+ Post New Thread
Results 1 to 11 of 11
  1. #1
    Advanced Member level 5
    Points: 31,585, Level: 43
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    6,602
    Helped
    478 / 478
    Points
    31,585
    Level
    43

    Dimming LEDs..via PWM dimming or Analog dimming?

    Hello,
    Ive just spent the day dimming “warm white” and “cool white” LEDs via analog dimming, and then pwm dimming…

    The intended product is eventually for a desk lamp.

    I cannot see any difference between the colors no matter what dimming method is used…if anything, the analog dimming looks better. Please can we here now debunk this “myth” that "PWM dimming" gives a better looking light for the LEDs when dimmed?..it is surely nonsense..if it were true , then why do the led manufacturers not demonstrate it on a video?…it is not shown anywhere..it Is surely a myth?

    •   AltAdvertisment

        
       

  2. #2
    Super Moderator
    Points: 28,814, Level: 41
    andre_teprom's Avatar
    Join Date
    Nov 2006
    Location
    Brazil
    Posts
    8,587
    Helped
    1087 / 1087
    Points
    28,814
    Level
    41
    Blog Entries
    6

    Re: Dimming LEDs..via PWM dimming or Analog dimming?

    As far as I know, each LED is designed using specific components to bright a very small range on the light spectrum, actually defined by a discrete transition of the electron between two energy levels of the atom, so that, assuming a high enough PWM frequency, should not make any difference in the human perception of the color shift.
    --------------------------------------------------------------------------------------------------
    Part of the world that you live in, You are the part that you're giving ( Renaissance )


    1 members found this post helpful.

  3. #3
    Super Moderator
    Points: 253,898, Level: 100
    Awards:
    1st Helpful Member

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    44,245
    Helped
    13463 / 13463
    Points
    253,898
    Level
    100

    Re: Dimming LEDs..via PWM dimming or Analog dimming?

    The color shift of white LEDs isn't a myth, it's specified by manufacturers, see e.g. http://www.osram-os.com/Graphics/XPi...GaN%20LEDs.pdf, diagram 8 and 9.

    It's however small, and I believe that you may have missed it when the test setup isn't sensitive. But there are surely LED applications where the color shift matters.


    1 members found this post helpful.

  4. #4
    Advanced Member level 5
    Points: 31,585, Level: 43
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    6,602
    Helped
    478 / 478
    Points
    31,585
    Level
    43

    Re: Dimming LEDs..via PWM dimming or Analog dimming?

    But there are surely LED applications where the color shift matters.
    You mean like in Ultraviolet LED applications?..surely not in room or space lighting?

    Also, I am sure that anyone would agree, that if this effect of wavelength shift is not discernible to the human eye, then its a waste of time thinking about it for eg office or home lighting applications. And as far as I know, there is not one single video on the internet which demonstrates a discernable color shift in LED light when the current in the led is reduced by "analog dimming".

    Not one video...not one demonstration...anywhere....if anyone is doing office or home lighting, then choose either pwm or analog diming, -please yourself, they both look just the same....well, in my opinion, "analog diming" just seems to look slightly better!
    Last edited by treez; 4th June 2015 at 22:33.



  5. #5
    Super Moderator
    Points: 253,898, Level: 100
    Awards:
    1st Helpful Member

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    44,245
    Helped
    13463 / 13463
    Points
    253,898
    Level
    100

    Re: Dimming LEDs..via PWM dimming or Analog dimming?

    I'm exclusively talking about LED lighting and applications where color temperature is a critical parameter. Graphic industry, product design, textile industry and many more.


    1 members found this post helpful.

    •   AltAdvertisment

        
       

  6. #6
    Advanced Member level 5
    Points: 31,585, Level: 43
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    6,602
    Helped
    478 / 478
    Points
    31,585
    Level
    43

    PWM dimming or Analog dimming give same appearance for CXA3050 LED?

