LED garden lighting

[storrence]

I found this schematic at http://www.discovercircuits.com/H-Corner/AC-Powered.htm

and was looking for some help with it.

I'm looking to use about 50 of these LEDs.

http://www.mouser.com/Search/ProductDetail.aspx?qs=sGAEpiMZZMuus4pmcqQnzpzxtnP9dh%2Fpn0dtqnyf%2FeI%3



D

I plan to make a outdoor lamp out of bamboo by drilling holes the same size as the LED and spacing them about 1 inch apart. The wiring will be done inside the bamboo by breaking it vertically and them putting it back together.

My question is how many watts will these 50 LEDs consume used in a circuit with the .47uF 250V capacitor and 1K resistor. I'm looking to create about a 6-10 watt lamp to replace a 60 watt incandescent.

It says they are 30mA each but I'm not sure what voltage you would calculate that with to arrive at watts.

TIA!

[ghurd]

Those LEDs are not the best choice.  4000MCD and 60 degrees is a pretty poor choice.

Shop around for higher MCD, and maintain at least 60 degree viewing angle.  90 degrees or more is better.  

MCD is a function of viewing angle.  It can get confusing without alot of reading.

Avoid driving them at 30ma.

The circuit you linked to says 20ma. Good, if it is correct.

And 50 5mm LEDs will not replace a 60W bulb.

G-

[scottsAI]

4000 mcd, Viewing Angle: 60 deg

Convert mcd to Lumens:

http://led.linear1.org/lumen.wiz

Approximate luminous flux: 3.367 lumens each LED.

4volts at 30 ma is 0.12w. 50X = 6 watts.

Lumens efficiency is 3.367 / 0.12 = 28 lumens/watt

If you drive LED below rating efficiency will be less and will be dimmer. Rated current will give good life if properly heat sinked. Heat is the killer long term.

Twice as good as incandescent, yet the current limit resistor will waste a lot of power if your not careful. Many times negating the benefits of LED.

Like ghurd said, better LED are out there, 60lumens/w not too costly.

The 50 LEDs are not as bright as 60w bulb (800 lumens). 3.37 * 50 = 168 lumens Lots of difference.

Compact fluorescent 13w rated at 800 lumens, twice the power 4x the lighting.

Beware of false claims: http://cgi.ebay.com/30p-8mm-HP-0-5W-Multi-Chip-White-LED-100Kmcd-FREE-SHIP_W0QQitemZ350167395245QQcm
dZViewItemQQptZBI_Electronic_Components?hash=item350167395245&_trksid=p3286.c0.m14&_trkparms
=72%3A1205%7C66%3A2%7C65%3A12%7C39%3A1%7C240%3A1318%7C301%3A1%7C293%3A1%7C294%3A50

Device is 0.5w at 100,000mcd with 140 viewing angle.

According to above calculator: 400lumens/w or 200lm/w with 50% power at 140 deg.

Nothing is that good.

Have fun,

Scott Beversdorf.

[wooferhound]

You can convert any LED to a 120 degree spread just by using a wire cutter and cutting the round plastic end off of the LED leaving it flat.

This is an LED Array Wizard. Just input a few things about your power and LEDs and it will tell you how to build your array and calculate the total wattage for you too.

http://led.linear1.org/led.wiz

[richhagen]

I was thinking of working out the equation to lumens per Watt but you beat me to it.  Compared with the latest offerings from Cree, Lumileds and others, this is not too good.  I soldered a few long series of LED's together once.  The strings failed, and I found that many were failed to short circuit, which led me to believe that they failed and increased the forward voltage across the remaining, and thus the current through the limiting resistor and remaining led's to the point of causing failure in others.  IIRC four out of five strings failed. I am subsequently not in favor of running long strings of LED's in series, but would recommend a switching power supply to bring the voltage down so that a smaller number can be used in series.  Rich  

[dnix71]

Why not just buy a bright white Christmas light string and mount it inside the garden fixture? They are already properly wired and made for outdoor use.

http://www.ledropelightsandmore.com/led_string_lights/

[storrence]

Hi ghurd,

Well I already made that purchase so they will have to suffice for the first round of tests If you know of a better choice/source please tell! I just had a pending order at mouser for the capacitors so just got them at the same place.

