IIRC there was a discussion about that on the forum a while back, the answer was that ceramic should be OK, and they use electrolytic because they're cheaper.
what do you think?
use an electrolytic?
or maybe polyester?
or multilayer-ceramic?
Depends on the function of the capacitor and where its located. Quite often, ceramic or metallised film would do fine. But if can get electrolytic of equivalent capacitance then use that...
IIRC there was a discussion about that on the forum a while back, the answer was that ceramic should be OK, and they use electrolytic because they're cheaper.
Had replaced electrolytics with ceramics and film capacitors in many power supplies before (for example, that Creative GigaWorks S750 where I've used X8R ceramic because of the proximity to the heatsink), and so far there had been no complaints nor returns. Of course, touch wood though...
You have to be careful when choosing to use ceramic capacitors, because ceramic capacitors have very low ESR and because the capacitance of the ceramic capacitors varies with the voltage applied on them.
Some circuits rely on the output capacitor's ESR to make them stable, and too low ESR can make them unstable at some current loads. May work just fine when you start the device but at certain point when the current reaches a threshold, you may find the voltage starting to oscillate which could trigger a reset or protection inside the device and you don't know why that happens.
Some LDO regulators (linear regulators with low voltage drop, mostly the ones using a pnp transistor inside instead of npn or darlingtons) and a lot of switching regulators have such requirements.
For example see the Application information section for LM1117, a common regulator used everywhere (may not be LM, it's made by lots of companies) : (see page 9, esr recommended to be between 0.3-22 ohm)
As for the DC-bias issue, basically the more the voltage gets close to the maximum voltage rating of the capacitor, the capacitance will drop - this varies with the ceramic capacitor type... C0G/NPO has no problem with this, X(5-8)R has relatively small problems, y5v is ridiculous.
A minor problem with some ceramics is also microphonics / piezoelectric effect - some ceramic acts as small microphones and with strong vibrations or loud noise, a bit of voltage can be induced in the ceramic capacitor and naturally change the circuit behavior - for example you may not want to use a ceramic in a RC oscillator in a subwoofer (unless you pick a ceramic that's not susceptible to vibrations)
See page 4 from the AVX pdf above which mentions it.
You have to be careful when choosing to use ceramic capacitors, because ceramic capacitors have very low ESR and because the capacitance of the ceramic capacitors varies with the voltage applied on them.
Some circuits rely on the output capacitor's ESR to make them stable, and too low ESR can make them unstable at some current loads. May work just fine when you start the device but at certain point when the current reaches a threshold, you may find the voltage starting to oscillate which could trigger a reset or protection inside the device and you don't know why that happens.
Some LDO regulators (linear regulators with low voltage drop, mostly the ones using a pnp transistor inside instead of npn or darlingtons) and a lot of switching regulators have such requirements.
For example see the Application information section for LM1117, a common regulator used everywhere (may not be LM, it's made by lots of companies) : (see page 9, esr recommended to be between 0.3-22 ohm)
As for the DC-bias issue, basically the more the voltage gets close to the maximum voltage rating of the capacitor, the capacitance will drop - this varies with the ceramic capacitor type... C0G/NPO has no problem with this, X(5-8)R has relatively small problems, y5v is ridiculous.
For your information, I have already checked the datasheets and application notes on those PWM controllers (such as that TOP2xx series from Power Integrations, used in that Creative GigaWorks subwoofer). And even their sample schematics shows usage of non-electrolytic capacitor especially for capacitance around 0.1uF. Thus should be no problem at all. Furthermore, I usually prefer ceramics like C0G/NP0, U2J, X8R and X7R as they are more stable than cheaper (higher capacity) ceramics like Y5V, Z5U, etc.
A minor problem with some ceramics is also microphonics / piezoelectric effect - some ceramic acts as small microphones and with strong vibrations or loud noise, a bit of voltage can be induced in the ceramic capacitor and naturally change the circuit behavior - for example you may not want to use a ceramic in a RC oscillator in a subwoofer (unless you pick a ceramic that's not susceptible to vibrations)
See page 4 from the AVX pdf above which mentions it.
Yes, I'am aware of Class 2 ceramic capacitor microphonics, however there is very little or no effect at all on subwoofers. In fact, there are lots of ceramic caps on the subwoofer itself, at the power supply board in the switching power section and especially those SMD ones used on the controller board as well as some parts of the power amplifier boards. The only main problem with microphonics is self-failure due to cracking from excessive mechanical stress (piezo-electric effect) by the ceramic material itself. Can look around for information on self-failure modes of ceramic capacitors, which happens mostly for SMD type (even the way the SMD capacitor is soldered on to the board has some effect). Those leaded type are much less susceptible...
The 0.1uF and 0.01uF are decoupling capacitors, which are supposed to be ceramic because they're meant to filter high frequencies and react faster than electrolytic capacitors. You're not supposed to use electrolytic capacitors in the first place for such purpose.
I'm talking in general about cases where you may be tempted to replace an existing electrolytic (let's say a 2.2uF 50v rated electrolytic) with a ceramic capacitor - it may be electrolytic for a good reason.
0.47uF is a bit too big to be used for decoupling, so it has another purpose in that circuit. Figure out what that's used for before replacing with a ceramic. And pay attention to the voltage that's on the capacitor while the circuit works.
The 0.1uF and 0.01uF are decoupling capacitors, which are supposed to be ceramic because they're meant to filter high frequencies and react faster than electrolytic capacitors. You're not supposed to use electrolytic capacitors in the first place for such purpose.
I'm talking in general about cases where you may be tempted to replace an existing electrolytic (let's say a 2.2uF 50v rated electrolytic) with a ceramic capacitor - it may be electrolytic for a good reason.
0.47uF is a bit too big to be used for decoupling, so it has another purpose in that circuit. Figure out what that's used for before replacing with a ceramic. And pay attention to the voltage that's on the capacitor while the circuit works.
Not talking about the decoupling capacitors (which can be either film, ceramic or electrolytics depending on the circuit configuration), but rather areas where non-electrolytics can be applied. About those larger capacitance you've mentioned, another place where I've replaced electrolytics with non-electrolytics is the speaker crossover system. Quite often I found those (bipolar) electrolytic capacitors to be way off tolerance, making the speakers "un-balanced" on either sides. Rather than ceramics, for replacements I've used film type (either polyester or polypropylene) which often were much larger than the original electrolytic capacitor...
Comment