I purchased a Wisblock Unify Outdoor Enclosure (Unify Enclosure IP67 150x100x45mm with pre-mounted M8 5 Pin and RP-SMA) which I plan to deploy as a stand-alone, solar-powered unit on a vertical mast. But first, I did some unit testing on a product sample.
The PV panel provided with the enclosure is the HX-140X90. At 25°C, the specified open-circuit voltage (Voc) of this model at 1000 W/m^2 irradiance is 6.1V. I measured the Voc myself on a product sample in bright sun at 25°C, and I measured 6.2V. The short-circuit current (Isc) is specified at 400 mA, but I measured 250 mA. This is about what I expect, given the concurrent performance of the array on my roof.
The datasheet max-power voltage (Vmp) for the HX-140x90 solar panel at 25°C is 5V (±10%) in full sun. This is exactly the Vmp that the CN3165 (which is the charge-controller chip on the DFRobot DFR0559) MPPT expects, so this PV panel is well-matched in this sense. Note that the Imp (max-power current) in this case is 380 mA. For a 4200 mAh battery, this corresponds to a charge rate of 0.09C, which is a very slow charge rate for such a battery. An upside to this slow charge rate is that it can extend the battery life, particularly when charging at sub-freezing temperatures. (The permanent damage that occurs to lithium batteries when charging them at sub-freezing temperatures is proportional to the charging rate.) Thus, having an over-sized battery in this application might make the battery last longer.
The CN3165 input-voltage tolerance is up to 6.5V, so that looks OK at first glance. However, the Voc of the solar panel increases significantly at low temperature, well beyond the rated voltage limit of the CN3165.
I chilled my PV panel in my kitchen freezer, which is set at 0°F (-18°C). This is a reasonable/expected winter ambient temperature where I live.
I then moved the PV panel immediately to bright sun, and I measured the Voc to be 6.9V, and the Isc to be 285 mA. Unfortunately, the Voc in this case significantly exceeds the max-allowed solar input voltage (6.5V) of the CN3165. This is a serious design flaw for the CN3165; a standard PV panel with an expected 5V Vmp at 25°C has a Voc of 6.9V at realistic cold winter temperatures. Note that such low temperatures are not a quick, transient condition; they could persist for hours on a cold winter day. The CN3165 should have a higher upper-voltage limit on its solar input.
One possible solution is to add a shunt power zener diode (1N5341B–6.2V) to the PV panel to clamp the PV panel voltage below 6.5V. This is a simple, single-component solution that is rated to 1.5W (@ 25°C ambient) power dissipation, far above the expected max of 620 mW. Note that the 6.2V zener clamp voltage would be well above the 5V Vmp operating voltage, meaning that the zener would have no effect during battery charging. The leakage current for the zener is 1 uA at the nominal battery voltage, so the zener does not present a significant battery drain at night.
A fortuitous benefit of using a zener diode is that it can provide battery heating in cold weather–the zener generates heat only when the Voc is high (Voc > 6.2V) due to low ambient temperature, and when the battery is not charging. When the ambient temperature is above 25C, Voc will drop below the zener voltage, stopping the zener heating.
With a Voc of 6.9V (at 0°F) clamped to 6.2V, I measured the current through the zener diode to be 100 mA! This results in 620 mW being dissipated by the zener, which provides significant heating inside the box. Note again that this heating does not at all interfere with battery charging–the charge controller engages constant-current charging by lowering the PV output voltage to 5V, which is well below the 6.2V clamp voltage.
Note that the Vmp of the solar panel also increases at low temperature, potentially increasing energy production, but since the CN3165 uses a fixed-voltage MPPT algorithm, you won’t get the increase in power that you could get if the Vmp were properly tracked.
I lined the bottom of the box with aluminum tape to act as a heat spreader/sink. I wrapped the battery holder and the zener diode with aluminum tape, which allows the heat from the zener to warm the battery uniformly. I also used a strip of aluminum tape as a heatpipe between the DFR0559 heatsink and the bottom of the box.
The DFR0559 fortuitously happens to have battery screw terminals for directly connecting the zener and the PV panel in parallel. The cathode (banded side) of the zener connects to the ‘+’ terminal.