Deep battery discharge – hot and cloudy.

I thought this graph might be of interest to those who want to see what happens when you run a large Nissan leaf pack down to zero state of charge. (SOC)

It’s been consistently 90 degrees during the day and still about 75-80 at night, the last few days have been cloudy, I left the air conditioning on at 72 degrees F, and the bus isn’t plugged in to shore power to augment power consumption.

You can see the cell voltage taper off sharply after the knee. This is a reason it’s a good idea to keep a BMS watching over the cells – you can see the voltage divergence as they drop, and you want to catch 0 percent SOC on the first cell to hit bottom voltage.

The reason the data stops is because the inverter cut out, turning off data collection. I could fix that by moving the power source for the docker farm that this data collection stuff runs on to run directly off the battery. The Victron Venus GX runs directly from battery so it stayed energized, and relayed the cached data in the middle graph after turning the long term graph storage back on.

This day actually was a perfectly clear day, but the per-cell voltage for 100% soc is 4.09 volts which means the voltage had not quite recovered from the previous day of cloudy weather.

In the middle chart, you can see the consumption pulse when I turned the air conditioning on just before noon. The consumption then declines below solar production, which is basically the net solar in, which you can cross reference with the orange graph on the bottom – around the same time it’s just slightly net positive then by 3:30 PM the consumption begins to exceed the production, drawing on battery as the sun moves along.

An interesting theory I have for incrementally increasing performance, and something I’d love to try, is cooling the panels somehow. You can see that the production curve doesn’t match solar angle exactly. On this day’s graph, there’s a very linear production suppression from thermal response in the middle of the day. If I were to gently spray the panels with some light water mist, you would see spikes of production. Water cooling (thermal harvesting) the backs of the panels would likely increase the production by several percent.

The spikes are typical of the MPPT controller measuring and adjusting the conversion curve.

One Reply to “Deep battery discharge – hot and cloudy.”

  1. I’m wondering if there would be a not crazy difficult way to build a pre-heater for the water based on circulation to the back side of the panels. I’m trying to imagine what that look would look like that isn’t a nightmare long term. Maybe if you could loop the pex as one giant continuous loop so that you’d only have fittings at the start and finish, but then you’re also piping a decent amount of wait to the roof and trying to secure it tightly to the back of the panels.
    I’ve been following your build (and am using it as a template for my own build) for years now… so please go ahead and give it a try and I’ll attempt V2 after you’ve got some of the bugs worked out! 😛

Leave a Reply