Over the past few weeks I did some extensive tests for battery life of
my Nixie Watch.
When I started with the design, I took David's numbers as a reference.
An average of 50 readings per day at 1mA tube current (about 185mA
battery current) results in 4 to 6 months battery life. Say 5 months,
that would be 7,500 readings for 1 battery.
I wanted my battery to last for 1 year, so it needed to be able to
last for say 20,000 readings. I started off with making the HV circuit
as efficient as possible which resulted in the fact that I could drive
the tubes with 1mA at around 95mA battery current. I could use this
advantage to go from say 7,500 readings to 15,000 readings (or 5
months to 10 months). However, I decided to increase the tube
brightness to go from 1mA to 2mA so that the tubes would be better to
read when you're outside.
This meant that to increase battery life, I had to be able to adjust
the tube brightness according to the amount of ambient light, which
would also make it much more convenient to read the time when it's
dark.
The high efficiency of the total circuit resulted in a minimum of 6mA
battery current when reading the tubes at night and around 185 - 200mA
in (very) bright daylight.
To test battery life, I implemented a feature that the watch can
automatically display the time with intervals of 5 seconds. This comes
down to around 600 readings per hour. I also implemented a 16 bit
counter
that would keep track on how many time readings had been triggered. I
had estimated that the number could be anywhere between 20,000 and
50,000 so that's why I choose 16 bits.
To estimate battery life, and check how the watch would behave when it
would run out of juice, I then installed a new battery and set the
watch to display the time 600 times per hour and 24 hours per day.
Wherever I went, I was wearing the watch and I only took it off at
night.
Now the interesting part. After about 4.5 days, my 16 bit counter
overflowed, passing the maximum of 65,535 readings. And it didn't just
do that once, no, several days later it overflowed again, clocking a
total of 131,000 readings! But, it didn't stop just there because the
battery still had enough juice to go on for another 100,000 readings.
Then, the display faded and in the end at 270,000 readings the battery
was dead!!
Now, that's what I call an increadible result!
Sure the tubes are not at their brightest, and the low current (6mA)
during the night would compensate for the higher current the watch
would use during the daytime.
Then I did a second test, basically the same but now with the tubes
set to a constant 100% brightness, 600 times per hour. Not
surprisingly, the battery could not supply the required power anymore
after about 14,000 readings. I built a feature into the software that
monitors the battery performance and automatically reduces the tube
brightness to a minimum when the battery is nearly flat. This way, the
time will still display at low tube brightness (which is 25% of
maximum brightness) and you can change the battery before it is
completely flat.
This also means that if we would set the watch to a constant 50%
brightness (1mA tube current), the battery will last for about 25,000
readings, enough for 1 year.
http://xiac.com/Images/Steampunk_Nixie_Watch.JPG
Michel
http://facebook.com/nixiewatch