View Full Version : Discussion LED power help from wall socket
Troy
Sep 23, 2008, 02:39 PM
This should be a fairly easy one. I'm looking to power a bank of LED lights using a Wall Wart style transformer. I have sourced some LEDs that are pre-configured with resistors to handle straight 12V source:
http://www.allelectronics.com/index.php?page=search&search_query=led-12w&x=40&y=15
http://www.allelectronics.com/index.php?page=search&search_query=led-12b&x=22&y=10
So the question is which power supply to use. I plan on using a bank of these in the qty 48-52 range. I calculate I will only need about .62A if I run these all in parallel. Imagine a christmas light display/configuration. Would I be good using a Switching Power Supply like this?:
http://www.allelectronics.com/make-a-store/item/PS-12170/12-VDC-1.7-AMP-SWITCHING-SUPPLY/1.html
I have wired up LEDs on RC aircraft using DC power source but never from an AC wall outlet converted to DC via transformer.
Thanx!
Ron W3FJW
Sep 23, 2008, 04:30 PM
Should work ok with that WW. Can't see why it wouldn't.
rich smith
Sep 27, 2008, 05:28 PM
You might like to know that 30-40 LEDs (in series) run fine off 110vac with no power supply or current limiting required at all.
This should be a fairly easy one. I'm looking to power a bank of LED lights using a Wall Wart style transformer. I have sourced some LEDs that are pre-configured with resistors to handle straight 12V source:
http://www.allelectronics.com/index.php?page=search&search_query=led-12w&x=40&y=15
http://www.allelectronics.com/index.php?page=search&search_query=led-12b&x=22&y=10
So the question is which power supply to use. I plan on using a bank of these in the qty 48-52 range. I calculate I will only need about .62A if I run these all in parallel. Imagine a christmas light display/configuration. Would I be good using a Switching Power Supply like this?:
http://www.allelectronics.com/make-a-store/item/PS-12170/12-VDC-1.7-AMP-SWITCHING-SUPPLY/1.html
I have wired up LEDs on RC aircraft using DC power source but never from an AC wall outlet converted to DC via transformer.
Thanx!
Zlatko
Sep 27, 2008, 06:37 PM
You might like to know that 30-40 LEDs (in series) run fine off 110vac with no power supply or current limiting required at all.
Please put a current limiting resistor in there and a fuse ...
I also hope that it is well insulated ....
PLEASE !!!
rich smith
Sep 27, 2008, 09:34 PM
Current limiting is only required if the supply exceeds Vfmax. A resistor is required to run a red LED off a 5v or 12v supply but you will find no current limiting device at all in a cheap LED flashlight or LED replacement light bulb.
Several years ago I was hired to design LED home lighting system and quickly discovered there were some products already available. None used current limiting of any sort and neither did the designs I submitted to my client. My products made it to market and are probably still being sold. More recently the chinese cloned these 110vac LED light bulbs and they are available now from places like Dealextreme, Computergeeks, and even my local Family Dollar Store. If you dismantle one you'll find no resistors or inductors (4.3v*30=129vac). Also since Vr typically exceeds Vfmax there is no need for a rectifier either.
It's common for cheap LED flashlights to power from 3 AA or AAA w/o resistor too. Pushing it to run gallium arsenide type from 3 cells but no problem for white silicon carbide because they have higher forward bias voltage (2v vs 4.5v). The little keychain ones have no resistor either.
BTW there is actually no such thing as "white" LED. Except for the rare RGB triple diode types they are actually tiny flourecent lights. A blue or near UV junction excites the phosphor which produces white light via secondary emission. Just like mercury plasma in the 4' tubes.
You are absolutely correct about isolation from high voltages. Without knowing more about Troys application hard to tell if it needs that.
Please put a current limiting resistor in there and a fuse ...
I also hope that it is well insulated ....
PLEASE !!!
Tomapowa
Sep 27, 2008, 10:24 PM
If you dismantle one you'll find no resistors or inductors (4.3v*30=129vac).
