# Voltage regulators

I would like to ask the function of the capacitors arranged in parallel to the LM7805 voltage regulator and if i can get an explanation of the resistive divider coonected to the LT1528 voltage regulator

• Have you had a chance to read the datasheets yet? – Ignacio Vazquez-Abrams Mar 28 '14 at 14:07
• A capacitor on the output of U1 appears to have two names: C2 and C6. – Kaz Mar 28 '14 at 18:54

The output capacitors are because the 7805 can only supply a limited peak output current (average or transient) whereas a capacitor is only limited by its internal resistance. For instance, an electrolytic capacitor might have an effective series resistance of 1 ohm and when connected across a 5V supply, clearly the peak current it can supply will be about 5A - this is much bigger than what the 7805 can supply BUT remember we are talking transient demands not average demands by the load.

As a footnote, it's always best to put a capacitor like this where it is needed - at the point of load because transients of several amps up and down long copper traces can cause other problems.

The circuit you posted doesn't show a particularly important capacitor this being the input cap to the 7805: -

(source: antihero.org)

Note the 0.33uF cap shown at the input - if you have long feed-wires connecting the power to the 7805, a capacitor helps peak demands taken by the 7805 - this is because feed-wire has inductance and the capacitor prevents instabilities.

The LT1528 is typical of some "adjustable" voltage regulators - the output can be set by a potential divider from output pin to the feedback/sense pin. Normally the LT1528 runs at 3.3V on the output and the sense pin can help the 3.3V be reproduced at some distance from the regulator (say close to a load). If you look at page 1 of the datasheet it shows you various settings for these resistors that allow an output greater than 3.3V.

The divider is for setting the output voltage of that regulator. It internally feeds the SENSE input into its error amplifier circuit and compares that with its internal reference.

Given the divider ratio of $\frac{330+51}{330} = \tilde{} 1.15$, then provided the output voltage is 3V3, it's internal reference has to be around $3.3V/1.15 =\tilde{} 2.8V$ which btw. coincides with the IC's type and it is also the minimum voltage at which the regulator can be set (directly feeding the output to SENSE.)

Anyway, all this can easily be found in a device's datasheet.

Capacitors between a power line and ground are like water towers, they provide the extra energy that a circuit might need for a few micro seconds, so that when the current the circuit uses changes rapidly, the variations are evened out by the capacitors, so that the regulator can just supply a steady current and not worry about constantly adjusting every nanosecond, just like what a water tower does for the pump.

This means that whatever is "filling" the capacitors doesn't see the high frequency current. The shorter length you run high frequency current, the better, because all wires are antennas and we don't want to radiate RF and make an accidental radio transmitter.

The tons of different capacitors are used because capacitors have some built in inductance and resistance that the manufacturers can never fully get rid of, so each capacitor is better at dealing with a different type of current. The big electrolytic deal with the large load spikes that are relatively low frequency, and the little ceramics deal with the small but high frequency oscillations that a microcontroller might make.

The voltage divider is probably because that IC tries to keep its sense pin at a fixed voltage, the feedback voltage. If the voltage divider divides by 5 and the feedback voltage is 1v, the output voltage will be 5v, because the chip will regulate the output such that the feedback pin is at 5v. Different devices have different feedback voltages.

• The bypass caps on a linear regulator serve a third purpose as well. – Ignacio Vazquez-Abrams Mar 28 '14 at 14:24
• And scattering a load of different values in parallel as per the example is usually a sign of laziness / poor design. – John U Mar 28 '14 at 15:00
• @JohnU Can you clarify your remark about "a load" of different values being a sign of laziness? What do you mean by "a load", and what do you see as the correct approach? – Joe Hass Mar 28 '14 at 15:06
• As it was explained to me: A big electrolytic for smoothing - OK. Plus a small tantalum or similar to catch high frequencies/transients - OK. A big pile of assorted caps can actually lead to creating something that oscillates or has very odd characteristics. So, scattering caps around the place without good reason is un-necessary and can create problems. if you have some particular noise/transient that means you need to add some specific smoothing/filtering then fine, do it, but don't just chuck stuff in & hope. – John U Mar 28 '14 at 15:42
• I never did anything that needed loads of caps like that, so I always assumed that people who used parallel capacitors did so because they did all the proper calculations etc. But Google seems to suggest that sometimes people really are just guessing, which is usually a bad idea in anything electronic. – EternityForest Mar 28 '14 at 22:07

The capacitors on a 7805 are only necessary as reservoirs to improve transient response, as others have explained in some detail.

The capacitor (particularly the output capacitor) on an LDO regulator (and most negative regulators) such as the LT1528 is necessary for stability. A minimum of 3.3uF and a maximum ESR of 2 ohms is required at the output (satisfied by most 100uF capacitors, as shown in your schematic). If that capacitor has too much ESR or too low capacitance, the regulator will likely oscillate.

The resistive divider on the LT1528 sets the output voltage, as is spelt out in the datasheet.