The LED lighting project

[bobofthenorth]

Some regulars here are aware that I have been working on an LED conversion project for our coach.  As of this afternoon that project is substantially complete.  For the 110 V. side of the coach the bulbs came from my favorite campground - Camp Wallymart.  For the 12 V. side it was considerably more complicated.  I ended up building my own modules and documented the process here:

LED Project

[viento1]

very interesting. I am surprised there is nothing comercially available.

[Len Silva]

Nice work Bob,

A friend told me about a similar project he was working on.  He mixed in a few yellow LED's with the whites and got a much warmer color without much reduction in light.  I think the ratio was about 1 in 5.

[bobofthenorth]

Viento - there's some marine stuff available - the Sailor Sam's link has some good stuff.  But its wicked expensive.  I've got about $10 of cash cost in each module.  Buying comparable luminosity commercially would be north of $100.  My time is worth something but its not worth that much. 

Len - we actually like the cold light.  I know that's not the norm but we prefer it to the "warmer" light that we had from the halogens.  We had them side by side in the sunshine ceiling in the kitchen for the past couple of weeks and found we preferred the LEDs.  But for somebody who wanted the warmer light you are absolutely right, it wouldn't be hard to mix in some yellow LEDs.  In my banks of 4 it would make the most sense to put 1 in 4 but it wouldn't be hard to rearrange the layout and there's other voltages available.  If you found a 2.5 nominal voltage then you could have banks of 5 with 1 in 5 being yellow.

[boogiethecat]

Oh gawd, here we go again.

Bob, your homemade board is nice and it may work, but I don't see any resistors in your concoction, and you are talking about adding LED voltages up to make them equal to your battery voltage... bad bad bad.

Why?  Led's are current driven, not voltage driven as incandescents are.  This appears to be a difficult thing for a lot of people to grokk, and I have to admit that if you don't have a concept of current sourcing, it's equally as hard to explain.  But I'll try...

An Led's operating point (and it's destruction point) is defined by the current you put through it, NOT the voltage you put across  it.  An Led's voltage is a printed spec that is a result of the physics of the material it's made of and the color it is, (and it's temperature) and this does NOT mean that this is the voltage it requires.  It's kinda like this- if you put the proper amount of current through it, and then measure the voltage across it, you'll see roughly the spec'd voltage.  But the drop across the LED has absolutely nothing to do with how you should drive it.   It's only an artifact of being properly driven, as a specification it's useful to a designer but it's NOT an operating voltage in the sense that it would be for an incandescent.

Case in point, your cool little boards.  You mentioned that they draw .32 amps (320Ma) per board, and looking at your board, you have 4 "strings" of four leds on each board... that means each string "section" is drawing .32 amps / 4 =  80 milliamps per set of 4 leds. This means that each LED on your board has 80Ma going through it!~!  Now if you look at your LED's data sheet, you'll see that it recommends an operating point of only 20 milliamps per led... this means that you are overdriving every led on your board 4 times more than the maximum recommended current!!!  This is a major no-no design-wise...you are way overdriving every LED and that will cause early failure, or at worst possible burn/fire because they will likely be running WAY too hot.  You need to take control of the current going thru each LED so it doesn't exceed 20ma...

How?

In short, to gain control of the current you are putting thru your led's you need a resistor in series with each "string" at best, or at least a resistor in series with your whole board.  In your case, since your boards are already done, you probably can get by with putting a resistor in series with each board whose value  makes the boards draw 80 Ma instead of 320.  Yes your Led's will get a lot dimmer than they are, but they won't fail because they will no longer have 4 times their rated current being rammed down their poor little anodes!. As it is you're headed for disaster someday.

IF you were doing this board over again, you'd probably want to put 4 resistors on it... one for each "string".
To calculate the resistor, you'd want to stack up your LED's so that their total "specification" voltage is maybe 3-4 volts less than your vehicle's system, and then use ohms law to calculate a resistor that will drop the remaining voltage whilst limiting the current thru the led's to that given in their spec sheet as the proper driving current (volts across resistor/current=needed resistance).  Then you again use ohms law to calculate the wattage requirement of the resistor so that it doesn't burn up (voltage across the resistor x current = watts of the resistor).
Being that your LEDs are typically 3.2 volts, I'd only put 3 in series instead of 4 (3x3.2=9.6v) and then calculate the resistor to soak up what's left between the 9.6 and your bus's 12-14.


Clear as mud?

Here's a webpage that tries to explain it.
http://white-leds.co.uk/led-wiring-guide.htm

A great calculator for all of this stuff is here:
http://www.theledlight.com/resistancecalculator.html
You can add up the voltages for the led's in the string and put the total in the "led volts" box and it will calculate the rest for you.
Quite nice....

