www.fairchildsemi.com ? 2010 fairchild semiconductor corporation www.fairchildsemi.com rev. 1.0.0 ? 12/23/10 AN-9720 power path implementation tradeoffs, featuring the fan5400 family of pwm battery chargers overview many battery-powered systems, such as smart phones, must be fully functional and their electronics up and running shortly after the user plugs a charger into the phone. the system electronics need to be functional regardless of the state of the battery, even if the battery is fully discharged or absent. the fan5400 family charger ic allows power to be delivered to the system when the charger is plugged in, whether the battery is absent or present, and allows the system to power up and be functional quickly after the charger is plugged in. other approaches, such as power path, add additional impedance in series with the battery. this application note describes how the fan5400 family of battery chargers distributes power between the system and battery to achieve similar results of power path, often without the overhead and power loss of an additional switch element. the block diagram in figure 1 is a typical implementation of power path using an ?ideal? diode. current flow is indicated by the arrows, which reveal that the ?ideal? diode (whether internal or external) helps to steer current appropriately. while this implementation meets the criteria for power path, an ideal diode is never truly ideal. for example, the internal ideal diode of one such ic is actually a pmos with a typical value of 180m . this means that there is a permanent 180m power dissipating series element between battery and system load that creates significant additional power loss during periods of high-current drain from the battery, such as gsm pulses. a parallel diode pmos switch can reduce this resistance; however, this increases the solution size and cost. the implementation in figure 1 differs from the fan540x approach in the block diagram in figure 2. although on the surface it may seem like fan5400 does not have power path functionality; it serves almost all the same needs and, in addition, can provide the benefit of having no power dissipating series element between system load and battery. q1 figure 1. typical implementation of power path using an ?ideal? diode
AN-9720 application note ? 2010 fairchild semiconductor corporation www.fairchildsemi.com rev. 1.0.0 ? 12/23/10 2 pwm modulator q1 q2 q1a q1b isns pmid sw pgnd pmid l1 vbat c mid + battery csin 1 h 4.7 f c bat system load vref sda scl otg/usb# vcc vbus c bus 1 f stat i2c interface logic and control power output stage pmid osc 30ma c out 0.1 f r sense q3 charge pump vbus ovp i_in control vreg c reg 1 f dac disable 1.8v / pmid reg figure 2. fan5400 block diagram system and battery power sharing power sharing between the system and the battery means that power can be steered or prioritized to go to the system in the case that the input power is not sufficient to power both the system and charge the battery. the typical configuration for the fan5400, as shown in figure 3, is one where the system is connected in parallel to the battery. the way in which this configuration can steer power similar to power path is sometimes confusing, so scenarios based on real-world battery capacity and input power numbers are provided below. figure 3. typical application circuit, system parallel with battery example 1: 1500mah battery (1c maximum charge current capability of ba ttery is 1500ma), input power source usb 2.0, 5v 500ma scenario a) partially charged battery at 3.6v and system load turns on at 400ma before the system load turns on, the charger is already in cc mode. because the input power source is 5v 500ma and the battery is at 3.6v, there is ~632ma of current available to charge the battery. this is computed by accounting for the charger conversion efficiency and also the output current multiplication factor achieved when bucking down a voltage: efficiency i v v i in out in outmax ? ? = / (1) using the values in this example reveals that 5v/3.6v?500ma?91%=632ma. the 91% efficiency data point can be found in figure 4. 84% 86% 88% 90% 92% 94% 2.5 2.7 2.9 3.1 3.3 3.5 3.7 3.9 4.1 4.3 efficiency batter y volta g e , v bat ( v ) 4.5vbus 5.0vbus 5.5vbus figure 4. fan5400 conversion efficiency vs. battery voltage vs. v bus voltage
AN-9720 application note ? 2010 fairchild semiconductor corporation www.fairchildsemi.com rev. 1.0.0 ? 12/23/10 3 once the system load turns on, 400ma is diverted to the system and only 232ma is left to charge the battery. this is the equivalent of power steering; the charger is prioritizing the system over the battery. once the system load turns off, the full 632ma once again flows to the battery. the benefit of fan5400, as configured in figure 3, is that there is no power- dissipating series element between the system and load. scenario b) partially charged battery at 3.6v and system load turns on at 2000ma before the system load turns on, the charger is already in cc mode and using all input power to charge the battery at 632ma, similar to scenario a. when the system load turns on, 632ma is diverted to the system and the remaining system load of 1368ma comes from the battery. this is the equivalent of power steering; the charger is prioritizing the system over the battery. once the system load turns off, the full 632ma once again flows to the battery. again, the circuit in figure 3 has an advantage of no dissipating element between the system and load. scenario c) battery is fully charged at 4.2v and system load turns on at 400ma before the system load turns on, the charger is off. when the load turns on, all the system power first comes from the battery. as soon as v bat < v oreg - v rch , the charger turns