Lighthouse Circuit

[Hero999]

I can't read that ASCII schematic, try using code tags.

 

[audioguru]

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.

With a 12V supply, the outputs of a 4017 will try to give a max current of about 20mA. But if a series resistor isn't used then the tiny output transistors will have about 9V across them and might melt from the resulting 180mW. Use a 330 ohm series resistor and each output's current will be about 12mA.

My Ultra-bright Chaser projects have a 74HC4017 driving about 25mA through its LEDs in very short blinks. They are blinding. The max supply for the 74HCxx series is only 6V to 7V.

 

[kcflyer]

In fact, I am using 330 ohm series resistors, so it sounds like I should be OK. I've thought about using a common 330 ohm resistor for all of the LED's as shown in some of Bill Bowden's designs referenced in this post, but my concern is that since I'm driving two LED's at a time, if one of them opens up I will then have twice as much current on the other LED.

BTW, how do you wire 4017's for more than 10 outputs? Do you connect the clock inputs in common and then connect the carry output to the clock-enable of the next 4017 in the series?

 

[audioguru]

The Carry Output of a 4017 doesn't do what you think. Bowden shows how to cascade 4017's.

 

[bigkim100]

The Carry Output of a 4017 doesn't do what you think. Bowden shows how to cascade 4017's.

Ive cascaded 10, 4017 units together to create a super long flashing direction arrow.
Personally, instead of white leds, would use ultrabrite yellow leds. Some are very "washeds out" colour as they are so bright, and can be mostly white with just a tinge of yellow.
I was at a surplus store today, and they were selling 12 volt rotating beacons for $5.00 each. I just had to buy one, for halloween.

 

[kcflyer]

Here's the circuit diagram for the lighthouse.

 

[Roff]

Here's the circuit diagram for the lighthouse.

I think you got your 22mF smoothing cap on the wrong pin.

 

[philba]

A couple of points:
- you should put a filter cap on the regulator input. Probably 470 uF or so. also, I'd put a 100 nF on the output of the regulator. see the datasheet for more details.
- I don't know if the 4017 can handle that much current on the output pins. the TI data sheet says 100 mW max per pin and at 12V that's 8 mA. since you are driving 2 LEDs, that means each should be 4 mA or lower to stay within the datasheet spec.
- I think your LED resistors are too low. you will get a bit less than 12V out of the 4017, subtract off the diode drop and call it 11V. With an LED Vf of 2V, your resistor will see 9V. Ohms law says it's 9/330 or 27 mA. Doubling this up, your 4017 pins will see 54 mA. pretty high. 470 would get you to 20 mA per LED but 40 mA through the 4017 pins still seems high.

I'd use a driver chip like a ULN2803 or transistors to drive the LEDs.

Now, sorry if this isn't what you wanted to hear, I'd rather you know in advance. The effect you will get from this circuit when the errors are corrected is that the LEDs will, in turn, come on to full brightness and then go off. I'm not sure that will be very realistic. Having lived around lighthouses, the effect is a fast ramp to full brightness, several hundred mS duration of that and then a fast ramp to dark. Lighthouses also often have a dim flash 180 degrees out of sync - reflection off of the glass, I think.

 

[kcflyer]

No problem. I appreciate the feedback.

I added the smoothing capacitor off the 555's clock output because I was picking up quite a bit of noise and the 4017 was counting erratically. Adding that capacitor eliminated that problem. If that's the wrong solution I'm open to other suggestions.

As for the resistors on the LED's, here's how I arrived at 330 ohms. I measured about 0.7 volts drop across the diodes. The data sheet for the super-bright LED's (Calrad 92-120-WH) provides the following formula for computing the resistor: R = (Vin - Vf)/If, where Vin in the input voltage (12.5v minus 0.7v across the diode), Vf is the forward operating voltage of the LED (3.4-4v for this particular LED), and If is the LED current draw, which is listed as 30ma average. So R = (12.5 - 0.7 - 3.4)/.03 = 280 ohms. I took it up to 330 to be safe.

I was wondering about the need for output transistors. I've operated the circuit for several days more or less continuously without any noticeable problems, but if it will increase the longevity of the circuit I'll add the transistors. Which side of the diode array would be the best place to put them?

What I'm doing with the LED's is arranging them in a circle with a mini-maglite reflector behind each one. It may not be a perfect approximation, but I think it will be close enough for my purposes. I'm still waiting for the reflectors, so I can't test that assumption until then.

 

[philba]

with an If of 30 mA, then you are correct. The 4017 spec may be very conservative - I was just going on that. I know that the 74HC parts are pretty conservatively spec'd when it comes to putput drive, maybe the 4xxx series is too though I thought it wasn't. The cd4017 DS says 200 mW for the chip and you will probably be under that. My guess is you are heating each output driver but the duty cycle is long enough that it cools down. Keep an eye on it, though. If the chip runs cool to the touch then I wouldn't worry about it but do watch the ambient heat (i.e. vent the assembly housing).

