f battery analysis - circuit...

8
34 CIRCUIT CELLAR ® www.circuitcellar.com September 2011 – Issue 254 Battery ratings aren’t always accurate. In fact, many batteries don’t meet their rated capacities. This microcontroller-based battery analyzer enables you to sort the good from the bad. It’s designed for use with single-cell li-ion batteries with a nominal 3.7-V rating, and you can use it with various cell types. Battery Analysis R F EATURE ARTICLE by Richard Pierce (USA) ecently, I started wondering why some of my li-ion batteries seemed to perform differently than others even though their rated capacities are the same. It’s true that batteries change over time and as they are cycled, but these were all “new” batteries. So, I designed a battery analyzer to figure out what was going on. Batteries are usually rated in mAh (or mA-h, Ah, A-h, or something similar). A mAh is one milliamp hour, and an A-h is one amp hour. A battery rated at 1 A-h (or the equivalent 1,000 mAh) means that the battery should be able to deliver the equivalent of 1 A for a period of 1 hour. For example, if the battery drain is only a 0.5 A, it should last 2 hours. If the drain is 2 A, it should last a half an hour. During the battery drain time the battery voltage is not constant; it drops from a fully charged voltage to a minimum voltage that is specified by the battery manufacturer. The discharge rate (load current) for the battery capacity rating is usually specified by the manufacturer. So you might have, for example, a 10-A-h bat- tery with a 1-A discharge rate. If the battery is discharged at a higher rate than 1 A, it probably won’t be able to provide the full 10-A-h energy rating. My battery analyzer—which is based on a Microchip Technology PIC18F4525 microcontroller—is intended to be used with single-cell li-ion batteries with a nominal 3.7-V rating (see Photo 1). It can be used with var- ious types of cells including 18650, 16340, 14500, 17500, 18500, and 17670 with capacity ratings of 500 to 3,000 mAh. The charge and discharge currents can be up to 1 A, maximum. The process of analyzing the perform- ance of a battery consists of charging the battery to a “full” level and then discharg- ing the battery to an “empty” level while measuring and totalizing the amount of current produced by the battery. The full and empty levels are configurable, and the values can usually be found on the bat- tery’s datasheet. When it’s done, the total milliamp hour discharged is shown. This value should be, in theory, pretty close to the battery rating. There are two ways that the battery can be discharged: one is for maxi- mum energy and one is for maximum power. The maxi- mum energy mode discharges the battery at a constant cur- rent until the empty voltage is reached, then reduces the current while maintaining the empty voltage until the bat- tery isn’t able to supply 50 mA of current. The maximum power mode discharges the battery at a constant current and then stops when the empty voltage is reached. The maximum power mode is intended to be representative of how a battery would be discharged in a real-life application. It is also possible to charge (or discharge) a battery to a spe- cific level. This enables a battery to be prepared for storage, for example. There are also some preset charge levels and the ability to set a custom charge level. The presets include stan- dard (4.2 V), military (3.92 V), and storage (3.5 V). The stan- dard level is the maximum energy that can be safely stuffed into a li-ion battery. The military level is a more conservative value that increases the battery life somewhat and enables the battery to be used over a wider temperature range while fully charged. The storage level is for storing a battery for a long period of time. Storing a li-ion battery with a high level of charge tends to reduce the battery’s lifespan. a) b) Photo 1—The front (a) and back (b) of the finished battery analyzer unit Build an MCU-Based Analyzer Unit

Upload: dodung

Post on 10-Jun-2018

216 views

Category:

Documents


0 download

TRANSCRIPT

34 CIRCUIT CELLAR® • www.circuitcellar.com

Septe

mber

2011

– Is

sue 2

54

Battery ratings aren’t always accurate. In fact, many batteries don’t meet theirrated capacities. This microcontroller-based battery analyzer enables you to sortthe good from the bad. It’s designed for use with single-cell li-ion batteries with anominal 3.7-V rating, and you can use it with various cell types.

Battery Analysis

R

FEAT

URE

ARTICLEby Richard Pierce (USA)

ecently, I started wondering why some of my li-ionbatteries seemed to perform differently than others

even though their rated capacities are the same. It’s true thatbatteries change over time and as they are cycled, but thesewere all “new” batteries. So, I designed a battery analyzer tofigure out what was going on.