on. v rch is the recharge threshold and has a value of 120mv. because the input power source is 5v 500ma, the maximum available current the charger can provide is computed as 5v/4v?500ma?92%=575ma (battery is assumed to have to 4v for the sake of this exercise). the charger turns on, attempting to charge the battery at 575ma. however, because the system load is still present, only 575ma-400ma=175ma is flowing into the battery. this is the equivalent of power steering; the charger is prioritizing the system over the battery. once the system load turns off, the full 575ma flows to the battery until the battery enters cv mode; at which point, the charge current decreases. again, the circuit in figure 3 has an advantage of no dissipating element between system and load. scenario d) battery is fully charged at 4.2v and system load turns on at 2000ma before the system load turns on, the charger is off. when the system load turns on, the power first comes from the battery and almost immediately the battery charger turns on and goes into cc mode. this is because li-ion batteries typically have an output impedance of 150m , which almost instantly forces v bat < v oreg - v rch . similar to scenario c, the charger attempts to charge the battery at 575ma. (actually slightly higher than 575ma because battery voltage is lower in this case, as compared to scenario c, and the multiplication factor is slightly higher. for the sake of this exercise, it is inconsequential.) the charger attempts to charge; however, because the system load is 2000ma, the 575ma flows to the load and the remaining system load of 1425ma comes from the battery. this is the equivalent of power steering; the charger is prioritizing the system over the battery. once the system load turns off, the full 575ma flows to the battery until the battery enters cv mode, at which point, the charge current decreases. again, the circuit in figure 3 has an advantage of no dissipating element between system and load. example 2: assume a 700mah battery (1c maximum charge current capability of battery 700ma), input power source ac/dc adaptor 5v 900ma or usb 3.0, 5v, 900ma scenario a) partially charged battery at 3.6v and system load turns on at 400ma before the system load turns on, the charger is in cc mode. because the battery is at 3.6v, there is 5v/3.6v?900ma?91% =1138ma of current available to charge the battery. however, the battery is limited to a maximum 1c charge current of 700ma and, therefore, the charger is set to charge at 700ma. what is unique in example 2, in comparison to example 1, is that the input power supply can supply more power than the battery is able to accept in 1c scenario. when the system load turns on, 400ma is diverted to the system and only 300ma is left to charge the battery. some designers might object because the input power supply is not being used to the full extent that the combination of the battery and system load can accep t. one solution is to connect the system load to csin, as shown in figure 5. figure 5. application circuit with system load connected to csin
AN-9720 application note ? 2010 fairchild semiconductor corporation www.fairchildsemi.com rev. 1.0.0 ? 12/23/10 4 connecting the system load to csin allows the ac/dc adaptor or usb source to supply current up to maximum power level even if that is higher than the battery?s 1c level. in this configuration, before the load turns on, the battery is charging at its 1c maximum charge capability of 700ma. when the 400ma system load turns on, the entire 400ma system load is supplied by the charger and the battery continues to charge at 700ma. one drawback to this configura tion is that there is a power dissipating 68m series element always in the path between the battery and the system load. this constant power dissipating element between battery and system load is similar to the one found in the power path implementation in figure 1. however, the 68m in fan5400 is substantially less than the 180m found in some products with power path. scenario b) partially charged battery at 3.6v and system load turns on at 2000ma before the system load turns on, the charger is in cc mode and charging the battery at 700ma, similar to scenario a. when the system load turns on, if the configuration from figure 3 is used, 700ma is supplied to the system load from the charger and the remaining 1300ma by the battery. if the configuration from figure 5 is used instead, 1138ma is supplied to the system load from the charger and the remaining 862ma is supplied by the battery. both configurations are the equivalent of power steering; but in the configuration of figure 5, all the input power is being used. the tradeoff is the 68m -series element between battery and system load. once the system load turns off, 700ma flows to the battery in both configurations. scenario c) battery is fully charged at 4.2v and system load turns on at 400ma before the system load turns on, the charger is off. when the load turns on, the system power first comes from the battery. as soon as v bat < v oreg - v rch , the charger turns on. v rch is the recharge threshold and has a value of 120mv. because the input power source is 5v 900ma, the maximum available current the charger can provide is 5v/4v?900ma?92% =1035ma (assume the battery has dropped to 4v for the sake of this exercise). the charger turns on, attempting to charge the battery at 700ma. however, because the system load is still present, if the configuration in figure 3 is used; only 300ma is left to charge the battery. if the configuration in figure 5 is used; when the load turns on, 635ma flows to the battery and 400ma is supplied from the charger to the system load. this makes sense because there is a total of 1035ma charger