6
\$\begingroup\$

Why is that when I use electrolytic capacitors (the cylindrical ones) and measure the end with respect to ground, there is some DC voltage appearing (where in fact it should be 0 volts since capacitor passes only AC signal but blocks DC signal), but when I substituted a non electrolytic capacitor, like tantalum capacitors, they work well. In a simulation program, electrolytic and tantalum shows no difference and they block DC, but in real life, electrolytic fails to function correctly. What's happening?

\$\endgroup\$
15
\$\begingroup\$

Electrolytic capacitors are well known to pass a small DC current. How much they pass is a complex function of temperature, capacitance, the age of the capacitor, and the DC voltage across it.

This is because the insulating film is formed by electrolysis; as the current flows, it electrolytically deposits an insulator on the anode (which is aluminium : this is the same process as anodised aluminium) and the insulator eventually stops the current flow.

So for example an electrolytic that has been unused for several years will draw quite a large current (milliamps) for a while, then the current will reduce as the insulator re-forms itself. The capacitor is healing itself.

The result of this is:

  • do not use electrolytics in series with high impedances (say, as coupling capacitors between amplifier stages)
  • do not use them with no (or reverse!) DC across them (they will not heal, and reverse DC will break them down)
  • Ideally you want 2/3 their rated DC voltage across them to keep them formed (at least 50% anyway)
  • do not use them in circuits like R-C timer circuits where leakage will change the time constants
  • remember that at high temperature (usually 85C) their lifetime is rated at 8000 hours - which is only a year!

Tantalums can be a better choice but have their own problems

  • Don't use them across a high current supply. Across a low current supply they can self-heal too, but across a high current supply they can turn into fireworks...
\$\endgroup\$
  • \$\begingroup\$ self heal sounds new ..... \$\endgroup\$ – WantIt Dec 2 '12 at 11:11
  • 5
    \$\begingroup\$ @vvavepacket Not new at all. It's a well-understood phenomena. See section 2.6 of this Nichicon app note. \$\endgroup\$ – Adam Lawrence Dec 2 '12 at 14:30
  • 1
    \$\begingroup\$ +1, nice detailed answer with long term reference value, thank you. \$\endgroup\$ – Anindo Ghosh Dec 2 '12 at 15:49
  • \$\begingroup\$ Follow-up question: Electrolytics are commonly used in audio paths since, in addition to some other reasons, they come in large enough values that reactance is low enough for the low frequencies which must pass. "Audio-grade" capacitors tend to be electrolytic. Is this why low ESR in caps is crucial in these applications - to avoid significant losses between stages due to leakage? (I ask because it never really clicked for me why those caps were considered such a must-have and are much more expensive.) \$\endgroup\$ – tjbtech May 10 '17 at 16:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.