Have a client who would like LED strips to fade on and fade off on a motion sensor. I found on eBay a controller that has the PIR motion sensor but it doesn't fade. Can anyone direct me to where I might find something like this?

Thanks.

Craig.

I'm assuming you already have a lighting system installed? If so can you not connect the PIR to this system and program accordingly?

No lighting system. Guy just asked me if I could look into some lighting strips on a motion sensor that turn on and off slowly. I'd like to keep it simple. There is no desire to install a large lighting or automation system or control system.

He has a long bench in his hallway. There is a power receptical that's not switched. He thinks it would be cool if it had some funky lighting under it.

Use an automotive LED strip that runs from low voltage. Put the strip in parallel with a capacitor and add a resistor in series with the DC power supply that charges the capacitor. The capacitor will charge slowly when power is applied and discharge slowly when power is removed.

I understand how that would allow it to fade to off, by varying the discharge rate of the cap, but how does that allow for slow turn on. The cap doesn't limit anything passing through.

That's not my area of expertise (hence my question) but I've done just that make constant lighting circuits in my model trains and they come on immediately and then turn off slowly after I remove track power.

Craig

Hook up a PIR to hot/neutral to switch out the hot to the outlet. Outlet output feeds a standard 12v transformer/wallwart. 12v transformer powers an LED ramp up/down unit such as this:

[Link: oznium.com]

buzz suggests something that's a really good idea, for which I've been quite criticized in the past: use old technology (capacitor charging) and don't give any suggested parts values. Still, I'd definitely try this myself.

The capacitor will probably have to be in the 4700 mfd or larger range. The resistor will make the voltage ramp up, but it will also lower the maximum light level. You might find that you have to use a 24 or higher voltage power supply because the resistor will always be in series with the LEDs. The resistor also might get hot.

A better idea is basically the same thing, but with a stage of amplification added. Charge a capacitor through a resistor, but have the capacitor voltage fed to the base of a transistor wired in an open collector configuration with the LEDs as a load. This could then be done with a much smaller capacitor.

This is more complicated because switching is not so simple since the "power" switch would turn on and off voltage to the capacitor. The actual power supply would stay on all the time.

The difference between buzz's idea and the use of a transistor is that with buzz's idea you have to make a capacitor charge up slowly to 12 volts with quite a load on it, then fade slowly, which is even more difficult without a HUGE capacitor. The transistor idea has you charging up a capacitor to only about 0.7 volts, then draining to zero. Using an FET makes this even easier as the gate current is tiny.

Time constant calculations: [Link: en.wikipedia.org]

The trick here is to work from the current draw of the LED strip and convert this to an equivalent resistance that will be used to draw an equivalent circuit while you calculate possibilities. You'll need to fuss with things because the human's response to the change in light value is nonlinear, as is the voltage ramp, and the LED's will not turn ON until the voltage crosses their forward drop (probably in the 1.8 - 3V range, depending on color). I would start with a charging current that is twice the rate of the LED current draw.

If the DC supply is regulated, this will mess with the timing because the regulated supply probably has its own capacitor and the supply's output will not collapse instantly. However, you can work around this by loading the supply, thereby forcing it to collapse promptly after power is removed.

If the capacitor does not discharge fast enough, place a reverse diode across the charging resistor, and a resistor in series with that diode. After the charging supply collapses, this reverse R/C diode network, along with the LED draw, will control the discharge rate.

Wow awesome info fellas. Thanks very much for taking the time to describe all of that in so much detail.

Craig

Actually, just looked at the suggestion by tweeterguy above, This seems interesting. From the description I think that it would be best to have the dongle powered at all times and use the PIR strictly as a trigger.

Yes based on the description that prevents flickering on start up. That would be pretty easy to accomplish.

Would you please tell me the operating voltage and current of the LEDs?

I've seen several approaches to this and none have been satisfactory as far as the lowest couple dozen of light levels when dimming down.

I'm about to buy one of these: [Link: oznium.com] to see if this company has solved this weakest point of LED dimming.

If that doesn't work I'd use the approach I spoke of before: use a transistor to turn the LEDs on and off, and put a resistor and capacitor on the transistor to make it turn the LED voltage up and down over several seconds instead of rapidly.

Hey Ernie.

This is the kit I'm hoping by to use.

[Link: m.ebay.com]

It's a 12v 2A power supply for the 5m strip. The bench is 3m long. 3/5 of the length so maybe 1.2 A current draw?

The idea would be to dig into the PIR circuit and find constant 12v, tap it to power the fader box and then tap the PIR out put to feed the trigger.

Craig

Ernie,

You'll probably have better luck at the lower levels if you start the ramp at or just below the forward junction voltage of the LED's. Otherwise, you'll have an apparent delay while the voltage ramps up from zero. LED light output vs current is relatively linear. Fortunately, if you are operating in the DC domain, you won't need to deal with crazy intensity "snap" issues associated with some phase control designs.

Craig,
That sounds right.
buzz brings up a bit of reality here: let's say each LED has two volts of drop across it; the LED strips I've seen all have three LEDs in common. As the oznium device brings up the voltage from zero, no LED will come on until the voltage has ramped halfway up to twelve volts! If the ramp period is two seconds, that mean you'll have to wait a second before there's any light. That doesn't sound like much but it seems unacceptable to me.

buzz,
On the other hand, if the oznium device has been designed to ramp up LEDs, maybe those guys have heard of the forward voltage drop and designed the unit so it doesn't start with zero volts but comes on at, say 5 volts and ramps from there.



And current does not flow in an LED strip until the applied voltage equals the sum of the forward voltages of the LEDs; again, about six volts.


Right. Oznium informed me that they shipped my unit today, the day after Thanksgiving. I'll let you know what I find.