Setting up LEDs on Timer + Dimmer circuit, help needed

RyanBrucks · #1 11/30/2007, 05:11 PM

Hey All,

Not too long ago I posted a thread asking about backdrops, plexiglass, light transmission, and LEDs.

I finally figured out that part of my backdrop project

View the above link to see pictures of how my LED lightbox turned out.

The LEDs are powered by a 30V DC, 1amp adapter. They are fully dimmable 0-100% (via 5k potentiometer), so naturally I would like to put them on a timer that can slowly fade the backlight on/off for sunrise sunset.

I have a 5k potentiometer already, but I think trying to find a mechanical way to 'turn the knob' on a timer would be pretty crazy (though I did run across a project where some guy made a potentiiometer timer out of old watch gears and microswitches!!!). I wonder if there are any electrical circuit products out there that are already set up to do something like this?

I wonder if I could use standard timers for the on/off part, but use some kind of device that would A) make it take X amount of time to allow full voltage through and B) continulously supply power for X amount of time ramping down after the source is cut off.

I can't just use a standard wall-mounted dimmer for this, as it has to be a potentiometer/rheostat (standard dimmers turn the circuit on/off very fast, which won't work for LEDs, the actual voltage needs to change).

Anybody have any suggestions for me?

Thanks,
Ryan

pescadero · #2 11/30/2007, 05:24 PM

physically turning the knob isn't the best approach. a better way to do it would be to electronically regulate the amount of current supplied to the LEDs. granted, you need to know a bit of electronics to do this sort of thing.

so how does your dimmer work? are you changing the voltage, or are you maintaining regulated voltage while adjusting the current? the best way to get a good result with a dimmer circuit for an LED is to keep the voltage constant while you vary the current. there are a number of ways to do this, but microprocessor control is the best if you want programmable dimming times. unfortunately, at a minimum you need to know a bit about designing electronic timer circuits to make this happen.

RyanBrucks · #3 11/30/2007, 05:29 PM

"so how does your dimmer work? are you changing the voltage, or are you maintaining regulated voltage while adjusting the current?"

Well I have not yet wired up my potentiometer, but I thought they changed voltage while maintaining current?

My 30V DC adaptor has a switch on the front that makes it either 13V or 30V. If I set it to 13V, it perfectly dims the LEDS to about half brightness, so I'm assuming that voltage is all that is needed to change.

Seems I'm already learning alot, so I might as well keep learning on the subject until I can figure it out

Does anybody have any good recommended reading on learning how to set up timer circuits with a microprocessor?

Ryan

Donw · #4 11/30/2007, 05:35 PM

Do you want a slowly progressive type dim of just go from bright to dim based on time. The later is easy.

Don

RyanBrucks · #5 11/30/2007, 05:43 PM

Linear fading from bright to dim would be fine.

By 'slowly progressive' do you mean like a non-linear brightness ramp? I suppose that would be more realistic but if it adds complexity there's no need to worry about it.

pescadero · #6 11/30/2007, 06:04 PM

Quote:

Originally posted by RyanBrucks
"so how does your dimmer work? are you changing the voltage, or are you maintaining regulated voltage while adjusting the current?"

Well I have not yet wired up my potentiometer, but I thought they changed voltage while maintaining current?

That's not really how LEDs work. they're diodes. they won't illuminate until you exceed the minimal forward voltage drop that's required to light the diode, and then the diode lights up. varying voltage isn't really an effective way to dim an LED with any degree of controllability. what you're supposed to do is maintain the voltage at a constant (regulated) value that exceeds Vf, and you're supposed to modulate the amount of current being applied to the device in order to perform dimming functions. typically this done with a voltage regulator IC like an NS LM317.

if you look at the tech data sheet for your LED, it will list its operating parameters. knowing exactly what you're dealing with would help a lot. The app guides for the regulator ICs will have helpful information in them. You could always google for more info.

pescadero · #7 11/30/2007, 06:05 PM

oops.

pescadero · #8 11/30/2007, 06:10 PM

Quote:

Originally posted by RyanBrucks
Does anybody have any good recommended reading on learning how to set up timer circuits with a microprocessor?

The type 555 is a timer IC that is old as dirt and well documented online. that would be a good place to start reading.

RyanBrucks · #9 11/30/2007, 06:18 PM

I'm using 2 of these LuxeonStar Light engines

Each one has a 'buck puck' driver

the driver supplies constant 350mA of power to each LED, and voltage must be at least 2 volts higher than the minimum voltage for each LED.

That light engine spec doesn't list the exact info for the LEDs, it just says that it supports different LEDs. I'm actually not completely sure which LED mine is using, but I believe the minimum forward voltage for them is about 1.75, since I got a support email from luxeon saying I'd need at least 23volts to power my 12 LEDs.

Trying to track down the exact info.

"typically this done with a voltage regulator IC like an NS LM317."

sorry I'm a bit of a n00b when it comes to electronics..... but a voltage regulator regulates current, not voltage?

