Dash-cam irritation avoidance circuit (How to use a sledgehammer to crack a nut). A guaranteed supply.

My dash-cam has a irritating issue. If I stop at the traffic lights, and switch the engine off in the car to save a little fuel, and hopefully reduce my environmental impact, when I restart the engine, sometimes (most times) the camera hangs up and stops recording. It's powered from a 12V to 5V USB adaptor plugged into the accessory socket. What's probably happening is the supply is momentarily dipping and the micro inside the camera is "browning" out, despite having an internal battery. No messing about with the camera's settings or firmware have solved this issue. 

What we need to do is make a little circuit that ensures the power supplied is OK. 

My idea is, if the main supply dips or even switches off completely, the supply is temporarily held up by a small LiPo cell for 3 minutes. After the that, power is removed, unless the 12V is restored within that time. This should be long enough if I'm sat in traffic, even if I've switched the accessory socket off. 

Here's the schematic:

12V arrives to one of those small buck converters (from eBay/Aliexpress etc) .. and is used to provide a 5.7V regulated supply via one half of D1 to supply the 5V for our camera and the ATTiny85 microcontroller. It's also feeding one of those small PSU boards for charging a LiPo cell & boosting it's output that I used in the Geiger counter project (albeit modified). This board is unmodified, and is adjusted to provide 5.2V output. 

Once the microcontroller has started up, it takes PB0 High, biases on Q1 via R2, and energises the relay. This connects the LiPo cell to the charger/boost controller. The 5.2V output is then connected to the other half of D1, which is currently reversed biased, and does nothing. The output from the 5.7V converter is monitored by the microcontroller PB1, via D2.

In the event of the 12V supply failing, PB1 will be pulled low by R3, as the 5.7V supply is removed, and a timer is started. 5V supply is now maintained, as the output from our 5.2V Lipo boost converter now feeds the 5V supply, as the second schottky diode in D1 is now forward biased. Once the 3 minute timer has elapsed, PB0 is taken low, which kills the LiPo supply by opening the relay, and, as there's no power supply anymore everything stays off, until the 12V supply is restored. 

R1, D4 and C3 form a reset delay circuit. C1& C2 provide some power supply filtration & decoupling. D3 prevents the back-EMF from the relay coil destroying Q1. 

The principal is the same as the circuit I used to write the mileage to the EEPROM on the Mini speedo project.

The code is simplicity itself, and is uploaded to an ATTINY85, set for a 1MHz internal clock.

# define OUTPUT_PIN 0

# define SUPPLY_OK 1

unsigned long timer;

void setup() {

  pinMode (OUTPUT_PIN, OUTPUT);

  pinMode (SUPPLY_OK, INPUT);

  digitalWrite (OUTPUT_PIN, HIGH);

}

void loop() {

  if (digitalRead (SUPPLY_OK)) {

    timer = millis ();

  }

  if (millis() > timer + 180000) {

    digitalWrite (OUTPUT_PIN, LOW);

  }

}

Possibly (!) far too complicated. I could simply use a high side drive FET circuit to drive the output, and use a simple RC time constant to drive the FET, but I suspect the actual parts cost would be about the same, and I don't have a suitable FET in the junque box!

Right, down to construction...

Having got a bunch of parts from the junque box, the circuit is constructed on a bit of perf-board.  

I added an LED across the 5V output, just so I could see what was going on during testing.

All mounted up in a *slightly* too big box, and I could have used a smaller relay (had the junque box provided one) . Input lead is connected to an accessory plug, and output via a USB socket. 

And finally put into the clutter box in the car, and connected up. It performs faultlessly ...

Which is a good job, because this is the state of the wiring on the old USB supply I'd just removed 😬