LED dimmer

[Cabwood]

Ugly, and incomplete, but hey, it's a work in progress!
This was the test circuit I messed around with. The biggest trouble with it is my opamp's limited output swing, making levels difficult to predict, and making it complicated to switch the transistor. Amongst other things.

View attachment 39407

 

[Cabwood]

Here's the last version of this that I'm going to do.

At the top left is a nearly-trianglular waveform generator. The peak and trough voltages are carefully chosen. The bottom left part is the ramp generator. I decided to be careful here because the level limits are important, given the range of the triangle wave it will be compared to - hence the use of two opamps.

The last opamp of the four available is the comparator, which switches TR1. R12 is there to compensate for the opamp output not swinging low enough.

The ramp output swings through the whole range (3V - 9V) in about 5 seconds. Change R8 or C2 to speed it up or slow it down.

It turns out that this was a waste fo time, because it doesn't do any better than the simple two-transistor dimmer I posted first. The perceived brightness of an LED seems to be proportional to the log of the power it dissipates, which means both circuits seem to change brightness quickly when dim, and slowly when bright.

So, I played with a log amplifier to do this, which worked OK, but it's severe overkill for a simple dimmer circuit.

View attachment 39426

 

[indulis]

You might try looking at the data sheets and app's notes for the Linear Tech LT3486 and LT3474 for ideas. They seem to control brightness by voltage not duty cycle directly.

 

[audioguru2]

Yes, the response of your vision to brightness is logarithmic, like your hearing.
The exponential charge and discharge voltages of your capacitor are the opposite to what is needed.

I made some slowly fading and brightening LEDs by including a coupling capacitor in the positive feedback loop of an opamp. The overall gain was 2 so when the cap charged 0.1V then the output was 0.2V. Then it charged another 0.1V to 0.2V and the output was 0.4V, etc. The opposite occurred during the discharge.

My Chaser projects use PWM as a brightness control for very long battery life. The width of the pulse changes more than 1000:1 for brightness control.