Reproduced this on Instax SQ1

[Lenart12]

I'm making a note here, huge success!

I have reproduced this on Instax SQ1 using ATtiny13A. Biggest deviation was that SQ1 has Vdd always active, even when the lense is turned to off, so the code was changed a bit for more deep sleep saving - using power down sleep mode (and adjusting the interrupts to support it), setting input pull-up for all unused pins, disabling BOD fuse, reducing the clock speed to 128 kHz and I ended up using internal pull-up resistors instead of an external one.

Hardware

After programming with Arduino ISP only 4 wires are required - GND, Vdd, PB0 (purple), PB1 (white).

Connected in much the same way as in the guide.

Software

I have "ported" the code for the Arduino IDE using the MicroCore framework. Here are the setting to burn the fuses:

After setting the fuses to the above mentioned settings I have flashed this sketch with maximum power savings:

Note: Arduino ISP needs to be reprogrammed with #define SPI_CLOCK (128000/6) to account for the reduced cpu frequency after burning the fuse that reduced the CPU frequency from 9.6MHz to 128kHz.

#include <avr/io.h>
#include <avr/interrupt.h>
#include <util/delay.h>
#include <avr/sleep.h>
#include <avr/wdt.h>

#define BUTTON_PIN PB1
#define OUTPUT_PIN PB0

#define SIGNAL_DURATION_MS 300
#define SHORT_BUTTON_PRESS_TIME_MS 1500
#define BUTTON_PROBING_INTERVAL_MS 100
#define LONG_BUTTON_PRESS_SIGNAL_DELAY_MS 10000

void send_signal()
{
    // PORTB0 is always set to 0 (output/low or input/high-z depending on DDRB)
    // Drive output low (output)
    DDRB |= (1 << DDB0);
    _delay_ms(SIGNAL_DURATION_MS);
    // Return to high-z (input)
    DDRB &= ~(1 << DDB0);
}

short is_button_pressed()
{
    return !(PINB & (1 << BUTTON_PIN));
}

short is_long_button_press()
{
    int delay = 0;

    do
    {
        _delay_ms(BUTTON_PROBING_INTERVAL_MS);
        delay += BUTTON_PROBING_INTERVAL_MS;
    } while (is_button_pressed() && delay < SHORT_BUTTON_PRESS_TIME_MS);

    return delay >= SHORT_BUTTON_PRESS_TIME_MS;
}

void wait_for_button_release()
{
    while (is_button_pressed())
    {
        _delay_ms(BUTTON_PROBING_INTERVAL_MS);
    }
}

ISR(INT0_vect)
{
    // Disable INT0 to avoid multiple triggers while the button is held down.
    GIMSK &= ~(1 << INT0);
}

void app()
{
    // Set PB0 as input (high-z) and rest as input
    DDRB = 0;
    // Set pull-ups on input pins except PB0 (which should stay high-z)
    PORTB = (1<<PB1)|(1<<PB2)|(1<<PB3)|(1<<PB4)|(1<<PB5);

    // Disable unused peripherals for power saving.
    ADCSRA &= ~(1 << ADEN); // Disable ADC.
    // Keep PB1 (AIN1/INT0) digital input enabled.
    DIDR0 = (1 << AIN0D) | (1 << ADC0D) | (1 << ADC1D) | (1 << ADC2D) | (1 << ADC3D);
    ACSR |= (1<<ACD); // Disable analog comparator
    PRR |= (1<<PRTIM0)|(1<<PRADC); // Enable power reduction Timer0/ADC

    // Enable INT0 in low-level mode (required for wake from power-down).
    MCUCR &= ~((1 << ISC01) | (1 << ISC00));
    GIMSK |= (1 << INT0);

    // Disable WDT
    wdt_disable();

    // Enable interrupts.
    sei();

    // Use deep sleep and wake via INT0.
    set_sleep_mode(SLEEP_MODE_PWR_DOWN);

    // Infinite loop.
    while (1)
    {
        sleep_mode();

        if (is_button_pressed())
        {
            if (is_long_button_press())
            {
                wait_for_button_release();

                _delay_ms(LONG_BUTTON_PRESS_SIGNAL_DELAY_MS);
            }

            send_signal();
        }

        // Re-arm INT0 only after the button returns high.
        wait_for_button_release();
        GIFR |= (1 << INTF0);
        GIMSK |= (1 << INT0);
    }
}

extern "C" void setup() {app();}
extern "C" void loop() {}

Its a shame Instax does not do this out of the box since they have such a capable mcu on board. Anyways I'm thanking you for your guide. If I have any issues I will update them here.

[TomaszRewak]

Oh, nice, thanks for sharing the info!

Indeed, the fact that Mini 12 just cuts the power whenever the lens is closed was a big simplification in case of my integration.

When it comes to the external pull-up resistor - I'm not sure how did I end up using it instead of the internal one. If you asked me yesterday - I would have confidently said that ATTiny only has pull-down and no pull-up resistors. Which is clearly not true. Nice improvement!

[Lenart12]

I ended up redoing it anyways since I left much of the power saving still on the table. Now pretty much everything is disabled, in maximum sleep mode and running at minimal cpu frequency. I have edited my initial post with the new sketch :^)

[Lenart12]

Update: this killed my batteries in 2 days -_- i will need to figure something out

[TomaszRewak]

Hmm. Is there no other pin on the board that gets cut off when the lens is closed? Maybe there is a similar circuit to the one in Instax Mini, just in a different location.

[Lenart12]

Revisiting this I have found the following: There is an enable and mode switch triggered by the rotating sled as show in the picture:

However the "enable" is just for the trigger enable since the mcu is as far as i can tell always on because it flashes the status light when the rotator is incorrect mode - e.g. not fully closed or in between the enable switch indents.

The enable switch is the orange wire connected to the SW2 label. It is pulled up with a 66k resistor and when closed (when camera is ON or SELFIE) it shorts to ground. So I have connected the GND of the ATTiny to that pin instead of the BAT-.

I will see how the battery life goes.

EDIT: Another finding - it seems that the floating PB0 (output pin) causes trouble when trying to let the main board sleep. This might have been the real reason for short battery life.

[Lenart12]

Finally! I have figured out the power drain. In sleep the main MCU expect the shutter input to be 0 but in reality it was driven high by the ATTiny, backfeeding the main MCU during sleep and waking it up. Instead of wiring the enable switch I have modifed the shutter code to be high-z/low during sleep instead of high/low and that stopped messing with the sleep states. It went from 1.98mA powered off to 7.8uA - about 280x reduction in power consumption! I have already updated the sketch in the first post. I suspect that the biggest power draw now is the fused on watchdog timer that is always enabled by the microcore framework.