6 Golf Cart Batteries

The current battery bank was installed in October of 1999 prior to leaving for Kwajalein and was primarily charged from wind power or the engine from December of 1999 until November of 2001. The charge rate was limited when on shore power until July 2004 due to minor fan issues with the inverter, while the batteries would get charged, the rate was a bit low (20 amps instead of 100).

The current specific gravity values for the charged battery bank:

Each of the colors in the table represent a 6V battery with three cells. Overall the variation of specific gravity is not too bad except for the forward pair of batteries. The most forward cell in the outboard battery is 1.190 while the most forward cell of the inboard battery is 1.260. The desire is that all cells be within 0.050 of each other. The inboard and outboard pairs of batteries are wired in series to create a 12V battery and then the three pairs are wired in parallel. The positive terminal of the bank is on the inboard-aft end of the bank, the negative terminal of the bank is on the outboard-forward end of the bank. This wiring scheme balances the resistance of the interconnection wires.

The current specific gravity values for the charged battery bank after two inverter/charger based equalization cycles (6amps continuous for 8 hours voltage not to exceed 16.0V or about 1% of capacity) The bank absorbed 61AH during the cycle:

The delta changes from before equalization are

From these data we can see that the very worst cell (outboard-forward) had the largest increase in specific gravity from 1.190 to 1.265 for an increase of 0.075 as compared to the more nominal value of about 0.030. Prior to equalization the range of specific gravity values was 0.070, after equalization the range was reduced to 0.025. The minimum value after equalization was 1.250 and the maximum was 1.275. From the chart below these data imply that the lowest state of charge of a cell in the bank is approximately 90% with the average being 96%. 

I am running an additional inverter based equalization cycle to see if there is any change. Two additional cycles have merit, the first is to use the engine alternator to equalize the bank, using the link 2000R controller the equalization cycle applies 26.4 amps (4% of the declared 660 amp-hour capacity) until the voltage reaches 16.3. It is unclear if the cycle is terminated at this point or if the voltage is held at 16.3 with tapering current. The other method would be to separate the bank into three smaller banks of 220 amp-hours each and use the inverter to perform a "proper" equalization cycle, at 6 amps applied the current would be 2.7% of the battery capacity instead of 1%. At 1% the voltage appears to never rise significantly above 14.5 volts resulting in essentially an extended absorption mode charge. While as we see above this significantly improves battery performance it will be interesting to see if a more aggressive equalization cycle has significant benefit.

The current specific gravity values for the charged battery bank after two inverter/charger based equalization cycles (6amps continuous for 8 hours voltage not to exceed 16.0V or about 1% of capacity) The bank absorbed 61AH during the cycle:

The delta changes from before this equalization cycle are

The batteries are improving dramatically, so another inverter based equalization cycle which absorbed 50 amp-hours results in the following readings:

The delta changes from before this equalization cycle are

From these data we can see that the average is 1.272 with a minimum of 1.265 and a maximum of 1.275 for a range of 0.010. The overall range for groups of 6 cells combined were approximately 0.015, suggesting a reasonably well balanced bank. Some of the cells consumed as much as 1cup of water during this process. All of the cells were topped up to half way up the splash guards inside the cells.

Note that the inverter based cycle is limited to 6 amps of continuous current, this only gives about 14.5V as the peak charging voltage during equalization. If the alternator on the engine is used to equalize the bank then 26.4 amps are applied continuously which would have brought this bank into equalization much more rapidly. The moral of the story is make sure that the equalization cycle for your charging system takes into account the total capacity of the battery bank or it will take a bit of work to get the bank equalized, however, if a constant low current which can bring the batteries up to the acceptance voltage and hold it here for several hours is all that is available it can be made to work.

To complete the battery testing I am applying a constant load of 30 amps on the battery bank, this should require approximately 22 hours to completely deplete the bank. I am applying the constant load from the inverter as it has a 10.5V cutout to prevent battery damage. The amp-hour monitoring system will then provide an excellent indication of battery capacity at the end of this discharge cycle, I will measure the specific gravity of each cell at the end of the discharge cycle. I will then charge the batteries with the Freedom 20 inverter (100 amps), verify the charging efficiency, and measure the specific gravity of each cell. Finally I will equalize the bank one last time and measure the specific gravity one last time to verify the state of the bank. 

The load testing started at 7am on 7/28/04. At 11:30pm on 7/28/04 the bank was still being discharged at approximately 26 amps, the voltage was 11.60 and 383 amp-hours had been consumed. At 6am on 7/29/04 the bank was still being discharged at 26 amps, the voltage was 11.15 and 552 amp-hours had been consumed.

Note that the above voltages are NOT the resting voltage but rather the voltage while being discharged and thus are different from the chart below, however, the specific gravity should be approximately correct.

Overall these partially discharged values show a battery bank coming down evenly without any cells struggling under the load. With a range of values of 0.020 and the six cell values ranging by only 0.015 the bank appears to be performing evenly. The following numbers represent the following deltas compared to the fully charged bank.

