Hi Folks,
I'm a newbee regarding the picking of certain part types for certain applications. The probably basic questions came up when I read "somewhere", that filter caps right after rectification should not have low ESR:
- In which cases/applications should low ESR caps be used?
- In which cases/applications would "higher ESR" caps be the better choice?
- In which cases/applications it simply "doesn't matter?
Thanks for some enlightment with explanations why.
Regards,
Winfried
I'm a newbee regarding the picking of certain part types for certain applications. The probably basic questions came up when I read "somewhere", that filter caps right after rectification should not have low ESR:
- In which cases/applications should low ESR caps be used?
- In which cases/applications would "higher ESR" caps be the better choice?
- In which cases/applications it simply "doesn't matter?
Thanks for some enlightment with explanations why.
Regards,
Winfried
Low ESR caps are needed on switching power supplies where the ripple frequency is much higher than 100hz or 120hz on linear supplies.
Capacitor charging current surges/spikes depend on both the capacitance and ESR. If a linear power supply has a very large smoothing capacitance charging surges will be large provided the rectifier bridge and transformer can support such surges. A very low ESR would make the current surges mostly dependent on C.[dV/dt]. These capacitors are recommended for switching power supplies.
As far as audio applications go, I have found low esr caps to work well decoupling class d amplifiers, and sometimes digital circuits in DACs.
Main filter capacitors also benefit from low esr, but usually not extremely low.
The main pitfall of very low esr parts is that there can be resonance within the circuit and the other components within that circuit.
Main filter capacitors also benefit from low esr, but usually not extremely low.
The main pitfall of very low esr parts is that there can be resonance within the circuit and the other components within that circuit.
Caps directly after rectifiers can have low ESR just that its not normally worth the effort/expense at such low frequencies as mains. Low ESR comes into its own at higher frequencies - say above 3kHz.
Where you shouldn't use low ESR is after certain voltage regulators - the DS will sometimes specify 'minimum ESR' for any cap hung on its output.
Where you shouldn't use low ESR is after certain voltage regulators - the DS will sometimes specify 'minimum ESR' for any cap hung on its output.
Hello everyone,
thanks so much for the many clarifying answers. It's interesting and reassuring, that there's so little difference in the expressed views!
So what I understand is, that low(est) ESR Caps are not necessary or sufficiently beneficial directly after rectification in linear power supplies.
I saw applications where low(est) ESR Caps are used in V+/V- rails buffering of analog OPA circuits, e.g. RC of 100Ohms / 330µF per rail plus 100nF/NP0 Ceramics near each OPA's power supply pins.
Which other analog applications do you recommend for low ESR Caps?
Thanks and regards,
Winfried
thanks so much for the many clarifying answers. It's interesting and reassuring, that there's so little difference in the expressed views!
So what I understand is, that low(est) ESR Caps are not necessary or sufficiently beneficial directly after rectification in linear power supplies.
I saw applications where low(est) ESR Caps are used in V+/V- rails buffering of analog OPA circuits, e.g. RC of 100Ohms / 330µF per rail plus 100nF/NP0 Ceramics near each OPA's power supply pins.
Which other analog applications do you recommend for low ESR Caps?
Thanks and regards,
Winfried
If you are building a B+ supply after a simple bridge rectifier - you maybe don't want low ESR at 120hz in the following CRCRC power supply. Low ESR, if I'm thinking this through might actually impede smoothing results?
It seems the conversation is just looking at charging the reservoir caps. Not those downstream. Like smoothing and decoupling, who are charged with dc
The question maybe changes. Audio is a niche thing for old men. The industry needs low ESR and about every consumer/IT device is full of low ESR caps so the cap manufacturers focus on this because of very high numbers. Low ESR caps have become the standard cap so to speak. Every trick in the book is done to produce them cheap and many.
So: where to even find good non low ESR caps? Those that can be found are generally by brands and types/series one does not pick for various reasons. Exceptions confirm the rule 🙂
A few TH and SMD series by Panasonic are both low ESR, 105 degrees and sometimes even ultra long life but they perform OK in audio and thus they are pretty universal.
For classic regulators this can be tricky at their outputs but then SMD tantalum caps with relatively high ESR perform fine.