Batteries are usually rated in mAh (or mA-h, Ah, A-h, orsomething similar). A mAh is one milliamp hour, and an A-his one amp hour. A battery rated at 1 A-h (or the equivalent1,000 mAh) means that the battery should be able to deliverthe equivalent of 1 A for a period of 1 hour. For example, ifthe battery drain is only a 0.5 A, it should last 2 hours. If thedrain is 2 A, it should last a half an hour. During the batterydrain time the battery voltage is not constant; it drops from afully charged voltage to a minimum voltage that is specifiedby the battery manufacturer. The discharge rate (load current)for the battery capacity rating is usually specified by themanufacturer. So you might have, for example, a 10-A-h bat-tery with a 1-A discharge rate. If the battery is discharged at ahigher rate than 1 A, it probably won’t be able to provide thefull 10-A-h energy rating.

My battery analyzer—which is based on a MicrochipTechnology PIC18F4525 microcontroller—is intended to beused with single-cell li-ion batteries with a nominal 3.7-Vrating (see Photo 1). It can be used with var-ious types of cells including 18650, 16340,14500, 17500, 18500, and 17670 withcapacity ratings of 500 to 3,000 mAh. Thecharge and discharge currents can be up to1 A, maximum.

The process of analyzing the perform-ance of a battery consists of charging thebattery to a “full” level and then discharg-ing the battery to an “empty” level whilemeasuring and totalizing the amount ofcurrent produced by the battery. The fulland empty levels are configurable, and thevalues can usually be found on the bat-tery’s datasheet. When it’s done, the total

milliamp hour discharged is shown. This value should be,in theory, pretty close to the battery rating. There are twoways that the battery can be discharged: one is for maxi-mum energy and one is for maximum power. The maxi-mum energy mode discharges the battery at a constant cur-rent until the empty voltage is reached, then reduces thecurrent while maintaining the empty voltage until the bat-tery isn’t able to supply 50 mA of current. The maximumpower mode discharges the battery at a constant currentand then stops when the empty voltage is reached. Themaximum power mode is intended to be representative ofhow a battery would be discharged in a real-life application.

It is also possible to charge (or discharge) a battery to a spe-cific level. This enables a battery to be prepared for storage,for example. There are also some preset charge levels and theability to set a custom charge level. The presets include stan-dard (4.2 V), military (3.92 V), and storage (3.5 V). The stan-dard level is the maximum energy that can be safely stuffedinto a li-ion battery. The military level is a more conservativevalue that increases the battery life somewhat and enablesthe battery to be used over a wider temperature range whilefully charged. The storage level is for storing a battery for along period of time. Storing a li-ion battery with a high levelof charge tends to reduce the battery’s lifespan.

a) b)

Photo 1—The front (a) and back (b) of the finished battery analyzer unit

Build an MCU-Based Analyzer Unit

2011-8-014_Pierce_Layout 1 08/09/11 4:50 PM Page 34

cabate
Text Box
Copyright Circuit Cellar, Inc. This article cannot be reprinted without permission: [email protected]. To subscribe: http://circuitcellar.com/subscription/

critical and is low-current,so a 78L33 does the jobnicely. A precision 5-V ref-erence chip provides theanalog voltage reference.

Normally, a 5-V logicsupply would be used, butthe 5.5-V logic supply volt-age level is necessary forthe digital potentiometerchips so their wiper volt-ages don’t exceed the supplyvoltage. The CPU andpotentiometer chips arerated for this voltage, butthere is no margin for error.

CPUA Microchip Technology

PIC18F4525 microcontrolleris the brain of this batteryanalyzer (see Figure 1). Ichose a crystal oscillatorsource so that reasonablyaccurate timekeeping wouldbe possible. The LCD endedup determining the requiredcrystal frequency. This is

because the optimal speed for communication with the LCDis with a 10-MHz SPI clock. The PIC18F4525 has a four-clockinstruction cycle (the instruction cycle is also the maximumSPI clock rate), an internal PLL that multiplies the 10-MHzcrystal to 40 MHz (resulting in a 10-MHz SPI clock rate), andan instruction cycle rate of 10 MIPS.

The SPI is used with both the LCD and the digital poten-tiometers. To simplify the software, two separate SPI chan-nels are used. The PIC18F4525 only has one hardware SPIchannel so the second channel is bit-banged with software.

www.circuitcellar.com • CIRCUIT CELLAR® 35

Septe

mber

2011

– Is

sue 2

54

BATTERY CHARGING STRATEGYA variety of li-ion battery charging methods can be used.

