LEDs... resistors not necessary with constant current driver, right?

Discussion in 'General Electronics Chat' started by s_mack, Aug 27, 2015.

  1. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    I'm only asking this because the LED panel manufacturer is being oddly persistent on the point, and I am by NO means an LED expert, so I'm asking you guys for your opinions on the matter.

    Background: I've been working on a project that is an add-on to an existing consumer product, which utilizes a non-backlit LCD screen. One of the features of my project is it adds a backlight. I provide an LED panel, which I am having made in China. Onboard my project is a constant current LED driver IC.

    When the Chinese factory produced my samples, I noticed tiny resistor chips in series with each LED "string", which were shown on their circuit diagram (shown below) but I managed to miss them.

    upload_2015-8-27_9-45-29.png

    I told them the resistors are unnecessary, since I'm driving it with a constant current. They insist I'm making a mistake.

    Am I?

    The driver I'm using provides this typical application circuit:

    upload_2015-8-27_9-57-14.png

    clearly no resistor is shown for each LED.

    But maybe the LED panel manufacturer knows something I don't?


    I tested it both ways. It was slightly more dim with the resistors in place, but not a big deal. I let it run for several hours without the resistor and no harm seemed to come of it.
     
  2. dl324

    Distinguished Member

    Mar 30, 2015
    3,237
    619
    If you're operating from a current source, the brightness should be the same with or without the resistors. Otherwise your current source isn't acting as a current source with the resistors.

    The manufacturer obviously intended for the LEDs to be driven from a voltage source and they probably selected resistor values that would match brightness between colors when operated in that manner. You can short the resistors and use your current source...
     
  3. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    And it may have been. I tested it, then shorted and tested it again and it subjectively seemed (slightly) brighter to me the 2nd time, but it wasn't a very scientific observation by any stretch. I may have been simply expecting it to be brighter.

    Thanks.
     
  4. dl324

    Distinguished Member

    Mar 30, 2015
    3,237
    619
    Leave the LED connected to the current source and short/unshort the resistor quickly. You should be able to detect any significant change in brightness.
     
  5. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    OK, did that. You were right, of course... no difference at all.

    Other than brightness, is there any other concern with leaving the resistor in? Battery drain?
     
  6. GopherT

    AAC Fanatic!

    Nov 23, 2012
    5,992
    3,744
    Battery drain will be the same or slightly higher if your current source can force the same current through the LED whether or not the resistor is there. Remember, that the current source will cause some heat/waste in the transistor of the constant current source to drop the voltage to the right level to keep the current constant. If it is not the transistor, it will be the resistor. Of course, it depends on how the constant current source is managed - I assume all linear components in my explanation. There may be a pwm constant current source where less heat is generated/lost but I don't see a big deal.

    My bigger concern, is the paralleled LEDs from your Constant Current source. How do you know that all of those LEDs have the same voltage drop? You generally do not have any control and you should have a small resistor in front of each LED to buffer the maximum expected variation between the LEDs or have a constant current supply for each LED, (or multiplex the LEDs to insure only one is on at a time).

    Otherwise, the LED with the smallest forward voltage will get more current than the others and appear brighter, a slightly different color, and possibly have a shorter lifetime.
     
  7. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    A LED driven with a constant current will dissipate I*Vf Watts. If it has a resistor in series with it, the LED dissipation will be I*Vf, but the resistor wastes an additional I*Vr Watts, where Vf is the forward LED voltage and Vr is the resistor voltage. Just wasted power in the resistor because it can have no influence on the LED power...
     
  8. Colin55

    Member

    Aug 27, 2015
    191
    19
    But you need the resistors if you are connecting all the colours to the same constant-current source.
    The resistors convert the LEDs to requiring the same VOLTAGE. Without them, each string wants a different voltage to be applied. That's the DIFFERENCE, which no-one has observed.
     
  9. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    while that does make some sense to me... what would be the real world application of driving discrete RGB from the same source? The point of RGB is to facilitate the representation of all the colours visible light has to offer. Driving them all the same... that would be what? A sloppy approximation of white? The driver provides 3 constant current sources, not one split 3 ways.
     
  10. GopherT

    AAC Fanatic!

    Nov 23, 2012
    5,992
    3,744
    That is not what we meant on our comment about issues. The problem is if you parallel all of the blue with one current source (with no resistors), all the green with another current source and all of the red with a third current source (as drawn in your first image on an early post). Even though three identical looking blue LEDs all have the same specs, they may have different Vf and you will have issues with the Led with the lowest Vf.
     
