Lighthouse Circuit

[kcflyer]

Before my father passed away a few years ago, he was working on a lighthouse project to set in front of our lake house. The project was partially completed, and the family and I would like to finish it in his honor.

I'd like to bring it to life with dusk-activated timer circuit that will turn the 120v incandescent light off and on in a smooth wave-like brightening and dimming with about a 2-second cycle. It looks like the dark-detector circuit (fig. 1) in Tony van Roon's will get me much of the way there, leaving me with two main questions:

1. How do I produce the gradual brightening and dimming of the light?
2. How do I actually control the light with the low-voltage control circuit? Would a triac be most appropriate?

 

[Hero999]

1. Use a ramp generator - a 555 will do this.
2. Use a voltage controlled dimmer - I've designed a circuit to do this but it isn't quite finished.

Or:
3. Build a circuit using a PIC which will perform both functions but I don't know how to do this so don't ask me.

 

[Rolf]

Before my father passed away a few years ago, he was working on a lighthouse project to set in front of our lake house. The project was partially completed, and the family and I would like to finish it in his honor.

I'd like to bring it to life with dusk-activated timer circuit that will turn the 120v incandescent light off and on in a smooth wave-like brightening and dimming with about a 2-second cycle. It looks like the dark-detector circuit (fig. 1) in Tony van Roon's will get me much of the way there, leaving me with two main questions:

1. How do I produce the gradual brightening and dimming of the light?
2. How do I actually control the light with the low-voltage control circuit? Would a triac be most appropriate?

Why not make it rotate like a real light house? It would look much more realistic. Mount the bulb horizontally in a reflector and turn it at the proper speed with a small motor. Problem is that I cant think of a way to mount it without using slip rings. Or maybe a rotating mirror placed at 45° would do it.
For the proper effect the cycle time should probably be much longer than two sec's.

 

[philba]

I agree that the rotating beam adds some realism. it would probably need to be some sort of hood to prevent light leakage.

the problem of PWM pulsing has come up before. It's an interesting problem by itself. I know how to do it with a microcontroller, triac and zerocrossing detector but the non-uC approach isn't as obvious to me.

 

[Prense]

Why not make it rotate like a real light house? It would look much more realistic. Mount the bulb horizontally in a reflector and turn it at the proper speed with a small motor. Problem is that I cant think of a way to mount it without using slip rings. Or maybe a rotating mirror placed at 45° would do it.
For the proper effect the cycle time should probably be much longer than two sec's.

I'd think it would be easier to rotate a housing around the light...focusing the light at an appropriate angle as the housing rotated.

 

[HiTech]

A rotating beacon is easy to construct but many of you folks don't realize that it's disturbing and annoying to any nearby neighbors who might be relaxing on their porches or whatever other annoyances it can cause to the neighbors... even if the lamp is kept dim. His initial thoughts to have the light blink on and off gradually are easier on the eye, eventually going unoticed and neighbors should welcome seeing such a novel lawn ornament.

 

[kcflyer]

I think I'd rather go the electronic route rather than the mechanical route. The electronics will be more reliable and a lot more flexible. As for the neighbors, that's a valid concern. I'm not planning on making this a bright beam, just a soft brightening and dimming, probably with a colored bulb.

One other consideration is that this will operate unattended for most of the time, as we're only there on the weekends during the warm months.

 

[ParkingLotLust]

Just use tp1 and tp2, omitting tp3. Now you just need a way to control the light. A TRIAC dimmer maybe?

 

[bigkim100]

Fading led circuit

**broken link removed**


Here is a circuit that slowly illuminates, then slowly extinguishes a pair of LEDS.
You will still need to find a schematic to interface this to a 120 volt lamp.
Its from Bill Bowdens website. Last time I referenced to a schematic on his page...Bill Bowden tossed a fit, even though I was very careful to state that the circuit was his.
Hopefully hes taken his medication, so that this reference wont cause him to wet himself.

 

[philba]

It's not like he came up with that one. It's been around forever. he has some nice circuits but they aren't THAT original. I wonder why he has them on a web site if he doesn't want people refering to them??

unfortunately, it doesn't pwm but, rather, varies the voltage to the LEDs. You could use that as the control voltage into a pwm circuit though you'd need to deal with the output voltage range being a bit low.

I think I would use a 12 VDC lamp (via a mosfet) instead of 120 VAC. then you dont have to deal with zero crossing sync for the PWM cycle.

 

[kcflyer]

Just for my education, what happens when you use PCM but don't synch to the zero crossing? And how in general is that accomplished?

Interesting idea about the 12v bulb. That might simplify things considerably. Where's a good source for 12v colored bulbs? How's their longevity compared with a 120v bulb?

If I go with a 12v bulb, what would the output side look like? A power transistor? I've built a lot of circuits in my lifetime, but they often tend to quit working after a while because I probably overdrive the components.

