Question about LED power

Discussion in 'The Projects Forum' started by JSWarner, May 20, 2015.

  1. JSWarner

    Thread Starter New Member

    May 20, 2015
    4
    0
    Hello everyone,

    I am working on an electronic circuit project, and I have a question about power supplies.

    The circuit uses a microcontroller with a couple different sensors, and a string of a few dozen LEDs.

    The microcontroller and sensors together draw only a couple hundred milliamps of current, while the LEDs draw approximately 1.5 amps. All components have the same operating voltage.

    A friend told me I may want to consider using separate voltage regulators...one for the LEDs alone, and one for everything else. I didn't understand his reasoning so I thought I'd ask here. Is this necessary? Are there any good reasons I would want to use separate regulators?

    Thank you,

    JSWarner
     
  2. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    You want to keep the ground return path from the microcontroller/sensors separate from the ground return from the high power circuits all the way to the negative terminal of the common power supply.

    If the sensors are ratiometric, then by all means use a separate low-current regulator for them and the A/D reference pin...

    Since the LEDs should be driven by a constant-current regulator, then there is no need to voltage-regulate the LED supply.
     
  3. JSWarner

    Thread Starter New Member

    May 20, 2015
    4
    0
    Thank you for replying MikeML.

    I should have been more clear about the LEDs I'm using; my apologies. They are programmable RGB LEDs, and each LED has its own embedded driver chip which handles the current regulation. The only requirement is that the logic-high level of the programming signal is the same as the supply voltage. Since the microcontroller I/O levels are 0-3.3V, I need to use a 3.3V rail to power the LEDs as well (not wanting to deal with level shifters for the programming signal).

    No issues there, the sensors are not ratiometric.

    So as long as I keep the ground returns separate, there shouldn't be issues using a single regulator?
     
  4. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    I doubt the RGB LEDs will work on 3.3V. I dont think there is enough headroom. Post a data sheet...
     
  5. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,791
    1,103
    Does the LED spec allow a supply voltage as low as that?

    Edit: Mike beat me to it ;).
     
  6. JSWarner

    Thread Starter New Member

    May 20, 2015
    4
    0
    They are rated down to 3.5V, but I've had no issues using 3.3V. Datasheet is attached.
     
  7. JSWarner

    Thread Starter New Member

    May 20, 2015
    4
    0
    Hi Alect_t, thank you for responding. The spec allows for supply voltage as low as 3.5V (datasheet is posted above). I've been using 3.3V with no adverse effects so far. Other than the LEDs not turning on or being dim, neither of which is a problem, are there any other concerns with undershooting that spec?
     
    Last edited: May 20, 2015
  8. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,386
    1,605
    Drawing a high current around a digital circuit has two prominant pitfalls. The first is it may load down your power supply, hence your friends recommendation to use two separate supplies (or two regulators).

    The next is ground bounce, or the effect of the LED current causing a voltage drop in the ground leg of the micro. The best fix for that is to join these two section right at the regulator so their currents have the shortest path to effect each other.
     
  9. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,791
    1,103
    Fine, if they work.
     
  10. hp1729

    Well-Known Member

    Nov 23, 2015
    1,950
    219
    You don't have to drive the LEDs at 20 mA. 5 mA or even less can light them enough for most purposes.
     
Loading...