Understand I'm a musician first and electrician second http://f11music.com/

This page had a lot of helpful info regarding LEDs but didn't answer one question I have.

http://wolfstone.halloweenhost.com/Lighting/litlpo_PoweringLEDs.html#UsingResistorsToLimitLEDCurrent

I know it's not good to use AC directly on LEDs as they talk about in that link above.

The circuit at http://www.discovercircuits.com/PDF-FILES/acwhiteleds.pdf seems to solve that by putting the inline capacitor.

I think with my 126V 60Hz line I will be limited to about 68 white LEDs so I will just double the circuit.

[ghurd]

You have a misunderstanding about the capacitor.

It sort of limits the current through the LEDs.  (I would Not trust it)

It does Not solve the problem of reverse voltage on the LEDs.

Standard White LEDs have a forward voltage of about 3.1V The reverse voltage rating is 5V, meaning more than 5V backwards blows them up.

If the forward voltage is 3.1V on the LEDs pointing to the right, then the reverse voltage applied to the LEDs pointing left is 3.1V.

The reverse voltage is kept below 5V because the "other" LEDs keep it at 3.1V.

Do Not trust the data sheets regarding the white forward voltage at 3.4V.  It is a sure way to ruin the LEDs.

V RMS is something to consider.  126VAC RMS line voltage peaks at 178V.  It is almost like 178VDC for a micro-second.  I believe that is why most home-brew high voltage AC long strings of LEDs fail.

Consider going all the way back to the basics if you want the LEDs to last for years instead of weeks or months.

I would recommend finding a 12 to 18V transformer (16V door bell transformer?), rectify and filter the output to DC, design the circuit with more accurate numbers, test the real ma in the circuit, and keep each LED under 18ma.

Or a switching power supply (laptop?), then repeat above.

I have a feeling most of the online LED circuits were tested for a day or "a whole week!" before the circuits were published.

If I built the circuit shown in the pdf, I would reduce the cap value to 0.22uF, change the 1K resistor to 3.9K, and remove the jumpers between reversed LED pairs.

I would want the AC current reading to be under 14ma, probably about 12ma.

Then test it.  A Lot.  Before making any changes.

G-

[storrence]

Thanks Scott.

If I use enough LEDs is it possible to do this without a limiting resistor or capacitor? Sort of like putting 10 12V lights in series on a 120V line.

I'm figuring i need about 100 LEDs to get the lumens I want. I know you shouldn't use a/c on LEDs but like the circuit in my original post I could put 50 in one direction and 50 in the other. Would I still need a limiting resistor?

http://www.discovercircuits.com/H-Corner/AC-Powered.htm

[storrence]

Nice site and maybe something there could work but one reason I wanted to do it myself is so I could have the LEDs perpendicular to the bamboo surface to maximize the light coming out. Also I thought it would be cheaper. I found a source for 4 cent LEDs from Vietnam if I get 2000 so that would make it very cost effective. 100 LEDs $3 and less than $1 for the rest. I need to make about 20-30 of these lights.

[Opera House]

I know it says that, but I have used AC on LEDs without a diode for years without problems. Never seen one damaged by reverse voltage yet.

[Opera House]

LEDs are a little like zeners.  If the voltage is too low, they won't light up, too high and hey will draw too much current.  Remember you are also dealing with a sine wave.  That will control how long a period the led is actually on.  A capacitor in series wonderfully solves many problems.

[ghurd]

I have experienced dramatic catastrophic high-decibel failures.  

That said, I ran a red LED direct from a stepper's AC output for ~9 months without smoke, and the Vr was certainly over 20V at times.

The stepper drive tire eventually got a flat, but the LED was fine.

Vietnam LEDs?  You get what you pay for if you are lucky.

G-

[Opera House]

This falls into stupid engineer tricks.  Over twenty years ago I saw a LED operated from line voltage with only a resistor.  At that time I did a test and could find no noticeable reverse leakage.  About a year ago this came up again on an engineering forum.  Again I performed with a dozen different LEDs and a 33K resistor in series.  Ran them over two weeks.  Then inserted a diode so LED only saw reverse voltage.  Leakage current was no more than you would see with a 1N4004.

[storrence]

Thanks everyone for all the suggestions!

I would say my objectives at this point are:

1.  Energy Efficiency. (the most important point)
   
2. Minimal parts which hopefully translates to minimal cost
   
3. life of LED is important
   
4. I prefer many smaller or weaker sources of light instead of a few higher power LEDs. This is more for the effect but I don't want to go over 100 LEDs unless I can get a good deal on them. I'm not trying to really light the area that much. It's more for ambient light and effect.
   