My device uses the constant current, constant voltage(CCCV) method, where a constant current is used to chargethe battery to the “full” voltage and then the charge currentis reduced while maintaining the final voltage. The chargingprocess stops when the charge current drops to 50 mA.Although it’s not the fastest method, it’s safe and gentle andresults in a fully charged and stable battery condition.

Most li-ion batteries specify that they should be charged ata 0.2C rate (the battery rating divided by two). So, the chargerate for the battery analyzer is fixed at 0.2C. The specifieddischarge rate can vary; it’s usually specified at somethinglike 0.5C to get the most energy out of the battery. But, youmay want to discharge at a higher rate, so this value is con-figurable. (Refer to the Sidebar, “Li-ion Essentials,” for moreinformation.) The maximum discharge rate is limited to 1 A.So, for example, discharging a 2,000-mAh battery at a rateabove 0.2C is not possible with this battery analyzer.

POWER SUPPLYA 9-V, 2-A AC/DC power adapter is the main power

source of the battery analyzer. Good old LM317 voltage reg-ulators are used to derive the 6.8- and 5.5-V supplies. Forthis application, I wanted these supplies to be pretty muchright on the target voltages, and the voltages on the bread-board version were perfect. But something was different onthe PCB version, perhaps due to the different LM317 pack-age type or changing from 1/4 watt to SMT resistors, sotrimming resistors were needed. The 3.3-V supply is not

Figure 1—The battery analyzer block diagram

Li-ion EssentialsBattery capacity ratings are usually specified at a dis-

charge rate of 0.2C (1/5 of the rated capacity). So, to getthe rated capacity from a battery, it should be dischargedat rate that’s less than 1/5 of its rated capacity. If you aregoing to discharge at a higher rate, then capacity shouldbe derated. For example, at a 2C discharge rate less than90% of the rated capacity will be available.

Keep the battery comfortably warm. To get the ratedcapacity from the battery, it should be operated at 23°Cor higher. At low temperatures (at or below freezing),battery performance drops off quickly, especially at highdischarge currents. Be careful not to drop the battery.Physical damage, such as dents or punctures, can causethe battery to fail, sometimes in a dramatic fashion.Never short the battery terminals. Doing so can causepermanent damage to the battery, and maybe you, too.

2011-8-014_Pierce_Layout 1 08/09/11 4:50 PM Page 35

cabate
Text Box
Copyright Circuit Cellar, Inc. This article cannot be reprinted without permission: [email protected]. To subscribe: http://circuitcellar.com/subscription/

36 CIRCUIT CELLAR® • www.circuitcellar.com

Septe

mber

2011

– Is

sue 2

54

HUMAN INTERFACEThe controls for the battery analyzer

are basically two buttons. A Reset buttonprovides a quick way to shut down every-thing, enter a failsafe state, and get backto the main menu. A five-way (up, left,right, down, and push) joystick buttoncontrols everything else. Although thereare enough I/O pins free to directly con-nect the joystick, it’s become a habit ofmine to encode the joystick position as avoltage and then use a single analog inputto read the switch state.

The display is a small (about 1” square)color LCD. The menu-driven applicationprogram consists mainly of text displays(see Photo 2). The graphical capability isused to show the voltage and currentgoing into (or out of) the battery overtime. The display contrast is controllableby commands to the LCD chip and thebacklight is controlled using the PICPWM output and a driver MOSFET.

A large 10-mm RGB status LED makesit easy to know what the battery analyzeris doing from across the room. When it’scharging or discharging, the battery theLED is red; when it’s done, it turns green.Blue is only used when the diagnosticsare running or to indicate that a safetyshutdown has occurred.

Lastly, there’s the obligatory bug indi-cator LED (a diagnostic LED). There areonly a couple of fatal errors possible inthis application: a watchdog timeout ora reset instruction encounter. Shouldone of these errors occur, the batteryanalyzer will shut down to a failsafestate and then blink an error code onthe diagnostic LED.

