I've had some time to do some testing and will make the following improvements to the control circuit. I'm going to hack these into the current PCB, so it will be messy. If it works as expected over time, I'll make another layout takeing these changes into account.
Failure Modes
- Power drops below 4.5 volts for the motor control 555 - can cause the feeder to miss deposited objects
- Deposit trigger may get jammed and cause the dispenser to run contantly
- Dispense trigger may fail to respond or the hopper is empty, causing the dispenser to run contantly
Inefficiencies
- Powerbank Keep Alive always running, even when the feeder draws enough power to keep a power bank on.
Fixes
- Power - Will only switch power to the sensor LEDs and phototransistors
- Motor control 555 will not suffer from a voltage drop below 4.5 volts as it does now
- Standby current usage will increase, but still be under 50mA
- Can reliably use a 5V USB charger or power bank again
- Deposit Trigger
- Will add a couple resistors, a cap and a diode to keep the trigger pulse short, even if the sensor is constantly on.
- If the deposit chute gets jammed or very dirty the sensor may cause the motor to keep running and dispensing peanuts.
- This will partially mitigtae this failure mode - if the sensor is jammed it will still start up the next time a bird triggers the PIR, but will stop the dispense motor from running indefinitely.
- Dispense Trigger
- Will turn the bistable 555 controlling the motor to a monostable with a period of ~20 sec.
- If the dispense sensor does not trigger for some reason within that time, the dispense motor will stop.
- Powerbank Keep Alive
- Put a 10K resistor between 555 pin 4 and the power rail so I can add an NPN switch to turn off the keep-alive pulses while the sensors are on (they draw enough power)
I should have these changes in place before the weekend. I'm eager to get the feeder back up and running. Once it's working I'll need to update the description, components and schematics. A new PCB layout will come shortly after everything is confirmed.
Discussions
Become a Hackaday.io Member
Create an account to leave a comment. Already have an account? Log In.