that you will come by one, or want to spend the money when it's not necessary.
Some regulators, like the LM317 and the 78xx family, have been around for 20-30 years. When their datasheets were written, you couldn't get cheap large value ceramic capacitors. It was assumed that for 1uF or more, you would use a polarised capacitor.
These days you can get 10uF ceramics which fit in 0805 packages. If you use them, you need to watch the voltage rating (small components often have low voltage ratings), and check that the dielectric (X5R, Y5V etc) has a low enough temperature coefficient that the capacitor will not drop below the minimum level of required capacitance at any expected operating temperature. Having said this, I have used ceramic capacitors with voltage regulators successfully.
I hadn't thought of the ESR but if John says it is a potential problem then it is a potential problem 8) I looked into using some tantalum (polarised) capacitors the other day, and found several types that had surprisingly high ESR's.
Wouldn't you say a non polarized cap is always more reliable than a polarized cap? If you can get by with say a 1uF ceramic, you will be better off than just brick-walling with a 4.7uF electrolytic, especially since the electrolytic has a big tempco. I guess my point is sometimes less is more.
When you reverse polarity in a polarized EL, you get a short.
In an NP El, you get one shorting through and the other working, regardless of which way it gets hooked up or excited, so the end result always merely appears as a properly working, properly attached cap.
This still doesn't answer the question regarding how the reverse bias is prevented if you don't use diodes.
I have a few crossover caps that are non-polarized. I think they are just freakin' huge to get the capacitance. I don't believe they internal contain two caps, though I have no way to prove this without cutting them open. They don't have the "crimp" in them like an electrolytic.
That hasn't been my experience. One the one occasion I've made this mistake the circuit seemed to work fine until suddenly there was a loud bang, and the cap in question deposited its contents around the available space.