   

Something I would like to understand better is how LEDs are different than incandescence lamps in regards to resistance. They generate heat so I assume resistance so shouldn't there be an amount of LEDs I could put in series that would bring the voltage down to the right voltage for the LEDs. I guess 3v. The same way you can string enough incandescence to get to the right voltage for each lamp. If each LED will be at 3 volts then in my case I would use 126V / 3V = 42 lamps.

This site someone suggested: http://led.linear1.org/led.wiz came up with:

126. Source voltage (I would need to first rectify to DC)
     
3. (diode forward voltage
   
20. (diode forward current (mA)
    
42. LEDs
    
    

The schematic is 42 leds in series with a 1 ohm resistor. Seems this would be a good solution. Would it still be a good idea to use an inline capacitor?

To rectify the A/C what would be a good diode to use to make a bridge? I have many 1N914 if that would work.

[storrence]

Schematic:

Solution 0: 42 x 1 array uses 42 LEDs exactly

    +----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>
|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>
|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>
|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>
|----|>|----|>|---///----+  R = 1 ohms

The wizard says: In solution 0:

  each 1 ohm resistor dissipates 0.9 mW

  the wizard thinks ¼W resistors are fine for your application

  together, all resistors dissipate 0.9 mW

  together, the diodes dissipate 3780 mW              

  total power dissipated by the array is 3780.9 mW      

  the array draws current of 30 mA from the source.

[Opera House]

"I have many 1N914 if that would work."

If you put about six in series, that would make an equivalent diode.  stll a very bad choice.

I did the diodes in series and got a much different number......

Of course I chose 140V.  That design wizard is just enough to be dangerous!

[wooferhound]

Here is Commanda's 5 part LED Master Class . . .

http://www.fieldlines.com/story/2007/4/20/95351/1614

http://www.fieldlines.com/story/2007/4/26/82929/6080

http://www.fieldlines.com/story/2007/5/1/74856/25992

http://www.fieldlines.com/story/2007/10/13/141930/41

http://www.fieldlines.com/story/2007/10/31/848/94375

I think that the cheapest, most efficient lighting can be gotten at a larger Home repair store, or even Walmart, in the replacements for fluorescent light bulbs. Little 1 foot long fixtures with fluorescent tubes in them that run on 12vdc or batteries.

http://www.fieldlines.com/story/2007/6/28/16314/5486

[storrence]

The mast class was very interesting but I think it applies more to battery system where you have varying voltage. In my situation my voltage will be very stable at about 126v.

Also the rope lighting will be less efficient as the leds will not all be pointing where I want. Putting individual ones in bamboo will allow me to direct the light outwards.

What I think I'm going to do for my first test is just run enough in series to not need a resistor.

Schematic:

Solution 0: 42 x 1 array uses 42 LEDs exactly

    +----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>
 |----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>
 |----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>
 |----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>|----|>
 |----|>|----|>|---/----+  R = 1 ohms

The wizard says: In solution 0:

  each 1 ohm resistor dissipates 0.9 mW

  the wizard thinks ¼W resistors are fine for your application

  together, all resistors dissipate 0.9 mW

  together, the diodes dissipate 3780 mW              

  total power dissipated by the array is 3780.9 mW      

  the array draws current of 30 mA from the source.

The 1 ohm resistor I will not use.

Then to deal with the AC can I just run another 48 in the opposite direction? Something like this:

~<::::::::::::::::::::::::::::::::::::::::::::::::>~

That would double my light output and also protect from reverse voltage on the LEDs?

[Opera House]

Can I come and watch the smoke?

The last time I helped someone with an LED I said, "use a 33-47K ohm resistor in series with the LED."  That was quickly interpreted as it being ok to use a 47 ohm 1/4 watt resistor.  It turned into a fireball when he tried it!  Constantly I need to remind myself to be careful of any gleaning I may give.  The interpretation my not be what I expect.  An understanding of the basics can not be obtained by a quick visit to a website r a few sentences by  mildly interested poster.  Your proposed circuit will draw too much current.

Yes, lamps and LEDs are a little like resistors but they are non linear.  The resistance of a lamp increases with voltage, the resistance of a LED decreases with voltage.  You absolutely have to have something to limit the current and low ohm resistors.  Line voltage is not stable and I have no idea where you are getting your numbers that are plugged into the wizard.  Please use a capacitor in any design you build.

