开云体育

One of the issues operators have with remote solar powered anything and lithium batteries is cold weather.

?

They don't like to be charged at below freezing temperatures.
Thanks to another user I discovered a possible method of preventing battery damage using a Thermal reed switch.

It appears to be a normal magnetic operated reed switch with a substance called Thermorite orientated as a magnet.

According to the spec sheet, Thermorite is a ferromagnetic substance who's magnetism decreases with rising temperature.

Depending on the recipe, it can be made to operate at specified temperatures.

I've ordered one to switch open at freezing (0C).?

It will be in contact with the lithium battery and the switch in series with the 5W solar panel. Specs say it can handle 10W.
Anytime the battery is too cold, no charging will occur.

If it works out, the next thing will be to see how much storage I need to get thru the sunless days.

?

Here's a link to the data sheet. Digikey carries them. Not cheap at $30 ea.

?

?

73

Mike N0QBH

The thermal switch arrived and I tested it with a multimeter and the freezer. It appears to work as advertised (opens at freezing, closes above).

The switch is encapsulated and enclosed in a aluminum case with 8" leads.

I soldered the switch in series with the solar panel's positive lead and taped it's case to the battery.

When the battery temperature drops to freezing the switch opens and prevents further charging until it warms up.

The lithium battery is insulated from the case by a layer of foam.

?

Solar Mesh node update.

The thermal switch has worked exactly as expected opening at freezing thus isolating the solar panel from the charge circuit and the battery.

This is done to prevent damaging the LiPo battery.? Charging when the cell is below freezing is a no-no.

?

So far, the longest "off" period was almost a week before the weather warmed and the sun recharged the 3.7 volt 3 Ah LiPo battery.

Fortunately, the node is lightly used and very frugal with it's Watts, so it stayed on with some power to spare (30%).

The next goal is to use the solar electricity to heat the battery up, allowing it to charge when the sun is shining on those below freezing days.

?

After some trial and mostly error, I believe I have a circuit simple enough for others to make.

The idea is to sense when the thermal switch has opened isolating the solar panel from the charging circuit.

This, in turn, shunts the solar panel current through a 20 ohm resistor at 5 volts generating 1.25 Watts of heat.

The resistor is a larger surface mount variety laying on a double sided copper clad circuit board which becomes a heat sink.

Some short wires soldered to both sides transfer the heat to the pcb bottom and the battery.

?

By using a P channel MOSFET I was able to do it all with one device, an IFR9530 3 legged TO-220.

It's gate senses the volts on the thermal switch out, turning on hard when the switch turns off and vice versa.

The 100k pull down resistor insures the MOSFET gate goes low while a blocking diode built in the node's solar charge circuit prevents any back flow from the battery.

?

Here's the yet to be populated circuit board as it would sit next to the thermal switch atop the battery.

?

?

---

Outstanding!!!

PCB with parts soldered on and wired for testing the actual circuit.

The main concern is how hot at full output (1.25W).

Tie wrapped to the battery and un-insulated brought the battery temperature up about 20 deg F above surroundings.

The pcb alone with only air to cool it rises to about 140 deg F after 10 minutes.?

?

The switch mimics the function of the thermal switch.?

In operation the thermal switch opens when the battery temp drops below 32 deg F.

?

Switch closed (on) MOSFET gate is high - no current - no heat.

?

?

---

Typo - should read almost 1 1/4 Watts output.

I wired the heater circuitry into the node's solar charging harness and mounted the thermal sensor on the long narrow side of the battery.

Both the sensor and heater pcb are held in place tight with a small tie wrap around the battery. Not the one shown.
Then I wrapped this assembly in two layers of thin foam for insulation.

The battery assembly sits atop a 1/4" foam pad to further insulate it from the case.

I also updated the node's firmware to the latest stable release while I had it in the shack.

The white connectors shown attach to the RAK 4631 mesh node circuit pcb.

Bottom to solar jack and battery's to the battery jack.

I'll get pictures next time it's in the shack.

?

?

Then I put the node back on the tower to wait for a sunny, freezing day.

Today (12/20) was a perfect test with temps in the single digits and bright sun part of the day.

The battery was down to about 60% from a stretch of cloudy days.

The simple circuit warms the battery as long as the temperature sensor is open (<32F), else all the current goes to charge the battery.

?

Meshtastic has some nice built in metrics for visualizing it's power status over time.

In it you can plainly see when the battery is being warmed (low half of the "square wave") and charging when the voltage is higher.

It also gives a good idea of the duty cycle which tells me the heating is gentle.

?

While this is only day one of testing, the results are very promising as a solution to the LiPo battery cold weather problem in solar powered remote nodes.

The first two hours were bright sun then filtered through clouds the rest of the day.?

Interestingly, it was still charging, tho not as fast.

?

?

73

Mike N0QBH

?

?

3rd Day Update

To coin a military term Bravo Zulu - doing great!

?

We had a fair part day charging Friday 12/20.

Almost no sun Saturday until about 2 pm.

Sunday is sunny (aptly named :).

As the chart shows, it was able to warm up and charge the 3 Ah battery in almost all conditions.

?

Now we can think about deploying year round mesh nodes in good locations without electricity available.

73

Mike N0QBH

I took the Meshtastic node down off of the tower's 10' spot and brought into the shack for a checkup.

My main concern was if anything got hot enough to melt anything.?

None, nada.

Because it offers a view of the final wiring and mounting I took a few pictures with and without the insulating cover.

?

?

Next will be to button it back up and get on the tower, this time at the 20' level.
I'm relatively confident I won't need to climb back up twice.

?

73

Mike N0QBH