Hi All,

I'm new to electronics, so currently I'm closer to a house plant than I am an electronics genius, but I'm a programmer in about 10 languages so I do learn fast.

I'm currently working on a project (not for work just personal) and I've got a question about using potentiometers in parallel to control lighting levels.

Basically, I am creating a desktop green house for my office, with artificial lights. I will be using a number of LEDs: blue, red, UV and IR. I will be using 4 of each and want to be able to control the brightness of each of these four types of LED. If there's too much blue light and the plants are growing too quickly, I want to be able to dim that down..

So I thought maybe four potentiometers in parallel, each of which will give some power to 4 separate circuits - one for each type of LED.

But the problem is, I think, that electricity prefers to take the path of least resistance, and so when other circuits are on full and I put the resistance up a bit on another, it just switches off, rather than dimming.

It could be that the pots are too high resistance, or it could be a thousand other things. At the moment I am just using simulation software whilst I wait for the rest of my components to arrive - so it could just be a bug in that.

Here is the circuit diagram:
http://i55.tinypic.com/9jkchz.jpg
(ignore the resistors, I set their value to get the current right on the LEDs)

Will my idea work? Is this the right way of doing it or are there better ways?

I appreciate any reply!

Regards,
Adam

 

Hello,

It is never wise to connect the leds directly parallel.
Each led should have its own current limiting resistor.
You can also put a couple of leds in series with a curretnt limiting resistor.

Take a look at this thread by Bill_Marsden on connecting leds and more:
LEDs, 555s, Flashers, and Light Chasers

Also the PWM circuits by him might interest you:
http://forum.allaboutcircuits.com/showthread.php?t=24028555 PWM Oscillator

In stead of the shown motor you can put the led strings.

Bertus

 

I think Bill Marsden's method may work best..

http://forum.allaboutcircuits.com/c...uits.com/picture.php?albumid=45&pictureid=321

But if I was to also put a variable resistor on the beginning each of the chains, would increasing the resistence on one of them mean that the current would instead just go through the other chains and ignore the slightly resistant one?

Thanks

 

That is not the way parallel circuits work - http://www.allaboutcircuits.com/vol_1/chpt_5/3.html

You link to the circuit did not work. You can post it on AAC by using the "Go Advanced" button, and the "Manage Attachments".

 

But if I was to also put a variable resistor on the beginning each of the chains, would increasing the resistence on one of them mean that the current would instead just go through the other chains and ignore the slightly resistant one?

Not usually. The only way one leg would affect the others is if the power supply is the limiting factor. Reducing the current in one leg also reduces the total current. If that allows the power supply (eg. battery) voltage to rise slightly, then a bit more current will flow through the remaining legs. BUT, if your power supply maintains a steady voltage, the various legs don't "see" each other; they just see a steady voltage. Picture 4 small pipes draining the bottom of a big water tank. Closing one just doesn't affect the others unless the tank is nearly empty.

 

Not usually. The only way one leg would affect the others is if the power supply is the limiting factor. Reducing the current in one leg also reduces the total current. If that allows the power supply (eg. battery) voltage to rise slightly, then a bit more current will flow through the remaining legs. BUT, if your power supply maintains a steady voltage, the various legs don't "see" each other; they just see a steady voltage. Picture 4 small pipes draining the bottom of a big water tank. Closing one just doesn't affect the others unless the tank is nearly empty.

Perfect way of explaining it. Thanks very much! Makes much more sense now.

That is not the way parallel circuits work - http://www.allaboutcircuits.com/vol_1/chpt_5/3.html

You link to the circuit did not work. You can post it on AAC by using the "Go Advanced" button, and the "Manage Attachments".

Thanks for the link - a big help, been working my way through the documents, they're great!

Thanks guys!

 

If I understand you correctly this may be what you want...

Controllable power supply circuit for high power LEDs...any ideas?

This is not one of my better efforts, it will get hot, very much so, because it is analog. A PWM system would generate much less heat.

 

Thanks for that Bill, but right now I think that's a bit beyond my electronics level. I have however saved it for the future and will hopefully be able to understand it a little better in a few months time.

I understand that LED's can get hot if they are constantly on, can they get so hot that they become unstable (i.e. burnout or blow) or are they able to run quite stable for their lifetime? (I'm going to be using ultrabright LEDs, but only 20ma as it's a small project)

Thanks again!

 

Just reading up on PWM now, am I right in understanding this: you control the LEDs in bursts, so say at 100 hz, the LED would light up 100 times a second, which would allow it to cool down during the off periods?

