Voltage regulator capacitors, what to use?

[ERMEGM]

Hi All,
Happy New Year to you all. I have a question regarding the capacitors on either side of a voltage regulator. I have seen multiple schematics with different values. What are these values based on? How are the values calculated? Just wondering if the values are figured by trial and error AND an oscilloscope or is there a formula one uses to calculate them?



My circuit is for a vehicle application. I’ve seen this schematic for a vehicle and they use 100uF caps on either side. I have another schematic from Ford where they show them using 10 uF caps on each side and another schematic using 0.1 uF caps on either side along with a 10 uF cap on the input side and a 220 uF cap on the output side. Are they just connecting a basic voltage regulator to their intended circuit, using an oscilloscope and adding caps until they get the desired results?

The other thing I’ve seen is a Zener diode connected directly to ground as in the circuit above, but another circuit has a Zener going to a resistor and then to ground. Any reason why? I know you want to protect your circuit using the Zener, but is the resistor really necessary?

I’d like to make sure my circuit is properly protected, but don’t want overkill. What are your thoughts?

Thanks,
Tony

[malc-c]

Hi Tony,

On most of my projects where I've used a 6v ac transformer, my psu consisted of four 1N4001 to rectify the ac, then a 7805 to regulate the dc to 5v, and place a 100uf across the dc in to the regulator, and a 10uf across the dc out. This works fine for me.

I don't know if there are any formula regarding the "correct" values and any further requirements to use Zener diodes to further regulate the supply. Basically the higher value of uf the capacitor are the smoother the dc will be, but there can be a trade off as really large values tend to be physically larger.

[ERMEGM]

Thanks for the reply, but again, how did you come up with that? Was it just something you've seen others do? Did you test it out with an oscilloscope or other equipment? I'd like to know how to make it rather than copying what other people have made. All scenarios are different, so each setup won't be the correct one.

As for the cap sizes, I've read that smaller caps filter out the low frequency noise while larger caps filter out the high frequency noise. How do I see (or hear) this noise? Can it be calculated or does it require machines? Any programs out there that can simulate this?

I'm assuming machines are needed, but is this all done AFTER the circuit is completed? This way, you know what the circuit is generating for noise. It seems logical, but sometimes, there is always a simpler path.

Thanks,
Tony

[Acetronics2]

Hi, Tony

My thoughts ...

if you want to place your supply in a car ( and have fear of what COULD happen ) ... you should use regulators dedicated to automobile use, that are specially designed to whithstand the spikes and other funnies you can find on car 12v lines ...

ST ( between others ) has developped a whole line of such rugged regulators.

So just browse their catalogs, automotive applications sector ... and see what you can get from your dealer. ( Farnell, Mouser, Digikey ...)

or simply READ what manufacturers show into their Datasheets ...

Alain

[malc-c]

Tony,

Mostly just browsing the web and seeing what others used, or generally guess work.. most of the stuff I needed just a simple 5v supply - If I found a problem with ripple I just substituted the capacitors for higher value ones.

[Plcguy]

The filter capacitor cleans up the ripple to give you a flat DC supply. There is an equation for it (ripple and load based), but since the supply is DC, not AC the ripple is minimal. A 780 amp hour battery is an amazing cap, so a 10uf input cap for an auto system is fine.
As Alain alluded though, higher frequency stuff sits on the DC that is VERY bad. Those intermittent wipers keep firing a DC relay and kicking pulses of CEMF on the line. so, add a .001 and a .1 cap in parallel with the 10. Also, I'd put a .1 ohm resister in series with the regulator and hang a 25v zener on the regulator input. The output needs a 1uf to prevent oscillation FB. A typical 3 terminal regulator can handle up to 35vdc, though only for a short time without heat sinks.

[BrianT]

Read the data sheets for the regulator you are using. Most specify a maximum distance between the regulator and the input and output capacitors to ensure no oscillations from the regulator. If in doubt go for two large capacitors (100 uF or more) adjacent to the regulator input and output and make the PCB traces as fat as possible for minimum impedance. Make sure you have a big safety margin on the input capacitor voltage rating. Automotive +12 volts should have a minimum 50 volt rating on the input capacitor, preferably 100V.

HTH
BrianT

[ERMEGM]

Hi guys,

Thanks for the inputs, but the original question is not getting answered. How are you coming up with these values? Other than what other people have used in the past, are the values just trial and error using an oscilloscope and changing out the different values, or is there a formula?

Does this have to be done AFTER I build my circuit, or if I follow this setup, I can use it on all of my different setups? Here's an example from another company for their product:



Can I use this, or is it specific to their circuit? This is what they're using to power their circuit which is an amplifier. Mine is just to supply power to my pics and transistors controlling relays.

Thanks,
Tony

[Plcguy]

C*ΔV = q = I*t, where ΔV is the voltage sag (ripple), C is the capacitance in Farads, q is the charge lost in Coulombs, I is Amps, t is time in seconds.


for a true DC supply t = 0 -> C = 0
for a small I, the ripple is small.
If you have a 3 phase alternator, you may have 180hz freq, but prob no ripple due to the battery acting like a massive capacitor.

generally, this is for ac rectifiers at 60 hz.
the ripple freq is 120hz for 60hz sine wave
Example:
Given full wave rectification, a load current of 3.5A, and a capacitance of 6800µF, calculate the ripple voltage.
From equation (3)
Δv = 3.5A 120Hz x 6800µF = 4.29V