General > General Technical Chat
Simple voltage cutoff schematic for solar charger, any tips?
Zero999:
--- Quote from: Psi on April 23, 2022, 04:18:51 am ---
--- Quote from: Zero999 on April 21, 2022, 12:37:31 pm ---
--- Quote from: Psi on April 21, 2022, 11:29:18 am ---Thanks for the circuit.
Yeah, it's an option, but i think just disconnecting the panel has more advantages.
--- End quote ---
Which is exactly what the circuit I posted does.
Do you need to switch the positive side of the panel for any reason?
--- End quote ---
Sorry, i got confused and thought your circuit was one that shorted out the panel.
Yeah, i've been thinking about this and doing some testing but the main problem with the comparator + nfet idea seems to be that it will need quite a bit of hysteresis added.
--- End quote ---
It should disconnect the battery nice and cleanly, without the need for hysteresis. When the MOSFET starts to turn off, the voltage on the solar panel will rise, causing the voltage on the TL431's reference to increase further. This positive feedback results in latching action, disconnecting the solar panel from the battery until the of the day, when it's reset. A night the battery powers the load via the MOSFET's body diode. It might become unstable, when resetting, but I doubt that much hysteresis is required. You really need to test any solution with a real battery, load and panels.
Try a feedback resistor between M1's drain and the TL431 reference? That should provide positive feedback, as the TL431 is an inverting amplifier and so is the MOSFET.
You'll probably need another resistor between the TL431's output and the MOSFET's gate. I forgot the minimum voltage across a real TL431 will be around 2V, which could exceed its threshold voltage.
Zero999:
--- Quote from: Psi on April 18, 2022, 09:36:02 pm ---
--- Quote from: Zero999 on April 18, 2022, 04:11:22 pm ---
--- Quote from: Psi on April 18, 2022, 10:42:50 am ---The more I think about it the more i like the MCU option. That was until I looked at stock levels of sub $1 MCUs at digikey.
--- End quote ---
How many do you need?
--- End quote ---
Not that many. Maybe 1000 a year, but I like to design products using a MCU that digikey/mouser stock at least 30k of, which is getting very hard of late.
I just really hate having to redesign a product every few months to use a new MCU, its starting to get annoying.
--- End quote ---
--- Quote from: Psi on April 23, 2022, 04:18:51 am ---The MCU idea will work great, and is simple.
--- End quote ---
How about the ATTINY10? Digi-Key has >50k in stock and they're cheap. The only downside is it only has 4 IO pins, but that should be enough for this application. A thermistor could be used to measure the battery temperature and perhaps a look-up table could be used to simplify calculating the cut-off voltage?
https://www.digikey.co.uk/en/products/detail/microchip-technology/ATTINY10-TSHR/2051008
Psi:
The constraints put on me for this project have been changed.
I can now locate the circuit somewhere else and so burning off the excess voltage as heat is doable. I'll probably just go with an LM317 or similar and a PCB large enough to heatsink it.
Thanks for all who helped and put forward ideas.
Psi:
Out of intertest, D1 is only there for protection from the solar panel being connected backwards, correct?
It's not doing anything special other than that?
Zero999:
I've used that circuit before, in the same application and it works.
Yes, D1 can be removed, if you can guarantee the solar panel won't be connected backwards. How much current can the solar cell produce? If it's >1A, use a larger diode, as the LM317 might not current limit until 2.2A.
I would also recommend adding a 33V zener diode across U1. I had problems with the LM317 failing, which I put down to lighting strikes and the zener fixed it. Obviously it doesn't protect against a direct strike, but it will suppress shot spikes, capacitively or inductively coupled from a nearby strike.
I thought you wanted to vary the battery voltage, depending on the temperature? You could make part of R1 or R2 a thermistor.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version