We are using BQ76952 in a 16S configuration and testing for voltage accuracy during cell balancing. We know that the AFE runs voltage measurements while balancing by turning off the balancing operation of the cell and its adjacent cells for a certain duration. For example, if Cell 2 is to measured, the AFE will turn OFF the balancing for Cell 1 and Cell
- we have practically observed an error of over 70mV in voltage measurement while the balancing operation is active. Is this behavior expected and if not what can be done to increase accuracy here.
- we have considered the possibility that the cell filter RC time constant might affect this (though we have used TI recommended values) . what is the time interval between stopping balancing operation and voltage measurement of the measured cell and the adjacent cells?
Design notes: we have used 100 ohms (TI recommendation) and a differential mode Cap of 1uF on the cell sensing lines. Circuit snippet below.
Below attached are the results and circuit diagram.
Unit: mV |
No Balancing |
Even cell Balancing(C2,C4,...C16) |
||||
V-Actual cell |
V-AFE |
Diff |
V-Actual cell |
V-AFE |
Diff |
|
C1 |
3802 |
3798 |
4 |
3792 |
3813 |
-21 |
C2 |
3906 |
3905 |
1 |
3885 |
3877 |
8 |
C3 |
3842 |
3841 |
1 |
3833 |
3894 |
-61 |
C4 |
3903 |
3902 |
1 |
3883 |
3887 |
-4 |
C5 |
3812 |
3811 |
1 |
3802 |
3862 |
-60 |
C6 |
3576 |
3575 |
1 |
3550 |
3554 |
-4 |
C7 |
3863 |
3862 |
1 |
3855 |
3916 |
-61 |
C8 |
3920 |
3919 |
1 |
3891 |
3893 |
-2 |
C9 |
3688 |
3688 |
0 |
3677 |
3749 |
-72 |
C10 |
3929 |
3929 |
0 |
3900 |
3905 |
-5 |
C11 |
3688 |
3688 |
0 |
3678 |
3742 |
-64 |
C12 |
3779 |
3778 |
1 |
3756 |
3760 |
-4 |
C13 |
3674 |
3675 |
-1 |
3662 |
3727 |
-65 |
C14 |
3832 |
3833 |
-1 |
3811 |
3816 |
-5 |
C15 |
3888 |
3889 |
-1 |
3881 |
3938 |
-57 |
C16 |
3679 |
3677 |
2 |
3661 |
3655 |
6 |