LED lights dimming question

[Curt Harms]

I have a project replacing halogen puck lights used for under cabinet lighting with LED. I've already added LED strips to a pantry which made a world of difference. My question is how to dim LED puck-style lights. The 2 lights are from Amazon, 3 elements/light, 12 volts, the blurb says 3 watts.

http://www.amazon.com/Lighting-EVER-...led+puck+light

There's nothing more in the packaging about specs. I hooked one up to the 12 volt switching power supply and those little suckers are BRIGHT!! I'm planning to put 2 strips under the cabinets wired through a 3 position switch. One position lights one strip, center position off, 3rd position lights both strips. It'd be nice to be able to reduce the light output of the two puck lights. The common method seems to be to use a PWM dimmer. These things seem to be pretty good sized boxes that I'm having a hard time finding a place for and I don't really need variable light output. Just 100% in the 'bright' position and something like 50-70% in the 'dim' position. Google turns up some things about using one or more inline resistors to accomplish dimming. I can't find anything about what values to use. Does anyone have any thoughts about how to dim 2 LED puck lights using small inline bits such as one or more resistors or other devices? My electronic knowledge is quite limited but I do know which end of a soldering iron to hold .
LED_Lights.jpg

[Art Mann]

Here is a little math for you. The total power to the fixtures at full brightness is 6 watts. Since current is equal to power divided by voltage, the current is 6/12 = 0.5 amps. The equivalent resistance of the two fixtures together is equal to the voltage divided by the current or 12/0.5 = 24 ohms. If you want to half the power to the fixtures then add a series resistor (the "Device" in your diagram) equal to the equivalent resistance of the fixtures --> 24 ohms. These calculations aren't quite right for a couple of reasons. First, the voltage across the diodes is not linear with respect to the current. Secondly, the light emitted by a diode is not linear with respect to the current. Still, 24 ohms would be a place to start. With the resistor switched in to the circuit, the current will be half the original value or about 0.25 amps. The power requirement for the resistor is equal to the current squared times the resistance or 0.25 * 0.25 * 24 = 1.5 watts. To be on the safe side, I would use a 2 or 3 watt resistor. Unlike a PWM controller, the resistor will get hot so you need to mount it where air can circulate around it and nobody will touch it. 5% power resistors of this type are pretty cheap. You can buy several values near 24 ohms and see which one you like best.

By the way, you do realize that your power supply is gross overkill. Those fixtures will run nicely with a small 1 amp 12Vdc power brick.

[Curt Harms]

Thanks Art. The power supply is powering other lights as well and it wasn't much more money than the wall wart types. I'll get a few resistors as you recommend and experiment. In researching this stuff, I found it interesting how PWM (pulse width modulation) dimmers work and that LEDs will stand up to that sort of cycling. Different than what we're used to, huh?

[Art Mann]

Pulse width modulation is a standard technique used in the automotive industry to dim dome lights, instrument clusters and other interior lighting. It is also used in voltage regulation circuitry. Automotive electronics is the industry in which I spent a career so I am pretty familiar with it.

[Erik Loza]

I assume you have these controlled by a standard wall switch? Why not just install a dimmer switch at the wall plate? I have several LED's in our house set up that way and it works fine. Unless I am missing something?

Erik Loza
Minimax USA

[Curt Harms]

I assume you have these controlled by a standard wall switch? Why not just install a dimmer switch at the wall plate? I have several LED's in our house set up that way and it works fine. Unless I am missing something?

Erik Loza
Minimax USA

Hi Eric

Naw, that'd be too easy. These are under-cabinet type lighting. I'm looking for a '2 speed' dimming solution that is very compact. For most of it I'm running two parallel strips of stick-on LEDs controlled by a single pole double throw center off switch. One position lights one row, the second position lights both rows. This question arose because I can't use strip lighting on one segment and have it look right. Puck lights seem to provide the best solution. If I were in the mood, I've found plans to build your own PWM dimmer that could be assembled in any shape desired. I could find space easily for a thin - like 1/2" thick slice of bread shaped device. The most compact PWM dimmers I'm finding are 1 1/4" X 1 1/4". Yes I could probably work something out with that and may, it'd just be cooler to not have any lumps to deal with.

[Matt Meiser]

Check ebay for PWM dimmers. I bought kits of LED tape, power supplies, and dimmers and surely they sell just the dimmers. They're adjustable too so you could tune the light to where you want in the "dim" position of your switch.

[Dan Hintz]

I assume you have these controlled by a standard wall switch? Why not just install a dimmer switch at the wall plate? I have several LED's in our house set up that way and it works fine. Unless I am missing something?

Unfortunately LEDs don't work that way... yours dim because they have electronics in the base that convert the incoming AC to a low-voltage DC that can be used by the LEDs. Those electronics also allow standard dimmers (which chop the AC waveform) to work, but using the same method of chopping will not work on LEDs without those driver electronics.

[Erik Loza]

Ah, thanks for the explanation.

Erik Loza
Minimax USA

[Thomas S Stockton]

try www.outwater.com they have a good selection of led lights and might have a dimmer that works. I havealso found that their descriptions are pretty good

[Curt Harms]

Well, I have the puck lights in place and wired up temporarily. I went to Radio Shack and rummaged through their drawers. They had wirewound 50 ohm resistors rated for 10 watts. Those were the only high wattage resisters I saw. They come in a package of two, I figured I could experiment with series or parallel if necessary. I wired one where I have 'device' in the diagram above. So far so good. I used a clamp on digital ammeter to check amperage. With the resistor inline it read .01 amps, I don't know how reliable such a low reading would be. Straight thru, it read .06 amps. The light output when set to dim is what we have in mind, full bright is .. pretty bright. These lights are advertized as replacements for 20 w. Xenon puck lights, I think they're brighter. We'll use it in this configuration an see how we like it. I expected the resistor to get warm after being on for a couple hours. So far, temperature rise is barely perceptible - or it could be my imagination.

I'd have preferred to use a PWM dimmer but I couldn't figure out how to mount it and not have it look strange. I plan to use one of Radio Shack's mini or micro SPDT toggle switches mounted to the backside of one of the stiles. It'll be invisible to everyone except kids or little people.

[Dan Hintz]

At 10-60mA, you won't feel any warmth... simply not enough power being dissipated.

[Art Mann]

A clamp on DC current probe that accurately measures in the milliamp range is likely to cost several hundred dollars. If you want to know how much current is flowing, just use a cheap multimeter with a milliamp range setting and route the current directly through the meter. I agree that a 10W power resistor won't get very warm. If it is a size you can live with then that is your solution. I sometimes forget that not everybody lives in a town where all the major electronics suppliers have a well stocked store.

[Dan Hintz]

A clamp on DC current probe that accurately measures in the milliamp range is likely to cost several hundred dollars. If you want to know how much current is flowing, just use a cheap multimeter with a milliamp range setting and route the current directly through the meter.

Or just measure the voltage across the power resistor