Single-Cell Gauging 101

1

2

What is Fuel Gauging Technology?

• Fuel Gauging is a technology used to predict battery capacity under all system active and inactive conditions.

• Battery capacity – Percentage– time to empty/full– milliamp-hours– Watt-hours– talk time, idle time, etc.

• Other data can be obtained for battery health and safety diagnostics.• State of Health• Full Charge Capacity

73%Run Time

6:23

3

Outline

• Battery chemistry fundamentals• Classic fuel gauging approaches

– voltage based– coulomb counting

• Impedance Track and its benefits

Single-Cell Gauging 101Part 2: Classic fuel gauging approaches

5

Goal: Full Use of Available Battery Capacity

+

0%

20%

40%

60%

80%

100%

Capacity

Charging Voltage Tolerance

Shutdown Uncertainty due to inaccurate gauging

• Only 80-90% of Available Capacity may actually be used!• High Accuracy Gas Gauge Increases the Battery Run-time

Actual Useful Capacity

6

PACK-

Battery Pack

TSProtector

GasGauge

PACK+

I2C/HDQHost ProcessorTI OMAP

Power Management

GPIO

Traditional Battery Pack-Side Gas Gauge

7

Host PTI OMAP

Power Management

GPIO

Gas Gauge

Portable Devices

PACK-

PACK+

TS

Battery Pack

Protector

System-Side Impedance Track Fuel Gauge

8

What does the Fuel Gauge do?

• Communication between battery and user• Measurement:

– Battery voltage– Charging or discharging current– Temperature

• Provide:– Battery Run Time and Remaining Capacity– Battery health information– Overall battery power management (Operation mode)

9

How to Implement a Fuel Gauge?

• Voltage Based: SOC = f (VBAT)

• Coulomb Counting:

• Impedance Track: Real time resistance measurement

dt iQ

BATOCV RI-VV •=

10

?

Voltage Based Fuel Gauge4.2

3.4

3.0

Bat

tery

Vol

tage

(V)

Open Circuit Voltage(OCV)

EDV

Battery Capacity QmaxQuse

3.8

• Applications: low end cellular phone, DSC,…• Pulsating load causes capacity bar up and down• Accurate ONLY at very low current BATOCV RI-VV

I•RBAT

11

Battery Resistance

Impedance = f( Temperature, State of Charge, and Aging)

Resistance doubles after 100 cycles

10-15% cell-cell resistance variation

10-15% resistance variation from different manufacturers

?BATOCV RI-VV •=

12

Impedance is strongly

dependent on temperature,

State of Charge and aging

Impedance Dependent on Temperature and DOD

Fully DischargedFull Charged

DOD=1-SOC (State of Charge)SOC=1 (Full charged battery)SOC=0 (Full discharged battery)

SOC: State of ChargeDOD: Depth of Discharge

maxQ

QSOC =

13

Impedance Differences for New Cells

• Low-frequency (1 mHz) impedance variation 15%

• At 1C rate discharge, 40-mV difference, causes maximum SOC error of ±26%

0 0.062 0.084 0.11 0.13 0.15R(Z) -

- Im

(Z) -

0.05

0.025

0

1 mHz

1 kHz

Manufacturer 1

1 mHz

1 kHz

0 0.042 0.064 0.086 0.11 0.13R(Z) -

- Im

(Z) -

0.05

0.025

0

Manufacturer 2L C1

RSER

Rhf R1 R2

C2

14

*C/3 rate current used for both tests

• Complete relaxation takes about 2000 seconds

• Different voltage at different instants

• Voltage difference between 20 and 3000 seconds is over 20 mV

Battery – Transient Response

3.830

3.855

3.880

3.905

0 1000 2000 3000 4000 Time (Second)

Bat

tery

Vol

tage

(V)

L C1RSER

Rhf R1 R2

C2

0 500 1000 1500 2000 2500 3000 3500Time (Second)

3.250

3.275

3.300

3.325

Bat

tery

Vol

tage

(V)

Load Removal

Load Removal

15

Voltage Relaxation and State of Charge Error

• ±20mV difference

• Error depends on particular voltage at the moment of estimation

• Maximum error reaches 15%, average error 5%

3.2 3.45 3.7 3.95 4.3

100

50

0

-50

SOC %

Battery Cell Voltage (V)

SOC % Error

3.2 3.45 3.7 3.95 4.3Battery Cell Voltage (V)

20

10

0

-10

-20

16

SOC Error of Voltage-Based Fuel Gauging

Error for a New Cell Error Evolution with AgingS

OC

Err

or

– %

20

15

13.13

11.25

9.38

7.5

5.63

3.75

1.88

0

40 60 80 1000

SOC – %

Relaxation ErrorCell-to-Cell VariationTotal Error

15

11.25

7.5

3.75

0

0 20 40 60 80 100 SOC %

SOC

Err

or %

• 20-mV relaxation measurement error• 15% cell-to-cell resistance tolerance• Battery resistance doubles every 100 cycles

SO

C E

rro

r –

%

20

100

87.5

75

62.5

50

37.5

25

12.5

0

40 60 80 1000SOC – %

0 Cycles

300 Cycles

100 Cycles200 Cycles

SOC

Err

or %

0 20 40 60 80 100 SOC %

100

75

50

25

0

?OCV BAT•V = V - I R

17

• Advantages– Learning can occur without full discharge– No correction needed for self-discharge– Very accurate with small load current

