flipflipclock

a self-calibrating GNSS powered mechanical split-flap display clock

Ingredients & Facts

• ancient split-flap display flip clock from a radio alarm
• with custom «calibration sensors» (micro switches)
• ATmega328p (8bit, 32kb flash, 2kb RAM, 16 MHz)
• Atomthreads OS
• http://www.nongnu.org/avr-libc/ AVR libc
• battery-backed u-blox NEO-M8N GNSS module (configured for GPS + ГЛОНАСС operation, with «AssistNow Autonomous» enabled)
• stepper motor
• home-made GPS helix antenna à la http://home.iae.nl/users/plundahl/antenne/helical.htm (dead link, archive.org and local copies)
• multi-function 6 bit LED display (status, seconds)
• …

Source Code

• This is part of my Arduino Uno Stuff.

Pictures

Video

What happens is (mm:ss):

• 00:12 power on (the GNSS receivers is battery-backed and has been on before, hence it does a warm-start)
• 00:13 the status display shows the configured UTC offset (+2, i.e. 0b00010) for a few seconds
• 00:17 the status display blinks (start of calibration)
• 00:18 the split-flap display moves a bit slowly (for show, and to assert none of the triggers are active)
• 00:20 the stepper moves the display quickly to the trigger on the hours wheel (~12.00) and to the trigger in the minutes wheel (~xx.26)
• 00:43 it moves the display just beyond the the minutes wheel trigger (12.26), this is the «calibrated zero position»
• 00:44 some status display blinking (again for show, and to indicate the end of the calibration process)
• 00.45 since the system doesn’t know the time yet (the GNSS has not yet acquired a fix) it will move the clock to 00:00
• 01:05 the GNSS receiver has a position fix and it’s status (time pulse) LED starts blinking (barely visibly) at 1 Hz, the clock/system doesn’t care and continues heading for 00:00
• 01:30 the clock display «arrives» at the «minutes zero» (~xx:26) just before the target time (00:00), re-calibrates itself (by slowly moving just beyond the trigger, and then continues «slowly» to the target
• 01:35 the clock arrives at the targeted 00:00, now the system checks again for the GNSS time (which is now available) and sets for a new target (18:49)
• 02:52 again the clock arrives at the «minutes zero» just before the target time (18:49), re-calibrates and then slowly approaches the current time (which is by now 18:51)
• 02:56 the target time is reached and the display stops moving, the LED status display now displays seconds (in binary, compare the time with the DCF77 clock)
• 03:41 it’s now 00:52 and the display advances a minute accordingly
• 04:41 again.. :-)

Watch the video on Youtube.

Attic

Old stuff. A very old video of a first attempt at driving the display.

Watch the video on Youtube.

created: 2008-11-20, updated: 2020-07-12