LED Outdoor lighting question

Thread Starter

trader007

Joined Feb 27, 2010
249
LOL! Yeah that would be hilarious to see. I love looking at hackjob installs.

I have been an electrician for over 10 years, with extensive training in safety. I don't do anything questionable when it comes to electrical work. I always over-do it a little. Also why I am getting into LED accent lighting, as it's inherently safe as long as you use the proper power supplies and install techniques.

The only electronics outside besides the diodes themselves would be the photo sensor and the PIR sensors.

I am going to try these sensors from china. I would rather have something from honeywell or whatnot, but the price can't be ignored with these- http://www.ebay.com/itm/221236158512

and the PIR sensors here- http://www.ebay.com/itm/350792365698

I was looking on digikey and noticed I could buy the actual photo and PIR sensors themselves, but I would need to add the circuitry to go with and that's just more time consuming. These come with what they need to work, in a nice small package. The PIR sensor has a 3min timer built in too, and I figured I will probably be able to add some capacitance to extend that to about 10min.
edit- again, sorry but this is trader007! Somehow I am logged in as trader007 yet it posts as wirednuts. I must have two accounts... very weird.
 

LDC3

Joined Apr 27, 2013
924
so the question is, do I find a larger transistor or should I just double up a couple BD679's? If I have to look for a larger variant, is there a special kind of NPN transistor that I need or are they all generally the same?

Edit- I believe these will work. 135w max, so thats well above what I need for still a good price- http://www.mouser.com/ProductDetail...=sGAEpiMZZMshyDBzk1/WixbhOhHbn/druthM5RMK1TE=
Go with the larger transistor. It is very difficult to get 2 transistors to share the load equally. Usually, one of them ends up providing all the current and overheats, then the other one overheats. Then you replace the circuit board.
 

Thread Starter

trader007

Joined Feb 27, 2010
249
OK sounds good.

Another question. In reference to the earlier statement about using computer power supplies. Not outdoors of course, but I am intrigued about using them for two reasons-

1) low cost with long warranty
2) low-current on switch (ground the green wire out and it turns on the supply). This would allow direct control from the PIR using a transistor.

Of the MeanWell supplies I have seen on Jameco, they do not have a low-voltage on/off switch. The benefits to the Jameco supplies-

1) Single-output. All I need is 12v and that's all they do. KISS, and keep the clutter of wire out of it
2) Passive cooling. Should last longer by keeping the dust from clumping up, especially when mounted in dirty basements.

So what do you think I should pursue? I wouldn't mind opening up the MeanWell supplies and finding a way to switch them on and off, however that would void the 3 year warranty...

If anything, I probably should just use the MeanWells, and switch them with a relay, but that is more parts/cost/labor as well... I just don't really know what route to go with.... and I don't want to leave the MeanWell supplies on 24/7 because even at 'no load' they draw about 7 watts. Then again, I have never even tried measuring what a computer supply draws in standby, so maybe that point is moot and I just need to switch the DC load side of the MeanWell.
 

Thread Starter

trader007

Joined Feb 27, 2010
249
I believe you could even keep a dim level by putting a power resistor in parallel to the darlington transistor. It would limit current to the entire LED strip when the transistor is off. When on, the transistor would take the current and bypass the resistor.
I am close to building the first one, complete and installed on a deck.

I have decided, instead of a power resistor in parallel with the darlington transistor, I will use a PWM driver. This will allow me (or the home owner) to change the brightness level when in "dim" mode, or even eliminate the dimming function by just turning a knob.

It should also reduce power consumption, and be a little safer.

edit- just built the circuit on the breadboard. Works great. I am going to need to increase the capacitor about 10 fold, though. I want a much slower fade in and out. Noticing I can't adjust the resistors too much either, or the transistor doesnt fully sink.
 
Last edited:

Thread Starter

trader007

Joined Feb 27, 2010
249
Question- I have been doing more breadboarding. I have the fade on/off working pretty well. What I can't figure out is something very simple.



I want to bridge the collector and the emitter with a 5K resistor to allow for a "dim" mode when the switch is off. I have done some bench testing with just one set of diodes to find that 5K is a good resistance for this.

How do I figure out how many amps I will be drawing on a full string in this mode? My multimeter wont measure the minute current of just three diodes at this dim level. The ropelights consume up to 6a@12v, or 72w. I am hoping a 7w resistor will be enough as they get much more expensive when they go larger then that in the 5K range...
 

wayneh

Joined Sep 9, 2010
17,495
Your test shows that (12V - 3*Vf)/5kΩ= ~0.5mA is the right current for "dim". That sounds about right to me.

