Long Clock Prototype

I can't believe that it has been exactly one year since I last posted on my 10000 Year Clock.

Well I finally got the low power LCD and put together a prototype of the clock. So progress is being made.

The LCD is a 40-pin 4-digit module 40LCD from Futurlec.

The LCD driver is MAXIM ICL7211A.
Here is the datasheet in pdf.

I have programmed the MSP430G2553 so that the display cycles through the following display sequence:


The entire circuit is powered by a 1F/5.5V super cap. The solar panel shown was salvaged from inexpensive garden lights. In bright sunlight the output voltage is about 3.5V, just perfect for this project. It is connected to a 1N5817 Schottky diode and then to the super cap.

The average current draw is about 3μA. In testing, I found that the super cap can power the clock for about 10 days.

So now it is on to Stage 4 - synchronization with the sun. This is the tricky part. For now here is what I have.

I found a piezo buzzer, XL-450 that is just the perfect size. I took it apart and removed the internal components. Then I glued an IR photodiode, Honeywell SDP8276, to the inside bottom of the buzzer. On the retainer ring, I glued two thin brass shims to make a very narrow aperture.

The idea is to attempt to synchronize the clock to the midday sun. The algorithm is going to be challenging, to say the least.

For the moment, I am collecting test data using MATLAB and will later decide what is going to be my strategy.

Here are the challenges:

1) I need to ignore data from days that the sun doesn't shine.

2) I need to watch for false peaks, i.e. when there is cloud cover at the critical moments.

3) I do not have an ideal south facing window and hence I need to accommodate other locations.

4) What happens when someone moves the clock, even by the smallest amount?

Blog entry information

Last update

More entries in General

More entries from MrChips

Share this entry