Wisdom comes from knowledge.

Knowledge comes from experience.

Experience comes from bad decisions.

[wooferhound]

If you Rectify line voltage of 120 vac and put a filter capacitor on it as you should, then you will have a voltage close to 170 vdc, so you are inputting the wrong voltage into the Array wizard.

It makes much more sense to use a 12 vac transformer, after converting to DC you will end up with about 17 vdc. This is much easier to work with and way safer than running 170 vac inside of your bamboo poles

[TomW]

House;

Constantly I need to remind myself to be careful of any gleaning I may give.  The interpretation my not be what I expect.  An understanding of the basics can not be obtained by a quick visit to a website or a few sentences by  mildly interested poster.

Good points on giving one liner advice.

Lots and lots of gotchas in electronics that an experienced person just "knows" that leads to  simple yet catastrophic errors an old pro would have simply known was not going to have the expected results.

Thats one of the reasons I have pretty much stopped offering circuit advice. That and the fact that I have lost a good bit of knowledge due to not using it and health issues over the years affecting long term memory.

In this case the "back to back" string would probably NOT have the hoped for result.

Tom

[Bruce S]

I looked at the circuit too.

I didn't see where anyone mentioned the circuit does show a 1 WATT resistor.

The person writing up the drawing seems to at least at first glance understands that there's going to be some current trying to fry those LEDs :-). Forward voltage not taken into account, by me just grazing the circuit.

For safety reasons, I'm thinking a quick way to get this down to safer(er) AC voltages by using a step down xformer. Get the AC down to say 9Vac, using an old cell phone charger or like Ghurd said an old house bell ringer xformer, then rectify.

small xformers are cheap at electronic resale shops, or free depending on dumpster diving-ability.

Just a Thought

Bruce S  

[storrence]

I thought running them in series would result in lower part count, less energy lost at heat.

Here I found a simple version of the same thing:

http://www.instructables.com/id/LED_Lightbulb/

[storrence]

Looks like I tingled some nerves

I've done some more studying in the mean time at 'other sites' and see the error in my ways and why incandescent lamps deal with current and voltage differently than semiconductors

Now I just need to see if this will play nice with my dimmer which is an x10 home automation dimmer.

I'm hearing that a 'reactive current limiter' will have problems when the dimmer starts messing with the sine wave Any comments on that? I was really needing this to work with my dimmers :-(

[Opera House]

Many triac dimmers will not work unless they have a resistive load.  You can sometimes get around this with a capacitor across the load like a .47uF.  Use a 100 to 220 ohm resistor in series with the capacitor.  That still may not be enough.

[Bruce S]

That is a nice instructable, does give you something to do with all those burnt out light bulbs :-)

The parts count can certainly be lower, just that with LEDs more time needs to go into the design phase to accomplish the end result.

LEDs are still Diodes, no matter what very useful/pretty lights they put out, so by designing with that part in mind as well and respecting the voltages one is working with, a simple solution can be found.

ME: I like to bring voltages down to safe levels to work, I've been bitten more than once by the 110, 220 and even as high as 660Vac, not fun, messes with your whole day's thoughts.

In today's throw away mind set, there's hundreds of ways to get the circuit that you want by stuff people are pitching. Take the 1st circuit you saw and others, along with ideas from here and see what you can come up with.

LEDs are great to work with, just keep the knowledge that long time users from here have said and don't over drive them with either voltage or current and once built the light will probably outlast the bamboo :-)

Above all else have fun and becarefull

Cheers

Bruce S

[storrence]

I found a circuit halfway down this page that is exactly what I've been thinking would work. It doesn't have the capacitor which would cause problems with my x10 capacitor and if I use enough LEDs then I will need just a small resistor which will not put off much heat. Seems perfect. I'm going to use the full bridge version and about 50 or so LEDs. I will start out with a little extra and eliminate one by one until I get near the right voltage for each and then take one more out and put in the proper resistor value.

http://ourworld.compuserve.com/homepages/Bill_Bowden/page10.htm#lineled.gif

[ghurd]

You need a resistor at start.

Do not look for "the right voltage for each".  They will not be the same, and it is not related to the current the way you seem to be thinking about it.  0.3V can be the difference between 10ma and 75ma.

Look at the current.  Use a resistor.  Start with low amps, high resistance, and work from there.

And be careful.  That circuit can bite.

More safe to use a transformer.

G-