Want a triac LED lamp dimmer circuit that can work with the new LED lamps; 20 LED array, 15 watts, 0.032 amps @ 120 VAC. On my powerstat, this lamp shows light ignition at about 36 VAC (.0025 amp) and is fairly linear all the way up to maximum, 120 VAC .032 amps.

Cheers, DPW [ Everyone's knowledge is in-complete...Albert Einstein]

 

The biggest problem I have run into dimming the LED strings with triac based dimmers is the low amount of current drawn and the heat developed as a result of this.

A few years back I ran across the patent (attached in pdf here) for circuitry that assisted the triac dimmers to function adequately when dimming LEDs.
the document helped a great deal with understanding the process.

With a bit of trial and error, success was not futile.


Surprisingly enough, a Light-o-Rama kit I built years ago does a pretty darn good job of LED dimming

http://www.lightorama.com/Documents/CTB-16KV6_Assembly_Manual.pdf

the light o rama runs small ac motor actuators at varying speeds also

 

Thanks for the patent (learning document, because that's what it is). It looks like this patented circuit acts as a shunt between the dimmer circuit and the LED load. My intent is to use an AC triac (if possible) to vary the phase (power) to the LED circuit, similar to what is done in incandescent lamp dimming. Maybe logic Mosfets might be a better choice in this instance. Anyway, the light-o-rama board would be an interesting study, if they included a schematic, but maybe I can glean some information by looking at the IC's. Thanks anyway.

Best regards, DPW [ Spent years making heaters out of op-amps.]

 

Have you considered a SCR circuit?
I breadboarded one of Bill Bowdens circuits for the heck of it to test it on a LED string. The first schematic here works fairly good with the LED string. I havent tweeked it or performed any measurements but the componentry was cool to the touch.

I havent tried the second circuit yet but I will sometime soon just to satisfy my curiosity.

Scince the low cost and duribility of the light o rama is such a bargain, I havent had the need to do much experimenting with programming LED dimming.

Seeing how these boards have been buried by snow and soaked by rain...and keep ticking.



120 VAC Lamp Dimmer

The full wave phase control circuit below was found in a RCA power circuits book from 1969. The load is placed in series with the AC line and the four diodes provide a full wave rectified voltage to the anode of a SCR. Two small signal transistors are connected in a switch configuration so that when the voltage on the 2.2uF capacitor reaches about 8 volts, the transistors will switch on and discharge the capacitor through the SCR gate causing it to begin conducting. The time delay from the beginning of each half cycle to the point where the SCR switches on is controlled by the 50K resistor which adjusts the time required for the 2uF capacitor to charge to 8 volts. As the resistance is reduced, the time is reduced and the SCR will conduct earlier during each half cycle which applies a greater average voltage across the load. With the resistance set to minimum the SCR will trigger when the voltage rises to about 40 volts or 15 degrees into the cycle. To compensate for component tollerances, the 15K resistor can be adjusted slightly so that the output voltage is near zero when the 50K pot is set to maximum. Increasing the 15K resistor will reduce the setting of the 50K pot for minimum output and visa versa. Be careful not to touch the circuit while it is connected to the AC line.

MenuVarying brightness AC lamp

In this circuit, an SCR is used to slowly vary the intensity of a 120 volt light bulb by controlling the time that the AC line voltage is applied to the lamp during each half cycle.

Caution:

The circuit is directly connected to the AC power line and should be placed inside an enclosure that will prevent direct contact with any of the components. To avoid electrical shock, do not touch any part of the circuit while it is connected to the AC power line. A 2K, 10 watt power resistor is used to drop the line voltage down to 9 volts DC. This resistor will dissipate about 7 watts and needs some ventilation.

Operation:

A couple NPN transistors are used to detect the beginning of each half cycle and trigger a delay timer which in turn triggers the SCR at the end of the delay time. The delay time is established by a current source which is controlled by a 4017 decade counter. The first count (pin 3) sets the current to a minimum which corresponds to about 7 milliseconds of delay, or most of the half cycle time so that the lamp is almost off. Full brightness is obtained on the sixth count (pin 1) which is not connected so that the current will be maximum and provide a minimum delay and trigger the SCR near the beginning of the cycle. The remaining 8 counts increment the brightness 4 steps up and 4 steps down between maximum and minimum. Each step up or down provides about twice or half the power, so that the intensity appears to change linearly. The brightness of each step can be adjusted with the 4 resistors (4.3K, 4.7K, 5.6K, 7.5K) connected to the counter outputs.

The circuit has been built by Don Warkentien (WODEW) who suggsted adding a small 47uF capacitor from ground to the junction of the current source transistor (PNP) to reduce the digital stepping effect so the lamp will brighten and fade in a smoother fashion. The value of this capacitor will depend on the 4017 counting rate, a faster rate would require a smaller capacitor.

The light o rama is indeed a great tool.
I have used dozens for light shows of course but the animation potentials are what is fascinating. The ability to control the speeds of actuators and the ability to pulse them make animated figures appear alive.

for the low cost...

 

Just curious, why not take a diode bridge, and use an MOSFET?

We're skirting territory that is taboo here, but I think you could do what you needed from there.

 

Or try a Joule Thief:

http://www.emanator.demon.co.uk/bigclive/joule.htm
http://www.prc68.com/I/JouleThief.shtml
http://www.evilmadscientist.com/article.php/joulethief

Add a variable resistor in series with the base resistor to vary the brightness.

 

The LEDs the OP is talking about takes 700ma or more, maybe much more. The new 3W LEDs are quite robust.