EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: SanctePieDecime on June 20, 2024, 09:56:30 pm
-
So, I am working on a small project involving a battery-powered system. The battery of this system is charged by some small solar panels. This project is a mail arrival notification system. The battery would be 3.7V LiIon and the solar panels would be something like 5V or 6V.
I have two questions:
The first is: should I use MPPT, only a buck/boost converter, or something else? Most of the sources I have found seem to indicate that, for low-power applications, MPPT is more complicated than it is worth and that I should just use a buck/boost converter. On the other hand, because the system is battery-powered it would seem that you would want to charge the battery using the solar panels as optimally as possible.
The second question is: Regardless of whether I use MPPT in my charge controller how well do you think the following MPPT conceptual circuit would work? Op-Amps measure the current and voltage. For any given current-voltage combination there is only one curve that will match the solar panel's power output. If we follow the maximum power point as the light intensity increases, we see that it follows a path. Now, I might be wrong about the function representing the shape of that path, but it looks like a logarithmic or exponential path depending on which way you have the axes (and that would make sense considering that solar cells are a form of diode). Using the measured voltage and current as well as a known maximum power point path (with the offset either preset or defined by a calibration) we could have an MPPT controller that doesn't need to do a ton of calculations and could be a simple mostly analog design. Would it even work? If so, how would it compare with existing solutions? And if I have no idea what I am talking about, please fill me in.
-
You can get something like 90% of the theoretically available power if you just use a buck converter at a fixed ratio to turn 5 volts into 4 to charge the battery from a 6v open circuit panel.
Now given you need it to work in the winter, your panel needs to be oversized which makes the mppt circuit rather pointless, unless the battery can last 2 weeks.
-
My battery should hopefully be able to last a while, I have designed a power-latch circuit to help with that, but I don't have any numbers for the power consumption for even 1 day since I am still in the designing phase. There will be a functionality that will turn on lights inside the mailbox if it is dark out to see the mail more easily (which will time out if opened for too long). I know that lighting the inside of the mailbox will use a good deal of power compared to the other parts (sensing, timers, and transmission). So, I hope it lasts long enough, but it depends in large part on the capacity of the battery and I could theoretically just use a bigger battery if needed.
I saw that most sources talked about using a buck converter. I feel like I am missing something but wouldn't a buck-boost converter be better since the output voltage could be lower than say 4V if it is overcast?
-
Solar panels are mostly current sources so a 6 volt panel, which might be more like 9v open circuit, (given they are often nominally sized to charge lead acis batteries) will never have a problem charging a lithium cell.
Cheap solar lights at walmart are like 4 square inches of panels and thsts probably all you need
-
For a small, low-power application like a mail arrival notification system, using a buck/boost converter is generally more practical and cost-effective.
-
How much power do you need? There are dedicated energy harvesting ICs that are plenty capable of charging a lithium battery from a solar panel, and do so with MPPT, however they generally operate in a boost configuration and work with < 3 V solar panels.
-
I don't have anywhere close to an exact number as to how much power I would need but it should hardly be any.
I need a radio to transmit the data and a microcontroller to communicate with the radio (unless I can find some sort of GPIO to I2C master IC that is inexpensive, readily available, and low-powered, or maybe I could find a radio that doesn't use I2C) aside from those, which should only be running for less than 5 seconds, the power consumption should be under 5 mA (probably about 1mA) when the mailbox light is not on. My very rough estimate for when the mailbox light is on is somewhere between 20mA and 50mA for about 30 seconds max.
Would it be better to go with lower voltage solar panels and use a boost charging configuration? I can see that using a pre-developed IC would be beneficial as well as using smaller solar panels, but I don't want to "under-do it" if it cannot charge the battery except under perfect conditions. It would also be nice if I could use dim ambient lighting (such as you would find in a forest) to still charge the battery. I don't live in a forest but it would be nice to have that level of confidence in my charging setup.
-
Can you give us more information about what you have in mind? Such as:
What kind of wireless connection will you use? Is it wifi to a router in your hearby home? And ultimately are you sending a notification to a phone?
Are you detecting a mailbox door opening and powering up when that happens? If so, how will you do that (the "power-latch circuit" you mentioned)?
