0
\$\begingroup\$

I would like to make a dimmer for several ikea desktop LED lights. This one is plugged on a LED driver (adapter) with specifications: Primary:220-240VAC 50/60Mhz, 0.06A Secondary: constant current: DC 500mA, 2.5...3.5V max 3.8V.

What's is the better? To interact with the DC side and low voltage or with the AC side?

We're going to make a DMX installation so there's a lot of work, and we would like to reuse it, so if it works with AC lights too it's better.

I've searched and would like to use this one: http://www.instructables.com/id/safe-and-simple-AC-PWM-Dimmer-for-arduino-Raspberr/step3/AC-PWM-dimmer-with-arduino/

Would it dimm the led? I understand that LED are not linearly related it terms of voltage/lumen relation, but would the driver change a lot? can i go from 0V to 3.5v in order to fade in/out it?

Thank you a lot for your help :-)

PH

\$\endgroup\$
6
  • \$\begingroup\$ Work on the low voltage DC side. It's safer, easier, lower cost, and has the advantage that it would actually work. Can you give us a link to the light? \$\endgroup\$
    – Majenko
    Apr 1 '15 at 11:14
  • \$\begingroup\$ Sure! It's this one:ikea.com/us/en/catalog/products/20169658 \$\endgroup\$
    – hachpai
    Apr 1 '15 at 11:21
  • \$\begingroup\$ I guessed it might be - I have a pair of them clipped to my monitors in front of me. I know them intimately. \$\endgroup\$
    – Majenko
    Apr 1 '15 at 11:22
  • \$\begingroup\$ As i've 20 lights like this to control, I can't power them with arduino. But I would like to dimm all of them with one arduino and PWM \$\endgroup\$
    – hachpai
    Apr 1 '15 at 11:22
  • \$\begingroup\$ Would you like to dim them individually or all at once? \$\endgroup\$
    – EM Fields
    Apr 1 '15 at 11:47
1
\$\begingroup\$

No, that method will not work. You would be chopping up the incoming AC waveform to a switch-mode power supply which, at best, would just get confused and not work.

You have to work on the DC side of things.

The lamps themselves contain a constant current sink (actually two of them in parallel) - the A375R. As a result, varying the voltage won't work - it'll just change the amount of heat dissipated and the brightness will stay roughly the same.

enter image description here

So you have to instead create a PWM waveform on the power or ground connection, which can actually be done quite easily by interrupting the ground connection with a powerful enough N-channel MOSFET. If you use a logic level MOSFET then it can be driven directly with an Arduino's PWM waveform.

schematic

simulate this circuit – Schematic created using CircuitLab

The presence of C1 in the circuit may cause some issues with the PWM signal, so it may be necessary to use quite a low frequency PWM signal to allow it to turn off properly in-between each cycle.

An even better solution would be to completely re-wire he interior of the lamp. Bypass the two regulators and the capacitor completely (remove them from the circuit board) and directly connect the power lines to the LED (move the GND connection from GND to one of the OUT pins where a regulator was). Then the whole lamp just becomes a plain LED for you to then control with a proper PWM LED power driver, such as the CAT4101, which will give you perfect control over the LED's brightness. Of course, the lamps can then never be used as normal stand-alone lamps again since they lack a driver of their own.

\$\endgroup\$
6
  • \$\begingroup\$ Thanks a lot, great explanation! Indeed, I was suspecting something like that. As the lights must return to their owner after the performance, I think that we'll just control the dc current with a NPN transistor and PWM... \$\endgroup\$
    – hachpai
    Apr 1 '15 at 15:20
  • \$\begingroup\$ FYI, your link for the A375R is bad - I get the datasheet for a Spansion flash chip. Googling A375R doesn't seem to give anything useful. \$\endgroup\$
    – DoxyLover
    Apr 1 '15 at 19:26
  • \$\begingroup\$ Oops, must have copied a link from earlier in the day. When I get home I'll update it. \$\endgroup\$
    – Majenko
    Apr 1 '15 at 20:48
  • \$\begingroup\$ @DoxyLover Ok, I have updated the link. Overflowing with datasheets today yet again... \$\endgroup\$
    – Majenko
    Apr 1 '15 at 23:25
  • \$\begingroup\$ Now I understand the logic of PWM through transistor. The PWM is extended through the ON/OFF logic of the transistor. This schema should work for me then. bildr.org/2011/03/high-power-control-with-arduino-and-tip120 \$\endgroup\$
    – hachpai
    Apr 3 '15 at 8:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.