BATTERY POWER CONTROLTwo programmable power regulator

circuits are attached to the batteryunder test (see Figure 2 and Figure 3).One circuit is for charging and the other

Photo 2—The battery analyzer’s menu-driven display

Figure 2—The battery analyzer processor, display, and power supply

2011-8-014_Pierce_Layout 1 08/09/11 4:50 PM Page 36

cabate
Text Box
Copyright Circuit Cellar, Inc. This article cannot be reprinted without permission: [email protected]. To subscribe: http://circuitcellar.com/subscription/

41.qxp 1/7/2009 3:07 PM Page 1

38 CIRCUIT CELLAR® • www.circuitcellar.com

Septe

mber

2011

– Is

sue 2

54

Figure 3—The battery analyzer powercontrol circuitry

Save 20% on all titles when you purchase from www.newnespress.com.Enter promotional code Newnes20 at checkout.

Register for our e-news at newnespress.comFind us on Facebook!

By: William Kafi gISBN: 9781856177047

By: Tammy NoergaardISBN: 9780750684552

By: Bruce DouglassISBN: 9781856177078

By: Hazarathaiah MalepatiISBN: 9781856176781

N e w n e s P r e s s

Education Never Ends. Everything you need to know to get started.

2011-8-014_Pierce_Layout 1 08/09/11 4:50 PM Page 38

cabate
Text Box
Copyright Circuit Cellar, Inc. This article cannot be reprinted without permission: [email protected]. To subscribe: http://circuitcellar.com/subscription/

www.circuitcellar.com • CIRCUIT CELLAR® 39

Septe

mber

2011

– Is

sue 2

54

resistor R27 introduces an offset of 50 mA, so the actual con-trollable range is 50 mA to 1.05 A.

Diode D1 blocks current backflow when the battery ana-lyzer is not actively charging the battery. Diode D2 ensuresthat the output of regulator U8 does not exceed about 5.3 V.This is because digital potentiometer U6 could be damagedif the regulator output was allowed to go above the chip’sVCC supply voltage of 5.5 V.

Resistors R26 and R27 prevent excessive currents fromflowing through the digital potentiometers. The LT3080 reg-ulator, like many low-dropout regulators, has a nasty behav-ior characteristic at very low voltages—that is, it startsdrawing a relatively large amount of current. In this case,the current would flow through the Set pin of the regulatorand potentially damage the digital potentiometer chip.

Resistor divider network R1 and R12 adjusts the batteryvoltage to a level that’s convenient for the A/D converter touse, which is 5 mV per A/D count. At least that’s the theo-ry. Actual empirical results showed a need for an additional100-kΩ trim resistor and a couple of small filtering capaci-tors in order to get an accurate voltage reading. ResistorR28 prevents large currents from flowing in through thePIC18F4525’s A/D input pin when the power is off but abattery is present.

THERMAL CONSIDERATIONSI like my circuits to run cool, and in the breadboard test

circuit the regulator chips were getting pretty hot. So, the

is for discharging the battery. Normally, only one of thesecircuits will be engaged at a time. Power MOSFETs are usedto completely shut off the power regulator circuits. TheMOSFETs are driven in such a way that they are normallyoff, so if the microcontroller is not controlling them (as dur-ing a reset) they automatically shut off.

Linear Technology LT3080 voltage regulator chips areused to regulate the battery charge and discharge currents.These voltage regulators are a little different in that thevoltage Set control pin can be actively driven and the out-put voltage is equal to the Set pin voltage. The Set pin alsocan be passively used because it provides a 10-µA sourcecurrent. So, to get a 5-V output, a 500-kΩ resistor from theSet pin to ground is all that’s needed.

For charging, a series current sense resistor (R10) and adigital potentiometer (U6) are used to set the charge cur-rent. The 50-kΩ digital potentiometer in combination withthe 10-µA source current from the LT3080 establishes acontrol voltage range of 0 to 0.5 V. Since the current senseresistor is 0.5 Ω, this equates to a current range of 0 to 1 A.Current-limiting resistor R26 introduces an offset of 50 mA,so the actual controllable range is 50 mA to 1.05 A.

For discharging, a constant load resistor (R2) provides aknown resistance so the load current is a function of the regu-lator (U9) set voltage. For example, the load resistor is 1 Ω, soto get a 1-A load current the regulator set voltage is 1 V. Usinga 100-kΩ digital potentiometer gives a controllable voltage setrange of 0 to 1 V, which equates to 0 to 1 A. Current limiting

Pico Technology Celebrates 20th Year In Test & Measurement

Providing High Performance PC Based Data Acquisition Solutions Since 1991

Pico Technologyhttp://www.picotech.com/[email protected]

Give Us A Call At: 1-800-591-2796

4 Channel oscilloscope with spectrum analyzerBuilt-in arbitrary waveform generatorUART, SPI, I²C and CAN bus decodingAdvanced triggeringMask limit testingFull math and measurement functions

