# What if a capacitor exposed to slightly higher voltage than rated?

I'm building my 5 V circuitry for my bike's dynamo rated 3 W 6 V.

Today I went for testing peak voltages without load and capacitors, just with diode bridge 4 x 1N5819.

Unfortunately my multimeter doesn't have peak function, so I made peak detector from LM324N:

Capacitor used: 100 nF ceramic. Meter leads were connected to Vout and GND pin of LM324N. Maximum readings were 24 V at maximum speed I could do.

I tested this peak detector on my ATX PSU (bridge rectifier omitted), I got:

real 5V -> Vout = 3.9 + 0.1V (Vdrop of 1N5819) = 4V, delta = 1V
real 11.1V -> Vout = 9.8 + 0.1V = 9.9V, delta = 1.2.


Considering that real peaks are higher then 24V, real peak I'd say +2V for rough estimation = 26V

Can those peaks be damaging for 25V electrolytic capacitor with no load? And with load? I personally think if there is a load, the cap cannot be fully flooded (charged), hence cannot be damaged that way. But I'm not sure.

UPDATE: Just in case I'll use 24V transil

• Check the datasheet, chances are that a 25V cap has an "absolute maximum rating" of quite a bit more. – PlasmaHH Jun 26 '17 at 12:19
• I think ceramic caps are pretty good in terms of over voltage. (Tantalum not so much... sometimes not even good to rated voltage.) – George Herold Jun 26 '17 at 12:21
• @GeorgeHerold, I use only electrolytic caps for 5v supply, ceramic cap I used only in peak detector – Qeeet Jun 26 '17 at 12:23
• @Qeeet, Aluminum Electrolytic? I'm not sure about those, I tend to use Tant's at 1/3 their rated voltage and Al at 1/2. – George Herold Jun 26 '17 at 12:30

Your open circuit dynamo - actually it's an alternator since the output is AC - may be irrelevant to the real-world use. The unit will contain internal series resistance and inductance.

As desribed in my answer to Non-led simple bicycle dynamo light system, the impedance of an inductor is given by $Z=2ωL=2πfL$. This shows that the impedance is proportional to the frequency which, of course, is directly related to the speed of the bike. If designed correctly the lamps will turn on to a reasonable brightness at quite low speed and will be noticeably brighter at high speed but without blowing the lamps - the reason being that the inductors and lamps form an L-R voltage divider. The inductance helps keep the voltage more constant over a wide range of speeds than if it was minimised.

Figure 1. Bicycle "dynamo" equivalent circuit.

If you place your regulator circuit after the switch shown in Figure 1 it will never run with a no-load situation and the alternator output voltage should be pulled down to a safe value for your capacitor.

Can those peaks be damaging for 25 V capacitor with no load?

You are right to be concerned but given the short duration of exposure to the higher voltages it's unlikely to fail. If going into production where warranty costs could become an issue you might take the cautious approach.

And with load? I personally think if there is a load, the cap cannot be fully flooded (charged), hence cannot be damaged that way.

Agreed.

Try your speed tests again with various loads connected to the output and I suspect the voltages obtained will be much less.

• For critical (product design) applications: check the datasheet and ensure in your design that the voltage across the capacitor doesn't ever exceed the absolute maximum rating.
• For non-critical applications (e.g. prototypes, testing): try it and see what happens. A bit overvoltage is usually going to be OK, since there will be a certain safety margin built into the design of any capacitor. When overvolting capacitors from sources that can potentially source a lot of current, be sure protect yourself and your environment against shrapnel from exploding caps.
• for example Rubicon 25V 3300uF link. Cannot see absolute maximum voltage rating, only see voltage rating. Rubicon is well known caps manufacturer – Qeeet Jun 26 '17 at 12:30
• In such a case, you should consider the rated voltage to be the absolute maximum. Above that, all bets are off. What goes perfectly fine on one capacitor may go ballistic on another one even from the exact same batch. – Mels Jun 26 '17 at 12:32
• That's also why I made the distinction in my answer between product design and prototyping: if you're whipping up a quick prototype, all you've got is a 25V cap where you expect it to see 26V, by all means give it a go - chances are it'll work. If you're going to make multiples or exposing the circuit to others who don't know about the blast/fire hazard, you may not want to take the chance. – Mels Jun 26 '17 at 12:40
• I'm that kind of person who prefers to spend some time in constructing something rigid, reliable and then have no any problems with it – Qeeet Jun 26 '17 at 12:47