SMPS Tutorial (3): Charge Pumps, Buck Converters, Switched Mode Power Supplies

hello this is part three of my video tutorial series about switch mode power supplies and power conversion techniques in general in part one and two I dealt with AC to AC AC to DC and non switching DC to DC topologies and if you haven't watched those two videos yet I recommend you to do so this video as well as the following videos deals with the idea behind the actual topologies of switching DC to DC as well as DC to AC converters if you came here to receive knowledge about mains powered AC to DC switched mode power supplies like the PSU of your computer you are still at the right place why is that well because what you might know is AC to DC switched mode power supplies are actually not switching converters for AC to DC but they are usually systems comprised of a non switching AC to DC topology plus a switching DC to DC converter that is true for ATX power supplies charging adapters for mobile devices the power supplies inside printers scanners DVD and blu-ray players Welding inverters and the list goes on I will deal with these real-world applications once I explained all the basics to really understand the operation of these complicated circuits you first have to know the topologies which I explained in parts 1 2 3 & 4 DC to DC switching converters at the end of part 2 I mentioned that the two main disadvantages of linear regulators are first the bad efficiency which increases with a voltage difference between input and output voltage second the fact that V out is always smaller than V in as a third disadvantage it could be added that V out can never have the reverse polarity of VN all switching converters on the other hand overcome at least one in many cases two or even all three of these limitations of the linear regulators actually the first point efficiency might be the biggest advantage which actually all switching converters are aiming at in fact you could say that the switching converter with ETA equals 1 or 100% efficiency is the holy grail and the ultimate goal of power supply design even though it will never be possible to reach that final goal modern switching converters usually operate at efficiencies over 90% the reason for the bad efficiency of the linear regulator lies in its very nature the fact that the series transistor is acting as a variable resistor over all resistors however a certain amount of thermal energy is dissipated once current passes through them that amount of energy is then lost from our system therefore you can say in general that an ideal converter with ETA equals 1 cannot employ any resistive elements at all therefore basically only three of the basic electronic components could be used for that kind of circuit first capacitors second inductors and third switches keep in mind that in all modern real world applications the switches are of course not real switches but diodes and transistors but for better understanding of the principle ideas behind the switching topologies I will use switches in my drawing for now every switching converter employs at least one switch and at least one type of component that acts as an energy storage the one converted type solar relying on capacitors as storage elements is the charge pump the charge pump sometimes called switch capacitor voltage converter or similar is based on a very simple idea let's say you have a real voltage source with the supply voltage vs and the internal resistance RI you now want to obtain an output voltage vo which is bigger than VI and which could theoretically be applied to a load RL now how could you do that with just capacitors and switches well imagine you just connect two capacitors in parallel to the real voltage source and you wait until the two capacitors are charged to a certain voltage VC 1 and VC 2 even though the capacitors can never reach the actual voltage of the source they can easily be charged to 99.9% of the source voltage which we will call vs for the sake of easier understanding once the capacitors have reached that voltage you disconnect c1 and c2 from the source in this way then you quickly connect them in series now the two voltages VC 1 and VC 2 add up the two capacitors can now be discharged in a third capacitor acting as an output capacitor this capacitor co can then act as a voltage source that has a source voltage of up to two times V s now that the principle idea should be clear you can talk about how real charge pumps work in any real charge pump the capacitors are most likely electrolytic capacitors and the connecting and disconnecting is done by switching elements of some kind the first capacitor is called CI because it is constantly connected to the input voltage the second cap is called CP for pump capacitor while the cap at the output is called Co the four switches as one two three and four are activated in the following sequence before any switching begins the input capacitor CI is already charged to round the value of V s now phase one begins the switches s2 and s3 are closed connecting CI CPE and the source in parallel while phase one the pump capacitor is then charged to a certain value usually very close to the supply voltage of the source once that value has been reached phase two starts s2 and s3 open directly followed by s 1 and s 4 closing effectively connecting CI and CP in series as well as connecting the output capacitor to those two caps the output capacitor is then charged to a value of up to two times vs the output capacitor can now act as a voltage source for the load once the capacitor is discharged to a certain point the whole process starts anew this circuit can also be used to obtain output voltages lower than two times the source voltage that is done by going into phase two before the pump capacitor