ESP8266 Community Forum

- Sun Jan 31, 2016 5:35 pm #40120 I just wasted half an hour writing a detailed reply that involved a dirty hack with a 3.9V, 5W Zener diode and Lithium battery but decided to delete it for fear of a lawsuit

Jim, your power budget looks something like 0.17A*3.3V*20/3600 which appears to be under 5mW if you are only waking from deep sleep for 20 seconds in every hour. A small solar panel (say 0.3W) getting 3 good hours of sunlight would give roughly 1W/hour which is amost 10 times your power budget... about what you should be shooting for I'd say.

- Sun Jan 31, 2016 9:47 pm #40133 I recently put together a remote sensor project. Mine wakes up from deep sleep for 2 minutes every hour. I wanted to keep costs down, so I powered it with 3 C batteries. I expect about 5 months of battery life, but my sensor has only been in place 3 weeks, so I haven't been able to verify that yet.

- Mon Feb 01, 2016 5:52 am #40146

Jim, your power budget looks something like 0.17A*3.3V*20/3600 which appears to be under 5mW if you are only waking from deep sleep for 20 seconds in every hour. A small solar panel (say 0.3W) getting 3 good hours of sunlight would give roughly 1W/hour which is amost 10 times your power budget... about what you should be shooting for I'd say.

That's what I was hoping for, I'm thinking about this on the power scale of a solar garden light. However, it's been pointed out to me in another thread that most charge controllers work on a minimum of 500mA input (because they are designed for charging via USB at the highest speed they can, presumably) this would not be suitable for the trickle charging that I was thinking of.

Would using Nicads be a better solution?

- Mon Feb 01, 2016 7:36 am #40156 Jim, unfortunatley the typical Chinese garden solar lamp with two NiMh batteries in series falls short of getting enough volts to run the ESP8266

With NiMh ranging from 1.1 to 1.6 from depleted to fully charged, there's too much variation over the cycle to be useful. You don't have to charge Lithium at 500mA and charge controllers (at least the ones I've used) will attempt to do their job with limited current going in. Their main task is to limit the final terminal voltage at 4.200V

However, if you consider the way some Laptops implement "battery saving" mode which limits charge to 60% when plugged in permanently, this represents keeping the cells at some lower potential than 4.2V This was the inspiration behind the idea of a Zener clamp to simply shunt excess power and keep the volts below 4.2V at all times. If you look at the design of solar regulators for charging 12V Lead Acid batteries, they typically consist of a transistor shunt that limits the voltage to 13.8V (LA float) When 13.8V is reached, the battery is considered to be charged and the transistor starts to conduct to shunt the panel current to ground (just like a fancy Zener really)

The problem with the Zener diode is it's a bit elastic so the volts across it will rise as it's "pushed" harder by the Solar Panel. If the panel can be guaranteed to only "push" so hard, this is OK and a suitable Zener diode can be chosen. 3.9V 5W (1N5335) seems suited to small "garden" style panels that normally come with two NiMh in series. Open circuit these put out something around 4 to 6V. With a Lithium battery in circuit all we have to do is keep it below 4.2V (starting with a battery that's discharged below our Zener clamp voltage - otherwise it will rapidly discharge!). It's still about as cheap a solution as you could wish to get though.

Of course a linear regulator like the LM317 could be used to regulate the voltage from the panel to the battery, however it will cost a little bit more. The Battery charger circuit on the first page of the data sheet would be what you could use - with the addition of a 1N5817 (Schottky) diode between output and battery to block the current flowing back into the regulator when it's dark. Using 270 Ohm for R1 and 620 Ohm for R2 would set the output to 4.12V (before the Schottky Diode drop) and bring the battery up to about 3.9V where it would happily sit all day. A 1N4001 (silicon) diode between battery and ESP8266 would feed it 3.3V max and just about 3V min. (at which point the battery, at 3.6V, could really do with charging to keep it happy).

The 0.2 Ohm resistor in the example is there to limit the maximum current that can flow into the battery. No garden solar panel is going to push enough to worry about so this can be left out. If someone wanted to use a panel that could deliver an Amp or more then this resistor might be a good idea!

The critic would say that cycling from 3.6 to 3.9 is using only yielding half the capacity (hence run time) of the battery - true, but batteries of varying capacities are all peanuts to buy and this scheme trades battery$ against component/complexity$ (and also, a Lithium taken up to 4.2V shouldn't be left there for long periods of time)