Cree P4 under-cabinet lights with dc power supply

moon lander

Enlightened
Joined
Feb 8, 2007
Messages
287
Location
boston
Im creating under-cabinet lighting, and my plan is to run a single Cree P4 on a dc power supply. heres the setup: radioshack dc power supply, set to 4.5 volts, 1 amp rated maximum, run thru 1 ohm 10 watt resistor and the led. im actually using 2 .5ohm 5 watt resistors in series. the led is mounted with thermal paste on a big heatsink.

i couldnt find a power supply with a voltage closer to 3.7v, so i had to work with what i had. Obviously 4.5v is too much for the led, without the resistor it drew 1.5 amps (too much for the led or the power supply). i added a .5ohm resistor and it went down to 1.2 amps. added another resistor (total of 1 ohm) and got 800ma to the led. perfect. at that power, it shouldnt get too too hot (especially with the heatsink) and the power supply is happy. and let me tell you my kitchen counter lighting is now beautiful (and efficient too)!

the QUESTION is: is this setup safe? according to my knowledge it is, tho my knowledge about these things is extremely limited (i read electronics for dummies). the power supple should have no trouble because it is rated for 1 amp (1000ma) and the led is drawing 800ma (measured with my multimeter in series). the led itself cant start a fire, can it? i cant think of any reason (after some testing) not to leave it on unsupervised. can you?

my kitchen counter has never looked better. the light is perfect for about 4 square feet but i think it would be better with 2 leds spaced out about 3.5 feet. i think to do this, i could wire 2 leds in series and use something like a 7.4v power supply (give or take a little and/or add a little resistance) that should give me around 800ma to each led (total of 1600ma so the power supply would have to have a high amp rating as well). makes sense right? any thoughts?
 

bascerballer4

Newly Enlightened
Joined
Mar 29, 2007
Messages
53
If the dc power supply was set to 4.5 volts @ 1A, how did you draw 1.5 amps to start off with? I think you meant 1A. Correct me if im wrong.
 

walkabout

Newly Enlightened
Joined
Dec 8, 2006
Messages
59
moon lander said:
the QUESTION is: is this setup safe? according to my knowledge it is, tho my knowledge about these things is extremely limited (i read electronics for dummies). the power supple should have no trouble because it is rated for 1 amp (1000ma) and the led is drawing 800ma (measured with my multimeter in series). the led itself cant start a fire, can it? i cant think of any reason (after some testing) not to leave it on unsupervised. can you?

If you're using a UL-listed power supply, there shouldn't be reason to worry.

BTW: There are some great threads on LED-based home lighting in "Homemade and Modified Lights." Take some time and browse, it's definitely worth it.
 
Last edited:

moon lander

Enlightened
Joined
Feb 8, 2007
Messages
287
Location
boston
bascerballer4 said:
If the dc power supply was set to 4.5 volts @ 1A, how did you draw 1.5 amps to start off with? I think you meant 1A. Correct me if im wrong.

actually it did draw 1.5amps. the maximum rating for the power supply is 1 amp but i exceeded that for a very brief moment (luckily nothing happened). the 1 amp rating is the maximum recommended current before it starts to melt. apparently its flexible for very brief power.
 

riffraff

Enlightened
Joined
Mar 3, 2007
Messages
243
Location
Londinium
Many supplies will exceed their current ratings as they approach a shorted condition, but their voltage will sag at the same time.

BTW, when you say, "Radio Shack power supply set to 4.5VDC," are you talking about the wall warts (wall transformers)?
 

Biker Bear

Enlightened
Joined
Sep 9, 2006
Messages
279
Location
The Greater Los Angeles Metropolitan Sprawl
moon lander said:
my kitchen counter has never looked better. the light is perfect for about 4 square feet but i think it would be better with 2 leds spaced out about 3.5 feet. i think to do this, i could wire 2 leds in series and use something like a 7.4v power supply (give or take a little and/or add a little resistance) that should give me around 800ma to each led (total of 1600ma so the power supply would have to have a high amp rating as well). makes sense right? any thoughts?
Your calculation of the needed current is wrong.

