capacitance related to amperage...

[whiz115]

can someone explain to me the relationship between those two?

let's say i want to make a power supply how i decide the capacitor i'm
going to use for smoothing? someone told me that 1uF equals to 1mA
so if i want to smooth 1A power supply then i need 1000uF capacitor.

is it possible to draw 500mA from 1000uF capacitor without get the capacitor empty?

and a second question...
if we put a very big capacitor can we totally extinguish ripple noise?
if not why?

 

[bananasiong]

[latex]C = \frac{I{LOAD}}{2FV_{RIPPLE}}[/latex]
Where:
Vripple = maximum allowable ripple, the peak transformer output voltage (1.4 times RMS) minus the diode drops and regulator droput.

F = Line frequency, 50/60Hz.

You might use this method to find.

 

[Hero999]

That formula does tend to oversize the capacitors especially when a larger ripple voltages are acceptable. You could try the formula in the catalogue posted by cadstarsucks (see this thread) but it tends to undersize capacitors slightly when givin a constant current load like an LM317.

I still think my formula is better because it ears on the side of caution, there again buying larger capacitors than you really need in't much fun so I'm looking for a better formula. I'm looking to design a spreadsheet to calculate the minimum capacitor value, transformer rating and resistors values for an LM317/217 when given a voltage and current. If anyone here knows of a more accurate formula then I'd be glad to hear about it.

 

[whiz115]

unfortunately as a newbie...the above answers don't helping me much...

 

[Sceadwian]

This is what simulators are good for, you feed in your paramaters and load conditions and it shows you exactly what kind of ripple you're going to get. Try LTSpice. Especially if you're new to electronics it takes a while to get used to but it's very useful.

 

[cadstarsucks]

That formula does tend to oversize the capacitors especially when a larger ripple voltages are acceptable. You could try the formula in the catalogue posted by cadstarsucks (see this thread) but it tends to undersize capacitors slightly when givin a constant current load like an LM317.

It does quite well but it does state that it is shooting for minimums. One critical detail that it goes into that most hobbyists are not aware of is the ripple current ratings... many green engineers die young designing power supplies that fail in short order.

D.

 

[bananasiong]

unfortunately as a newbie...the above answers don't helping me much...

?? Do you mean your questions are not answered?

if we put a very big capacitor can we totally extinguish ripple noise?
if not why?

Depends on your load current. I've tried simulating, when the load current is small, say 10 mA, with a 2200 uF filter cap, i don't see any ripple from the simulator.

 

[kchriste]

and a second question...
if we put a very big capacitor can we totally extinguish ripple noise?
if not why?

No. If you draw any current off the supply there will always be some ripple because of two things:
The capacitor will always discharge a little between pulses from the rectifier. There is no such thing as an infinite capacitor so there'll always be ripple.
All real world capacitors have some resistive/inductive element to them made up of the leads, plates, etc. Think of it as a small resistor and inductor in series with the capacitor. This resistor will always have some ripple across it as the capacitor charges via the rectifier and discharges into the load between pulses from the rectifier.

is it possible to draw 500mA from 1000uF capacitor without get the capacitor empty?

A 1000uF cap will be fine for most 500ma supplys but there will be some ripple still.

 

[Sceadwian]

If you're not overly concerned about exact voltage regulation but just want to get rid of the ripple you can use a transistor in a voltage follower configuration with a zener refrence. Or an LDO regulator.

 

[cadstarsucks]

No. If you draw any current off the supply there will always be some ripple because of two things:
The capacitor will always discharge a little between pulses from the rectifier. There is no such thing as an infinite capacitor so there'll always be ripple.
All real world capacitors have some resistive/inductive element to them made up of the leads, plates, etc. Think of it as a small resistor and inductor in series with the capacitor. This resistor will always have some ripple across it as the capacitor charges via the rectifier and discharges into the load between pulses from the rectifier.