How would you increase power, more blinks? (would 1000 mini blinks/second be equal to 1 half-second blink?)

Or would you control the period it's on for, e.g:
On for 50ms - Off for 10ms - On for 50ms - Off for 10ms?

Regards
Adam

 

PWM varies the percentage of a pulse, when it is on.

About the image on post7...

It is not the LEDs that will get hot, but the transistors regulating the current to them.

Talking complexity, PWM control is a bit more complicated than that circuit on post 7. It is about as simple as it gets, if you are wanting to control light levels that is. The simplicity is why I drew it. I think you are looking at the LED count, and not the actual circuit.

Here is a single channel PWM intensity control, you be the judge...

The two would need to be merged. If you don't care about hitting maximum brightness or dimness you could use this PWM circuit to control the bank of LEDs and MOSFET...

The motor would be replace by the LEDs/transistor.

BTW, this would also work...

I would be happy to refine a schematic to whatever you are wanting.

There are about 3 threads that cover the subject of grow lights. I was active in the ones I know about.

How many LEDs are you wanting? Remember, individual LEDs are not that bright, that is why there are so many in the other images.

 

Thanks for the diagrams and explanations, I've dabbled in electronics before but never had the time to properly focus in it until now, so a few of these things are new to me - one question that occurs to me is the use of transistors in the Medium Current Constant Current Regulator: What is the difference between using a resistor and three LED's in series and using three LED's then a transistor and a resistor?

Also, in the last diagram, CR1 CR2 & CR3, what purpose do they serve?

About my grow light project:

As a first project, I don't want to do something too advanced, just a simple perspex box with lights in it - with the ability to control intensity of certain lights.

The size of the box is 150mm cubed, so only small.

My plan is to use 4 of each type of LED, placed evenly - so 4 clumps of leds in 4 points, with one of each LED there.

The LED's I will be using are: Red, Blue, UV (395nm) and IR (940nm).
The UV are 3000mcd (3.2v - 3.8v, 20mA)
The Red are 5000-6000mcd (1.8v - 2.0v, 20mA)
The Blue are 5000-8000mcd (3.0v - 3.6v, 20mA)
The IR are 1.5v - 1.6v, 60mA - not sure about the brightness of this LED.

The reason for using UV & IR is that I want to see if they have an effect on growth, as I know blue & green do.

I did think about using a bar graph driver to control the number of LEDs rather than the brightness of the LEDs, but I would like (if it is of course possible) to control the brightness of individual LED's, for the first project at least. That way I can learn about ways of controlling them.

Coming back to your final diagram, that does look like the idea I have in mind. I'm at work now without any simulation software to make a diagram but my idea was/is to have LEDs in series with a resistor first, and then have those in parallel, and put a potentiometer/variable resistor before the initial resistor for each line of LEDs.

Am I heading in the right direction? I would definitely be up for using transistors though, I do have some general use ones in my selection of goodies, and once I get home I will read up on them and their uses.

Thanks again for all your help, it is appreciated!

 

The reason you use a transistor is to allow a small current (thru the controlling resistors and pot) to control a much larger current - the LEDs. High-current pots are expensive and hard to find. Low-current pots and transistors are dead cheap and available everywhere.

Little practical facts that don't show up "on paper", only when you go to actually make something.

 

Ah that makes sense. Thanks!

 

So I've put together a little prototype tonight, just going to post some images before bed!

I decided to use 4 LEDs, one of each type, and use an old aquarium food pot to see a few things: 1) how hot the components would become, 2) how hot the enclosure (pot) would become, 3) to practise my soldering skills.

I put some holes through the lid, and pushed through the 4 LEDs, I used some 5mm Bezel clips to hold them in place. I then put some holes in for the resistors - so they would be on the inside for extra heat.

Rather than using wire, I just soldered the components directly. My soldering does look like it's been done by an OAP hahaha, but I will use my beginners get out of jail card for that one..

The final product is a bit bright! - but time will tell if it is too bright or not. I've put some cress seeds in to test, they like warmth and are easy and quick growers.

And to keep an eye on the temperature, a cheap digital thermometer on a lead..

I will keep you posted on the results, whilst I plan my next, more complex version..

Thanks all!

 

You are on the right track, figure out what the geometry of the LED source will be, and the mix of colors.

That is what this OP did...

http://forum.allaboutcircuits.com/showthread.php?t=36238

 

Interesting thread, I will read up on it in a couple of minutes. Thanks!