    Hello,
    Is it true that the frequency spectra of the CXA3050 LED will be the same whether it is PWM dimmed or Analog dimmed?…

    CXA3050 LED Datasheet:-
    http://www.cree.com/~/media/Files/Cr...%20CXA3050.pdf


    I ask because page 4 of the following shows that this LED is constructed of a phosphor coating over the actual LED die, as shown on page 4 of the following…

    Page 4 shows phosphor coating over LED:-
    http://www.cree.com/~/media/Files/Cr...aintenance.pdf


    -so is it true that there will be no difference in the light spectra of this LED whether PWM dimming or analog dimming this LED.?



  7. #7
    Advanced Member level 5
    Points: 34,805, Level: 45
    Achievements:
    7 years registered Created Blog entry Created Album pictures
    SunnySkyguy's Avatar
    Join Date
    Sep 2007
    Location
    Richmond Hill, ON, Canada
    Posts
    6,702
    Helped
    1672 / 1672
    Points
    34,805
    Level
    45
    Blog Entries
    7

    Re: PWM dimming or Analog dimming give same appearance for CXA3050 LED?

    ALmost. x,y spectral shift always occurs when exceeding max avg current or high temp or aging with Lumen maintenance by a drop in phosphor emissions or commonly called Blue shift.

    If the peak current is not much more than the average analog current, there will be no difference, otherwise there will be, but how much depends on your sensitivity or tolerance to blue shift, governed by Ipk/I avg.

    Note that saturation of photo emission causes a declining slope of Lumens or Luminosity vs Current. THis part due to BLue LED saturation and partly due to reddish-yellow phosphor saturation. Same with temp rise and aging.

    If the ratio of red & yellow to blue changes. then that CCT changes. That's explains everything in COlour temperature using CIE eye corrected intensity for each spectral peak. However instruments use either x,y or Cx,Cy colour coordinates. where 0.36, 0.36 is somewheres about neutral white at 4500K
    A good design question lists your overall requirements™ The best question deserves a better answer. ™
    ... Tony Stewart EE since 1975


    1 members found this post helpful.

  8. #8
    Advanced Member level 5
    Points: 13,005, Level: 27

    Join Date
    May 2015
    Location
    Melbourne, Australia
    Posts
    2,213
    Helped
    752 / 752
    Points
    13,005
    Level
    27

    Re: PWM dimming or Analog dimming give same appearance for CXA3050 LED?

    Quote Originally Posted by treez View Post
    Hello,
    Is it true that the frequency spectra of the CXA3050 LED will be the same whether it is PWM dimmed or Analog dimmed?…
    All bare LEDs are monochromatic, in other words they emit only a single specific colour clustered around a specific single wavelength.
    White light is made up of a wide spectrum containing many colours, so a LED cannot do that all by itself.

    The way white LEDs work, is that an ultraviolet LED irradiates a white phosphor which creates the visible light. The LED can only produce light in the ultraviolet region, about 450nM.

    Fluorescent tubes work in the exact same way.

    No matter how you drive that LED, it can only produce energy at 450nM anyway.
    So nothing changes between driving it with dc or pulsing it with PWM.

    The characteristics of the phospor are something else. This does change over time and different batches of LEDs, or LEDs from different manufacturers.

    But dimming, either by analog or digitally will not look any different provided the PWM is fast enough to avoid flicker, which is not all that fast.
    Cheers, Tony.


    1 members found this post helpful.

  9. #9
    Advanced Member level 5
    Points: 34,805, Level: 45
    Achievements:
    7 years registered Created Blog entry Created Album pictures
    SunnySkyguy's Avatar
    Join Date
    Sep 2007
    Location
    Richmond Hill, ON, Canada
    Posts
    6,702
    Helped
    1672 / 1672
    Points
    34,805
    Level
    45
    Blog Entries
    7

    Re: PWM dimming or Analog dimming give same appearance for CXA3050 LED?

    Actually LED's are similar to FL tubes except the White LEDs use 450nm = BLUE with 1 or 2 phosphors and tubes use UV = 315nm with 2 or 3 phosphors.

