I am making a circuit for a car and a regulator is used to get from 12V to 5V. This regulator is surrounded by two 10 uF electrolytic capacitors rated at at 25V. These capacitors have a 'lifetime' of 2000h. Does this mean that there is a 50% probability that the capacitors will fail after 2000h of usage?

Since the circuit is for a car the longevity and integrity of the circuit is important.

This is the datasheet.

  • 9
    \$\begingroup\$ If you know the life time, you must have a datasheet. If you have a datasheet, share it so we don't have to guess. \$\endgroup\$ – pipe Sep 14 '16 at 10:13
  • \$\begingroup\$ I think that this is an info you can find in the datasheet. By the way, I would say that 'lifetime' is the time span where performance is guaranteed if specified maximums and minimums are respected. So I would not think they will fail, only start to be less performant, possibly failing. \$\endgroup\$ – Vladimir Cravero Sep 14 '16 at 10:15
  • 1
    \$\begingroup\$ I'm voting to close this question as off-topic because of a failure to read the documentation. \$\endgroup\$ – Leon Heller Sep 14 '16 at 10:31
  • \$\begingroup\$ It sounds to me like you're interpreting capacitor "lifetime" as MTBF perhaps? If the question is "What does the term 'lifetime' mean for capacitors", then maybe that's clearer? \$\endgroup\$ – Roger Rowland Sep 14 '16 at 10:33
  • 2
    \$\begingroup\$ For serious engineering refer to the IEC standard! For a hobby project, roughly the life doubles for every 10°C below the rated temperature (including self-heating). So get a high temperature rated cap with long life and run it much cooler than rated. The mode is typically high-ESR failure due to electrolyte drying out. \$\endgroup\$ – Spehro Pefhany Sep 14 '16 at 16:14

The "lifetime" is the total time the capacitor can be in service while subjected to the maximum stresses specified in the datasheet, after which the manufacturer no longer promises anything about the cap. In some cases, the manufacturer even tells you which parameters it no longer guarantees to what level. Generally the ESR (equivalent series resistance) goes up.

Electrolytic caps wear out rather quickly at maximum voltage and temperature, but the higher the allowed voltage and temperature, the more possible uses, and therefore sales, there are for a capacitor. Manufacturers are therefore stuck with a tradeoff of specifying low voltage and temperature or low lifetime. Most of the time, they push the voltage and temperature and quote very low lifetimes, like 2000 hours (not even 3 months).

Notice that caps specifically sold as having a long lifetime have lower voltage and temperature specs for the same size case.

Lifetime goes up quickly as the voltage and temperature are lowered, although most datasheets unfortunately don't give you a lot of guidance on this. A good rule of thumb is to keep the steady voltage on a electrolytic cap at about 2/3 of the maximum. It's OK to go near the maximum occasionally, but don't ever exceed it. Similarly, you want the steady temperature to be 20 to 50 °C below the maximum.

The only way to get realistic lifetime values is to ask the manufacturer or find the rare cap where useful data on this is included in the datasheet.

  • \$\begingroup\$ By keeping the steady state voltage at roughly 2/3 of the maximum and the steady state temperature at 20-50 °C below the maximum, how will the lifetime scale exactly? Am i just guaranteed the X hours specified in the datasheet or can i expect to see orders of magnitude difference in lifetime? \$\endgroup\$ – Kafros Sep 14 '16 at 11:04
  • 5
    \$\begingroup\$ @Kaf: That's the problem I was expressing. Lifetime goes up rapidly (non-linearly) with reduced voltage and temperature, but datasheets often give you little guidance. \$\endgroup\$ – Olin Lathrop Sep 14 '16 at 11:12
  • \$\begingroup\$ Is there any reason other than space on the datasheet why they can't specify a lifetime under multiple conditions instead of just in a torture test? \$\endgroup\$ – Dan is Fiddling by Firelight Sep 14 '16 at 15:39
  • 1
    \$\begingroup\$ I vaguely remember that lifetime doubles for every 10°C lowered. \$\endgroup\$ – Agent_L Sep 14 '16 at 16:46

The datasheet guarantees that capacitors will keep their characteristics within limits after spending 2000h in quite extreme conditions (maximum rated voltage and 105°C). This says nothing about MTBF or expected lifetime, that information is simply missing in the datasheet. Caps lifetime is often limited by chemical electrolyte degradation, which is simply not going to manifest itself in 2000 hours. You'll have to ask the manufacturer to provide such information if you have MTBF requirements on your product.

Also note that the datasheet says:

This product qualify for AEC-Q200, but it has some deviations

So if you plan to use it in automotive applications, you'll have to contact the manufacturer anyway to know what those deviations are.


Datasheet states:

Endurance: 2000 h to 5000 h at 105 °C

That means that parameters will be within specification defined in datasheet: enter image description here

  • \$\begingroup\$ 30% Capacitance change seems like a pretty big change for a capacitor that is still "Good" \$\endgroup\$ – BenPen Sep 14 '16 at 16:20
  • 1
    \$\begingroup\$ I don't judge. That's just what's in datasheet. \$\endgroup\$ – Chupacabras Sep 14 '16 at 17:30
  • 1
    \$\begingroup\$ Yes, and I just make the implication that you really need to watch your data sheets. :) \$\endgroup\$ – BenPen Sep 14 '16 at 17:33
  • \$\begingroup\$ If the cap is specified as +80%-20%, would the capacitance change spec indicate that after 2000 hours the cap could be 50% below the nominal value, or would a cap only be allowed to drop 35% if it had initially been higher than the nominal value and was thus at least 80% of nominal even after the change? \$\endgroup\$ – supercat Sep 15 '16 at 2:46

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.