When you run LEDs in series, the needed voltage goes up, but the amperage is unchanged. When you run LEDs in parallel, the needed voltage remains the same but the required amperage goes up.

You could look at LEDs as water valves that only open (emit light) when the water in the pipe is of a certain pressure (voltage). Thus, if you have two in the same pipe (in series), your input pressure/voltage will have to be higher than with just one - but the two of them will pass the same amount of water (current/amperage) because it's only one pipe.

If you have two pipes side by side - one valve in each - that are fed from a bigger pipe, you're going to need just enough pressure for one valve - but you're going to need to deliver twice as much water.

Granted, these analogies can only be pushed so far - but I hope that makes what's going on a bit clearer.
 

riffraff

Enlightened
Joined
Mar 3, 2007
Messages
243
Location
Londinium
The problem with LEDs and water analogies is that water acts in linear fashion, whilst LEDs (being a semiconductor junction) are exponential. That's why the OP needed a resistor from the very beginning, even if he was to find a 3.7VDC supply that would "work." A change in voltage across the junction results in a very large swing in current, without a resistor.
 

moon lander

Enlightened
Joined
Feb 8, 2007
Messages
287
Location
boston
walkabout said:
If you're using a UL-listed power supply, there shouldn't be reason to worry.

BTW: There are some great threads on LED-based home lighting in "Homemade and Modified Lights." Take some time and browse, it's definitely worth it.

im sorry, but i dont know what UL-listed means. its a radioshack wall wart that can be set to 3, 4.5, 7, 9, or 12 volts. 1 amp rated.
 

moon lander

Enlightened
Joined
Feb 8, 2007
Messages
287
Location
boston
riffraff said:
Many supplies will exceed their current ratings as they approach a shorted condition, but their voltage will sag at the same time.

BTW, when you say, "Radio Shack power supply set to 4.5VDC," are you talking about the wall warts (wall transformers)?

yes. wall wart with a switch on the front to set the voltage, rated 1 amp max.
 

moon lander

Enlightened
Joined
Feb 8, 2007
Messages
287
Location
boston
Biker Bear said:
Your calculation of the needed current is wrong.

When you run LEDs in series, the needed voltage goes up, but the amperage is unchanged. When you run LEDs in parallel, the needed voltage remains the same but the required amperage goes up.

You could look at LEDs as water valves that only open (emit light) when the water in the pipe is of a certain pressure (voltage). Thus, if you have two in the same pipe (in series), your input pressure/voltage will have to be higher than with just one - but the two of them will pass the same amount of water (current/amperage) because it's only one pipe.

If you have two pipes side by side - one valve in each - that are fed from a bigger pipe, you're going to need just enough pressure for one valve - but you're going to need to deliver twice as much water.

Granted, these analogies can only be pushed so far - but I hope that makes what's going on a bit clearer.

right you are! thanks for the correction. i really like your analogy. so i can use the same supply for 2 leds, ill just set the voltage higher and check that the current doesnt exceed 1amp. thanks!
 

riffraff

Enlightened
Joined
Mar 3, 2007
Messages
243
Location
Londinium
UL = Underwriters Laboratories. Required for just about anything that connects to house current (certainly all consumer goods).

Most of the wall warts are not regulated, so you may not be getting a true 4.5VDC.

You said you saw it go up to 1.5 Amps...I assume you have a multimeter? Do you have two? So you can monitor the current and the voltage at the same time?
 

moon lander

Enlightened
Joined
Feb 8, 2007
Messages
287
Location
boston
riffraff said:
UL = Underwriters Laboratories. Required for just about anything that connects to house current (certainly all consumer goods).

Most of the wall warts are not regulated, so you may not be getting a true 4.5VDC.

