Power Filter Caps, how do you know how big they should be?

[joecool85]

In my lm3886 amp I have two 10,000uF 35v caps for power filtering, and I know my 14w practice amp has two 2,200uF caps. So how do I know how large the caps should be? I'm working on building a SLA charger that will have a 300ma transformer and charge at 13.5vDC, someone said I don't need power filter caps. If thats the case, would I just go transformer, rectifier, lm317 regulator -> battery?

 

[ljcox]

It depends upon how much ripple you can stand in the rectified output.

For an audio amp, it is necessary to minimise the ripple otherwise you may be able to hear it.

For a less critical application, the caps can be smaller as you can usually withstand more ripple.

The cheap commercial a battery chargers have no filtering. But it is difficult to calculate the charging current. I'm not an expert in battery chargers so I don't know whether you would need to consider the average current or the RMS current.
Quote
"If that's the case, would I just go transformer, rectifier, lm317 regulator - battery?"
Assuming you can resolve the current question I raised above; you may need small, caps one either side of the LM317 for stability. See the data sheet for the LM317.

I suggest you do a search for "charging" in this forum. There should be plenty of hits.

 

[hyedenny]

Hi Mr Cool
The reason it was suggested that you dont need filter caps for a battery charger is that some say that the pulses help the battery plates from becoming sulfated. Not all battery experts agree, but I dont think it can hurt. (Im NOT an expert!)
For a LINEAR power supply (like your charger) up to 10 amps, a very rough rule is to have 5000uF for every 1amp drawn. It can be calculated, and it depends on the load current, the degree of smoothing necessary, and the frequency of the ripple.

 

[joecool85]

Ok. Well, I will plan on using the 2200uF cap I have laying around then. Sounds like it should work out fine. The datasheet shows a 0.1uF cap before the regulator, but only if the lm317 is more than 6 inches away from the main power supply cap, so I won't need that. It also shows a 1uF cap on the other side of it, which I will use.

 

[hyedenny]

HI again.
The cap BEFORE the regulator is to stabilize the regulator... Dont leave that one out. Make sure the voltage ratings on the caps are plenty big (double the voltage youre making) too!

 

[joecool85]

It has a note in the datasheet that says "Needed if device is more than 6 inches from filter capacitors." I don't see why they would say that if it was needed.

 

[RadioRon]

To calculate an approximate value for the filter capacitor, it helps to make some assumptions first. Lets assume that your power supply rectifier is a full wave type. This means that your ripple is dominated by the 120 Hz component. Let's also assume that your load current is constant and not jumping up and down rapidly. This assumption may not be spot on for an audio amp, but if you start with a load current that represents the maximum output power of your amp, at lower output powers your ripple will be better since the current draw will be lower.

Now, to estimate your capacitor value, use the equation
CxE=Ixt
where C is the capacitor in microfarads (not farads)
E is the voltage droop from the peak of one rectifier one cycle to the next. We approximate by calling this the peak to peak ripple voltage.
I is the load current in amps
t is the time between the peaks coming out of the rectifier, in milliseconds.

For example, let's say you have an existing capacitor you want to use and you want to know what ripple to expect. Let's say your output voltage is 13.8VDC and your current draw is 2 amps and you have a filter cap of 5000 uF. Use the formula to find that E, the ripple, will be 3Vpp (since t is 7.5 mSec in this case). This is a lot of ripple, so a larger cap is needed. However, if you included a linear voltage regulator after the cap that had an output of 10 volts, then you would get a very smooth 10V.

This is just a way of estimating where to start. After this, some experimentation is the easiest way to fine tune things.

Of course, in order to estimate your capacitor, you have to guess at how much voltage ripple is acceptable, and this will depend on the "power supply rejection ratio " of your audio amp.

By the way, this theory and example are taken from the ARRL Handbook 2001 edition, page 11.11.