output current. once the system load turns off, the full 700ma flows to the battery until the battery enters cv mode; at which point, the charge current decreases. the tradeoff is the 68m -series element between the battery and system load. scenario d) battery is fully charged at 4.2v and system load turns on at 2000ma before the system load turns on, the charger is off. when the system load turns on; if the configuration from figure 3 is used, the power comes first from the battery and almost immediately the battery charger turns on and goes into cc mode. this is because li-ion batteries typically have an output impedance of 150m , which almost instantly forces v bat < v oreg - v rch . similar to scenario c, the charger attempts to charge the battery at 700ma. however, because the system load is 2000ma, the 700ma flows from the charger to the system load and the remaining system load of 1300ma comes from the battery. if the configuration in figure 5 is used; when the load turns on, 1035ma flows from the charger to the load and the remaining 965ma is supplied by the battery. once the system load turns off, 700ma flows to the battery until the battery enters cv mode; at which point, the charge current begins to decrease. again, the tradeoff is the 68 -series element between the battery and system load. charge termination when system load is connected to csin when the system load is connect ed to csin as it is in figure 5; if the system load is larger than the power that can be supplied from the usb port with the i bus limit (e.g. 500ma), the battery current is reduced below the termination threshold. normally, when the battery current is reduced below the termination threshold, charge is terminated. however, if the charger input is in current limit, the fan540x does not allow charge termination. charge termination occurs only if i bat is less than i term and the charger input is not in current limit for at least 32ms. this prevents false termination from system loading. after charge termination occurs, the charger stops running and system power is drawn from the battery. charge termination can be disabled by disabling the te bit through i 2 c. charge termination may, however, be desired if the phone or data card is in a low-power mode and a drain on the usb port is not desired. in that case, recharge occurs automatically when the battery voltage drops v rch (120mv) below v oreg or if it is desired to turn on before 120mv, the system can manually turn on the charger as soon a drop in v bat is sensed. powering the system with no battery the fan5402 and fan5405 continues charging after vbus por with the default parameters, regulating the v bat line to 3.54v until the host processor issues commands or the 15- minute timer expires. in this way, the fan5402/05 can start the system without a battery. the fan5400 family?s soft-start function can interfere with the system supply with battery absent. the soft-start activates
AN-9720 application note ? 2010 fairchild semiconductor corporation www.fairchildsemi.com rev. 1.0.0 ? 12/23/10 5 whenever v oreg , i inlim , or i ocharge are set from a lower to higher value. during soft-start, the i in limit drops to 100ma for about 1ms, unless i inlim is set to 11 (?no limit?). this could cause the system processor to fail to start. to avoid this behavior, use the following sequence. 1. set the otg pin high. when vbus is plugged in, i inlim is set to 500ma until the system processor powers up and can set parameters through i 2 c. 2. program the safety register 3. set i inlim to 11 (no limit). 4. set oreg to the desired value (typically 4.18) 5. reset iolevel bit, then set iocharge. 6. set i inlim to 500ma if a usb source is connected or any other level that is preferred. during the initial system startup, while the charger ic is being programmed, the system current is limited to 340ma for 1ms during steps 4 and 5. this is the value of the soft-start i charge current used when i inlim is set to no limit. if the system powers up without a battery present, the cv bit should be set. when a battery is inserted, the cv bit clears. powering a system when the battery is deeply discharged when the battery voltage is lower than the voltage required to power the system load, the power routing implementation ( error! reference source not found. and figure 5) cannot bring up the system. in contrast, power path, as implemented in figure 1, can still power the system even if the battery voltage is substantially lower than the system load needs to operate. this is the primary advantage of the circu it in figure 1 as compared to fan5400. however, it is important to realize that the slope of the curve during this phase of the charge cycle is incredibly steep, which means that the battery voltage is brought up to the minimum system load requirement within a matter of seconds. to understand how quickly the battery voltage rises, it is necessary to examine the internal operation of the battery pack. inside every battery pack is a protection ic, as shown in figure 6, that features two back-to-back mosfets and an analog control circuit to prevent over-charging and over- discharging by monitoring the cell voltage and discharge current. the protection circuit is also referred to as ?secondary protection? since the charging system must also ensure that the battery is not over-charged. the protection circuit provides a back-up safety circuit where overcharging is concerned. protection circuit esr + cell control ? + q1 q2 figure 6. li-ion battery pack q2 in the protection circuit opens if the cell is deeply discharged (v cell < 2.7v). charging is therefore still possible by driving current into the pack through q2?s body diode. refer to figure 7 for the different states.
