Ordering parts for a 12V experiment

I've posted this in "Prius Technical Stuff" and here looking for comments and suggestions.

For my airplane project, I need a light-weight, 12V battery, that has 'drop-in' characteristics. There are light-weight, 12V LiON batteries but gosh they are expensive and are sensitive. But I have experience with our Prius NHW11 modules and feel comfortable with them.

GOAL: Configure NiMH cells as a 12V, drop-in battery

I checked the specifications for D-sized, 10Ahr cells and found:

1. 1.0V minimum charge per cell
2. 1.25V maximum charge per cell
3. 30A maximum "rated" discharge

Typical 12V, lead-acid operating range:

1. ~10.5 to 11.0V minimum lead-acid battery voltage to start Prius
2. 13.8 to 13.9V typical charger limit

So I did the following calculations:

• minimum voltage cell count 11V/(1V/cell) = 11 cells
• maximum voltage cell count 13.9/(1.25V/cell) = 11.12 cells

It looks like 11, D sized, NiMH cells have a voltage range consistent with the minimum and maximum voltage levels of our Prius. The 30A maximum discharge rate (per specs) is in the range that have been reported for Prius operation. The 10Ahr is about 1/3d the capacity of the lead-acid battery BUT I suspect the lead acid battery is sized to survive a month or so with the parasitic loads.

There is a problem of parasitic load, the keyfob receiver and ECU memory which as been reported in the 30-35 ma range. The parasitic load can over-discharge the NiMH battery just as it does a lead-acid battery. Lead-acid cells are a little more tolerant, less instantly killed, by the effects of fully discharge. So I need an active system to protect the NiMH battery from a complete discharge.

To address the parasitic load, I'll install a 12V solar panel with Schottky block and Zener voltage regulator. The Arduino will measure the solar panel, charging current, expected to be in the 100s of ma in direct sunlight. Hopefully the daily solar charge will be larger than the daily discharge losses and parasitic loads. If not enough, I can add a second panel.

PLAN:

I have ordered, 11, D size, NiMH, tabbed cells rated at 10Ahr. When they arrive, I plan to work up a minimum volume package and install it in our 2003 Prius. It will also be instrumented so we can get some metrics.

I will be using an Arduino nano; a 30A, Hall effect current sensor and; a resistor network to monitor the voltage. Temperature will be measured from the Arduino in the middle of the pack. "Heat is the enemy" and the Arduino will be placed perfectly.

From an economic standpoint, the 11 cells are coming in about $7.15/cell, just under $80. I still have to work up the housing and add the 12V, 1.5W, solar panel charger, $15. Add the housing and we're looking at about $100 for a 12V battery that should weigh significantly less than a lead-acid battery. I did not include the instrumentation as this is part of my experiment, getting facts and data.

I'll probably use a semi-log scale to analyze the data:

• 1, 2, 5 days - preliminary (will carry 12V backup, battery)
• 10, 20, 50 days - steady (optional, the 12V backup battery in trunk)
• 60, 60, days for a year - long term

Depending upon how well the storage works, we should have fairly accurate data showing Prius 12V operation with the solar charger.

We are entering the warm seasons, which curiously is the time of maximum heat, the greatest stress on the NiMH cells. As for handling parasitic loads, the system won't get a proper, winter test for at least 9-10 months. But after a year, we'll have metrics on how well the 12V, drop-in, NiMH battery works out 'in real life.'

Considered:

The battery could be built using spring-loaded, cell carriers so a single failed cell could be replaced. But welded tabs have the lowest resistance and are significantly better at handling vibration. Also, 11 cells in series are going to have similar operational characteristics. If one fails early, the others are likely to soon follow.

There are F size, NiMH cells with 15Ahr capacity but I have not found them with welded tabs. Worse, they are significantly more expensive than the D sized cells. Still, they may be needed in higher latitudes where there is less sunlight in the winter. Getting tabs welded on them is not that hard.

In theory, a voltage operated, low-voltage, isolation relay (aka., MOSFET) could keep the NiMH from completely discharging. But there needs to be a 'reset' so you can get in the car and boot it right away. If the initial battery/solar cell works, throwing a light on the solar panel might work. I may reconsider this as part of the initial test.

Anyone see a problem or suggestion?

Bob Wilson

 

We see the same problem using 'back of the envelope':

• 35 ma-24 hr day / 125 ma -x hr sunlight -> ~.28 hr-day, ~7 hrs daylight needed

I suspect we'll be OK on a single panel but these are ROM estimates. It does not include shade and cloud effects. This is why detailed metrics are needed. <grins>

Our Prius modules easily handle forced charging of 70A and they are six, 7.5Ahr cells. I'm using 10Ahr cells so there should be more margin. The vendor reports the maximum discharge of 30A, 3C, so I'm not too worried. But you bring up a good point, maximum charge rate.

The worst case should be pre-dawn when the NiMH pack has reached minimum voltage. The Prius 12V system can easily provide 75A based upon my earlier, inverter testing. Testing will give us hard numbers and I may reconsider active charge management but understand even lead-acid batteries suffer similar risks just to a lower degree.

Thanks for the comments. There is always a possibility that I (or any other engineer) might have overlooked something. Better to address the risks early than get into testing and have an "Uh" moment.

Bob Wilson