As a Birthday celebration, we went to Chef Todd's with Lloyd, Barbara (SolMates) and Dave (Fawkes). Chocolate Torte Cake! Dinner with friends. Joy.

Now for the concerns. On Friday, the house battery voltage was 10V. 10V is—effectively—stone dead. A level of discharge from which batteries rarely recover.

On Labor Day, Monday, one of the labors was to being the installation of the new digital ammeter. That required inserting the ammeter shunt into the battery returns. It's a bulky and exasperating (but simple) movement of a few wires. Instead of ground-to-battery, it's ground-to-shunt and shunt-to-battery.

What killed our batteries?

The shunt? It's a high-amperage resistor that measures about 0.01Ω or maybe even less. While it is the most obvious major change to the boat's systems, it's a benign change.

The smarter explanation is that something on the boat was left running.

Except. The odds of that are vanishingly small. We're very disciplined about turning all of the circuit breakers off except for two: the bilge pump and the panel meters themselves. We test the bilge pump and reset the counter when we leave. That was the way I found the panel on Friday night: all off except for pump and meters.

"Elementary" (Holmes said). "How often have I said to you that when you have eliminated the impossible, whatever remains, however improbable, must be the truth?" What's left running must have been the bilge pump.

Except. We have a cycle counter on the bilge pump . It was firmly displaying zero. Voltage was low, but not too low for this simple circuit to continue working.

The Vampire Question

Time to look at the vast gulf of "other battery drains". Trot out the trusty copy of Nigel Calder's Boatowner's Mechanical and Electrical Manual. [You do have this, don't you?] Start with the diagnostic procedure for locating battery drains.

1. Turn off everything. Everything. All the circuit breakers. Plus the main battery switch itself.

2. Disconnect the positive lead from the battery. Check the voltage from battery to wire. 11.8V. After attempting to charge the batteries, 11.8V is the expected potential between battery and boat.

3. Check resistance between the positive wire and the negative terminal (which includes the new shunt). In effect, measure the resistance of the boat as a whole. I get about 700,000Ω resistance. If this was a very small resistance, that would mean a dangerous short circuit. If this was a medium resistance, that wound indicate a device left running. This kind of super high resistance is just stray current.

4. Since it's not a short circuit, we can safely check current flow between positive battery post and positive wire. I get 0.04mA.

What do we have? Great resistance. Microscopic current flow. Bottom line? There's no uncontrolled drain.

From what I read on the web, jamming copper and zinc nails into a potato can make about 0.5mA of current; 10 times what's leaking from the batteries. Note. Red Ranger has bronze and zinc fittings floating in salt water. That might create 0.04mA of current.

Checking Our Work

I need to to confirm that I'm actually doing this right. At $150/battery, I don't want to buy new ones and kill them, too.

Let's switch the house bank back on. The entire breaker panel is off, so the results with the battery switch on should match the results with the battery switch off. Right?

They don't. With the switch on, I see only 80KΩ of load and a jump to 25mA of current. A 625× jump in current flow must mean something. We've found an active device! Joy! I found something that's pulling down current from the batteries.

What uses 25mA of power? Concern. It's something really small. A pump uses 5-10A. A lightbulb uses 1-2A.

What I left on is so small that it's irrelevant: that device would take 18,000 hours (about 2 years) to kill the batteries. The boat was unattended from Monday night to Friday night, about 96 hours. A device that uses an average of 4.7A would be required to suck down all 450AH. That's all the navigation lights at once. Or all the interior lights left on. Not one or two lights in a hanging locker. A single 15W interior light would take two weeks to drain the batteries.

A pump (running for about 40 hours) could drain the batteries. But the water pump was not burned to a crisp from the low voltages (and the tanks are still at expected levels.) The bilge pump counter was zero. The macerator would merely trip it's breaker because it's jammed with poo.

Let's check the panel again. A more careful examination shows that the new meters were left on. Okay. That's about right. They're specified to use 0.6W ("Variable with voltage, display intensity, segments illuminated, and sleep mode"). This 50mA spec is close enough to the actual 25mA measurement that it doesn't matter. And when I switch the meters themselves off, the current draw drops back to the 0.04mA of negligible trickle.

I try this with the bilge pump switch set to on. When not actually running, the pump doesn't draw any current. This is a good thing. It means there's no current leakage from a faulty connection in the float switch.

What's Left?

Okay. We've got a long list of battery drains that are impossible. There simply is no load. None. What improbable things are left? Battery failure. Research shows us four standard failure modes for wet-cell deep-cycle batteries like the T-105's.

1. Depth of Discharge. Draining a battery past the 80% level can only be done a few times. In this case, "few" is about 300. We try to be careful. The 80% level is 11.66V. I don't think we've ever seen a reading that low before Friday. But. The boat was sitting on the hard for a year, waiting for us to buy it. The self-discharge rate is 5% to 15% per month. That would create some wear.

2. Low Water. Checked this. Water is full.

3. Sulfation. This happens over time and (eventually) renders a battery unable to hold a charge. The Trojan T-105's should have a 3- to 7-year service life (some folks say 4 years; some folks brag about getting 9 years). Maybe this is part of it. These batteries are over 10 years old.

4. Internal Short. This is a sudden failure. Also, "A short circuit in the battery will reduce the voltage and capacity from the overall battery bank, particularly if sections of the battery are connected in parallel..." Maybe adding and removing battery cables when installing the ammeter shunt knocked things around internally and lead to a short.

From the maintenance logs, we suspect that battery "#2" (starting battery) was replaced in 1990. The manufacturer's label, however, suggests it's new in 2008. So the logs are incomplete.

The log claims the house battery ("#1") was new in 2000. 11 years of service. The post is stamped with "B-5", which may mean 2005, and only six years of service.

What killed our batteries?

The cause seem to be the vague "old age": all the numerous little failures that accrete over time.

At 62 pounds apiece, the hardest part about replacing them will be actually moving the damn things out of the engine room.