The Empirical Battery Test

By Jim Shell

When they were new, the four Rayovac 6-volt golf-cart (GC2) batteries on Phantom, our Pearson 365 ketch, had plenty of electrical capacity to provide all the power we needed to go three or more days between recharging, perfect for the kind of local cruising we enjoy. As the batteries reached the 5-year-old mark, I wondered whether they still had what it takes, especially given that our need for power consumption is probably greater than it was a half decade ago. How could I determine their capacity from a fully charged state?

The guy at the battery store told me to bring them in. “We can simulate the quick load from an engine starter and measure the cold-cranking amps and the internal resistance. But, frankly, you’re better off replacing them as five years is about the limit on how long these batteries last.”

I thanked him and left. I knew enough to know that his test was relevant for measuring the health of a car battery (by simulating the starting demands placed on it), but not so relevant to measuring the health of my house bank. The electrical demands placed on my house bank are very different.

I needed something that would tell me whether our batteries retained enough capacity to keep up with our electrical demands during the 2- to 3-day cruises (unplugged) we enjoy taking during the season. What I really needed was a sustained-load test over a 20-hour period. This kind of test would accurately measure my battery bank’s capacity. But these batteries are too heavy to lug around to be tested!

Then I came up with the idea of performing a “real-life” load test at the dock. It wouldn’t be the standardized controlled 20-hour load test, but I would unplug for 40 hours and over that time period, place the same loads on the batteries as we would during life at anchor — a custom capacity test, if you will. This was designed to replicate our usage, which, as empirical data, I think is more relevant than an estimate we’d derive from some capacity value, however precise.

The batteries were fully charged when I unplugged them. I noted the open-circuit voltage at that time as 12.8 volts (fully charged). An hour later, without my putting any load on the batteries, the voltage had dropped to 12.58 volts (87 percent of full charge). This “resting open-circuit voltage” measurement is a more accurate reflection of the batteries’ state (and general health) than the 12.8-volt measurement I got right after unplugging.*

I then began re-creating the electrical loads we’d place on Phantom during life at anchor, using refrigeration, the propane solenoid, lights, VHF radio, microwave (via an invertor), and fans as normal. My goal was to determine the battery-voltage drop over 40-plus hours under expected normal loads, and then extrapolate from this number to estimate the expected drop after 64 hours (closer to the amount of time we might normally spend unplugged). I wasn’t trying to quantify the capacity of the battery, but simply determine whether the capacity would be sufficient.

After 41 hours of imposing real-life loads, and with a trickle of charging power coming during daylight hours from our 20W solar panel, the house bank voltage measured 12.31 volts, approximately 66 percent of full charge. The slow, even discharge rate recorded indicates the batteries have good electrical capacity for our purposes. I can extrapolate that another 24 hours of use would drop the voltage another 12 to15 percent, to around 50 percent of full charge (about as low as we’d want to discharge, for the sake of battery longevity). We will do this custom capacity test as part of our yearly maintenance regimen. This is a stress test that truly indicates whether the batteries are good enough for our intended use. Batteries will not last forever, but with this test in our arsenal, we will neither go on cruises with inadequate batteries nor replace good batteries prematurely.

Jim Shell and his wife, Barbara, sail Phantom, their Pearson 365 ketch, off the coast of Texas.

* Editor’s footnote: One hour is probably not a long-enough resting period to get an accurate measurement of a battery’s open-circuit voltage. Battery manufacturer LifeLine recommends a four-hour resting period for an accurate voltage measurement. Trojan recommends six hours. Additionally, when measuring resting open-circuit voltage, it’s important to be sure there is no draw (or charge, in the case of solar panels) on the batteries. The best way to insure this is to disconnect at least one set of battery cables, but this can be avoided if the batteries are isolated.

Additionally, the values in this article, and in the accompanying voltage chart from Trojan, are accurate only for new batteries. A battery that has been in service for a few years is likely unable to maintain these voltages, and so they are not representative. Jim learned empirically that the voltage of his fully charged battery is somewhere south of 12.58 volts, for example. Further complicating things is the fact that older batteries are likely to suffer to some degree from sulfation on the plates. One characteristic of sulfation is that the open-circuit voltage may appear pretty good, not reflecting even severely diminished capacity that sulfation can cause. For this reason, Jim’s patient approach to determining that his batteries do have the capacity to meet his needs, is smart. And not drawing them down more than 50 percent is a practice that should help them outlive expectations. –Eds.




Join Our Community of Sailors. Subscribe Now!

Complimentary monthly supplement

  • featured articles
  • news from the helm
  • mail buoy
  • book reviews
  • sailor photos
  • and much more

Good Old Boat News Flash!

Our website is getting some long overdue improvements! will soon be merged with

Thanks for your patience while our website is under construction.
Menu Title