I have 3 Ikea Samtid lamps. They’re affordable, useful and the design is timeless.
One is in my living room and started to bore me. That has to change. As the lamp can be dismantled very easily it’s begging for some upgrades with uncountable masses of RGB LEDs ;-)
Preliminary list of parts:
- 1x ATmega 168-20PU
- 1x 16MHz quartz
- 2x 22pF ceramic capacitor
- 1x 7805 voltage regulator
- 1x 100nF capacitor
- 1x 47µF capacitor
- 2x UDN2981A source driver
- 2x 74HC595 8bit shift register
- 3x tpic6b595 8bit high power shift register with open collector outputs
- 24x matched resistors for R,G,B
- MASSES of RGB LEDs !
- LED diffusors
- 2x 2200µF capacitors
- 1x 12V 2.5A power supply
- 1x perfboard – cut to size
- 2x 9302 pushbutton switch
- 1x potentiometer
Some readers may remember that I had abandoned the UDN2981A source driver, as there were some issues with ghost images showing up on my RGB LED Matrix projects. The chip has a turn off delay of 5µs which might be responsible for that. A few days ago I was experimenting with inserting some delays in the PWM code to give it a chance to be fully off, but the effect didn’t disappear completely. I suppose this afterglow won’t be a big issue with the Samtid MOD, as the LEDs will be covered by diffusors and one cannot look at them directly. I tried to get hold of the drivers used for the Rainbowduino (MBI5168 / M54564FP), but they are absolutely impossible to get in small quantities if you don’t happen to live in Shenzhen. The MBI5168 has faster turn on/off times and the M54… is a shift register with high current sink capability and more importantly integrated constant current regulation. That way I could have done with just 3 pots to balance the colors. Without these chips I’m limited to the conventional LED + matched resistor approach.
This is a simple LED test to check if the brightness will be acceptable, even in normally lit rooms. The “Piranha” RGB LEDs can take a maximum current of 50mA (100mA pulsed). I tested one with 3x 100Ω resistors on my Diecimila board. Color balance is a bit off, but brightness is acceptable. In the finished version the LEDs will be driven with a pulsed current (PWM) and this time I’ll actually push the current above the safe DC region. So I’ll better make sure my PWM code runs rock-solid. Otherwise I’ll fry quite a lot of LEDs.
Enough blabla, here are some images and a video.
Right now I’m waiting for some more parts to arrive. I’ll also need to add a new set of tools to my growing collection – in case I’ll get a reply from the ebay shop that sells a hollow punch set I could use.