End-Of-Service Determination
26
SLUUBE8 – September 2018
Copyright © 2018, Texas Instruments Incorporated
General Description
1.3.3.5
Learn Post-Relax Phase
After the Learn Discharge Phase completes, the bq34210-Q1 device will disable the learning current and
enter the Learn Post-Relax Phase, whereby the device again waits for relaxation of the cell, during which
Voltage()
is monitored for 100-s time windows. While in this phase, the device sets the
[LRLX]
flag. If the
change in the consecutive 100-s averages of
Voltage()
is less than 4 µV, then relaxation is deemed
achieved. If not, the device continues monitoring for another 100-s time window. It is important that the cell
be fully relaxed before the algorithm can continue. The status of relaxation can be seen by the
[REST]
bit
in the Gauging Status register. If a learning fault occurs while in the Learn Post-Relax Phase, the status
flag (
[LRLX]
in this case) remains set to enable the host to understand what mode the system was in
when the fault occurred.
As in the Learn Pre-Relax Phase, if a discharge current in excess of
Discharge Detection Threshold
or
a charging current in excess of
Charge Detection Threshold
is detected, then the
[LUCD]
flag is set.
This detection also terminates the learning phase, and this flag will remain set so the host can recognize
why the learning phase was stopped. This bit will be reset to 0 when charging is terminated or the register
is read.
At the completion of the Learn Post-Relax Phase, if in CHARGE-BEFORE-DISCHARGE mode and
Voltage()
> the charging voltage determined by the selected charging algorithm and
[LVR]
= 1, then the
device will set
[LDSG]
and enable the
Learn Discharge Current
until
Voltage()
≤
the calculated charging
voltage. At which point,
[LDSG]
is reset and the
Learn Discharge Current
is disabled. The
[LRSTOR]
bit
will be set while the device is continuing discharge to reach the calculated charging voltage.
If in DISCHARGE-BEFORE-CHARGE mode and
[LVR]
= 0, then the device will set
[LDONE]
and
complete learning. If
[LVR]
= 1, then the device will initiate a new charging session to charge the battery
back to the appropriate charging voltage determined by the selected charging algorithm using a similar
sequence as described above. In this case, the device sets the
[LCHG]
and
[LRSTOR]
flags to indicate to
the host that charging can begin. If charging does not terminate within
Learn Charge Time Limit
, then
[LCTO]
and
[LFAULT]
is set. If a learning fault does occur, the status flag (
[LCHG]
in this case) will remain
set to allow the host to understand what mode the system was in when the fault occurred.
NOTE:
It is possible that a Learn Discharge Phase has completed, and a new
Rcell
value was
calculated and stored, but during a Voltage Restore phase, a fault could occur. This would
cause
[LFAULT]
to be set, and the device would retry a new learning phase after
Auto
Learn Retry Time
.
1.3.3.6
Cell Resistance Calculation
The device uses the values obtained during each learning phase to estimate a value of
Rcell
for the
battery. Note the actual value of this
Rcell
estimate is not critical here, rather its change as the battery
ages is what is important. The
[LRES]
bit indicates when a new
Rcell
value was acquired and stored by
the device.
The change in value of
Rcell
must be normalized relative to differences in temperature when they are
captured. Therefore, the bq34210-Q1 device uses resistance temperature parameters in
Rcell High &
Low Temperature Coefficients
to calculate an expected value of resistance at the
Learn Target
Temperature
.
The
Rcell High & Low Temperature Coefficients
are obtained during a preproduction test whereby the
customer initiates
Initial Rcell
calculations at temperatures of 5°C, 25°C, and 40°C. These
Initial Rcell
values are put into a spreadsheet provided by TI, which then calculates the values of
Rcell High
Temperature Coefficient
and
Rcell Low Temperature Coefficient
. These values are then loaded into
data memory and used by the device to normalize effective
Rcell
measurements back to an equivalent
value at the
Learn Target Temperature
.
