Connect with us

Powering 104 LED's from Molex connector, possible?

Discussion in 'LEDs and Optoelectronics' started by Skar, Jul 22, 2016.

Scroll to continue with content
  1. Skar

    Skar

    4
    0
    Jul 22, 2016
    I am currently working on a case mod and a part of it requires 101 white LED's. They are standard 5mm LED's,

    aRAUl61.jpg

    I plan on powering these LED's from the 12 volt molex connector in 26 groups 4 LED's in series each with a 1ohm resistor in each group via this wizard here: http://led.linear1.org/led.wiz

    My concerns are is this practical and safe without harming my PSU or computer. About how much current, voltage and watts will this array consume and is it it within spec for Molex to do this safely. Also need to come up with the best and safe way to actually wire this all togehter, perhaps a centralized point for both ground and voltage and run a wire out to each array from both or find a way to wire them as shown in the schematic the wizard produces.
     
    Last edited by a moderator: Jul 22, 2016
  2. Harald Kapp

    Harald Kapp Moderator Moderator

    11,008
    2,506
    Nov 17, 2011
    1 Ω is bad advice by this wizard (we've had a similar discussion a short time ago). The resistor needs to effectively limit the current when the LED voltage or supply voltage changes. but at 1 Ω a small change in voltage will result in a comparatively huge change in current. When the supply voltage is near to the LED string's total voltage, a controlled current source should be used (see our resource section).

    We cannot answer this question. You must have answered it before when you entered a value for the LED current into the wizard.
    Assuming a good current source, each string of LEDs draws power as calculated by Pstring = 12 V * I(LED).
    26 strings of LEDs the will dissipate Pto = 26 * Pstring.
    Total current will be 26 * I(LED). Look up the specs of the connector to check whether it is capable of carrying that current.
     
    Arouse1973 likes this.
  3. BobK

    BobK

    7,682
    1,687
    Jan 5, 2010
    You need to go to 3 LEDs in a string to make this work. Assuming 20mA LEDs, that would be 101/3 = 34 strings x 0.02A = 0.68A. Quite doable. I calculate 150Ω for the resistor.

    Bob
     
  4. Sunnysky

    Sunnysky

    483
    123
    Jul 15, 2016
    Check your supply V and test a few LEDs with 10Ω resistor.

    Each LED in UB variety >5Cd tend to be 3.1 Nom. +-0.1. Very very few in 5mm are less than 3V, while in >=3W chips, many are 2.9V from lower ESR.

    5mm LEDs are mostly ESR of 13 to 15Ω nom, so if you are operating at 12.5V say , then the series R might could still be anywhere from .0Ω to 0.5V/4. /0.02A =6.25Ω

    Testing the current of .random picks of 4 LEDs in series with 0R and 12V within 5% is safe and then you can determine if your choice is valid.

    IMHO 3 LEDs and a big R is wasteful and old school.

    I distribute 5mm LEDs for 10 yrs now and I can justify all my comments.
    If anyone is interested I have stock in 10 ~20 CD LEDs in Amber Yellow, daylight White (4500K), Red in 200 ~ 500 pc bags. top quality and ESD protected. Clear, 30 deg
     
  5. Sunnysky

    Sunnysky

    483
    123
    Jul 15, 2016
    101 ? why not 100 or 104 or why not 200?
    no big deal to bias 1.
     
  6. Herschel Peeler

    Herschel Peeler

    401
    65
    Feb 21, 2016
    Theories aside, what voltage do the LEDs actually drop across them at the 20 mA, or whatever you intend to run them at? Can you run 4 in series from 12 V? The difference between 1 ohm and no resistor at all is minimal,.Yes, good designs should have a resistor. That being said I have a flashlight that is just 3 AA batteries (4.5 ) and four LEDs in parallel (3.7 V) . No resistor at all.
     
  7. Skar

    Skar

    4
    0
    Jul 22, 2016
    The outline pattern that I am using these in calls for exactly 101 LED's. Going to design the circuit with 104 for evenness sake and cover up the extra 3.

    Your flashlight comparison gave me an idea, rechargeable 12V Li-Ion battery. I don't need to have these LED's powered all the time, and I don't want to risk damaging my power supply. Tried out out 4 LED's in series with 4 CR2032 batteries and they work just fine without a resistor. Planning on picking this up, http://www.ebay.com/itm/3000mAh-12V...614163?hash=item21076eab93:g:6cwAAOSwhkRWeOxX

    Is this a viable solution to project?
     
  8. Harald Kapp

    Harald Kapp Moderator Moderator

    11,008
    2,506
    Nov 17, 2011
    That's because the internal resistance of the CR3032 limits the current available to the LEDs (more precisely the limited rate at which the CR3032's chemistry can generate current).
    The situation is completely different with a power supply that has a low internal resistance (e.g. a PSU or a good Li-ion battery).
    Also, the voltage drop across an LED varies with temperature, therefore operating an LED without current limiting is definitely not recommended.

    You still haven't told us at which current you are going to operate the LEDs.
    Note that an LED works completely different from a converntional light bulb. A Light bulb is defined by the operating voltage and will draw as much current as is required to light it up.
    An LED is defined by an operating current and will drop a voltage acoordingly. It therefore needs a controled or at least a limited current source, not a voltage source.