  11. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    I understood yours and Colin's as two different issues. He said " all the colours to the same constant-current source", not "all of one colour". His concern, as I understood it, seemed moot by the fact each channel is distinct.

    As for your concern, I do see where you're coming from Gopher. I didn't really consider this because this is just the way these panels are constructed. I gathered up quite a few devices with "add-on" type backlights during my (limited) R&D, and they ALL have multiple LED's in parallel. The difference is mine is RGB and all those are monochromatic, and mine has 7 vs usually 2 or 4... but isn't it the same thing, in principle? So is this perhaps a case where your theory is sound but in practice it doesn't matter? I'm not arguing... I'm asking.
     
  12. GopherT

    AAC Fanatic!

    Nov 23, 2012
    5,992
    3,744
    That is a good question. It may be that, in the earlier days of LEDs, and even the mad rush to make high power/high brightness LEDs in the 2008-2012 timeframe, the manufacturing consistency just wasn't there. The LED manufacturers may have solved this issue in the past three years (since I have stopped following closely). If it works and you have no issues, then it seems to work. If you see one LEd that is more bright or if you see one fail at a time, then you have a problem. Most people do not operate at rated current anyhow so it may be a very long time before you see a failure.

    Good luck and check in if you learn anything (positive or negative) on this, I am interested (just not motivated enough to check for myself.). Cheers
     
  13. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    I really have no idea if things have improved recently. As I said, I'm no expert. I really started by tearing out backlight panels from random devices I had similar to the one my device is for (r/c radio controllers, GPS units, toys) and looked at how they were done. I also ordered up a couple purpose built aftermarket LED backlight panels (ie. ones meant for retrofitting, as I'm doing). Each one was single color, so I don't have an RGB reference, but they were all wired in parallel (except one toy that had a single LED) with no resistors on the backlight themselves. Every single one had a single current limiting resistor on the upstream circuit (device's PCB or, in the case of the standalone backlight, inline with the cabling). None of them were driving by a constant current source, even though I think I understand that's the ideal way to drive an LED.

    So when the factory I'm contracting with suggested 7 LEDs in parallel... I didn't think anything of it. The only pause I had was when I saw the three resistors directly on the backlight (rather than assuming the controlling circuitry would take care of it).

    But when I think about it... how would one do it any differently? I suppose ideal would be 7 driver ICs but then that would mean a 22 wire loom coming off a 2" backlight!

    These things are cheap enough that I think I'll just go ahead and give it a try. I'll learn pretty quick if the 1000 people I send them to are having problems.
     
  14. Colin55

    Member

    Aug 27, 2015
    191
    19
    You haven't made it clear if you want to control the LEDs independently to create the full spectrum of colours or simply increase and decrease the brightness of the 3 strings as a group to produce white back-lighting.
    The resistors do not matter if each colour is driven by a constant current source but since the characteristic voltage-drop across a red LED is much more than a blue LED, the resistors were included so that each string could be driven by a simple PWM signal consisting of an ON/OFF signal.
    Since the resistors are included, the circuit will be much simpler and much cheaper with 3 transistors or FETs and 3 signals from the micro.
     
  15. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    To be honest, I don't think my circuit could be any simpler or cheaper :) Or if so, it is so minor I really don't care about the cost difference. I think I'm at $0.08 each for the circuitry. The panel itself is several dollars. Far more "expensive" would be using 3 signals from the micro. Those suckers are at a premium! ;)

    Sorry I wasn't clear. I suppose I figured if the goal was white light, that I'd be using a white LED. I'm sure that's my ignorance making that assumption. The goal is indeed to create a spectrum of colours. A serial line exists between the micro and the driver and the driver then controls each of the three channels accordingly.
     
  16. Colin55

    Member

    Aug 27, 2015
    191
    19
    What is inside the rectangle with outR on it?
     
  17. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    That's the driver IC
     
  18. Colin55

    Member

    Aug 27, 2015
    191
    19
    How come it only takes data to drive 3 lines PWM? Where is the clocking?
     
  19. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    Its special :)
     
  20. Colin55

    Member

    Aug 27, 2015
    191
    19
    If you send a long-enough stream of data you can perform this function.
     
Loading...