[capacitor] Different between value, type, time response

[jack_yls]

Currently, I have some design problem.

Analog design, power supply. 1 of the important component in power supply is capacitor. So, as we all know, capacitors have at least 2 basic functions, store charge and filtering the noise (particular frequency). Capacitors have different capacitance, time response and impedance.

For bypass capacitor, I use “Try and Error” method and basic calculation of impendence (capacitor). I have found some alternative capacitor for replacement (for cost down/ Value Added). However, there are some concepts I’m not sure.

FYI. The graph below is the voltage drop when the circuit ‘ON’. When the circuit ON, the voltage of the VCC circuit drop a little bit; then, capacitor discharge and provide the charge (current) to the circuit and the voltage will become stable after certain time.



The questions here are, if I use bigger value of (capacitance), the voltage drop should be better (not so much voltage drop), as the big value of capacitor can store more charge? How about the time response of different capacitor? Ecap capacitor is response faster than ceramic capacitor?

 

[marko_benigno]

From my understanding, it seems that your using just a simple power supply to pwer up some components. If you're using this just to power up active devices without timing constraint, then the speed of transition is not significant. Using a higher value of capacitor is the best option to minimize the drop. Ecaps have higher capacitance values than ceramics but they have higher parasitics. For the initial stages of power supplies, Ecaps are the ones being used. Also, it would be better to use regulators to minimize the drop during power up. A simple zener diode will help if no active regulator is available.

 

[jack_yls]

Yes, you are right. I just use the capacitor (2 capacitor in parallel) to start a simple motor. However, from the graph(experiment); we will find that when the internal resistance of power supply increasing to certain level, the result will not parallel.

So, my thinking is when the internal resistance is not so high value, the combination of small value of capacitor will give better result, as the t=RC, when C is smaller; the t will smaller. The time response is small.

However, when the internal resistance is increasing to big enough value, the combination of big value of capacitor will give better result, as 1uF|1uF cannot store so much energy.

But, I still lack of this kind of info or technical report to support my thinking. [/img]

 

[Kral]

jack_yls,
Ecaps have a very high volumetric efficiency. That is, they provide a large capacitance-voltage product in a small volume. However, they also suffer from high ESR (equivalent series resistance). To overcome this problem, designers usually use an Ecap or Tantalum capacitor for "bulk" filtering, and sprinkle lower value ceramic capacitors around the printed circuit board. This gives you the advantage of large energy storage in the Ecap, and the ability to deliver short duration current pulses into the load with a very small voltage drop due to the low ESR of the ceramic capacitor.
Regards,
Kral