Is this going to be a 3.3V system?
How cold will things get in the winter? Remember that rechargeable lithium batteries have limitations as to the temperatures they can be charged at, or even discharged at.
I ask about these things because except for lighting the interior of the box, the load of this system will be very low. You can set it up so that the entire system is powered down all the time except for a few seconds twice a day. And it may be possible to power it with primary batteries such as the Energizer Ultimate Lithium, which are 1.5V primaries. Those batteries have the advantage of operating at very low, or high, temperatures (-40 to 140F). I would think three of those in series would last several years, and you wouldn't need to bother with a solar panel or a charger.
Edit: If you do go with solar recharging, I don't think there's any point in using any kind of switching circuit. Just use a TP4056 charger and a linear 3.3V regulator.
-
I would be using a 915 MHz radio to communicate to send data to a base station that I am designing. That is the extent of the radio communication.
I am going to use something like a reed switch to turn on the power latch circuit when the door opens and the microcontroller will turn the power latch circuit off when the data has been transmitted to the base station.
Yes, this is going to run on 3.3V so I will need to regulate that (and perhaps I should be more concerned about power dissipation from a voltage regulator than the rest of my circuit).
I haven't thought too much about battery specs yet, so your advice is very helpful.
Maybe I really don't need solar charging, but I am just a bit skeptical that any battery will stay charged. Even if it is disconnected from everything, I just think of how many batteries I have come back to only to find them dead for no apparent reason. Now maybe they don't use a power latch circuit, but I have a pair of walkie-talkies that I put fresh batteries in, and used like two times and the batteries are dead when I come back to them about 3 months later.
However, you are not the first person to recommend not worrying about the solar charging aspect because the batteries should last a while. So, I will consider not solar charging, but I am fairly certain I will still do it and still keep the power latch.
The lighting inside the mailbox is really my biggest concern (even though I will be implementing safe-guards to prevent the lighting from staying on if the mailbox is open abnormally long). Over time, I can see it really depleting the battery capacity (the same reason I try to keep the flashlight on my phone on as little as possible).
-
I've done one of these mailbox notifiers, and posted the project on Github. You might find something there that's useful.
I used a magnetic reed switch to ground the gate of a P-channel mosfet that switched the power on and off. The processor can also hold the power on until the radio transmission is completed. I didn't want to run the power directly through the reed switch because the postman might close the door before the transmission could be done.
https://github.com/gbhug5a/Mailbox-Notifier (https://github.com/gbhug5a/Mailbox-Notifier)
-
My project is a little more complicated. I could have built my own one of those or bought one for dirt cheap off of amazon, but there is a fundamental limitation of existing solutions that I am trying to overcome: knowing the status of the mail. The closest thing I could find to what I wanted to do was Ben Heck's https://www.youtube.com/watch?v=x7AiV7yg-yw (https://www.youtube.com/watch?v=x7AiV7yg-yw). Even that is not quite what I wanted. His layout is constantly detecting while the mailbox is open and only has sensors towards the front of the mailbox. My system will be detecting mail anywhere in the mailbox as well as the status of the flag. Because the flag status may change after the mailbox is closed there will need to be a delay before the final sensor statuses are transmitted.
I already designed a power latch circuit that is very similar to the one you used and built out a proof of concept on a breadboard. So it is good to see that I am on the right track.
I am getting into more details about the specifics of my project than I intended to get into here since this thread was about a specific question. I would be willing to do a full detailed explanation of what I am trying to do in a separate thread.
-
Well getting back to your original question, I think it's hard to find a convincing reason to use anything more elaborate than a linear charger like the TP4056. Your system will use very little power for very short periods of time, and your battery will probably be good at least for weeks with no charging at all. Then on bright sunny days, the panel will provide a lot of power for charging, perhaps more than your constant-current charging current setting. In those cases, it doesn't really matter how efficient it is. Even on cloudy days, the panel will provide as much charging current as it can at just above the battery voltage, and it's hard to do any better than that. So for this setup I would say that an MPPT or PWM charger would provide so little benefit over a linear charger that it wouldn't be worth the effort.
But for any charging, you still need to work out the temperature situation, unless you just won't ever have any really cold weather.