Can any electrolitic capacitor always B replaced with eq. ultra low ESR version?

[Kjelt]

Short introduction why I ask this question:
I see a lot in audio fora that people blindly replace all capacitors in their audio equipment with low ESR versions.

Now I am not an expert that is why I ask this question but I can imagine that some electronic circuits do not like to have a ultra low ESR capacitor and can start to oscillate or operate outside the spec.
All other parameters like temperature and lifetime etc. etc. are considered equal for this question 

So that is why I ask this question: can any electrolitic capacitor always be replaced with the equivalent ultra low ESR version or are there exceptions?
And if there are exceptions which are they?

thanks! 

[dannyf]

Yeah. The output capacitors on a voltage regulator for example are quite specific and may not work well with low ESR types.

High voltage applications for example are dominated by electrolytic.

I generally think ultra low ESR to be more of an evil than a friend. You just need it to be reasonably low, not ultra low.

[coppice]

If the difference in ESR is big, you might get an unexpectedly high inrush current when things switch, which could cause trouble.

[f5r5e5d]

you will still find op amp datasheets with 30 year old boilerplate rec for 1u Ta electro local bypass - this is way out of date

the original idea was that the recommended old sintered Ta caps of yesteryear had low ESL and fairly reliable ~1 Ohm ESR

today modern Ta Caps are likely to have much lower ESR, as will Polymer Electrolyte Al and will not provide the local power supply damping/impedance control

there can also be a resonant peak with the parallel ceramic high frequency bypass C and lead/trace inductance

many recent decade Vreg chips datasheets discuss, recommend local bypass cap specific ESR range for stable operation - it may be necessary to add series R


solid Ta caps used in PS bypass need not be replaced if not failed (which is usually spectacular unless the supply has current limiting - think flames)

Ta isn't recommended for audio signal coupling - any film is better if it physically fits, Bipolar Al electros with full thickness Oxide grown on both foils is the best that can be done with Electros - supersizing F and upping V rating help too

[janoc]

Well, readers of those "audio fora" also often replace their power cords with special "deoxygenated copper" ones, buy $1000 speaker cables and similar nonsense ...

Low ESR electrolytic caps are most often encountered in switch-mode power supplies (99% of electronics today), where the low ESR helps prevent the cap from heating up and failing when there is a high ripple current passing through it. That is probably the most common failure mode in consumer electronics - bulging and leaky cheap electrolytic caps and replacing them with good quality low ESR types is a good idea.

In other cases replacing caps blindly is likely a waste of money for no gain - low ESR caps are more expensive than "normal" ones and you certainly won't hear a difference if you replace the filter elco in a linear PSU of your high-end amplifier or the caps in an audio equalizer with a low ESR type. However, your wallet will certainly feel it!

If you get unlucky, they can even cause problems - e.g. some LDOs don't like low ESR caps on the output and will oscillate.

[Conrad Hoffman]

Capacitor design, like anything, is probably a compromise. I worry that a cap optimized just for low esr might sacrifice something else, low leakage or long life for instance. As an example, I measured some very long life caps, 12k hours, for LED lamp use, and was surprised at the high esr, at least compared to standard caps. Thus, I tend to go for middle of the road commercial grade parts, not parts really intended for switching supplies or other targeted applications. I do derate voltage, as higher voltage caps will tend to have lower esr and be more reliable. Don't carry this too far- probably best not to use 160 volt electrolytics in 5 volt circuits! An easy test is to see if there's any series resistance in the circuit. If there is, esr pretty much irrelevant since circuit impedance dominates, though in some locations like amplifier feedback circuits, esr might be worth reducing. Those can be troublesome locations because there's no real DC voltage present. Often 6.3 volt caps are used, which have inherently high esr and poorer performance.

As an aside, people didn't talk much about esr before the advent of switching supplies. One often chooses cap value there entirely by whether it delivers the necessary low esr. For most other applications one talks in terms of dissipation factor, not esr. That tends to be more illuminating, being a ratio and very easy to compare between parts. You can calculate between the two if needed.

[Neilm]

Some SMPS topologies will include the ESR of the output capacitors in the design. This will then affect the control loop characteristics and could degrade the performance of the PSU.

[Kjelt]

thanks all for your input  , so there is much more to it and each case should be evaluated before blindly changing the electrolitic caps.

[ShawnD]

Now I am not an expert that is why I ask this question but I can imagine that some electronic circuits do not like to have a ultra low ESR capacitor and can start to oscillate or operate outside the spec.

Anothe rplace it can be a probelm is the main input filter cap on DC powered equipment, especially if the supply cable is long.  The inductance of the supply cable forms an LC tank with the filter cap.  With low ESR there isn't much to dissipate the energy and you can get high voltage ringing.

Here is a good app note that describes the issue and a few solutions: http://cds.linear.com/docs/en/application-note/an88f.pdf