• Disadvantages– Inaccurate due to internal battery impedance– Impedance is function of temperature, aging,

and State of Charge

Voltage-Based Fuel Gauge

18

• Battery is fully charged

• During discharge capacity is integrated

• Qmax is updated every time full discharge occurs

dt iQ

Coulomb Counting Based Gauging

EDV: End of Discharge Voltage

Q

0 1 2 3 4 5 6

3.0

3.5

4.0

4.5

Capacity, Ah

Li-Ion Battery Cell Voltage

EDV

Qmax

0.2C Discharge Rate

Example: bq27010, bq27210

19

Learning before Fully Discharged

• Too late to learn when 0% capacity is reached

• Set voltage threshold for given percentage of remaining capacity

• True voltage at 7%, 3% remaining capacity depends on current, temperature, and impedance

0 1 2 3 4 5 6

3

3.5

4

4.5

Volta

ge (V

)

EDV0 0%

3%

7%EDV1

EDV2

Capacity Q (Ah)

20

0 1 2 3 4 5 6

3

3.5

4

4.5

Volta

ge (V

)

7%CEDV2 (I1)

Capacity Q (Ah)

Compensated End of Discharge Voltage (CEDV)

CEDV = OCV(T,SOC) - I*R(T,SOC)

30%CEDV2 (I2) 7%

• Modeling: R(SOC,T), good for new battery

• Calculate CEDV2 (7%) and CEDV1 (3%) threshold at any I and T.

• Not Accurate for Aged battery

21

Coulomb Counting Based Gauging

?BATOCV RI-VV •=

Key Parameter related to Aging: Impedance

Advantages• Not influenced by distortions of voltage measurement• Accuracy is defined by current integration hardware• Gauging error: 3-10% depending on operation conditions and usageDisadvantages• Learning cycle needed to update Qmax

– Battery capacity degradation with aging • Qmax Reduction: 3-5% with 100-cycles

– Gauging error increases 1% for every 10-cycles without learning

• Self-discharge has to be modeled: Not accurate

22

• Very accurate gauging from OCV without load (relaxation)

Advantages for typical gas gauges

4.2

3.6

3.0

2.4

Bat

tery

Vol

tage

(V) Open Circuit Voltage(OCV)

EDV

Battery Capacity Qmax

I•RBAT

Quse

• Very accurate gauging with Coulomb Counting with load

23

Issue Review

4.2

3.6

3.0

2.4

Bat

tery

Vol

tage

(V) Open Circuit Voltage(OCV)

EDV

Battery Capacity Qmax

I•RBAT

Quse

• Voltage Based gas Gauge: V = OCV(T,SOC) - IR(T,SOC, Aging)

• Current integration gas gauge: CEDV = OCV(T,SOC) - IR(T,SOC, Aging)

Problem: Battery Impedance

Finish

25

Back Up Slides:Impedance Track Reference

26

Single Cell Impedance Track (IT) Basic Terminology and Relationships

• OCV – Open Circuit Voltage• Qmax – Maximum battery chemical capacity

(SOC1/SOC2 is correlated from OCV table after OCV1/OCV2 measurement)• SOC – State of Charge

• (* From Full Charge State)• RM – Remaining Capacity

• (SOC start is present SOC, SOC final is SOC at system terminate voltage)

|SOC|SOC

PassedQQ

21max -

=

max

*

Q

PassedQ1SOC -=

maxfinalstart Q)SOCSOC(RM ×= -

27

• FCC – Full Charge Capacity is the amount of charge passed from a fully charged state until the system terminate voltage is reached at a given discharge rate

• FCC = Qstart + PassedQ + RM

• RSOC – Relative State of Charge

FCC

100RMRSOC

Single Cell Impedance Track (IT) Basic Terminology and Relationships

28

Single Cell Impedance Track (IT) Fuel Gauge Introduction

29

Gauging Error definition

• Reference points– at charge termination SOC =

100%– at EDV SOC=0– Charge integrated from fully

charged to EDV is FCCtrue

• From these reference points, true SOC can be defined as

SOCtrue= (FCCtrue-Q)/FCCtrue

• Reported SOC at all other points can be compared with true SOC.

• Difference between reported and true SOC is the error. It can be defined at different check points during discharge.

15%

3%

0%

Error check-points

0 1 2 3 4 5 6

3

3.5

4

4.5

Capacity, Ah

Vol

tage

, V

EDV

Qmax

Q

FCC

Check point at 0% is not meaningful – EDV is the voltage where system crashes!

30

Single Cell Impedance Track (IT) Error Definition and Calculation

• Relative State of Charge (RSOC) Error

RSOC Error = RSOC calculated - RSOC reported

(RSOC reported is the RSOC reported by bq275xx Impedance Track TM algorithm)

100

FCC

PassedQQFCCRSOC start

calculated

31

• Remaining Capacity (RM) Error

RM calculated = FCC - Qstart - PassedQ

(RM reported is the RM reported by bq275xx Impedance Track TM algorithm)

FCC

RMRMErrorRM reportedcalculated

Single Cell Impedance Track (IT) Error Definition and Calculation

32

Example error plotsTrue vs reported RSOC RSOC error

0 20 40 60 80 1008.5

9

9.5

10

10.5

11

11.5

12

SMB RSOCtrue RSOC

RSOC

Vol

tage

0 10 20 30 40 50 60 70 80 90 1004

3

2

1

0

1

2

SMB RSOC

RSOC

RS

OC

err

or

Remaining capacity test Relative RemCap error

0 1 2 3 4 58.5

9

9.5

10

10.5

11

11.5

12

SMB remaining capacitytrue remaining capacity

Capacity, mAh

Vol

tage

0 12.5 25 37.5 50 62.5 75 87.5 1003

2.13

1.25

0.38

0.5

1.38

2.25

3.13

4

SMB RSOC

SOC

Rel

ativ

e R

emC

ap e

rror

, %