is fully charged when this circuit is used to double the supply voltage its efficiency is however at the maximum value in theory it could reach ETA equals one therefore it is normally used as a voltage doubler thus this topology is called charge pump in voltage doubler configuration this charge pump can therefore overcome the disadvantages one and two of the linear regulator but the charge pump can also be operated in another configuration to produce a negative output voltage to do that the topology is changed in the following way switch s3 is moved down here and the polarities of the electrolytic cap CP and Co are reversed the switching sequence must also be altered in phase one s1 and s3 close connecting the pump capacitor in parallel to the input capacitor 1 CP is charged up phase 2 begins s 1 and s 3 are opened rapidly followed by s 2 and s 4 closing the pump capacitor is now connected in parallel to C O which can now deliver an output voltage of up to V s but with reverse polarity with respect to chassis ground like with a voltage doubler configuration this process starts anew as Co is discharging keeping vo as constant as possible if you want to build a charge pump circuit you can use the integrated circuit IC l 7660 the 7660 is intended to operate in the negative output voltage configuration here you can see a very simple circuit based on the IC l 7660 which i built up for this video it only consists of the IC plus - 10 microfarad electrolytic caps and an additional resistor which simulates a load as you can see here the charge pump converts a +5 volts input voltage to an output voltage of around the same value but with reverse polarity you asked what that would be good for well it allows you to power certain devices which need two symmetrical voltages from a single rail power supply or a battery typical applications are therefore rs-232 interfaces and other digital devices like dynamic RAM or 8080 micro processors you can also build up a charge pump in voltage doubler configuration with the ICL 7660 the efficiency is however not as good because to external diodes over which voltage drops occur are necessary for that same reason you can always only get a maximum output voltage of two times the input voltage minus the two forward voltages of these diodes it is also possible to cascade two or more circuits to boost the output voltage further this is possible for positive as well as negative output voltages to increase the maximum current two or more circuits can also be paralleled but talking about the allowed range of currents the biggest drawback of the charge pump becomes obvious a single charge pump based on the ICL 7660 can only deliver a couple of milliamps and even with several such devices in parallel this topology could not be used to convert electrical power of more than even a few watts therefore the charge pump is only used in very low power applications and you most probably will not see a major power supply based on a charge pump topology as of today the same thing is true for any converter that relies on switched capacitors alone therefore the charge pump is a rather exotic topology which doesn't play an important role in power supply design thus I will not make another video about it I simply wanted to show you that such a switched capacitor design is possible all other topologies that follow now involve inductive components in some way all of which play a bigger role than the charge pump in the cause of this series the upcoming topologies will be dealt with in detail in special videos that concentrate on only one topology per video therefore this video will only supply you with a basic information to get an overview over existing switching converter topologies the back converter the converter we will now talk about is one of the most simple abandoned and most popular switching converter topologies to understand its operation let us again return to the linear regulator as you know the linear regulator uses a variable resistive element to create a constant output voltage smaller than its input voltage the power dissipated by the resistive element is pedis equals I times VI minus vo which equals I times Delta V PDS and thus the energy loss is here from the electronic system increases as I + or Delta V increase to build an ideal converter with efficiency ETA equals 1 however that product of I times Delta V must be 0 as you know the product of two factors always becomes zero once at least one of the factors is 0 therefore the resistive element in this circuit must be replaced by another component over which always either I or Delta V is 0 what kind of magic component would that be well the simplest one you can imagine a switch an ideal switch can only exist in two states when an ideal switch inside any loop is opened the voltage across the switch is at the maximum value while the current through it is zero when the switch is closed however the voltage across the switch becomes zero while the current through it becomes maximum so as you can see the power dissipated over an ideal switch or waise must be zero now of course a switch which is in one of its two states constantly cannot be used to regulate the voltage across any load when the switch is closed the output voltage becomes maximum and when the switch is opened vo becomes zero so how could you use a switch to change the value of the output voltage well you have to close and open the switch periodically in this way you can control the mean output voltage which is nothing but the average value over time the mean voltage is defined as V equals 1 over the switching period t times the integral of V of T DT which can be simplified to V equals 1 over capital T times V T when the voltage vo has constant values within the on and off times of the switch which here