You said you saw it go up to 1.5 Amps...I assume you have a multimeter? Do you have two? So you can monitor the current and the voltage at the same time?

will it say UL on the back? im not home so i cant check until 1am.
i checked the voltage (not under load) and it reads high. the 3v setting read 3.4 and the 4.5v setting read 4.8. i only have 1 multimeter so i cant check that. i havent checked the voltage under load. not sure how to do this. do i place the test probes in series with the led/power supply circuit? or parallel across the led? thanks again.
 

riffraff

Enlightened
Joined
Mar 3, 2007
Messages
243
Location
Londinium
Have a look here for a series/parallel LED calculator.

If you have two LEDs drawing 800mA each, that's 1.6 Amps. You gonna need a bigger boat, er, supply, rather.


:popcorn:
 

moon lander

Enlightened
Joined
Feb 8, 2007
Messages
287
Location
boston
riffraff said:
Have a look here for a series/parallel LED calculator.

If you have two LEDs drawing 800mA each, that's 1.6 Amps. You gonna need a bigger boat, er, supply, rather.


:popcorn:


i think in parallel i would, but in series i should be ok right?
 

riffraff

Enlightened
Joined
Mar 3, 2007
Messages
243
Location
Londinium
Should be okay, but there's a greater likelihood of the brightness being different (due to differences in the performance of the junctions, a difference which is minimized in a parallel circuit).

You're looking at a 2.2 ohm resistor of about 3 watts (to be safe; don't want to set fire to the cabinets). This is assuming a true 9VDC supply, a forward voltage of 3.7VDC, and 800mA draw.
 

LEDite

Enlightened
Joined
Dec 6, 2002
Messages
324
Location
Dallas, TX
moon lander;

Do you have a good heat sink on the Cree ??

If it gets hot, the voltage accross the LED will fall and the current will go up.

You need to test it @ warm temperatures to see if the current increases as the LED heats up.

Run it about 10 minutes at least to check for higher current.

My Cree heatsink is shown on this post:

https://www.candlepowerforums.com/threads/155211

Larry Cobb
 
Last edited:

moon lander

Enlightened
Joined
Feb 8, 2007
Messages
287
Location
boston
LEDite said:
moon lander;

Do you have a good heat sink on the Cree ??

If it gets hot, the voltage accross the LED will fall and the current will go up.

You need to test it @ warm temperatures to see if the current increases as the LED heats up.

Run it about 10 minutes at least to check for higher current.

My Cree heatsink is shown on this post:

https://www.candlepowerforums.com/threads/155211

Larry Cobb

yes im using a big laptop cpu heatsink, and no it is not sufficient. the light does warm up considerably. after about 15-20 minutes it approaches "too hot to touch". i guess ill have to add more resistance. it is currently running at 870ma, i bet it will be ok around 750ma. i did not check the current when it was hot but i bet your right. also the resistors heat up a lot. i wonder if any of this could get hot enough to ignite something? Larry i love your triple cree house light!
 

balazer

Enlightened
Joined
Aug 22, 2005
Messages
486
riffraff said:
Should be okay, but there's a greater likelihood of the brightness being different (due to differences in the performance of the junctions, a difference which is minimized in a parallel circuit).
The difference in brightness will be minimized when the two LEDs are wired in series. That ensures that the same current is going through both LEDs.


A couple of things: the power supply is rated at 1 amp, meaning it can safely supply that much current. It is not a regulated supply, meaning that if you ask it fore more than 1 amp, it might deliver it, unsafely. It's up to you to make sure that the load does not draw too much current.

If you're going to use a meter to measure the current, don't just put the meter in series with the power supply. The voltage drop across the ammeter will reduce the total current. The proper way to measure the current is to put a low value resistor (0.1 or 0.05 Ohms) in series, and measure the voltage drop across the resistor.
 

balazer

Enlightened
Joined
Aug 22, 2005
Messages
486
riffraff said:
The problem with LEDs and water analogies is that water acts in linear fashion, whilst LEDs (being a semiconductor junction) are exponential.
Well it's not the water that's non-linear in this analogy. Think of the LED as a pressure release valve that has a non-linear spring in it.
 
Top