Debugging power gremlins

When I went on board earlier this year, I noticed that my batteries were well below a good resting point. Blue Opal has a 55 Wp solar panel hooked up to a Votronic MPP 250 Duo Digital. With the panel left on deck when I’m not sailing, it should have kept the batteries trickle charged. The fact they hit 12.4 volts (and lower) tells me that it wasn’t doing this.

I took the batteries ashore to put them on the mains-powered chargers for 24 hours, and contemplated whether now was a good time to convert to LiFePO4 (lithium iron phosphate – not as energy dense as the lithium cobalt mix used in cars and other lithium-powered devices, and not as “I want to break free” either). I ended up on Nordkyn Design’s pages about lithium battery systems, and concluded that for now I probably don’t want to think about it – I’d need to rejig the entire setup to accommodate the peculiarities of lithium cells.

With the batteries charged, I put them back on board, and just let them sit for two weeks, with everything disconnected. No change in voltage, so it has to be a parasitic drain. I re-worked the arrangement of distribution buses over the last few weeks, and left the solar regulator disconnected (since it draws power, even if it’s very little amounts).

  • The “always on” distribution panel in the battery compartment was moved to the load side of the house battery switch. This means the voltmeter and ammeter are now only on when the battery switch is on.
  • The shunt for measuring house amps was moved from downstream of the house battery load fuse to downstream of the engine battery negative terminal.
    • The engine battery to engine cable runs straight from the terminal, so the shunt won’t see the cranking amps (which it probably couldn’t handle).
  • A few items that connected directly to the negative terminals of the batteries were moved over to the distribution bus that sits on the load side of the shunt. Fewer wires to disconnect from the batteries now.
  • The charge cables from the mains charger were completely replaced and uprated by 2mm2 (6 -> 8). They went directly on to the load side of the batteries (at the battery switches), and thus were not fused. They’re now hooked to the charge fuses (70A) for each battery bank, same as the alternator is via the ProSplit-R.
  • The water pump finally received 6mm2 direct from the battery bank (via the distribution panel and negative bus), with the switch at the chart table only energising a relay (same type as the engine solenoid). Pump sounds better, and peaks at 9 amps. It has a 15A fuse on the panel.

This left just the solar aspect. With a voltmeter on the terminals, and the 55Wp panel in full sun, I wasn’t seeing a charging voltage from the regulator. It should produce at least 13V, and preferably 14V, with a full belt of sun (midday sun yesterday was pretty strong, no clouds – optimum conditions for a solar panel). The voltage at the terminals from the solar panel was sitting at about 15.6V, but I noticed that it would cycle down to about 13V, the regulator would turn off the MPP light, the voltage would come back up, and the MPP light would come on again. Rinse and repeat. The to-battery terminals showed 12.6V (the voltage at the house battery terminals agreed).

I brought the regulator home, discombobulated my UPS to get the 12V battery out, and rigged up my bench power supply as a “it’s a solar panel, honest!” input to the regulator. At 15V input, I could see the same cycling pattern, and the >80% light on the regulator never lit up (the UPS battery was sitting at 13.3V). The to-battery terminals showed 13.3V.

At 17V, the >80% light on the regulator lit up, and I got a constant 14.4V on the to-battery terminals.

My conclusion is thus that my solar panel has either aged to where it can’t provide enough power (but at 15V shouldn’t be hooked directly to a lead acid battery), or has some other fault. Reading the spec, it tops out at 17V, so perhaps it was never quite big enough to keep the regulator happy, though I do recall in the first weeks of use that it would indeed illuminate the >80% light.

Guess it’s time for a new solar panel.