 

[philba]

the way AC dimming works with a triac is that the triac is turned on at a certain point through each half cycle. for example, if you always turn it on 1/2 half way through the cycle, you get 50% brightness but you need to start timing at the beginning of the cycle. so the 50% period would be 1/240 th of a second starting at the start (zero cross) of the AC cycle. otherwise, you will turn the triac on at a random point in the AC waveform and thus get a random brightness. You will also need the PWM period to be pretty close to the half cycle period (1/120). technically, this is called phase control rather than PWM but the principles are quite similar. This app note is a good intro to phase control: **broken link removed**

automotive bulbs are pretty rugged though you might need a colored lens as I have only seen them in clear. I think they will last a fairly long while though given your usage (on/off all the time), I don't know if it will have longer or shorter life. You may avoid thermal stress since it will get hot and stay hot so the life may be very long.

yes, a power transistor would work but I'd use a mosfet since it will probably run cooler. If you have an NPN power transistor lying around, go for it. just calculate out the current and then the wattage dissipated by the transistor to be sure you won't emit smoke. definitely over size the transistor/mosfet. see the attachment for a simple nmosfet circuit. the resistor R1 is not necessary and can be fairly large (10K or more). The IRL520A can handle a 25W 12V lamp no problem. cost a buck from mouser. At a 50% duty cycle it will generate 250 mW or so - relatively cool. if you want belts and suspenders, use a modest heatsink on it. you could replace themosfet with a power NPN but need to calculate out the resistor to ensure proper current. you might need a drive transistor if the gain of the power transistor isn't high enough (or use a darlington).

 

[kcflyer]

I think I've decided to go a different direction with this. Now I'd like to have a ring of 10 bright white LED's fired in sequence to simulate rotation. When I was a kid, I remember building that exact circuit with an IC that counts pulses from a 555 timer and energizes one of ten outputs in sequence. I forget the number of that count-to-ten IC. Does anyone know the one I'm talking about?

If the light output from lighting one LED at a time is insufficient, I may use a layer of OR gates so that two or more ajacent LED's are lit at the same time.

Any thoughts on this idea? Do the super-bright white LED's draw enough current to require a driver transistor, or can they be driven directly off the IC?

 

[kcflyer]

I think I've answered my own question. A 4017 decade counter is the chip I'm thinking of. And this is the circuit I was talking about:

**broken link removed**

It's like finding a long-lost friend! ;^D

 

[bigkim100]

John Fields
Professional Circuit Designer


+12V>---------+--------------+-----+-------+--------------+
| | | | |
| | | | [510]
[10K] [1M] | | |A
| | | | [LED]
| +--[10M]--+ |R2 | |
|R1 | | [10K]<---|-\ E
[10K]<-+-|+\ | | | >-+-[10K]--B 2N3906
| | >----+-----|------|+/ | C
IN>--[100K]---|------|-/U1A | | U1B| |+ |
| LM393 |+ | | [22µF] |
| [10µF] | | |C2 |
| | | | | |
GND>----------+--------------+-----+-------+---+----------+


Mr. Fields was kind enough to design this circuit for me, to simulate the gradual ramping up and down of a incandescent lamp, instead of the instant on and off of a LED.
Once again Mr. Fields...thanks
Working with the 4017 and a 555 is tons of fun, and can be adapted in MANY ways...Ive built MANY lamp chasers with them (I have several MODEL AND PROP making buddies left over from my days in the film business, and they love to use this for interior lighting)...and they always work. I have a tattered photocopy of a Popular Electronics tutorial on working with the 4017, with many many schematics of different effects that an be made. My favorite is the fact that you can daisy chain them together, and easily make a 10, 20,30.....etc lamp chaser.
I once built a 4017 Squib detanator for bullet hits, to simulate machine gun fire in movies. It ws used hundreds of times in many different movies, by may different effects guys...much better than the traditional "Nail Board" method. I built many customed designed units for FX buddies, including tiny ones that could be carried in the costume of an actor, to detonate bullet hits worn by the actor... again, a huge technalogical advancement, as prior to this, often a cable was run down the actors leg to a "Nail Board" to detonate the Squibs all over the body... a cumbersome, and difficult to hide from the camera way of doing things.
You can drive low voltage incandescent lamps instead of using LEDS thereby simulating a nice delayed dimming of each bulb, just add a cheap switching transistor to each output to drive the lamp directly

 

[HarveyH42]

Guessing this was all twenty some years ago... Text based schematics are reformatted, not sure if you have any options on that. 4017s are great chips.

 

[Roff]

I think I've answered my own question. A 4017 decade counter is the chip I'm thinking of. And this is the circuit I was talking about:

**broken link removed**

It's like finding a long-lost friend! ;^D

You don't need a current limiting resistor for each LED, because they are only on one at a time. See **broken link removed** to see what I mean, and maybe for some more ideas.

 

[audioguru]

With a supply voltage of only 6V to 9V, a CD4017 can drive LEDs directly without a current-limiting resistor. The LED current with 6V will be only about 4mA and about 12mA with 9V.

 

[kcflyer]

I'm going with the circular array of LED's driven by the 4017. And I'm also using an array of diodes so that two neighoring LED's are lit at the same time. I've also found a place where I can get mini-maglite reflectors for about a buck each.

But I am interested in Mr. Fields' circuit diagram. Can you repost in a more readable format? Looks like the forum software "helpfully" eliminated all the extra spaces.

 

[kcflyer]

So how much current can a 4017 safely output? These super-bright white LED's are rated to draw about 20-40mA average, 200mA peak. Can the 4017 handle it in the long run? I'm running this off a 12v power supply.