    Got a question about driving LEDs?
     
  9. Skar

    Skar

    4
    0
    Jul 22, 2016
    I see, thanks for the information about the batteries and all of the variables that affect how LED's operate.

    As for current, going by what I see is the common current for 5mm LED's, would be 20- 35 mA across each LED? Would like to dim them a little bit if possible by using a higher ohm resistor.


    Looked through the guide you linked and I am thinking a 33 3 LED array with a 25mA current and 120 ohm resistors would be the best approach?
     
  10. Harald Kapp

    Harald Kapp Moderator Moderator

    11,008
    2,506
    Nov 17, 2011
    LED currrent varies widely. It depends on the LEDs you use. Where you got the LEDs from there should be a datasheet stating the required current. Or you can just experiment using different resistors, measuring the resulting current and selecting the current that gives the best match in brightness to your expectations.
    One minor point: that's "through", not "across".

    3 LEDs are 9 V ... 10.2 V according to the photo in post #1.
    That leaves 3 V ... 1.8 V for the resistor.
    At 25 mA this makes for a resistor in the range 3 V / 25 mA = 120 Ω ...11.8 V / 25 mA = 72 Ω.
    With R = 120 Ω you're on the safe and dim side, just as you want it to be.
    The power dissipated in the resistor is P = I^2*R = 75 mW, so the resistors you use should be rated at > 75 mW (e.g. 125 mW or 250 mW are typcal values).
    Note that in total you will dissipate 33*0.075 W = 2.5 W as heat in these resistors, so put them in a spot where they are cooled by air passing over them.
     
  11. Skar

    Skar

    4
    0
    Jul 22, 2016
    Alright so 33 sets of 3 LED's with a 120 Ohm resistor @ 1/4W dissipation on my planned 12V source from a PSU is the solution for this?

    These resistors look to be exactly what I need, any objections? http://www.ebay.com/itm/100-x-Resis...026415?hash=item4ae9ea4d2f:g:8W8AAOSwBahVUF7m
     
  12. Harald Kapp

    Harald Kapp Moderator Moderator

    11,008
    2,506
    Nov 17, 2011
    These are fine for this job.
     
  13. Sunnysky

    Sunnysky

    483
    123
    Jul 15, 2016
    It seems these 5mm white LEDs are good direct across a CR2032 which has an ESR, when new, similar to the LEDs around 15Ω.
    upload_2016-7-29_2-54-35.png
    This pack starts at 12.6V not 12V
    Using my experience, if it looks OK on one CR2032, it probably is only using 10mA from a loaded cell around 2.9V.
    But that's Li not LiPo.

    I would make an educated guess

    Vf = [email protected] , ESR~20Ω
    Vf = 3.1V @20mA , ESR~15Ω
    and [email protected] ESR~12Ω

    Thus with some interpolation
    12.6Vmax - (3*3.2)V = 3V
    If = 3V/120Ω = 25 mA max

    10.8Vmin - (3*2.9)V = 2.1V
    If = 2.1V/120Ω=17.5 mA min , after which when LED dim battery must be switched off, well before to avoid accelerated aging below 2.8V/cell 0r 8.4V/pack
    Charge cycle life can reduce to <250 cycles when used beyond 100%DoD
     

    Attached Files:

  14. Herschel Peeler

    Herschel Peeler

    401
    65
    Feb 21, 2016
    Not suggested for long term use or reliability. "Works" does not imply "works well". A word of caution. On my flashlight about once every year or two I need to replace the LEDs in the flashlight. Not a big problem. I buy the LEDs by the hundred. It would be bad advice to say no resistor is a good design, but it does work. Actually I think the internal resistance of the battery works as the resistor.
     
  15. Sunnysky

    Sunnysky

    483
    123
    Jul 15, 2016
    Not quite.

    The LiPo cells have a much lower ESR (<0.1Ω) than the 5MM LEDS (~15Ω),
    However when LED's exceed their rated current, the exceed a junction rise of 30'C with a rate of 24'C/100mW

    Thus 30 parallel 5mm LEDs is now 0.5 Ω so the single resistor for the group reduces the risk of all going into thermal runaway, certainly the lowest ESR device will also share more of the current depending on the % of mismatched parts. Parts in the same batch are exceptionally well matched. (<1%)

    TO guarantee avoidance of thermal runaway 30 parallel or 30P1S, One should aim for 15 or less mA
    For momentary 20 to 30mA and avoid thermal runaway the LED's should be configured for higher voltage as done in large power array chips and one might use 3P10S or 3 parallel strings of 10 @ ~ 30V~33V with an ESR now of 150/3=50Ω
     
  16. Herschel Peeler

    Herschel Peeler

    401
    65
    Feb 21, 2016
    That sounds like wisdom to me.
     
  17. Sunnysky

    Sunnysky

    483
    123
    Jul 15, 2016
    IF you ever need 16~20 Cd 5mm LED's, I have a box of excess stock for one client. 300 per bag. x 50 bags

    It was a custom order for tight luminance , 4000-4500'K , ESD protected.
    Each batch of binned parts turned out to be within a few mV of each other, not by spec, just the result of high quality or coming from the same epiwafer. That means the ESR is perfectly matched at 20mA.
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-