I've made myself a little power box designed to use two 18650's to power something that needs 12v but is fairly low draw. The 18650's provide 6-8.4v so I am using a Step Up Boost Converter MT3608 to provide the required voltage. On it's own this is pretty simple, but I wanted battery monitoring as well so I setup an Arduino with a small buzzer, and single WS2812 to provide voltage feedback.
The voltage on the 18650's gets too low for the Arduino Nano (7-12v) so I decided to use my regulator already providing 12v, but the voltage I need to watch is the input (battery) voltage. I used two equal resistors as my voltage divider. I realize I should probably have done more but this should have (and confirmed with multimeter) divided it in half which is within the 5v tolerance of the analog pin.
I've wired it up like this (apologies as the TinkerCAD circuit diagram I used doesn't actually have the Arduino Nano nor a step up regulator). I think you can get the overall circuit though.
On the bench it tested fine. I was getting the expected voltage on the ADC read. The beeper and WS2812 worked and I showed 12v at the power box output.
When I plugged a load in (~1 amp) it briefly worked, then the buzzer sounded and then I smelled magic smoke. You can see the chip melted on the Arduino here (bottom left).
After disconnecting the Arduino the step up boost still works, and the circuit, minus voltage monitoring, works fine.
So...what did I do wrong? I figure I created some kind of loop because I had to power the Arduino from the regulators output 12v, but measure from battery (6-8.4v). I am suspicious of the extra ground I had from the input side of the regulator to the Arduino (grey wire in diagram), but I didn't show continuity from the regulators output ground to it's input ground so I figured I needed that for the ADC to have a reference.
Or did my voltage divider, being only 1/2 factor get a spike of some kind since it would have only taken a brief surge to go over the 5v pin limit. Should I change that to a 1:11 factor since I certainly don't need the precision and it would be tolerant of much higher voltages?
Tried it again with new Nano. Removed the extra ground wire, changed the voltage divider to a 1:11 ratio, and dropped the total voltage to 10v.
Adjusted code for new voltage multiplier.
Tried it with a simpler load, everything appears to work. Tried it with the desired product and it goes haywire. Showing low voltage (red LED and buzzer) then back to good, back and forth. In the desired device I can see the voltage its reporting is stable at exactly 10v.
Finally Arduino appears to die too, no light and no buzzer. No scorch marks, and it shows up as a USB device but won't let me upload anymore.
I'm at a loss.
The voltage regulator on the Nano is rated about 1 watt, it's dropping 7V, from 12 to 5 so the current that would cause a 1 watt dissipation at 7V is 143mA, and you're trying to pull 1000mA through it? That's 7 watts, a 700% overload.
That 3 legged part in your picture with the blister on it is the deceased regulator. If you are handy with a soldering iron you might replace it, but ?? Try feeding regulated 5V into the 5V pin, it may still work ??
JCA34F:
The voltage regulator on the Nano is rated about 1 watt, it's dropping 7V, from 12 to 5 so the current that would cause a 1 watt dissipation at 7V is 143mA, and you're trying to pull 1000mA through it? That's 7 watts, a 700% overload.
That 3 legged part in your picture with the blister on it is the deceased regulator. If you are handy with a soldering iron you might replace it, but ?? Try feeding regulated 5V into the 5V pin, it may still work ??
The input voltage is listed as in different places from 6-20v, to more commonly 7-12 depending where you read (official Arduino Store). Granted this is a clone, so maybe it has a weaker regulator.
I don't think I'm trying to pull 1000mA through it though. The Arduino is only acting as a voltage monitor. The main load is just in parallel with the Arduino. After I burnt it I just disconnected the Arduino and the rest of it still works.
The main goal was to not have to add another regulator to handle the batteries (which could be as low as 6v or even slightly lower).
I ran out of spare Nano's so now I've got a DigiSpark ATTINY running off of the batteries (8.4 -> 6v) instead of in parallel from the regulator. So far it seems to work ok, but I won't know for sure until I get a chance to run the batteries through a full cycle. I'm worried that when the voltage drops past 6v, the regulator on the Digispark will drop to less then 5v which will affect the ADC accuracy.
Read the specs on the regulator. Input of 20V is only when output is 12V. Certainly not 5.
CrossRoads:
Read the specs on the regulator. Input of 20V is only when output is 12V. Certainly not 5.
Ok...but I was giving it 12, not 20. Which is within the specs.
From the documentation the Nano itself is pulling 19mA, I've got one WS2812 running on a single color so less the 60mA, and the buzzer which says 30mA. That's a grand total of 110mA @ 5v. Unless I'm missing something that should work for the official board regulator. It blew again at 10v as well, so there should have been more headroom yet.
I normally have a separate 5v regulator with most projects since I know the onboard regulator isn't meant to power much, so I admit this isn't something I've had to deal with much. However, for space and compactness I thought I could get away with 2 output components.
Anyway...someone suggested I might have gotten a voltage spike since it worked just fine until I plugged in the load which makes a lot of sense. Now that it's running off the battery side of the boost circuit it should be safe as that has reverse current protection. Working so far anyway, though again the ADC may be inaccurate if the 5v onboard regulator can't actually maintain 5v when the battery dips too low.
