Forums Power and Sound Power and Sound DCC
  • Topic: LED headlight

    Back To Topics
    (0 rates)
    • October 15, 2017 12:33 AM EDT

      •  
      • Posts
        9,393
      • Thanks
        249
      • Thanked
        862

      Yes, but the off time is miniscule, not what people are talking about with PWM... That miniscule off time is not sufficient to make a difference... remember this phenomenon is color shift by a different temperature in the LED junction.

      This miniscule off time is maybe a 99% duty cycle, instead of 100%... the thing you are reading about is using PWM to dim, so you are pulsing it with full track voltage but maybe like 50% or less on time. Just not the same thing.

       

      And you guys are still making the incorrect assumption that the power from the decoder to the light is raw unfiltered DCC as opposed to rectified, and normally there is a small filter cap there.

      And also the incorrect assumption that you have PWM output on the headlight function.

       

      The difference in voltage is from probably 7-10 volts (I'm sure you were not running the loco as full speed on DC) to 16 on DCC... so you have a difference in applied voltage of 9 to 6 volts... that is significant.

      Just to be sure, you are running the LED from the decoder blue and the headlight wire.

      So, on DC, unless you were running flat out, you weren't running the same voltage/current between DC and DCC.

      Take that loco and put it on rollers and feed it 16 volts DC.... I'll bet the LED is the same color.

       

      Greg

       

      p.s. If you looked at the waveform Dan, you surely noticed that the square wave had some ringing and overshoot also... all that can affect the max voltage the decoder sees. Look at 43 seconds on... realize this is NOT rectified, so you cut the signal in the middle, and flip the negative parts to positive

       

       https://www.youtube.com/watch?v=zJL2TEyNCwQ,

       

       

       

      This post was edited by Greg Greg Elmassian at October 15, 2017 12:51 AM EDT
      ____________________________________

      Be sure­ to visit ­my site, l­ots of tec­hnical tip­s and modi­fications,­ and you c­an search ­for topics­ and key w­ords.


      ­Click HERE for Greg­'s web sit­e
      PLEASE NOT­E: Please do NOT use private messaging, i­f you have­ a questio­n, feel fr­ee to emai­l me priva­tely, u­se regular­ email onl­y: greg@el­massian.co­m

    • October 15, 2017 6:46 AM EDT
      • Kittery, ME
         
      • Posts
        670
      • Thanks
        46
      • Thanked
        274

      I think we can all agree that putting a cap across the LED isn't going to do any harm, and is worth trying.  As intriguing as it may be, we're probably never going to know what the waveform looks like.

    • October 15, 2017 7:44 AM EDT
      • Pittsburgh, Pennsylvania
         
      • Posts
        11,130
      • Thanks
        127
      • Thanked
        919

      Greg Elmassian said:

      Yes, but the off time is miniscule, not what people are talking about with PWM... That miniscule off time is not sufficient to make a difference... remember this phenomenon is color shift by a different temperature in the LED junction.

      This miniscule off time is maybe a 99% duty cycle, instead of 100%... the thing you are reading about is using PWM to dim, so you are pulsing it with full track voltage but maybe like 50% or less on time. Just not the same thing.

      And you guys are still making the incorrect assumption that the power from the decoder to the light is raw unfiltered DCC as opposed to rectified, and normally there is a small filter cap there.

      No, I am not making that assumption.

      And also the incorrect assumption that you have PWM output on the headlight function.

      You made that assumption, I didn't.

      The difference in voltage is from probably 7-10 volts (I'm sure you were not running the loco as full speed on DC) to 16 on DCC... so you have a difference in applied voltage of 9 to 6 volts... that is significant.

      Just to be sure, you are running the LED from the decoder blue and the headlight wire.

      Yes

      So, on DC, unless you were running flat out, you weren't running the same voltage/current between DC and DCC.

      Take that loco and put it on rollers and feed it 16 volts DC.... I'll bet the LED is the same color.

      I ran it on DC, not 16 volts, but at normal speeds, and the pale yellow LED was pale yellow. On DCC its green. So there is something about the rectified DCC that is shifting the colour. And I have read reviews in publications like MR and NMRA's magazine that state the yellow headlight in production locomotives also glow green. That tells me that I am not the only one experiencing this phenomenon.

