Capacitors serve a few functions in power supplies / circuits.
They smooth out a recitified signal, making it look more like a DC voltage.
They store charge, providing some extra juice when peaks in power usage comes along.
And they are used to filter out ripple in the DC voltage near where the power is applied.
My question is, are there any rules of thumb for how big a capacitor needs to be for these applications?
Also, capacitors used for power applications are usually electrolytics, due to the large values involved. Are there situations in which it's important not to use electrolytics? (I'm only referring to power applications here, nothing that directly relates to the signal path).
They smooth out a recitified signal, making it look more like a DC voltage.
They store charge, providing some extra juice when peaks in power usage comes along.
And they are used to filter out ripple in the DC voltage near where the power is applied.
My question is, are there any rules of thumb for how big a capacitor needs to be for these applications?
Also, capacitors used for power applications are usually electrolytics, due to the large values involved. Are there situations in which it's important not to use electrolytics? (I'm only referring to power applications here, nothing that directly relates to the signal path).
Here is an approximate formula:
It assumes that the capacitor is charging for 10% of the cycle; and discharging for the remaining 90%.
C = [output current x 0.9] / [frequency x ripple voltage].
Ripple voltage should be the peak to peak you desire. This all depends on what you are building... less is better, but only to a point.
Output current should be maximum your circuit will require..
Frequency should 120 Hz for a full wave supply and 60 for a half wave.
Choose your desired ripple... run the numbers... Then double the cap value just for good measure. This will allow for capacitor tolerance and ESR (which will rise with age).
So... less ripple = bigger cap = more money.
Now as the cap get bigger, so does the inrush current, this means bigger diodes (= more money).
There are handy charts for this, but I can't find one to post.
Google "Duncan's Amp Pages" and download PSUD II (FREE!). It is a very handy program that will simulate and plot all this stuff for you. It won't actually solve your circuit, but you bang away with trial & error and get the performance you want.
It assumes that the capacitor is charging for 10% of the cycle; and discharging for the remaining 90%.
C = [output current x 0.9] / [frequency x ripple voltage].
Ripple voltage should be the peak to peak you desire. This all depends on what you are building... less is better, but only to a point.
Output current should be maximum your circuit will require..
Frequency should 120 Hz for a full wave supply and 60 for a half wave.
Choose your desired ripple... run the numbers... Then double the cap value just for good measure. This will allow for capacitor tolerance and ESR (which will rise with age).
So... less ripple = bigger cap = more money.
Now as the cap get bigger, so does the inrush current, this means bigger diodes (= more money).
There are handy charts for this, but I can't find one to post.
Google "Duncan's Amp Pages" and download PSUD II (FREE!). It is a very handy program that will simulate and plot all this stuff for you. It won't actually solve your circuit, but you bang away with trial & error and get the performance you want.
Let me add:
I have already posted a Capacitor Bleeder Resistor calculator spreadsheet. I just added another worksheet with Poobah's formula. Hopefully you guys don't rip it to shreds!
/edit: if you guys have more ideas I'll add them as I please, with credits given, and repost on the original thread where I first posted this.
I have already posted a Capacitor Bleeder Resistor calculator spreadsheet. I just added another worksheet with Poobah's formula. Hopefully you guys don't rip it to shreds!
/edit: if you guys have more ideas I'll add them as I please, with credits given, and repost on the original thread where I first posted this.
Where Not To Use Electrolytics?
Well, they tend to be a little on the noisy side when used as power supply bypass capacitors for IC op amps.
When the op amp data sheet says to put a small value capacitor in very close proximity to the op amp from the positive and negative supply pins back to the input ground, they don't mean to put an electrolytic. They mean to use a very low internal inductance capacitor. Ceramic disc caps and cut layered film caps are the usual methods. Read the data sheets to find out who has the lowest internal inductance, and buy that.
Why? Because the internal inductance determines the cutoff frequency where the capacitor stops acting like a capacitor and starts acting more like an LC resonator. Below that frequency, it is removing high frequency noise from the op amp's voltage supply.
Why? Read the data sheet for the op amp. The supply noise rejection graph will show that the op amp is very good at ignoring low frequency ripple in the supply voltage, but has a hard time ignoring high frequency noise in the supply voltage.
The bypass capacitors are intended to get rid of the high frequency noise, not the low frequency ripple. So using a big electrolytic sort of defeats the purpose.
Well, they tend to be a little on the noisy side when used as power supply bypass capacitors for IC op amps.
When the op amp data sheet says to put a small value capacitor in very close proximity to the op amp from the positive and negative supply pins back to the input ground, they don't mean to put an electrolytic. They mean to use a very low internal inductance capacitor. Ceramic disc caps and cut layered film caps are the usual methods. Read the data sheets to find out who has the lowest internal inductance, and buy that.
Why? Because the internal inductance determines the cutoff frequency where the capacitor stops acting like a capacitor and starts acting more like an LC resonator. Below that frequency, it is removing high frequency noise from the op amp's voltage supply.
Why? Read the data sheet for the op amp. The supply noise rejection graph will show that the op amp is very good at ignoring low frequency ripple in the supply voltage, but has a hard time ignoring high frequency noise in the supply voltage.
The bypass capacitors are intended to get rid of the high frequency noise, not the low frequency ripple. So using a big electrolytic sort of defeats the purpose.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.