ACHIEVE LOW ESR CAPACITORS...

[Kain]

Hello again, I was reading lately about swithing PSUs and noticed that they require a low or ultra low ESR capacitors in general. I have checked on different manufactors but I read that a ultra low ESR can be achieved by connecting smaller capacitors in parallel. I looked in google for some more information about it but either there isn't much or I am not looking at the right place. Can somebody here explain how ESR decreases with connecting capacitors in parallel and how we would calculate it? My guess is that the rules are the same as for resistance with the difference that we are talking about impedance of capacitors in AC waveforms and their reactance. Any help is appreciated  :)

[audioguru]

Hi Kain,
A switching power supply needs an output filter cap with a low ESR so that it can charge quickly by a high current and supply a high discharge current into a load. With the high currents, an ordinary cap would get hot and explode!
Adding smaller caps in parallel with an ordinary large one is different, it reduces the total impedance at high frequencies but usually at low currents.

An electrolytic cap is made by winding its foil and insulator around and around like an inductor. Therefore it has a high inductive impedance at high frequencies. Even a 0.1uF ceramic disc cap has a very low impedance at medium frequencies but resonates with its own inductance at high frequencies so is ineffective. A 0.001uF ceramic disc cap has an impedance of only 1.6 ohms at 100MHz. ;D

[Kain]

Ok I understand this :) Should I take this responce as "ESR cannot be decreased by connecting capacitors in parallel?" And also, is ESR portion of the real resistance or capacitor? By our discussion so far it seems like the only way to get low ESR is to go find a speciffic one. I wonder who might have one with ESR 0.01 at C > 470uF?  ;D

[Kain]

Ok, how about this? If the impedance of a capacitor is the sum of the real and imaginary part then if we know that the impedance of a capacitor that is given in catalog is 0.01Ohm then this means that the sum of the ERS and the imaginary are 0.01Ohm at most, meaning that ESR has to be even less than that. Is this correct?

[audioguru]

No, The capacitor's ESR is in series with its impedance (capacitive reactance) at a certain frequency.
To be a good power supply filter the phase shift of the capacitor would be very low.

[Kain]

Oh, so they mean capacitors reactance by saying impedance... Ok, makes sence.  ::)

[Almost-retired]

This still doesn't, and nowhere in this thread has it been mentioned exactly what the definition of ESR is.

Simply put, its the Equivalent Series Resistance, the equ of sticking a resistor in series with the capacitor.

This is, or was 10 years ago, a difficult to measure item, and often the old eyetube cap checker was relied on to test this, but its a very poor test, done at an applied frequency of 60 hz.

Many times better is a meter I ran into about 10 years back, fairly pricey at about $175, is called the Capacitor Wizard.  It measures the capacitors resistance with a few millivolts of a 100 khz signal, and can indicate clear down to an effectively pure capacitor by shorting its leads.  Its very very good at measuring the first 10 ohms of the caps ESR.  It will let you repair stuff that no other failure indicator can find.

Trust me, you don't want a 470uf capacitor in your critical circuit that doesn't measure under 2 ohms with this meter.  <0.3 ohms is good, but some new ones won't quite hit that low a figure.  Even 2 ohms can lead to serious heating and poor performance of the system in some cases. Serious enough heating of course leads to alu foil, antifreeze and smelly kraft paper all over the place, hard to clean up in tight places...

--
Cheers, Gene