True... but only for LED placed in SERIES... the original poster mentioned placing LEDs in parallel... just be careful. Just remember, current gets shared/split amoungst LEDs when in parallel, while voltage gets shared/split amoungst LEDs when in series.
Also since Vr typically exceeds Vfmax there is no need for a rectifier either.
Can you explain this theory.... ?
rich smith
Sep 27, 2008, 11:13 PM
Yes, that's why I used the term "in series" on my original post. Maybe Troy just wants to string a buncha cool lights around his bedroom in which case parallel with a wall wart is better.
In any case it's not theory. Thousands of products are being sold and used based on these principles. If reverse breadown voltage is less than forward bias voltage as with certain rare and exotic LEDs then a rectifier (no cap/filter tho) would be needed to prevent destruction. Fortunately common LEDs have relatively high Vr (10-30v) and, thanks to leakage resistance, volts are equalized and prevented from building up in any one device. So put that 1n4007 back in the parts bin.
One bright idea I had when running a single LED off 110vac did not work out so well. I figured the reactive impedance of a cap would be more efficient than a power wasting resistor. Resistors to drop 106v out of 110v get very hot and cost a lot. I was puzzled because sometimes when I powered up the LED blew. Then realized if switch closes during peak of sine wave you're putting 160v at infinite current across a 4v diode. Not good! This is why cheap dollar store 110vac LED nightlights use a power resistor and not a cap.
BTW current does not always get shared equally in parallel. That's why it's important to use LEDs from the same batch so some won't hog the current and be brighter than the others.
True... but only for LED placed in SERIES... the original poster mentioned placing LEDs in parallel... just be careful. Just remember, current gets shared/split amoungst LEDs when in parallel, while voltage gets shared/split amoungst LEDs when in series.
Can you explain this theory.... ?
Tomapowa
Sep 28, 2008, 10:03 AM
BTW current does not always get shared equally in parallel. That's why it's important to use LEDs from the same batch so some won't hog the current and be brighter than the others.
So true hence the reason I said shared/split... not "equally" shared as you mentioned. You just need to be sure to use LEDs with simlar forward voltage specs. Constantly reverse biasing an LED at a voltge close to its Vr (@ 60Hz no less) will eventually damage them... trust me I know, not good design practice. A 5 cent signal diode goes a long way.
http://wolfstone.halloweenhost.com/Lighting/litlpo_PoweringLEDs.html#AvoidRunningLEDsStraightO ffAC
Also... the original poster mentioned he was using LEDs with internal resistors for running direct from a 12V source. You would then only need 10 or so of these LEDs in series, not 30.
rich smith
Sep 28, 2008, 01:29 PM
Unfortunately using spec sheets don't help much when matching LEDs. Even those from same lot may vary greatly. Die near the edge of the wafer can be quite different from those near the center. I learned this having designed more LED signs than can be counted on two hands and a couple production photometer/handler projects. Very cool stuff.
You are right about rectifier needed for single LED on house current because 3watts of avalanche will blow it. 30-40 in series don't need one because reverse breakdown never occurs for one thing and if it did power would be harmlessly distributed across many (mw like zener).
Must admit I really don't know what's up with those resistor LEDs. Silicon resistors? Hybrid fab? The "blinkers" are kinda neat.
Did you know that forward biased LEDs make excellent voltage references and reversed biased ones make good photocells?
So true hence the reason I said shared/split... not "equally" shared as you mentioned. You just need to be sure to use LEDs with simlar forward voltage specs. Constantly reverse biasing an LED at a voltge close to its Vr (@ 60Hz no less) will eventually damage them... trust me I know, not good design practice. A 5 cent signal diode goes a long way.
http://wolfstone.halloweenhost.com/Lighting/litlpo_PoweringLEDs.html#AvoidRunningLEDsStraightO ffAC
Also... the original poster mentioned he was using LEDs with internal resistors for running direct from a 12V source. You would then only need 10 or so of these LEDs in series, not 30.
vBulletin® Copyright ©2000-2009, Jelsoft Enterprises Ltd.