Battery testing
Posted: Thu Nov 01, 2012 9:29 pm
I am working on a conversion (a 2007 Civic) and have been doing some automated testing on 8 cell packs. The first two test series used Headway 38120S cells, and I am in the process of upgrading the rig (more amps) to repeat the test with 8 of the new CALB 40 cells.
The main objective of the testing was to establish whether or not these cells would drift apart in voltage and state of charge after multiple charge/discharge cycles as Lead -Acid cells in a pack do. Here is a typical results for cycle 500 in the first test series:
Cells 6 failed about 70 cycles later, terminating the first test series, and has been omitted from this chart. Anecdotally, Headway cells do seem to suffer from cell failures
There are videos and data available as follows:
A video introduction to the testing (mostly for a mixed audience): http://www.youtube.com/watch?v=U7-WHUTLbHY
Describing the test rig:http://www.youtube.com/watch?v=23kubKKcsiU
Results from test series 1 (two parts): http://youtu.be/KuVIfSimXQc and http://youtu.be/X9W80CS7q6A
A summary of the test data as zipped Excel files is available here: http://tovey-books.co.uk/attachments/File/Test_1_Summary_Data.zip
The raw data (warts and all): http://tovey-books.co.uk/attachments/File/Test_1_Raw_Data.zip (this is a set of CSV files with one row every 30 seconds for a couple of months or so: quite extensive).
The smoking gun: the cells were bottom balanced on assembly and were cycled with no subsequent cell balancing (automated or manual). The minimum cell voltage variance (the difference between the highest cell and the lowest) on each cycle did not change:
Cell 6 was omitted from this graph because this was the cell that subsequently failed
My conclusion is that for these cells (and therefore probably for most LiFePO4 cells), no cell drift occurs. If that is the case then any kind of automated shunting or balancing is unlikely to improve the situation unless you have unbalanced parasitic loads of the kind that you might get from voltage monitors or other instrumentation.
You are most welcome to re-use, re-purpose or re-publish the data but I would ask two things: firstly please don't change or amend the data for re-publishing unless you make what you have done abundantly clear, and explain the reasons why. Secondly an acknowledgement and a link back to http://tovey-books.co.uk would be nice (but not essential)
The main objective of the testing was to establish whether or not these cells would drift apart in voltage and state of charge after multiple charge/discharge cycles as Lead -Acid cells in a pack do. Here is a typical results for cycle 500 in the first test series:
Cells 6 failed about 70 cycles later, terminating the first test series, and has been omitted from this chart. Anecdotally, Headway cells do seem to suffer from cell failures
There are videos and data available as follows:
A video introduction to the testing (mostly for a mixed audience): http://www.youtube.com/watch?v=U7-WHUTLbHY
Describing the test rig:http://www.youtube.com/watch?v=23kubKKcsiU
Results from test series 1 (two parts): http://youtu.be/KuVIfSimXQc and http://youtu.be/X9W80CS7q6A
A summary of the test data as zipped Excel files is available here: http://tovey-books.co.uk/attachments/File/Test_1_Summary_Data.zip
The raw data (warts and all): http://tovey-books.co.uk/attachments/File/Test_1_Raw_Data.zip (this is a set of CSV files with one row every 30 seconds for a couple of months or so: quite extensive).
The smoking gun: the cells were bottom balanced on assembly and were cycled with no subsequent cell balancing (automated or manual). The minimum cell voltage variance (the difference between the highest cell and the lowest) on each cycle did not change:
Cell 6 was omitted from this graph because this was the cell that subsequently failed
My conclusion is that for these cells (and therefore probably for most LiFePO4 cells), no cell drift occurs. If that is the case then any kind of automated shunting or balancing is unlikely to improve the situation unless you have unbalanced parasitic loads of the kind that you might get from voltage monitors or other instrumentation.
You are most welcome to re-use, re-purpose or re-publish the data but I would ask two things: firstly please don't change or amend the data for re-publishing unless you make what you have done abundantly clear, and explain the reasons why. Secondly an acknowledgement and a link back to http://tovey-books.co.uk would be nice (but not essential)