Skip to main content

Sweeping Analog Clock Analysis

The subject of this analysis is the $2 Ikea Stomma wall clock. Unlike the Rusch wall clock that had been the object of my experimentation all along, this clock uses a sweeping clock mechanism, which means the second hand appears to move continuously instead of every second. This also means it is quieter and does not produce the familiar tick-tock sound.

Note: Both clocks appear to have been discontinued at the time of writing. There is only one $2 clock on sale at their website now, which is the Tromma clock. I don't particularly like this clock, because the second hand is missing, which makes it difficult to tell if the clock is working at any instant.

The hacking process of the Stomma clock is quite similar to that of the Rusch clock, so no suprises there. 

Remove front plastic cover

Remove hour, minute and second hands

Remove clock mechanism

Open up clock mechanism. As before, keep the spindle side up helps keep the gear formation in one piece.

Solder 2 wires to the solder points where the thin coil wires is connected to the circuit board. I am using a 2-pin JST connector here.

Reassemble the clock mechanism. As before, I use the soldering iron to create a small hole at the side so that the wires can exit.

I inserted a AA battery into the clock mechanism and hooked up the wires that were just installed to a logic analyzer to check out the signal produced.

So basically, for each second, 16 pulses (each 32ms long) are produced, 8 pulses for each tick pin. They alternate at 62.5ms intervals (62.5ms x 16 = 1 second). So I think there is an extra 1/16 gear added to the clock mechanism, and each pulse moves the second hand by 1/16 of a second. However, I suspect the trade-off is higher power draw, because the pulse width is similar in the 2 mechanisms, so the power draw should be 16x for the sweeping mechanism.

To test this out, I connected the clock to the ESPCLOCK3 circuitry, and wrote a simple ATtiny85 sketch to drive the clock. The observations are as follows:
  • In my test, the power draw is about 10x when driving the sweeping clock mechanism. 
  • The pulse width cannot be lower than 32ms if the clock is to be driven reliably.
  • However, the rest time between pulses can be as low as 8ms (versus the current 62.5ms - 32ms = 30.5ms). This means the maximum rate at which I can fastforward the clock is about 150% (62.5ms / 40ms). In contrast, the maximum rate for a non-sweeping clock mechanism is about a whooping 800% (1sec / 120ms).
  • The stay-in-place ticking trick (by pulsing the same tickpin repeatedly) does not work with the sweeping mechanism.
So it would not be practical to drive a sweeping clock using ESPCLOCK3 primarily because the fastforward rate is too low. If we are a few hours off, it will take forever to catch up. Also, due to the higher 10x power draw, the clock will last only weeks instead of months.


  1. It would be interesting if swapping the Tick pins would result in the hand moving backwards.
    If so, then this type of movement would be a great choice for Vetinari clock.

    1. I don't think that's possible. The order of the tick pins are irrelevant. What matters are the timing of the pulses. Swapping the pins will have no effect on the direction of travel.


Post a Comment

Popular posts from this blog

Update: Line adapter for Ozito Blade Trimmer

Update (Dec 2021): If you access to a 3D printer, I would now recommend this solution , which makes it super easy to replace the trimmer line. I have been using it for a few months now with zero issue.

3D Printer Filament Joiner

I have been looking at various ways of joining 3D printing filaments. One method involves running one end of a filament through a short PTFE tubing, melting it with a lighter or candle, retracting it back into the tubing and immediately plunging the filament to be fused into the tubing: One problem with this method is that you can't really control the temperature at which you melt the filament, so you frequently end up with a brittle joint that breaks upon the slightest bend. Aliexpress even sells a contraption that works along the same line. As it uses a lighter or candle as well, it suffers from the same weakness. I am not even sure why you need a special contraption when a short PTFE tubing will work just as well. Another method involves using shrink tubing/aluminium foil, and a heat gun: But a heat gun is rather expensive, so I wanted to explore other alternatives. The candle + PTFE tubing method actually works quite well when you happen to melt it at the rig

Attiny85 timer programming using Timer1

This Arduino sketch uses Timer1 to drive the LED blinker: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 /* * Program ATTiny85 to blink LED connected to PB1 at 1s interval. * Assumes ATTiny85 is running at 1MHz internal clock speed. */ #include <avr/io.h> #include <avr/wdt.h> #include <avr/sleep.h> #include <avr/interrupt.h> bool timer1 = false , led = true ; // Interrupt service routine for timer1 ISR(TIMER1_COMPA_vect) { timer1 = true ; } void setup() { // Setup output pins pinMode( 1 , OUTPUT); digitalWrite( 1 , led); set_sleep_mode(SLEEP_MODE_IDLE); // Setup timer1 to interrupt every second TCCR1 = 0 ; // Stop timer TCNT1 = 0 ; // Zero timer GTCCR = _BV(PSR1); // Reset prescaler OCR1A = 243 ; // T = prescaler / 1MHz = 0.004096s; OCR1A = (1s/T) - 1 = 243 OCR1C = 243 ; // Set to same value to reset timer1 to