Table of Contents
Advertisement
Impact of Stress and Vibration on Current Measurement
A cell's voltage can be altered by mechanical stress and vibration. Any change of the cell's voltage
impacts the self-discharge current. Static stress can induce a relatively fixed voltage shift while
vibrations can create substantial peak-to-peak deviations in the measurement. Take care that you:
Do not subject the cell to stress.
l
isolate the cell from any direct vibration that may exist.
l
Noise Removal Functions
In real world settings, self-discharge current measurement data shows significant noise, thereby
complicating interpretation of the measurement results.
The primary source of measurement noise is temperature-induced cell open-circuit voltage changes,
which translates into a noise component in the self-discharge current measurement. In the simplest
case, there is one common noise source affecting all channels equally. In most cases however, the
noise source may affect individual channels to a different degree and there may also be more than one
noise source.
The Keysight BT2152A/B Self-Discharge Analyzer implements two noise removal functions: noise
removal using median subtraction, and noise removal using median fitting. Refer to the
FETCh:CURRent:MEDian
Additional denoising utilities are available from Keysight. Contact your Keysight representative for
more details. For further information on available noise-removal processing techniques please
reference Application Note p/n 5992-3979EN
Discharge Data
Sets.
Median Subtraction
The common noise source across channels can be extracted by comparing the individual channels and
calculating the median signal at any given time, thereby canceling out the systematic variations of
each curve and preserving the common noise. By subsequently subtracting the median from all
signals eliminates the noise common to all channels, and centers the signals of the good curves
around zero.
Median calculation assumes that majority of the cells in a measurement are good quality cells. If the
bad cells approach 50% of the total, the median may start to fluctuate. Note that for median stability,
it is important that a minimum number of at least eight channels are available for recording.
Median Fitting
To preserve the overall shape of the original curve data, a fit of the median curve can be added back
into the median subtraction result. This effectively discards any high-frequency variation in the
median. The median fitting approach preserves most of the advantages of median subtraction, while
making the curves more visually similar to the raw data.
Keysight BT2152A/B Operating and Service Guide
commands.
Removing Noise in Lithium-Ion Battery Cell Self-
3 Operating the Analyzer
47
Advertisement
Table of Contents