So: where to even find good non low ESR caps? Those that can be found are generally by brands and types/series one does not pick for various reasons. Exceptions confirm the rule 🙂
A few TH and SMD series by Panasonic are both low ESR, 105 degrees and sometimes even ultra long life but they perform OK in audio and thus they are pretty universal.
For classic regulators this can be tricky at their outputs but then SMD tantalum caps with relatively high ESR perform fine.
Last edited:
Hi,
For answering each question, you need to consider what is the current level passing through the capacitor and what is the effect of the equivalent voltage drop you have in the circuit, as V=R*I, and/or power dissipation in the capactior, as P=R*I^2. Being R=ESR
Examples to illustrate:
-Low ESR application: in switch mode power supplies or high power inverters since the ripple current is high and it occurs in high frequency (ESR increases with frequency). If not using low ESR units, capacitor will get hot and have its life span seriously reduced. In addition, ripple filtering will be poor, since there will be a high voltage drop in the capacitors.
In audio, main concern is voltage drop in passive high pass filters for tweeters. Heat is not much a concern, since normally is not recommended to use passive crossover for high power systems such as in P.A. But if you intende to have high power passive crossover, then you need low ESR capacitors for that.
-High ESR application: there is not much sense in selecting a capacitor for its high ESR. If you want a resistance in series with a capacitor, you add it externally, since you can better control what you want. High ESR is not a "quality" of a capacitor, but a side effect of a non ideal component. Ideal capacitor would have ESR=0. There are many cases where you want a resistance in series with a capacitor - this is part of your circuit design.
On the other hand, if you happen to have to use a high ESR capacitor, if possible, you can consider the ESR as part or the circuit and live with that.
-Doesn't matter ESR application: Here is when voltage drop and/or power dissipation is insignificant. Suppose you have to decouple DC in a pre-amp input where the input impedance is 100kOhm. You select a 1uF electrolytic capacitor that has "high" ESR of 20Ohms. Make your calculation of voltage drop across the capacitor if it has ESR=0.1Ohm (low ESR) or ESR=20Ohm (high ESR) - it won't make much difference to use one or another. And power dissipation is totally negligible in this application.
I wonder how linear the ESR part of the capacitance is compared to the capacitive part? I presume high ESR due to dried out electrolyte is definitely non-ideal, but how different are cheap v. low-ESR caps in this regard.
I try not to worry about such things, but I do wonder if high ripple current handling is inversely related to ESR when appraising what to purchase. Right now I'm still pouring over values to find the best price/performance for my project. Never mind my audiophile nervosa interference. 😏 I'd be lying if I said I wasn't loosing sleep over this kind of thing. Late night music and digikey. I put a bunch of Cornell Dublier (digikey 800 344-4539) caps in my power amp PS and so far they seem to be doing the job.¯\(ツ)/¯ There's about 50mv pp AC ripple in the 22V DC loaded output. I'm guessing that's fine.
Now I'm building a single stage line level tube pre. So that's fun....
Now I'm building a single stage line level tube pre. So that's fun....
For 60Hz/120Hz power supply, there isn't, in fact, much problem with caps.
ESR for power supply is more important in switch mode (ps, inverters, nobreaks) etc, where high frequency and high ripple current is involved and, thus, high power loss exists. Caps really get hot and may fail if not properly engineered.
If we abuse semiconductors, in general they fail right way.
But if we abuse capacitors, if not too wildly, it's more a matter of reducing life span.
Look at the chart below, from regular TDK capacitors (not low ESR).
At 1:1 rated ripple current, if operated at 45C it lasts 100,000h
But if operated at 75C it will last only 5,000h.
At the same temperature of 60C, if operated at ripple current of 0.5x rated current, it lasts 25.000h
But if operated at 1.7x rated current, it lasts only 10,000h.
ESR for power supply is more important in switch mode (ps, inverters, nobreaks) etc, where high frequency and high ripple current is involved and, thus, high power loss exists. Caps really get hot and may fail if not properly engineered.
If we abuse semiconductors, in general they fail right way.
But if we abuse capacitors, if not too wildly, it's more a matter of reducing life span.
Look at the chart below, from regular TDK capacitors (not low ESR).
At 1:1 rated ripple current, if operated at 45C it lasts 100,000h
But if operated at 75C it will last only 5,000h.
At the same temperature of 60C, if operated at ripple current of 0.5x rated current, it lasts 25.000h
But if operated at 1.7x rated current, it lasts only 10,000h.