Bottom line, your led's appear to work but their lifetime is going to be severely degraded by the way you've hooked them up.  A resistor is
required to do this right, period....

...............................................
Philosophy:
 
Though I'm not a "don't do that- the sky will fall" kind of guy... and in general I love to go by "what works is what works"... isn't the whole idea of using LEDs mostly because they don't burn out? ~reliability~

My take: ALL lighting on a bus is for safety.  If it's for safety, it should be done correctly, and head-in-sand is not an excuse for lack of understanding or poor engineering.  A burned out LED could be the beginnings of an accident, be it a full on crash or only a stubbed toe in the dark,  but throwing ohms law out the window because  "it works" or you simply don't understand it is not a valid reason for doing a half assed engineering design on a safety related system.

So PLEASE Bob and all of you guys... take a moment to either learn ohm's law and how to drive your LED's correctly, or ask questions until you understand, or probably go back to light bulbs.  Bob, your method appears to work, yes, but it's a bad design.  I would NOT recommend others reading your post and trying it... even though your experience is that it works...

My apologies if I sound harsh... I'm not trying to scold, but I am trying to get you guys to understand that with LEDs there is more to it than just "making it work"... if you learn to do it right, you'll have a wonderful system.  If not, you will likely have failures, or worse, accidents, fires, etc.

[bobofthenorth]

It sounds like I have a bad design.  Which comes as no surprise since there was exactly zero time spent on design.  I'll take my chances - they don't run "hot" - I'll check that with my digital thermometer today just for giggles but they don't run noticeably hot.  If they fail too soon I'll have to deal with it. 

FWIW, the .32 amp figure was arrived at by multiplying the 16 LEDs x the 20 mAmp draw.  I'm not an electrical engineer (and don't want to be one either) but I think the correct calculation for my configuration would have been 4 x the individual draw.  I've now looked at both the websites cited and they clearly DO take into account the voltage drop across the LEDs.  There ain't no way around Ohm's law.  I'm not convinced that this "problem" isn't akin the bumblebee.  You know the story - engineers can prove that the bumblebee can't fly but he can't read the proof so he just goes happily bumbling along - flying.

[boogiethecat]

Oh come on... did I just waste my evening typing all of that?  This is one of those little things in bus-land that should be done right, or not done at all.  There's nothing wrong with learning, nothing wrong with doing it right, nothing wrong with making a mistake and correcting it.

Just go get yourself a resistor, stick it in series and have a good design, and don't take your chances.  You ARE overdriving them, and their temperature right now or any other day won't tell you a thing.  It's the temperature of the LED die itself that matters and you can't get to that. And there's a million reasons that I won't waste my time typing that the temp. doesn't tell you anything useful.

Look... try this..  get yourself a handfull of ten ohm half watt resistors and stick one in series with your board, then measure the current. See what it is.  If it's even close to 80ma, you're done.  It'll probably be very close depending on your battery state.

My Guess is this- you have 4 led's that are 3.2 volts nominal at 20ma, this means 12.8 volts total and your bus is probably somewhere between 13.2 and 14.6... but since you have 4 strings you need 80ma to get the individual current correct.
Typing those values into the calculator I gave you above says you need roughly 20 ohms when your battery is high and 5 ohms when it's low (this is because your led string total is too close to your battery voltage... had you used three led's this range would be a lot less)

I would simply use a 10 ohm 1/2 watt resistor- it's a compromise but SO much better than nothing... your led's will like you and I will applaud you.

If you can't figure out where/how to get some resistors, I'd be happy to send some to you for free, in the interest of seeing your project be done properly.  Just email me your address....

Cheers
Gary

[viento1]

Boogie,

I heard you loud and clear... Thank you for taking the time and will give a full report on my project when done.

[bobofthenorth]

I heard you loud and clear too.  And I don't take well to being lectured without a complete explanation of the theory. 

Boogie stated:
"get yourself a resistor, stick it in series and have a good design".  Actually I still wouldn't have a "good design" because a "good design" with respect to LEDs would require a current limiting circuit.  Since that isn't practical for most lighting applications we work in the realm of poor compromises and use resistors.  The real issue with LEDs appears to be that the relationship between current and voltage isn't linear (IOW they behave as if they have a variable resistance) and that they break down with temperature.  So the risk is that you get a runaway inside the LED that starts with a voltage spike and results in a non-linear decline in resistance which leads to higher current flow which leads to more heat ...........   