    At high lumen flux density the phospor efficacy reduces so there is a slight blue shift.
    Depending on the duty cycle the apparent blue shift will be higher for dim light but it depends on how hard you are driving them.
    If 10% duty cycle at peak Amps has a drop in phosphor efficacy of 10% compared to DC at 10% of the peak current, there will be blue shift. You would only notice the difference with two side by side or using a spectrometer.
    .
    But one thing for certain emitter intensity is not linear with current but it is mostly due to the BLUE LED but there at least 5% drop due to the phosphor, at 1.4A based on PWM which translates into some rise in CCT or blue shift.
    It depends on factors mentioned already, but I suspect as much 500'K rise in CCT at 1.4A or an XY drop of 0.005 of x,y from low current DC to 1.4A pulsed. See page 19


    The white LEDs have a thick BLue transparent substrate with a relatively thin phospor coating which must be critically balanced to achieve the desired CCT.
    Partial conversion of energy from blue @450 nmD to the orange red range depending on chip specs and colour temp CCT

    Cree make pretty good LEDs so dont push the peak currents beyond what you absolutely need , and you will be fine, but if you plan on using the max 1.4A, you better have your mechanical and thermal design match the requirements.

    If you only plan to use 330 mA, then it will be stable colour same as DC.
    A good design question lists your overall requirements™ The best question deserves a better answer. ™
    ... Tony Stewart EE since 1975


    1 members found this post helpful.

    •   AltAdvertisment

        
       

  10. #10
    Super Moderator
    Points: 253,898, Level: 100
    Awards:
    1st Helpful Member

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    44,245
    Helped
    13463 / 13463
    Points
    253,898
    Level
    100

    Re: PWM dimming or Analog dimming give same appearance for CXA3050 LED?

    I merged the recent discussion with the previous same topic thread because it has additional material.

    The Osram application note gives an idea of the expectable color temperature shift. The fact that the presently discussed Cree datasheet doesn't specify a shift doesn't mean that it doesn't exist. But it's probably small enough to be ignored in general lighting applications.


    1 members found this post helpful.

  11. #11
    Advanced Member level 5
    Points: 34,805, Level: 45
    Achievements:
    7 years registered Created Blog entry Created Album pictures
    SunnySkyguy's Avatar
    Join Date
    Sep 2007
    Location
    Richmond Hill, ON, Canada
    Posts
    6,702
    Helped
    1672 / 1672
    Points
    34,805
    Level
    45
    Blog Entries
    7

    Re: PWM dimming or Analog dimming give same appearance for CXA3050 LED?

    All LED's shift dominant wavelength with current , but the question depends on how much is noticeable?

    My client is fussy and expects the colour temperature within a 500'K range around 4500.

    BLue shift is the standard term to indicate CCT shift with rising current, aging or increased junction temp.

    THis occurs not just because of the BLUE substrate shift, which is small in the normal operating range 5~10nm but moreover the declining efficacy of the Red phosphors at higher light flux levels beyond maximum DC currents when used in flash mode.

    So long as the forward current through an InGaN LEDs remains constant, no wavelength or color shift occur with PWM. Stated in CRee's Design guide.

    The problem is of course if one uses PWM to regulate current and operates much lower than 90% duty cycle to get absolute max radiated power, then the crest factor or peak/average ratio will reduce phosphor emissions more than the Blue emissions resulting in "blue shift" or apparent shift towards bluish.

    In your case, I doubt you will PWM a (err) 50 Watt chip with this method. e.g. 36V at 1400mA is DC rated at 85'C with suitable heatsink size of CPU type.. But if you ran at limit of 2500mA with a duty cycle of 1400/2500 or 56% to get maximum brightness with occasional bursts of 2500mA for a flash camera, I guarantee you would notice the difference side by side in CCT color temp. Ideal temp for me is 4000~5000. Others prefer warm, nobody likes cool 6000+

    So there is no worry, but it is not a myth if you do.
    A good design question lists your overall requirements™ The best question deserves a better answer. ™
    ... Tony Stewart EE since 1975


    1 members found this post helpful.

--[[ ]]--