I suspect the VReg not having a filter cap was causing your glitches.

 

[Roff]

Here is one way of adding transistor drivers. Of course, you could instead use emitter followers between the junction of your diodes and the 330 ohm resistor.

 

[kcflyer]

That looks great. But I'm not sure if I can dispense with the diodes because expanding this to all the outputs would result in all the outputs triggering all the transistors, wouldn't it?

What's the purpose of the 10k resistors? Is that just to limit the current from the output of the 4017?

 

[Roff]

That looks great. But I'm not sure if I can dispense with the diodes because expanding this to all the outputs would result in all the outputs triggering all the transistors, wouldn't it?

What's the purpose of the 10k resistors? Is that just to limit the current from the output of the 4017?

What do you mean by "expanding this to all the outputs"?
The 10k resistors limit the base current into the transistors.

 

[bigkim100]

nice little circuit dagram. One if the posters is correct, you will need additional smoothing caps in your power supply, as tghe 555/4017 combination is fairly sensitive to noise in its supply line. Your diode matrix is particularly clever to overlap on the led outputs. Mini mag reflectors are also a good idea. I would still take a look at ultrabright washy yellow leds. As I said, they can often simulate the output of a bright incandescent.

Although it may overcomplicate a fairly simple, straightforeward design, I would like to hear from the group, in regards to simply simulating the ramp up/down of a incandescent. I have tried it before using a resistor/electrolytic/transistor combination, with mixed results. The schematic that I posted in a earlier thread was a attempt to simulate this....but its a lot of extra components X10.
DOES ANYONE HAVE A BETTER IDEA TO DO THIS????.

Please post pictures of your completed project, in a new message, so we can all take a look at the completed version. If at all possible, include a link to a video clip of it in action.

 

[philba]

If you use a microcontroller, you can do this with just a few components. For an AC incandescent lamp, you would use a TRIAC and an opto driver (moc30xx) and phase control it. For a DC lamp, control a MOSFET via PWM.

The nice thing about using a microcontroller is you can arbitrarily set the brightness curve.

 

[kcflyer]

I'll answer several posts at once here.

What I mean by expanding to all the outputs is this: output 0 should light LED's 0 & 1, output 1 should light 1 & 2, output 2 should light 2 & 3, etc. If another output was added to Ron H's driver section, you'd have another transistor with the base connected to Out1 & Out2 via resistors. The result would be that the bases of all the transistors would be connected in common, so a trigger to one would be a trigger to all. Hence the diodes are still needed to prevent that.

Also, does it matter whether the LED is wired to the collector or the the emitter of the transistor? I'm guessing not, but want to make sure.

I've already purchased the white LED's so I'll stick with them for now, but I'll try the yellow ones for future projects.

Yes, I'll be happy to post some pictures & video clips when it's finished. I've also had a local sheet metal shop make a copper roof cap, so it will look sharp, especially as it ages.

 

[Nigel Goodwin]

Also, does it matter whether the LED is wired to the collector or the the emitter of the transistor? I'm guessing not, but want to make sure.

It's VERY IMPORTANT that's it's wired in the collector, putting it in the emitter is a completely different type of design, and VERY inferior for this type of use.

 

[Roff]

What I mean by expanding to all the outputs is this: output 0 should light LED's 0 & 1, output 1 should light 1 & 2, output 2 should light 2 & 3, etc. If another output was added to Ron H's driver section, you'd have another transistor with the base connected to Out1 & Out2 via resistors. The result would be that the bases of all the transistors would be connected in common, so a trigger to one would be a trigger to all. Hence the diodes are still needed to prevent that.

Not true. I don't think you understand transistors.
The transistor circuit is an OR gate. The transistor will turn on if Out0 OR Out1 (or both, which can't happen with a 4017) is high. If you add another 2-input OR with Out1 and Out2 as inputs, it will only turn on when Out1 OR Out2 is high. Out0 will have no effect. The same will be true if you expand it to 10 OR gates.

 

[kcflyer]

OK, I think I understand what you meant now. I was assuming that the next transistor in the series would share the 100k resistor from the shared output. But what you're talking about would have two 100k resistors connected to each output like this, right?

So can someone give me a nutshell explanation of what difference it makes whether the LED is driven off the emitter or collector? I'm largely self-taught and this is an area where I'd like some more education.

 

[Roff]

OK, I think I understand what you meant now. I was assuming that the next transistor in the series would share the 100k resistor from the shared output. But what you're talking about would have two 100k resistors connected to each output like this, right?

So can someone give me a nutshell explanation of what difference it makes whether the LED is driven off the emitter or collector? I'm largely self-taught and this is an area where I'd like some more education.

Yep. You got it.
Actually, in this case you could put the LED in the collector or the emitter circuit. You just have to modify your current limiting resistor value slightly to get the same current. See below.