RyanBrucks · #10 11/30/2007, 06:24 PM

erggggg, double post

pescadero · #11 11/30/2007, 06:34 PM

Luxeon LEDs are nice LEDs. i use them. there's detailed info on them at the luxeon web site. from memory i can tell you that the 1W blue LEDs are spec'd to draw anywhere from 100-400mA across the dimmable range.

i am not familiar with the buck puck driver. i haven't read the spec sheet, but t looks like a nice pre-made device that simplifies things so that all you need to worry about is twisting the 5k pot.

better than twisting a 5k pot though, would be to replace the pot with something like a 555 timer that controls an optoisolator (voltage controlled resistance element), but we're heading into electronics geekdom now.

BTW the price on that regulation puck is huge. i've built regulation/dimmer circuits for infineon LEDs for all of about $12 in parts.

pescadero · #12 11/30/2007, 06:40 PM

Quote:

Originally posted by RyanBrucks
"typically this done with a voltage regulator IC like an NS LM317."

sorry I'm a bit of a n00b when it comes to electronics..... but a voltage regulator regulates current, not voltage?

yes, i can see how that would be confusing. sorry about that. you're right, Vreg chips are designed to regulate voltage, not current. but that doesn't mean that you can't design a circuit around them that regulates current instead of voltage.

if you look at the LM317 applications guide you'll see that example schematics for both voltage and current regulation circuits are provided.

board is really slow tonight. lots of double posts.

pescadero · #13 11/30/2007, 06:41 PM

oops. man, the board is making me look like a dufus tonight.

RyanBrucks · #14 11/30/2007, 06:46 PM

yes each LED is 350mA, as clear from the label on the puck driver.

I'm pretty sure this is the LED its using

Looks like the specs are the same for the lightengine I bought. Lumens and Amps match and it looks the same:

Lumens 45 lm
Max Continuous Current 350 mA
Forward Voltage 3.42 Vf
Dominant Wavelength 5500 K

If that is indeed true, it suggests that I would need at least 43 volts to power the 12 LEDs that I am currently powering (3.42x12 + 2).

However, the maximum rated voltage for the luxstrip engine is 32, and I am currently powering 2 of them (12 LEDs total) with only 30 volts and 1amp. Even stranger is that the lights still work when the adapter is in the 13v mode (dims them down alot).

Luxeon replied to me in an email saying that 23volts is enough for both strips so now I'm definitely confused.

Also, this must be another n00b question, but current doesn't drop with each LED like voltage, does it?

I'm looking into those 555 timers right now.

From this page

it says
"Some of the more attractive features of the 555 timer are: Supply voltage between 4.5 and 18 volt, supply current 3 to 6 mA, and a Rise/Fall time of 100 nSec. It can also withstand quite a bit of abuse."

Are they really a max of 18 volts? If so it seems like it wouldn't be useful for my application?

Thanks,
Ryan

stugray · #15 11/30/2007, 07:05 PM

I looked at the datasheet, and this is a nice unit.

You can control the intensity through:

Variable resistance - the Pot that you have now
Analog output device ( Digital to Analog converter )
Pulse width modulation ( PWM )

You can get a digital Potentiometer chip ( lookup: XICOR EEPOT )and set it using a microproc.
You can build a 555 timer circuit that takes a VERY long time to charge and use the Cap voltage but thats complicated.

Pretty much all the options require some sort of processor in the loop. Do you have a controller already?

Stu

RyanBrucks · #16 11/30/2007, 07:13 PM

Nope, I do not have a controller. I'm still researching what my options are. Without understanding all the details I wouldn't really know what parts to buy.

"Analog output device ( Digital to Analog converter )
Pulse width modulation ( PWM )

I'm not familiar with either of these techniques. How would a DAC allow me to control intensity here? I've only used DAC for audio. Is it just what I need to convert back to analogue after using a digitial solution (like digital pot+microproc)?

Also, I have no idea what pulse width modulation is. Is this an alternate method you're suggesting, or an integral part of doing it period?

I really appreciate all the help already in this thread

If I ask somebody 'normal' for any help on this subject, I just get a look that says 'are you insane???' but that probably goes for alot of threads on this forum

Donw · #17 11/30/2007, 07:58 PM

Quote:

Originally posted by RyanBrucks
Linear fading from bright to dim would be fine.

By 'slowly progressive' do you mean like a non-linear brightness ramp? I suppose that would be more realistic but if it adds complexity there's no need to worry about it.

What I was getting as is do you to want it to get dimmer a little at a time or do you want it to go from full on to dim in one step. If its the later all you need is your pot and a timer.