The range of deltas varies by as much as 0.025 between the cell which changed the least and the cell which changed the most. Note that the average specific gravity in the previous table (1.095) is below the fully discharged level in the specific gravity chart. The above is near the end of a deep discharge cycle and the cells have not yet come to equilibrium. Also, water was added at the beginning of the cycle which may not have diffused completely into the electrolyte. The inverter will auto cut-off at 10.5 volts and then will be recharged fully. 

Inverter shutdown occurred at approximately 10am on the 29th or about 27 hours after starting the discharge cycle. The battery bank provided a total of 625 amp hours before the inverter shut down at 10.5 volts. Two hours later (when I was first able to monitor it) the voltage had increased to 11.35V. I attempted to measure the specific gravity of the cells but found that every cell was off the low end of the scale, all were approximately equal and an estimate of the average specific gravity would be 1.075 grams/cc. (Water has a specific gravity of 1 gram/cc making conversions to a ratio of water easy!). The temperature of the electrolyte was warmer than the ambient but I did not have a thermometer available to determine the value. As the specific gravity changes by 0.003 per degree F (lower at higher temperature) and we assume a specific gravity of 1.120 is correct at 70 degrees F this would give a difference of 0.045 or a temperature of 70+0.045/0.003 = 85F which would be consistent with slightly warm to the touch electrolyte.  Of more interest is that all of the cells had almost exactly the same reading providing an indication that the bank remains roughly even even when deeply discharged. With additional rest time I would expect the bank to settle in with a voltage of approximately 11.76 (the expected discharged voltage of the cells).

The charging process ran for 8.5 hours and transitioned to the float mode from acceptance mode while charging at 54 amps suggesting a timeout although the total amount of charge read +90 amp-hours. (I happened to be watching as it transitioned to float) suggesting a timeout and so I reinitiated the charge cycle and it returned to charging at 52 amps after starting in the bulk mode for a few seconds. The bank absorbed an additional 293 amp hours! until it transitioned to float again at 7 amps. This suggests that for deep discharge the efficiency of 98% was considerably too high although it may be reasonably accurate for more reasonable charge cycles. Also note that the electrolyte temperature was significantly higher than the set point (72F) due to the protracted charging cycle. After settling into the float mode for several hours the electrolyte specific gravity was as follows:

These fully recharged values (even though the cells are hot i.e. the electrolyte in the hydrometer felt significantly warm to the touch) show an evenly recharged system, all of the values are within 0.010 of each other.

The above table gives the deltas from the fully discharged case to the fully charged case and show that all of the cells changed by a similar amount, the largest to the smallest was 0.025 and with the single case of 0.175 thrown out the cells all are within 0.020 of each other in their delta change. The final data point is the cold cell specific gravity taken after the bank was allowed to cool for 10 hours with no load and no charging.

The above values give the cooled specific gravity of the charged bank 

The average 72F specific gravity for the cooled cells was 1.272, the the average specific gravity of the hot cells was 1.254 for a delta of 0.018 at 0.003 per degree F this would suggest that the cells were 6 degrees F hotter at the end of the charge cycle than after cooling to the ambient temperature of 72F for a combined temperature of 78 degrees F which seems a bit cooler than I would have thought. Perhaps the initial temperature was really not a cold temperature...

After a final equalization cycle (inverter based, 6 amps for 8 hours) brought all of the cells up to a very even level:

The variation from the highest to lowest values is 0.010 and the charge level is at minimum 1.265, basically the same values as when the batteries were new!

The deltas for an equalization were about 0.010 and brought the cells to an even level.

The spreadsheet for calculating the deltas is available here.

The following chart gives the relationship between the percentage of charge (state of charge), the specific gravity, and the resting voltage for 70 degrees F. Note that various battery manufacturers will have slightly different numbers based upon the specific chemistry of their batteries and the specific gravity and voltage vary with the temperature when compared to the state of charge.

The recommendation is that all cells in a bank of batteries have a specific gravity within 0.050 of each other. i.e. the worst cell to the best cell should not differ by more than 0.050 to prevent self discharge and loading up a single cell more than the others and inducing failures.

Update February 10, 2005: Accidentally completely discharged house bank! Turned off the charger while working on some wiring and forgot to turn it back on. At least I got a complete discharge cycle and the amp-hour meter determined that 651 AH had been removed from the batteries at about 10.2 volts. Recharge cycle returned to approximately 0 (-7AH) suggesting that the calculation for efficiency is not too far off.

I will finish up this weekend with a equalization cycle and check of cell specific gravity to verify the status of the house bank before the start of the sailing season.

Other links about batteries:
     vonwentzel: Discussions of bank sizing and battery types
     Battery Maintenance: (1922)

Cached Articles
     Battery Voltage as an indicator of State of Charge (SOC)