First, recalculate the above using the accurate values. Multiply the calculated current by however many parallel strings you have. I'm guessing it's around 30? Recalculate the formula to find what the resistance should be to achieve that total current.

Check the power dissipation = I^2 • R. Choose a resistor rated to at least twice that amount, to ensure it lasts without burning up.
 

Thread Starter

trader007

Joined Feb 27, 2010
249
Well that explains why my 1/2w resistor works just fine for the "Dim" mode. I wasn't thinking logically, it makes sense to me now why the current is so low at that point in the circuit. Now, when I added all 600 diodes, the 5k needed to be about 150ohm to achieve the same results. Still, 1/2w resistor seems to be holding just fine.

I redrew the drawing as I have it. Its very different from the first diagram because the transistor I am using is different.



I also removed the resistor from the base to ground. I am not sure why, but it seems to work better without it.

The transistor I am using is a BU931p. I have it pushing two LED strips right now, so 600 diodes, and its getting warm but I have a good heatsink on it.

With the 2k resistors and 2000uf on the cap, it fades as I would like it to. Its about a 3 second fade time. If I raise the resistance anymore then 2k, the transistor doesn't work.

I also notice about a 1v drop in output, so I will need to modify the power supply to push more voltage in so I can get 12v out. Of course, this will likely change the behavior again and I will be changing numbers around more :D
 
Last edited:

wayneh

Joined Sep 9, 2010
17,495
...the 5k needed to be about 150ohm to achieve the same results. Still, 1/2w resistor seems to be holding just fine.
I recommend measuring the ∆V across that resistor, just to be sure the power dissipation (V^2/R = I^2*R) is less than 1/4W or so. Or just put a finger near it to check for heat. Be careful, it could be a LOT hotter than you expect.

.I also removed the resistor from the base to ground. I am not sure why, but it seems to work better without it.
That should be fine. There's plenty of path to ground for the base through the transistor itself.

I also notice about a 1v drop in output, so I will need to modify the power supply to push more voltage in so I can get 12v out.
Be sure to stay safely within the specs of your supply. If it's a regulated supply and your pulling it down a whole volt, that's a concern.
 

Thread Starter

trader007

Joined Feb 27, 2010
249
I am not sure how to measure delta voltage on a resistor in a live circuit. In any case, I have left it on in "dim" mode now for over an hour, and the resistor barely gets warm. Even when I parallel up the resistor to make it 60 ohms, it still hardly gets warm. Honestly it doesn't make a whole lot of sense to me, but with my basic knowledge of ac/dc theory I am not surprised. I suppose since the resistor is on the drain side of a direct current path is why it doesnt see much load...

The power supply I plan on buying is a meanwell with a voltage adjustment pot built in, spec'd to give me 13.4v max. I would imagine that should be enough.
 

wayneh

Joined Sep 9, 2010
17,495
I am not sure how to measure delta voltage on a resistor in a live circuit.
It's as simple as placing one lead on each side of it.
In any case, I have left it on in "dim" mode now for over an hour, and the resistor barely gets warm.
Well that's really all you need to know. Sounds good.
I suppose since the resistor is on the drain side of a direct current path is why it doesnt see much load...
Current is the same throughout a circuit, top to bottom, and current is all that matters for heating a resistor. In other words, it's staying cool because the current is low.

I got to thinking that, rather than having a minimum current under control of that resistor, you could also perhaps limit the minimum PWM duty cycle of the controller to achieve the same effect. Could be a bit more elegant than a resistor, but it sounds like the resistor solution is working fine and not really wasting much power. Hard to justify making a change.
 

Thread Starter

trader007

Joined Feb 27, 2010
249
You have been a huge help, thank you.

Still surprising to me how 30ft of ropelights use less then 1/2 watts to keep a dim output, especially when they can draw up to 140w total at full current.

I also thought about controlling the darlington itself. Wouldn't it work if I just added a diode between the base and the second resistor, then add a pot between + and base? Of course, the diode would again change my resistor values for the fading, but I think it would work?



Or will the diode create problems?
 
Last edited:

wayneh

Joined Sep 9, 2010
17,495
Or will the diode create problems?
I don't think so. It will require 0.7V more on the signal to deliver the same voltage to the base, but that shouldn't be an issue. May affect your timing.