High Speed USB connectionEasy-to-use softwareSDK for major third party applicationsSegmented memoryDigital TriggersFive Year Warranty

World’s Fastest USB Oscilloscope

2011-8-014_Pierce_Layout 1 08/09/11 4:50 PM Page 39

cabate
Text Box
Copyright Circuit Cellar, Inc. This article cannot be reprinted without permission: [email protected]. To subscribe: http://circuitcellar.com/subscription/

thermal design of the final version was definitely overkill.Both of the LT3080 regulator chips are mounted on a 5” × 2”aluminum plate that acts as a heatsink. A couple of con-ventional TO-220 heatsinks were added. As if that wasn’tenough, a small fan is also attached to the plate. The fanblows air through the plate which greatly increases itscooling capacity. The current sense and load resistors arelocated so that the air from the fan blows directly on them,which keeps them cool, too. As aresult, it barely gets warm to thetouch.

For safety, there are a couple ofNTC thermistors that monitor theheatsink and battery temperatures. Ifeither exceeds a modest temperaturelevel, it triggers a safety shutdown.

SELF-TEST FEATUREBy simultaneously turning on both of the power regulator

circuits, the battery analyzer does a little self-testing. A volt-meter attached to the voltage and/or current test points canbe used to independently verify the circuit operation. Thereis a diagnostic menu and a number of self-test choices avail-able to check out the charge current, discharge current, andconstant voltage functions.

SOFTWAREI generally use Assembly language with Microchip

processors in the 10, 12, 16, and 18 families. So, I wrotethis application in 100% Assembly.

There are only two interrupt sources: the A/D converterand the periodic timer. The high- and low-priority interruptsare used to separate the two interrupts. The periodic timerinterrupt kicks off an A/D read sequence every 2.5 ms andperforms human interface functions every 20 ms. An A/Dread sequence is where all of the A/D inputs are read andstored into memory. Most of the A/D values are handled bythe foreground process, except for the battery voltage sensor.

As soon as a new A/D reading comes in from the batteryvoltage sensor it’s immediately processed. The constantvoltage function is actually being performed by software,which means the software is part of the feedback loop. So,it’s important to keep the loop response time as fast as pos-sible to get a stable constant voltage output.

A watchdog timer interlock between the timer interruptand the foreground main loop ensures that the watchdogwill trip if either the interrupt or the foreground processescrash. The foreground main loop takes care of everythingelse. Although the timing is not critical, the foreground isable to do overall precision timekeeping because the timerinterrupt provides accurate “ticks” for the foreground toprocess. These ticks are queued up so that the foregroundcan stall for short periods as long as it eventually catchesup and processes all of the ticks in the queue.

The main foreground tasks are to run the various timers,control the fan, store the milliamp hour data-logging informa-tion, perform time and temperature safety checks, monitor forbattery insertion/removal, update the LCD, and handle button

Septe

mber

2011

– Is

sue 2

54

CIRCUIT CELLAR® • www.circuitcellar.com40

inputs. By doing most of the work in the foreground, it is easi-er to serialize all the processes and avoid potential conflictsthat can happen at the interrupt level. A couple of helperfunctions assist with safely accessing information that theinterrupt handlers update.

The total milliamp hour discharged is an average of thecurrent discharged during the period, which is then adjustedfor a 1-h period. So, the first calculation is average = total

(mA)/time (hours). To adjustfor a 1-h period, the averageis multiplied by the elapsedtime, so mAh = average ×time (hours). Combiningthese two equations, thetime term cancels out so theresulting calculation is mAh= total (mA-s)/3,600 (s). Forexample, if an average of

1,000 mA was discharged over a 2-h period, the total mA-swould be 7,200,000 (1,000 mA every second for 2 hours).Dividing this by 3,600 results in a value of 2,000 mAh.Once a second, the discharge current (in mA units) is addedto a total accumulator. This discharge current isn’t actuallymeasured; it is derived from the discharge current poten-tiometer setting. A 31/23-bit divide routine from theMicrochip Technology application library is used to dividethe total current accumulator by 3,600. The result is themAh value.