      Greg

      p.s. If you looked at the waveform Dan, you surely noticed that the square wave had some ringing and overshoot also... all that can affect the max voltage the decoder sees. Look at 43 seconds on... realize this is NOT rectified, so you cut the signal in the middle, and flip the negative parts to positive

      https://www.youtube.com/watch?v=zJL2TEyNCwQ,

       

      ____________________________________

      Shannon car Shops
      Home of the infamous leg lamp

      I.A.R.R.R. Member #12

      and King Butt Modeler

    • October 15, 2017 7:57 AM EDT
      • Pittsburgh, Pennsylvania
         
      • Posts
        11,130
      • Thanks
        127
      • Thanked
        919

      Eric Reuter said:

      I think we can all agree that putting a cap across the LED isn't going to do any harm, and is worth trying.  As intriguing as it may be, we're probably never going to know what the waveform looks like.

      Eric. I am going to try filtering the power to the LED on my 4-4-0 installs. While we may never have a definitive answer as to why the LED shifted its colour, I am a tech, I care about making things work, the why can come later.

       

      I just thought with the DCC experts we have here, and the fact that this phenomenon has been reported by others, I thought someone may know of a fix for this.  I didn't mean to start a poo throwing contest.

       

      Maybe HO DCC decoders don't have the same sophistication as large scale decoders. So maybe this issue hasn't been encountered by large scalers.

      ____________________________________

      Shannon car Shops
      Home of the infamous leg lamp

      I.A.R.R.R. Member #12

      and King Butt Modeler

    • October 15, 2017 11:38 PM EDT

      •  
      • Posts
        9,393
      • Thanks
        249
      • Thanked
        862

      Of course you did not compare apples to apples... DCC full voltage 16 volts all the time... and your DC track voltage is WAY below that.

      If you would read WHY there is a potential color change, you would see it is indeed the junction temperature. Gee, wonder what makes the temperature? Oh, it's the power dissipated... why would voltage have anything to do with it, no you gotta believe you have PWM which you do not. The modulated DCC signal is not getting "through" the full wave bridge on the decoder.

       

      Bottom line, running 7 to 10 volts DC as inputs is NOT the same as rectified 16 volt DCC.

       

      Good thing you are not working in my lab, my hardware guys would go nuts... you will still have apples to oranges until you try the loco on 16 volts DC. The voltage is the SIMPLEST thing you could try and you refuse to do it.

       

      Greg - over and out...

      ____________________________________

      Be sure­ to visit ­my site, l­ots of tec­hnical tip­s and modi­fications,­ and you c­an search ­for topics­ and key w­ords.


      ­Click HERE for Greg­'s web sit­e
      PLEASE NOT­E: Please do NOT use private messaging, i­f you have­ a questio­n, feel fr­ee to emai­l me priva­tely, u­se regular­ email onl­y: greg@el­massian.co­m

    • October 16, 2017 1:27 AM EDT
      • Pittsburgh, Pennsylvania
         
      • Posts
        11,130
      • Thanks
        127
      • Thanked
        919

      Refuse? No, I didn't refuse. I just didn't dig the locomotive out of the packed show box and set up a test track to see what happens.

       

      Bottom line. How do I make a pale yellow LED glow pale yellow on a DCC decoder? That is all I am trying to find out.

      ____________________________________

      Shannon car Shops
      Home of the infamous leg lamp

      I.A.R.R.R. Member #12

      and King Butt Modeler

    • October 16, 2017 7:57 AM EDT
      • Eastern Massachusetts
         
      • Posts
        1,220
      • Thanks
        9
      • Thanked
        107

      My thought is when comparing a light bulb to the LED, pulsed power of any kind will average out on a light bulb due to the filament being basically a resistor.

      The led will see the full pulse voltage and be driven with over its current rating, thus heating up the led.  This may change the color, but will certainly shorten the life.

      Using the CL2N3 will maintain a 20ma max. which is better for LEDs than a resistor on pulsed power.

    • October 17, 2017 1:54 PM EDT

      •  
      • Posts
        9,393
      • Thanks
        249
      • Thanked
        862

      I'm pretty sure that the output of the decoder is not pulsed... also no PWM option.