is the case the actual mean value of the output voltage can also be defined as vo equals VI times D with D being the so called duty cycle defined as the on time divided by the switching period according to these equations it is possible to control the mean output voltage across the load via controlling the value of the duty cycle that technique is called pulse width modulation abbreviated PWM we will use it many times in the future in all kinds of switching converters in reality the switch is replaced by a power transistor which acts as a switch when it is fed by a square wave generator of some kind the generator is usually working with a frequency in the kilohertz range meaning that the transistor is closing and opening thousands of times per second now of course you will think alright so by switching the transistor on and off I can control the mean value of the voltage across the load but that mean value is just a theoretical value while in reality the voltage across the load is not constant but just periodically jumping from zero to the input voltage that is right and sensitive electronic equipment cannot be supplied with such a pulse width modulated voltage directly but there are some loads which can for example the good old light bulb the light bulb will only be destroyed when over a certain period of time a current passes through it which exceeds its ratings causing the filament to overheat and burn through but when fed with extremely short pulses like in this case the light bulb is just fine with the mean value of its rated voltage the light bulb will also flicker in the frequency of the PWM but the human sense of vision is much too slow to even notice that high frequency flickering to demonstrate this let me show you a little setup for that I will power a 6 volt light bulb from a 12 volt LED acid battery similar to the example with a linear regulator in part 2 the duty cycle necessary for this task is d equals 0.5 which means that the transistor will conduct for exactly half of the time here you see the experimental setup on the Left DMM the input voltage across the battery is displayed and in the middle you see the 6 volt light bulb on the oscilloscope screen you can see the actual waveform of the voltage across the bulb and on the right multimeter the mean value of that voltage is measured and witness how the mean value of the brightness of the bulb increases as I dial up the duty cycle of the square wave generator and how it dims down as I decrease the duty cycle again but as I mentioned before many other devices unlike the lightbulb cannot be driven by a pulse with modulated supply voltage directly they need a truly constant voltage and that is why it is necessary to find a way of physically averaging the output voltage of the converter rather than relying on the purely mathematical mean voltage we've been using as of now that averaging process is accomplished by adding an LC filter and a second switch to the circuit once the on phase begins the current rush through the LC filter is slowed down by the inductor as the output capacitor starts charging up once the off time starts the second switch is immediately closed acting as a freewheeling path so that the voltage induced across the inductor by the interruption of the current path is discharged into the output capacitor in this way it can be accomplished that the voltage across the capacitor stays relatively constant over the whole switching period this setup completes the buck converter circuit in most real-world buck converters the second switch is realized by a diode which can act as a so-called freewheeling diode if you do not yet understand why this averaging process works you have to learn about the principle behavior of switched capacitors and inductors why a voltage is induced across the inductor once which one is opened will be explained in detail in the next video when I will talk about boost and fly back converter as I have said before there will be a in detail design video about buck converters once I have explained all the topologies on a basic level therefore it is enough for now if you understand that the buck converter is basically a rapidly closing and opening switch followed by an LC filter which electronically averages the output voltage what you can see here is a piece of sensitive electronic equipment namely a microcontroller board which is powered via a buck converter circuit this is only possible because of the averaging done by the LC filter inside the buck converter on the oscilloscope screen you can see now that the output voltage is constant advantages and disadvantages of the buck converter the buck converter is normally used only for battery-powered circuits or on the secondary side of a transformer it is not suited to step down the rectified mains voltage because it does not provide electrical isolation mains powered AC to DC switched-mode power supplies on the other hand rely on the flyback the forward and the push pull converter as I will explain in the upcoming parts of this tutorial this topology can theoretically reach 100% efficiency which is basically it's single big advantage over the linear regulator but like the linear regulator its output voltage is always smaller than its input voltage this limitation will be overcome by the next converter we will talk about the boost converter so if you like this video watch the other parts of this tutorial and please subscribe to my channel

Channel: The Post Apocalyptic Inventor

Views: 168,888

Rating: 4.9609876 out of 5

Keywords: Switched-mode Power Supply, Buck Converter, Charge Pump, Boost Converter, Flyback Converter, Forward Converter, Push Pull Converter, Switching Converter, DC to DC Converter, ATX, PSU

Id: r_AAdxwfim8

Length: 22min 2sec (1322 seconds)

Published: Sun May 04 2014