I believe that I am operating in a safe range but time will tell. As I have already indicated I will monitor the temps to see if there is an indication of a problem.  Is a better design possible?  ABSOLUTELY.  But there's a lot of bad designs in this world that function just fine, starting with the giraffe.  Engineers don't like to admit that every design involves some kind of compromise and safety margin - each of us who designs our own equipment has to come to terms with where we are on that continuum.

Here's an exercise for you Boogie:  plug the actual nominal voltage of my LEDs (3.3) into your calculator and tell me what size of resistor to use at my current coach voltage (13.0). 

[Sean]

I heard you loud and clear too.  And I don't take well to being lectured without a complete explanation of the theory.

Bob,

I have to go along with Gary on this.  To be fair to him, he did provide a link to the explanation of the theory, which is sort of second-nature to those of us who had to reduce the entire universe to Thevenin equivalent circuits for a whole semester.

I'll spare you any additional lecture, and simply say that I believe, if you have no resistors at all in your modules, that you will see LED failures in relatively short order.  Every minute you are running those, more material is migrating across the junctions.  I know you have a lot of work in these modules, and I'd hate to see you have to start all over -- even replacing one LED looks to be a bit of work (and, by the time one LED in a module fails, the others will have migrated so much material that their failure is imminent).  I will go further to say that the only reason they are working at all (and did not fail immediately) is that the length of wire in the lighting circuit represents, at least, a small resistance, even if it is in the range of milliohms.  BTW, relying on those few milliohms to drop the (varying amount of) leftover voltage in the circuit probably also means that you are dissipating a lot of unnecessary power in the wiring.

I'm not sure I agree that this is any kind of safety risk.  Even incandescents burn out, and any heat generated even by overdriven LEDs is not nearly the fire danger that the halogens you replaced presented.  I'd just hate to see all your hard work (and the cost of your LEDs) go down the tube in a matter of just a few months, when, properly limited, those LEDs should last the lifetime of the coach.

FWIW.

-Sean
http://ourodyssey.blogspot.com

[boogiethecat]

Thanks Sean
Bob, my apologies if I appeared to be harsh.

Bob,  I don't need the excercise, I've been an electronic designer for 40 years.  And I guess I have to ask you please don't start preaching to me about stuff you obviously don't have a clue about.  I didn't write you to lecture you. I wrote you inform you and help you understand something that you and many others obviously don't.

To respond to your last post:

For what it's worth, a resistor IS a "current limiting circuit".

>Since that isn't practical for most lighting applications we work in the realm of poor compromises and use resistors<

No Bob, the use of resistors is the way it's done in the industry.  When you get into cellphone backlights and higher powered devices, then some electronic current limiting becomes necessary-usually using switcher topology- to save energy.  But for almost any "common" led you'll find a resistor behind it.

>The real issue with LEDs appears to be that the relationship between current and voltage isn't linear<

No Bob, the real issue is that LED's are manufactured with a VERY SPECIFIC maximum operating CURRENT and you or anyone else in the world should design to meet that spec.  It has absolutely NOTHING to do with an LED's linearity etc.   

>I believe that I am operating in a safe range but time will tell<
...no Bob, you're not.  Without current limiting of some sort, there is no "safe range"...

I think you should take some Led taillights apart and look inside.  Good designs, using resistors. 
All I can offer at this point is, I certainly hope that you don't apply the same mentality to
tightening lug bolts...
I'm done....

[Dreamscape]

Gary, I saved your information for further studying. Thanks for taking the time to do that, I'm sure it wasn't easy.

I had not thought of any LED projects, but with this information I might consider a small one.

~Paul~

[Van]

Bob,the arrangement looks bitchin but these guy's are right,I know ,I did a similar design for a set of foot pegs on the M/C a few years ago.The end result was pre mature failure ,a simple test should solve any dispute you might have,hook them up to a steady power source equaling your coach's and leave them hooked up,and note the length of their performance.Hope this helps and all works out for ya.

  Van

[Nick Badame Refrig/ACC]

Hi Bob,

About a year ago, I purchaced 16- G4 LED clusters from Salior Sams to replace my 10 watt halogen bulbs in my coach.

2 problems I had. 1- the G4 clusters didn't fit very well in the puck housings.  2- I couldn't get used to the light that they emmited.

Needless to say, I removed them all and went back to the halogens.

Good Luck
Nick-

[bobofthenorth]

We found the same Nick - the so-called replacements didn't fit in the puck housings.  I did some considerable alterations involving the hot glue gun, the soldering iron and a trip to WallyMart.  We now have what I think are fairly attractive replacements for the pucks.