      Now, some of the older decoders where the "common" is one of the rails, I believe those were pulsed, because they did not use the internal rectified DCC.

      Greg

      This post was edited by Greg Greg Elmassian at October 17, 2017 1:56 PM EDT
      ____________________________________

      Be sure­ to visit ­my site, l­ots of tec­hnical tip­s and modi­fications,­ and you c­an search ­for topics­ and key w­ords.


      ­Click HERE for Greg­'s web sit­e
      PLEASE NOT­E: Please do NOT use private messaging, i­f you have­ a questio­n, feel fr­ee to emai­l me priva­tely, u­se regular­ email onl­y: greg@el­massian.co­m

    • October 17, 2017 6:10 PM EDT
      • Pittsburgh, Pennsylvania
         
      • Posts
        11,130
      • Thanks
        127
      • Thanked
        919

      Ok, I don't want to run a locomotive with a 12 volt motor on 16 volts DC. But at 12 volts DC, the pale yellow LED is pale yellow. Also, on my home DCC set up, with 12 volts going into the DCC base station, the LED is green. So, Greg, is that apples to apples enough for you???

       

      BTW, I was out of town for a few days with my job. If that translates into refusing to walk off my job, drive 4.5 hours, dig the thing out of the box in the back of my garage, and test it.......... 

       

      And if I was working in a "shop", I would hope it would be equipped with an oscilloscope. But being a home hobbyist, I can't justify the expense of getting one.

      ____________________________________

      Shannon car Shops
      Home of the infamous leg lamp

      I.A.R.R.R. Member #12

      and King Butt Modeler

    • October 17, 2017 8:30 PM EDT
      • Denver, Colorado
         
      • Posts
        2,305
      • Thanks
        356
      • Thanked
        656

      Greg Elmassian said:

      Of course you did not compare apples to apples... DCC full voltage 16 volts all the time... and your DC track voltage is WAY below that.

      If you would read WHY there is a potential color change, you would see it is indeed the junction temperature. Gee, wonder what makes the temperature? Oh, it's the power dissipated... why would voltage have anything to do with it, no you gotta believe you have PWM which you do not. The modulated DCC signal is not getting "through" the full wave bridge on the decoder.

       

      Bottom line, running 7 to 10 volts DC as inputs is NOT the same as rectified 16 volt DCC.

       

      Good thing you are not working in my lab, my hardware guys would go nuts... you will still have apples to oranges until you try the loco on 16 volts DC. The voltage is the SIMPLEST thing you could try and you refuse to do it.

       

      Greg - over and out...

       

          I've never understood why Greg E. is allowed to talk like this to decent people in our community. And I'm not making this comment privately to Bob behind anyone's back.

       

       

      ____________________________________

    • October 17, 2017 8:36 PM EDT

      •  
      • Posts
        14,795
      • Thanks
        2,751
      • Thanked
        1,612

      Me either considering it's Pumpkin and Squash season ? However you will need to also define "decent people in our community "

       

       

    • October 17, 2017 8:39 PM EDT
      • Denver, Colorado
         
      • Posts
        2,305
      • Thanks
        356
      • Thanked
        656

      " Rooster " said:

      Me either considering it's Pumpkin and Squash season ? However you will need to also define "decent people in our community "

       

       

        I may or may not put you or me in this category (!), but I would absolutely unequivocally qualify Maynard, the latest target, as decent people.

       

       

      ____________________________________

    • October 19, 2017 8:09 PM EDT

      •  
      • Posts
        14,795
      • Thanks
        2,751
      • Thanked
        1,612

      John Passaro said:
      " Rooster " said:

      Me either considering it's Pumpkin and Squash season ? However you will need to also define "decent people in our community "

       

       

        I may or may not put you or me in this category (!), but I would absolutely unequivocally qualify Maynard, the latest target, as decent people.

       

       

      It's all good John and you actually got Maynard to push the "Thank" button so you must have done something correct!

Forums Power and Sound Power and Sound DCC

    Icon Legend

  • Topic has replies
    Hot topic
    Topic unread
    Topic doesn't have any replies
    Closed topic
    BBCode  is enabled
    HTML  is enabled

Add Reputation

Do you want to add reputation for this user by this post?

or cancel

Ads by Google