Don

pescadero · #18 11/30/2007, 08:05 PM

the "supply voltage" for the 555 is the power supply voltage that you need to power up the chip. that's not going to be directly related to the voltage that you'll be providing to the LEDs.

as far as stacking the LEDs goes, if you wire them in series, you have to multiply each LED's Vf by the number of LEDs in the stack to get the total Vf for the stack. Because the LEDs are wired in series, the current flow through one of them would be the same as the current flow through any other. To answer your question then, the total current remains the same regardless of how many of them you wire in series, and only the Vf for the array will change.

you know, there might be a simpler way to do the timer thing that won't require a processor that's more complicated than the 555. use an optoisolator to convert voltage into resistance, and use the resistance to replace your 5K adjustment pot. then drive the opto with a voltage that ramps up very slowly and ramps down very slowly, just like the moon gradually rises and gradually and sets with the earth's rotation. all that you'd need are really large RC values to provide a long time constant for the 555.

RyanBrucks · #19 11/30/2007, 09:20 PM

"or do you want it to go from full on to dim in one step."

Ah, guess I misunderstood your initial question then. I definitely want it to take some time to turn on and turn off.

Just guessing, here's how I'd have it set up:
at night time, operate at ~10% brightness
~5-10 minutes before actinics turn on, slowly ramp up brightness to 100%, so that 100% is reached as actinics turn on.
Same thing in reverse after the actinics go off at night.

"all that you'd need are really large RC values to provide a long time constant for the 555."

http://www.uoguelph.ca/~antoon/gadgets/555/555.html

I found that page helpful to understand what you're talking about

t = R X C

Assume a resistor value of 1 MegaOhm and a capacitor value of 1uF (micro-Farad). The time constant in that case is:

t = 1,000,000 X 0.000001 = 1 second

Is that what you meant by large RC values? I would need a resistor with very high ohms to achieve a very long delay?

Ryan

pescadero · #20 11/30/2007, 10:18 PM

i was thinking about much longer time constants, so the power-up and power down could take an hour or so. nice and gradual, like a sunrise or a sunset (or a moonrise and a moonset).

RyanBrucks · #21 12/01/2007, 02:34 AM

that T=RxC was just copied directly from that article, not meant to indicate the amount of time I would use. I agree that a much longer time value would be better.

I was just wondering if that was the same varianble you were referring to. I know that all the people responding to this thread have a pretty good knowledge of electronic circuits. Please understand that I've been working with computers and programming since my teens, yet electronics is almost completely foreign to me. I guess what this means is that while I most likely understand many of the concepts of electronics, most of the terminology is still foreign to me.

obviously the value of the variables would be changed based on the specific application.

it seems like getting things timed on a 24 hour cycle where it matched up every day would be very difficult. would it be eaier to have a regular timer turn off everything in the middle of the night help to reset everything, so that there is no slowly creeping in error?

seems like this topic is far to complex too be summarized in a thread, so I definitely have alot of reading to do on my own. That said, this thread has given me several starting points in my research that will no doubt save countless hours down the road.

I hope that others can somehow benefit from this madness. I'll probably be asking more stupid question within the next day or so

Ryan

pescadero · #22 12/01/2007, 10:20 AM

Yes, T=RC was what I was referring to. Sorry that I was not clear about that.

I agree that an electronic control circuit is a complicated project, but there's no reason not to dedicate a thread to it. I have had the idea of doing what you're doing on the back burner, and the only reason that I haven't worked on it yet is because in setting up my system there are a number of other tasks that I need to take care of first.

pescadero · #23 12/01/2007, 10:29 AM

Quote:

Originally posted by RyanBrucks
it seems like getting things timed on a 24 hour cycle where it matched up every day would be very difficult. would it be eaier to have a regular timer turn off everything in the middle of the night help to reset everything, so that there is no slowly creeping in error?

well, it all depends upon the accuracy of any individual timer. the 555 chip is one of the first timers. its well documented, easy to use, and there are zillions of established circuits that you could look at for ideas. that isn't to say that its the most accurate timer chip, though. if you want watchlike timekeeping accuracy it may not be the best choice. it will drift, but drift isn't all that bad, because the time of sunrise and sunset changes from day to day.

to deal with clock drift you could just use a reset button to restart the timer at zero whenever you notice the drift being very far off. then you install the circuit you could start it at its designated on-time, and after a while if you notice some drift, just press the button to reboot and resync the time at t=0.

BeanAnimal · #24 12/01/2007, 11:05 AM

A few lines of code in a PIC or ATMEL will do all of the work. Add an simple I2C RTC and youve got a very simple dimmer.

Now mine does sunset/sunrise, moonset/moonrise/percent illuminated and does these things based on date/time and longitude and latitude. It is a bit more than a few lines of code....

But I think you get the idea

pescadero · #25 12/01/2007, 11:35 AM

Bean, although I'm familiar enough with electronics to build this sort of thing from scratch, I have to confess that I'm not at all familiar with the current crop of appliances that are designed to do this sort of thing as a turnkey solution, like the I2C interface stuff for PCs. Instead of re-inventing the wheel, I should probably look into this stuff. By any chance could you provide a link to some info on the the hardware options so that I could work on coming up to speed?