I'm not sure how that variable resistor will work, and I'm not sure how reliable a solution it will be, but some experimentation will answer that for you. I mean, it should work on paper, but may be affected by operating temperature of the transistor and such. Be prepared for a high resistance, I guess 50K-100K or more. Note that, unlike your drawing, you will leave one end of the resistor disconnected, and the wiper will attach to the base. Once you find the value that works, you could just replace it with a fixed resistor.
 

Thread Starter

trader007

Joined Feb 27, 2010
249
oops, I see what you mean. Fixed, I think.

I did a quick test with the resistor controlling the base, and it works fine. No diode installed yet, so I'm not sure how it affects the overall circuit but as far as controlling the brightness it works fairly well. I do see, though, why people say to use a PWM to control LED brightness, its much smoother when you do it that way.

I believe I need to keep a pot in there, because the more LED's I add the more the circuit changes. And I am sure everyone is going to want something different as far as the "dim" mode brightness is concerned.
 

wayneh

Joined Sep 9, 2010
17,495
You might consider using a fixed resistor AND a pot as sort of a coarse and fine adjustment. Otherwise I imagine setting the pot to just the right level might be too touchy, with too small a difference between too dim and too bright.

So for instance if 150Ω is just right, maybe use a 100Ω fixed and a 100Ω variable.

Oh! And if this is something a user can futz with, you have to allow for them setting the dial to 0Ω. Murphy's law. It WILL happen. This would risk frying the potentiometer, because current would no longer be limited. Another good reason to add the fixed resistor, to cap the worst-case current.
 

Thread Starter

trader007

Joined Feb 27, 2010
249
Yeah thats why I thought I would need one, otherwise the cap would slowly charge and become useless in action.

I already tried without the diode because I didn't take the time to figure out what one I could use with the diodes assortment I have... and just as I thought it doesn't fade. But the pot does adjust the brightness of the whole string, so that is promising.

Also, I noticed today that at full load, the circuit is only pushing 9v at the LED rope. I'm not sure why the voltage drop is so severe, but I think it has something to do with the transistor I'm using. With the BD679, I remember the voltage being 11v or so.

Tomorrow I will try bridging the collector and emitter to see if its a circuit problem or a power supply issue. It is a 150w power supply but I forgot that is total, and being a computer supply it has a few watts for 5v and 3v lines... It could be that it only has 100w or less available on the 12v rail and there's simply too many diodes drawing it down. Next week I will have the MeanWell supply from Jameco, which will also allow me to adjust the output voltage.
 

Thread Starter

trader007

Joined Feb 27, 2010
249
I did it! It's fully working!

Operation as follows- Turns on at night in "dim mode" automatically via day/night sensor. Then, when a person walks within 20ft, the ON FADE is activated, fading the lights to full brightness. After a preset time, the lights automatically fade back to DIM.

There is a slight problem with this as setup though. After the preset time limit has passed, the IR Sensor turns off, even if there is activity. It has to turn off before activity can turn it back on.

So, for now, this will suffice, but I will need to find a different IR sensor or find a way to modify these so the timer "resets" as soon as there is another activity reading. I am afraid I might have to add my own timer circuit.... one that stays latched as long as there is activity, even if the timer runs out.
 

LDC3

Joined Apr 27, 2013
924
I did it! It's fully working!

Operation as follows- Turns on at night in "dim mode" automatically via day/night sensor. Then, when a person walks within 20ft, the ON FADE is activated, fading the lights to full brightness. After a preset time, the lights automatically fade back to DIM.
Kudos, well done.
 

Thread Starter

trader007

Joined Feb 27, 2010
249
My first shot at pcb desgin! :D Tell me if its crap or won't work!

https://dl.dropboxusercontent.com/u/587032/LEDPros/Deck Lighting/Fade PCB Bottom Rev3.zip


I will go over it tomorrow to make sure it will work electrically speaking, and also get a proper schematic for everyone to use. Spent over 2 hours just designing the pcb! I have never done it before though. I really appreciate the help here, and this is a way for me to give something back. It will be complete with specific readily-available part numbers so anyone can duplicate what I am doing. There is a member here making pcb's, so I hope this will work out for him.

edit- OK over the last 2 nights I redrew this about 40 times LOL :D I think this is the way it's going to stay!


 
Last edited:
Top