Wednesday, November 30, 2011

Buck Regulator- Limitation in Output Voltage


Almost all of high speed digital PC boards wouldn’t be come without using Buck switching regulator. Because, in common, digital ICs need wide range of low voltage levels like 1.0, 1.2, 1.8 and etc for operation  and buck regulator will perfectly suit to generate these voltage from higher input voltage (Usually 5V or 12V). But in real, using buck regulator we can’t generate any voltage for a given input voltage. In simple words, high to very low voltage conversion will not be possible if you use buck regulator (Ex: 12V to 1V). In this post we see how this scenario comes into picture.
            To understand this, we first see some basic functions of a Buck regulator. It uses a PWM controller which runs at switching frequency of buck regulator (can be found in device datasheet). PWM controller’s duty cycle will vary based on our output voltage. PWM controller switches an inductor using MOSFET. Average current of inductor will generate a constant voltage with ripple (These ripple will be smoothen by using a capacitor). So we come to know that if a user wants to change output voltage, he has to change duty cycle (This can be done usually using resistor divider in many buck regulators). We can relate input voltage, output voltage and duty cycle by the following mathematical equation.
                        Vout  = D X Vin (D is duty cycle)
            (We have to do lot of mathematical jobs to arrive the above equation. For simplicity and to lack competency I / we avoid it)!!!!!
                        D  = Ton * Fsw (Ton is on period of PWM controller and Fsw is switching frequency)
            So Vout = Vin X Ton X Fsw -------------à (A)
            In the above formula, Ton is the variable and by changing it Vout can be varied for a given Vin. Now by rearranging the above formula,
            Ton = Vout / (Vin X Fsw)
            Now we come to know that Ton is directly proportional to Vout (For fixed Fsw and Vin). If we want to have lower Vout, obviously Ton also will be reduced. Now the question arises? How far the Ton can be reduced (means how much less Vout we can generate). This parameter will be defined in device datasheet as “Minimum Controllable On Time”. This limitation comes because, gated pulsing mechanism of PWM controller. If we have an on time (Ton, to generate required Vout) which is lesser than specified value, output signal of PWM controller will more likely a DC voltage and consequently switching in MOSFET will not function properly. Finally we don’t get output voltage what we really want.
            In datasheet of all buck regulators, they specify the minimum controllable on time parameter. For example, let take minimum controllable on time is 200ns and Fsw = 600KHz. In our board let take we use 12V as input voltage. By using above formula A, we can find minimum output voltage.
            Vout_min = 200 X10-9 X 12 X 600000 = 1.44V
            So any voltage below than 1.44V, it is impossible to generate using that buck regulator. Bu this applies in ideal condition and in real we have to take the loss in upper and lower MOSFETs of buck regulator. If we consider them also, Vout_min would be little bit higher. So for safer side, we can say that Vout_min = 1.5V. But in datasheet, you can’t find this 1.44V or 1.5V. They say Vout_min is Vref (usually 0.8V or 1.2 or 0.6). So as a designer, we have to find what the real minimum voltage we can generate by using the formula A.

            Summary: Never try to generate below than Vin X Ton X Fsw in buck regulator.

Best Regards,
Arun Alagarsamy...

Friday, August 26, 2011

PN Junction Diode

(In last blog we have seen about Semiconductors and now we are gonna look out PN Junction diode.)

        Normally, PN junctions are manufactured from a single crystal with different dopant concentrations diffused across it. Once PN junction is made it looks like this.


        On the left side, it is P type semiconductor and the right side it is N type semiconductor. In P type, we have more no of holes and and in N type we have more no of free electrons. So in any material, which carriers is more, they known as 'majority carriers' and which are less, they known as 'minority carriers'. In p-type, holes are majority carriers and electrons are minority carriers. This is reverse in N type materials. Holes are represented by unfilled circles and electrons are represented by thick circles. Now what represent unfilled circles with - and + sign?? They are new to us and we check them.
      Imagine we use Si-P (Silicon-Phosphorous) materials for N type and Si-B (Silicon-Boron) materials for P-Type
      When creating PN junction, holes and electrons of both P and N materials respectively near the boundary area diffuse into other material each other (P diffuse to N and N diffuse to P). At that time, free electron of Si-P atoms will fill the electron deficiency in Si-B atoms. So Si-P atoms lose one electron and thus gain one positive charge (represented by unfilled circles with + sign). As same, Si-B atoms gain one electron and thus gain one negative charge (unfilled circles with - sign). This is called 'Space charge Creation'. Atoms which have lost/gain electrons become charged and they posses an electric field. This electric filed will inhibit further diffusion of holes and electrons cross PN junction boundary and ultimately avoid more space charge creation. A PN junction is look like this:

  

         This is called as "PN Junction Diode" and basic part of electronics. We see biasing of diodes in next post.

Thanks and Regards,
Arun Alagarsamy @ arunvasan@gmail.com



Semiconductors- Short story

          All of you may aware of "Conductors" and "Insulators". It means things which conduct electric current well called 'Conductors', ex: Copper, Gold, Iron, etc and things don't called as 'Insulators', ex: Glass, Ceramic, Plastic, etc. Wait. A simple question. Which characteristic defines an object? Obviously it is "Electrical Conductivity". Objects have more conductivity become conductors (Silver has 63000000 S/m) and objects has lesser conductivity become insulators (Glass has 10e-11 to  10e-15).
            How conductors conduct electric current is another phenomena. To explain this thing clearly you have to aware of Energy Levels, Electron orbitals, valance band and conduction band. Please read here: http://hyperphysics.phy-astr.gsu.edu/hbase/solids/semcn.html
          Got something?? OK fine. Simply say, energy level of the outermost electron orbital is called as Valance band and energy level of electrons required for electrical conduction is called as conduction band. In atoms, electrons occupy in orbitals in according to the probability function (You have to read Quantum Mechanics guys !!!). So depending upon no of electrons in that conduction band energy level, we can say whether it is conductor or insulator. In another way, we can differentiate objects based upon where the conduction band resides. Look at this picture.


              So in conductor, there are plenty of free electrons in conduction band so it easily conducts. But look at insulator. For conduction, every electron needs to jump from valance band to conduction band and that's why it doesn't conduct electric current and it exhibits great electrical resistance!


           Now where 'Semiconductors' comes into the picture?? The name itself it has the meaning and the few basic characteristics  are
  • They have conductivity less than conductors and more than insulators
  • Band gap between conduction and valance band is less than insulator.
  • At room temperature, few electrons occupy conduction band but still it doesn't conduct electric current.
  • These semiconductors are called as "Intrinsic Semiconductors".
  • Ex: Silicon (Si), Germanium(Ge)
        But for conduction, these intrinsic semiconductors are not suitable and our aim is make the conductivity of semiconductor more means to have many free electrons. One of the way is to add impurities to intrinsic semiconductor. These semiconductors are known as "Extrinsic Semiconductor". We see bit brief about extrinsic semiconductor.
       Adding impurities in intrinsic semiconductor is known as 'Doping'. Lets take Silicon atom. It has 4 electrons in outer orbit. So if we add a Phosphorus (has 5 electron in outer orbit) atom with it , 4 electrons of Phosphorus and Silicon make covalent bond and remaining one electron of Phosphorus become free electron. At room temperature, we have electrons above valance band. This type semiconductor is known as "N-Type" semiconductor. Look at the following figures.


         If we add materials which has 3 electrons in outer orbit (Ex: Boron[B]), the outcome is "P-Type" semiconductor. 3 of both Si and B make covalent bond and deficiency in Si atom creates a "Hole". This hole is not permanent. Electrons from nearby atoms try to replace it but atom which replaces nearby hole will create hole. So at room temperature, "P-Type" material has free electrons at quite higher level than valance band. Look at this pictures.
 

          When compare with intrinsic semiconductors, these P and N type extrinsic semiconductors has better conductivity and obviously lesser electrical resistance.
           Now we came to know a bit about P and N type semiconductors. Our next goal how a PN junction diode is created and how it works..

Thanks and Regards,
Arun Alagarsamy @ arunvasan@gmail.com