Hi. Wonder if you guys can help me out. Let's say I have 12v 1A coming off a rectifier. This makes the light (load) pulsate, which tells me to add a capacitor for smoothing. I've read that you want a voltage rating on said capacitor to be higher than whats needed, as much as 3 times higher? What voltage and uf rating do I need for this capacitor? Please provide the answer and help me understand the formula used to get it. Thx.

 

Definitely not 3X. It depends a bit on the use of the capacitor and is always advisable to have a bit of "headroom" but not necessarily 3X.
Here is a good basic article on cap sizing for a smoothing service. https://electronicbase.net/smoothing-capacitor-calculator/

 

Hello there

What voltage and uf rating do I need for this capacitor? Please provide the answer and help me understand the formula used to get it. Thx

If you could post the rectified DC circuit. I can walk you through circuit analysis providing the answers you requested.

 

I've read that you want a voltage rating on said capacitor to be higher than whats needed, as much as 3 times higher?

No reputable source would over design that much. Electrolytic capacitors will withstand their rated voltage for the specified lifetime as long as maximum temperature is within specifications. Ripple current could also matter. Most will derate 10-20%.

What voltage and uf rating do I need for this capacitor?

What voltage are you rectifying? What type of rectifier are you using? What voltage and current does the load require?

 

You can try 1000μF 16V for starters.
In order to apply a formula you need to know the load (current) and how much ripple you can tolerate.

 

MINIMUM Voltage rating calculated as follows:
Multiply the transformer secondary voltage by SQR(2). That gives the peak output voltage.
Then, if your transfomer is rated at 230V AC, multiply by 253/230, because 253 is the maximum allowable mains voltage.
(This applies to Britain and EU countries)
Then add the transformer "regulation" which is quoted as a percentage.

A rule of thumb for the capacitance is C=It/V where I is the load current, t is the time between capacitor recharges (10ms for full-wave rectifier on a 50Hz supply), and V is your chosen maximum ripple voltage. This formula will slightly overestimate (which is a good thing) as it assumes that the capacitor charges instantaneously at the peak of the cycle whereas it actually takes longer, say 1-2ms, so the time between recharges is shorter. As there tolerance on capacitor values is generally +/- 20% it's accurate enough.

How to choose the ripple voltage.
Work out the peak voltage at low line.
So that's secondary voltage x SQR(2) x 216/230. Then subtract your chosen ripple voltage. If you power supply supplies a linear regulator (7805 etc), make sure that this voltage is higher than the minimum input voltage for your linear regulator.

 

What you will find is that there are common voltage ratings for capacitors, and they are not always available in exactly the value calculated. And for electrolytic filter-type capacitors the steps get larger as the voltage goes up. So normally you might see a 6.3 volt rating for one size cap made by Rubycon. while another will offer 10 volts and then 25 volts and 35 volts, but then 50 volts. And different manufacturers may offer different ranges. For rectified DC from a 12 volt transformer 16 volts would be an absolute minimum, I would choose 25 volts unless some in between value happened to be available. So you see that a big part of the choice relates to what is available.

 

Keep in mind, unless that 12 volts is already a peak, the voltage is likely to rise after adding the filter cap.

 

I forgot to subtract the diode voltage drop..
You’d have to be using a transformer specified at high-line with better than 3% regulation to get away with 16V caps.
12V x sqr(2) x 1.03 = 17.47V less two diode volt drops= 16.07V
if it is specified at 230V then it will be 17.8V at high-line (253V)

 

Keep in mind, unless that 12 volts is already a peak, the voltage is likely to rise after adding the filter cap.

I’ve never seen a transformer specified as the peak or peak-to-peak voltage. It‘s always the rms voltage.

 

In addition to the transformer rating being RMS or close to that, the TS is probably using a multimeter that gives an average value close to RMS. So really, if the load is not going to be permanently connected, a capacitor voltage rating of 20 volts will make sense. That will leave a bit of head room and yet be reasonably close.

 

Also there is possibility of the va rating of the transformer being compromised, where the present current draw approches the present Va rating, any significant increase in smoothing capacitance, increases the required transformer Va rating. requirement.
Max.

 

tons of great information guys. thnks. ill provide a few more details, and this is heading me in a good direction for testing.

 

ill provide a few more details, and this is heading me in a good direction for testing.

That means all you brainiacs scared him away, almost as if you were taking turns doing something explicit

 

Here http://www.hammondmfg.com/pdf/5c007.pdf is a really good place to start.

For simple rectification, you can use 1000 uF per Amp as a guesstimate. The voltage rating of a capacitor of 2x the max DC is usually pretty good to start.

Unfortunately, those rules can always be broken. Capacitors in the safety portion of a telephone have to be rated to like 1500 V, not 50 and not 100.

Capacitors from line to N and N to ground also have to be special.

Ripple will kill a capacitor. Capacitors used in switching power supplies are a little different.

The key point with that article is that the AC amp rating of a transformer is lower than the rectified DC current. It makes sense

Capacitors types vary. There are capacitors specifically for AC applications like motor start or motor run.

Temperature ratings play an important part.

I built an amplifier and I didn't understand sizing at the time. I doubled what was suggested. That wasn't enough.
I built the amp with 4x35VX3A windings, but I also built the amp using an 70 V CT 18A constant voltage transformer. Bass was incredible. Noise made by the transformer was horrible.

Capacitance was 9600 uF on each supply rail, 4 of them, which creates power-up problems.

So, I compromised. I have a 500 W AC voltage regulator on the power to the system.

 

That means all you brainiacs scared him away, almost as if you were taking turns doing something explicit

D. P. has it all wrong! The only purpose for asking questions is to be able to provide useful information instead of random guesses. Some of the responders are able to provide accurate explanations instead of just random comments that may not even be applicable or correct.

 

Here http://www.hammondmfg.com/pdf/5c007.pdf is a really good place to start.

For simple rectification, you can use 1000 uF per Amp as a guesstimate. The voltage rating of a capacitor of 2x the max DC is usually pretty good to start.

1000uF per amp is a bit mean on a 12V supply. That would give almost 10V ripple, but on a 100V supply it would be just fine.

 

1000uF per amp is a bit mean on a 12V supply. That would give almost 10V ripple, but on a 100V supply it would be just fine.

Need to explain that one!
Max.

 

Using an F-25X 12V CT secondary XFMR capable of 1.5A with a rectifier chip and 2200uF 50V smoothing capacitor I have near 0 detectable ripple. Especially after running it through an LM317 regulator. Cap size is overkill but I had one handy.

 

OF COURSE! As long as there is an adequate source voltage provided a linear regulator will remove all ripple at frequencies below it's frequency response limit. So that is an additional option available. BUT it does require a large enough capacitor to keep the voltage at least 3 volts above the regulator output setting.