AN-9720 application note ? 2010 fairchild semiconductor corporation www.fairchildsemi.com rev. 1.0.0 ? 12/23/10 6 table 1. internal batter protections scheme condition detection response reset over-charge v cell >4.25 open q1 v cell >4.10 over-discharge v cell <2.7 open q2 v cell >2.95 over i discharge v cell ? v bat >x open q2 v cell v bat in the extreme case that the battery is deeply discharged, the discharge protection switch (q2) is open. when the charger is plugged in the usb port; if otg = 0, the trickle charge of 30ma charges the system capacitor to 2v in 20ms time ( refer to table 2 ). the charger continues to run in ?auto? mode until the applications processor wakes up at 3.4v in about 15- seconds time. the applications processor then enumerates with usb and sets i bus = 500ma and i charge = 700ma. table 2. bringing battery out of deep discharge condition v bat i bat comment linear charge (20ms) 0 to 2.0 30ma v bat =emf+0.7 charger is charging system cap. t 15min (default) pwm charge with i bus set by otg pin (15s) 2.0 to 3.4 100ma v bat =(emf=+0.7+.1?esr) some charge going to vcell apps processor wakes up at v bat 3.4v high-current pwm charge 1s to q2 close 3.5+ 700ma higher charge current set by processor. this causes v cell to rise, which causes q2 to close at point a in figure 7. v bat drops to about 2.9 and charging continues from there. the charge cycle behavior is shown graphically in figure 7. the time it takes to charge a typical battery to a stable 3.4v is 40 seconds. this process is explained below. a processor wakes up when v bat > 3.4v, which occurs about 15s after v bus is plugged in. b q2 closes about 1s after the processor programs the ic for higher charge current. this causes v bat to drop (no more diode in series with v bat ). figure 7. charge characteristics for deeply discharged and dead batteries 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 0 20 40 60 80 100 120 140 ibat (a) vbat (v) time (seconds) vbat battery current (a) a
AN-9720 application note ? 2010 fairchild semiconductor corporation www.fairchildsemi.com rev. 1.0.0 ? 12/23/10 7 conclusion although there are situations where the fan 540x partial power path in figure 2 cannot immediately power the system load when the battery is very low or deeply discharged, this timeframe for a typical cell phone battery is only 40 seconds. it is important to weigh this against the benefits offered by the fan540x. this benefit was discussed in detail, showing that the fan5400, configured as in figure 3, does not have a power-dissipating series element between battery and system load and provides the two more critical aspects of dynamic power routing? power sharing between the system and the battery and powering the system with no battery. related resources fan5400 family ? usb-compliant single-cell li-ion switching charger with usb-otg boost regulator an-1721 ? li-ion battery charging basics, featuring the fan5400 / fan5420 family of pwm battery chargers disclaimer fairchild semiconductor reserves the right to make changes without further notice to any products herein to improve reliability, function, or design. fairchild does not assume any liability arising out of the application or use of any product or circuit described herein; neither does it convey any license under its patent rights, nor the rights of others. life support policy fairchild?s products are not authorized for use as critical components in life support devices or systems without the express written approval of the president of fairchild semiconductor corporation. as used herein: 1. life support devices or systems are devices or systems which, (a) are intended for surgical implant into the body, or (b) support or sustain life, or (c) whose failure to perform when properly used in accordance with instructions for use provided in the labeling, can be reasonably expected to result in significant injury to the user. 2. a critical component is any component of a life support device or system whose failure to perform can be reasonably expected to cause the failure of the life support device or system, or to affect its safety or effectiveness.
|