The graphic data logging is performed using a dynamiccompression technique. This enables data from short runs(just a few minutes) to long runs (several hours) to be record-ed using the same size memory buffer. When a run is started,data is recorded at a high rate of one sample per second. If thedata buffer fills up, it is compacted (half the data is discarded)and the data recording rate is reduced to one sample per 2 s.This process is repeated until the run is completed. Since themaximum run time is to the order of 8 hours and the datarecording buffer is 100 entries, the lowest sample rate endsup being about one sample per five minutes.

There are a lot of display screen pages, so a data-drivensubsystem is used to handle the display updates. Each dis-play page is built using macros that specify things such asscreen locations, text or variables to display, and the but-ton handler function to use for the page. To display a newpage, all that’s necessary is to set the current page indexnumber to a new value, then the display subsystem takescare of the rest. Displaying variable information is handledby a callback function for each page that enables the vari-able data to be set up just prior to the screen being sent tothe LCD. A special callback function is provided forscreens that need to draw something unique, such as thegraph display in this application. Library modules providesupport for the EEPROM, the LCD, the digital potentiome-ters, and mathematical calculations.

CONSTRUCTION & OPERATIONI constructed the final version as a naked two-layer PCB

mounted on an acrylic base plate. The base plate also ties

“To charge or analyze a battery,all you need to do is insert thebattery and follow the promptson the screen to select thedesired settings.”

2011-8-014_Pierce_Layout 1 08/09/11 4:50 PM Page 40

cabate
Text Box
Copyright Circuit Cellar, Inc. This article cannot be reprinted without permission: [email protected]. To subscribe: http://circuitcellar.com/subscription/

www.circuitcellar.com • CIRCUIT CELLAR® 41

Septe

mber

2011

– Is

sue 2

54

Richard Pierce ([email protected]) is a medical product design engineer whostudied Computer Engineering at the University of Illinois. He also designs homeentertainment/automation gadgets.

PROJECT FILESTo download the code and bootloader files go to ftp://ftp.circuitcellar.com/pub/Circuit_Cellar/2011/254.

SOURCESTO-220 HeatsinksAavid Thermalloy | www.aavidthermalloy.com

LT3080 Voltage regulator chipsLinear Technology Corp. | www.linear.com

PIC18F4525 Microcontroller Microchip Technology, Inc. | www.microchip.com

NTC ThermistorsMurata Manufacturing Co. | www.murata.com

LM317 Voltage regulatorsTexas Instruments, Inc. | www.ti.com

the heatsink backplate to the PCB.The battery holder is a standard18650 type that I modified by trim-ming the plastic holder tabs to makeit easier to remove the battery. I alsocut a small hole in the back of thebattery holder to accommodate theNTC temperature sensor that’s locat-ed directly behind the battery. TheLCD screen is held in place and pro-tected by a small acrylic windowthat’s anchored to the PCB with somedouble-sided tape.

To charge or analyze a battery, allyou need to do is insert the batteryand follow the prompts on the screento select the desired settings.

HAZARDSThis battery analyzer is not well

protected from reverse polarity (abattery that’s inserted backwards).There should be some degree of pro-tection provided by the fuse andreverse diode (D3), but I am too nerv-ous to actually test it. The idea isthat if the battery is inserted back-wards the reverse voltage to the cir-cuit will be minimized and the fusemay blow. But if sufficient reversevoltage makes it into the circuit,then some components will most

likely be damaged.This battery analyzer also won’t

prevent you from doing things to ali-ion battery that you shouldn’t,such as overcharging or overdischarg-ing the battery. So, read the manualfor your li-ion battery and be careful.Not following this advice couldresult in damaging the battery andpossibly causing it to explode. Usinga “protected” battery (one with anintegrated safety circuit) is a goodsafety measure, especially if you arenot familiar with the potential haz-ards of li-ion batteries.

YOU GET WHAT YOU PAY FORBattery ratings seem pretty opti-

mistic. I haven’t found a battery yetthat meets its rated capacity. And Ihave discovered some batteries thatmay be defective or perhaps even coun-terfeit products. Perhaps a battery rat-ing is like the top speed of your car’sspeedometer. On a good day, with thewind behind you and going downhill,you might be able to get the ratedamount of power out of a li-ion battery.But, from now on, I’ll only be buyingmy batteries from reputable sourcesand derating the battery specificationsin my battery-powered designs. I

2011-8-014_Pierce_Layout 1 08/09/11 4:50 PM Page 41

cabate
Text Box
Copyright Circuit Cellar, Inc. This article cannot be reprinted without permission: [email protected]. To subscribe: http://circuitcellar.com/subscription/