A 1000uF cap will be fine for most 500ma supplys but there will be some ripple still.

Don't forget the leakage!

 

[stevez]

I may be missing something but the formula for calculating or estimating ripple voltage that many people use just doesn't seem right (Vr=I/FC, or something close to that). With that formula one can calculate a ripple voltage that can't possibly happen - I am thinking in terms of a ripple voltage that exceeds the applied voltage.

It may be that the formula is useful within limits - just don't know what the limits are. I ran into this when I did the math for a 5 volt supply, large current and small capacitor - then stepped back and looked at the formula.

 

[Russlk]

If you use the equation correctly: C = I/Vr*F , you won't have a problem.

 

[eng1]

I ran into this when I did the math for a 5 volt supply, large current and small capacitor - then stepped back and looked at the formula.

Did you use a voltage regulator? The filter capacitor is not supposed to cancel the ripple, if it's followed by a 7805 for example and the input voltage is high enough.

let's say i want to make a power supply how i decide the capacitor i'm going to use for smoothing? someone told me that 1uF equals to 1mA so if i want to smooth 1A power supply then i need 1000uF capacitor.

You have to say what the input and output voltage are. Do you need a regulated outout? are you going to use an IC regulator?

 

[whiz115]

?? Do you mean your questions are not answered?

it's ok now...

 

[whiz115]

You have to say what the input and output voltage are. Do you need a regulated outout? are you going to use an IC regulator?

why? i think it's irrelevant if i use regulator or what is the input voltage,
i think a capacitor can charge to any voltage within it's limits,
as i said before, someone told me that 1uF=1mA...if we have 1uF capacitor
it can keep charge of 1mA and i just want to ask if this is the actual ratio...

if you need to know voltage...ok let's say 12V but i don't understand what's
the point...

 

[eng1]

If you're going to make a power supply, you may want to evaluate the ripple at the output? IC regulators have a figure that is called Ripple Rejection and the ripple at their input (where the capacitor is connected) is going to be attenuated.

as i said before, someone told me that 1uF=1mA...if we have 1uF capacitor it can keep charge of 1mA and i just want to ask if this is the actual ratio...

This is a general 'rule of thumb' for the choice of the filter capacitor. You can use more precise formulas, as suggested.

 

[Roff]

If you're going to make a power supply, you may want to evaluate the ripple at the output? IC regulators have a figure that is called Ripple Rejection and the ripple at their input (where the capacitor is connected) is going to be attenuated.


This is a general 'rule of thumb' for the choice of the filter capacitor. You can use more precise formulas, as suggested.

V=I*T/C
Using 1uF/mA at 100Hz (fullwave rectified line frequency)
Vripple=1e-3*1e-2/1e-6
Vripple=10V
The actual ripple will be less than this, due to the fact that the discharging capacitor starts recharging when the rising sine wave of the next half-cycle forward-biases the rectifier, causing the cap to start recharging. The actual ripple will be less, with the actual amount depending on the input voltage.
If the input voltage were a square wave at 50Hz, the ripple would actually be 10V.
I think a better rule of thumb might be 3 or 4uF per mA. It depends on how much ripple you can tolerate. If you want to run with very low headroom on your regulator, 10uF/ma might be more appropriate. This will always yield less than a volt p-p ripple.

 

[eng1]

Ron, I agree and I was not supporting that formula. It is commonly used though and I've seen it on documents or books.

Do you usually consider the high tolerance of electrolytic caps?

 

[eng1]

I ran a simulation with transfomer (0-12V) + bridge rectifier + 1000 uF cap + .5A constant load and this is the ripple

 

[Roff]

Ron, I agree and I was not supporting that formula. It is commonly used though and I've seen it on documents or books.

Do you usually consider the high tolerance of electrolytic caps?

I think you should calculate the minimum value of the cap and use that value, e.g, if you find you need 4000uF, and you are going to buy a ±20% cap, buy 5000uF.