DIY project problem, not enough current

[knord]

Hi,

 

 

I'm new to the forum, but I have taken the plunge into making my own light. I am hoping that there is someone here that can help me on a problem I have with finalizing my LED light.

 

Here is the components I use:

 

Nunus S-400-24 a 400W, 24V switching power supply.

TC420 a 5ch, 20A overall max, auto 12-24V Programmable LED Time Controller.

1.5W single chip (Advertised as 3W) LED's from Epistar 3-3.4V, 5-600mA

1.5W single chip (Advertised as 3W) LED's from Epileds 2-2.4V, 5-600mA

 

When I run the diode tester on my Fluke 18B+ the readings are only 1.6v on the Epileds and 2.5 on the Epistar LED's.

Taking this into consideration I have connected 13 pcs. of Epileds and 1 Epistar LED's together in series laid out as a row, stacking them battery style -+-+. The voltage turned out to be ~23.2V.

 

I first tried to connect the 14 LED's in series to the TC420 dimmer which again is connected to my power supply. Nothing happened (Maybe I need to upload a file threw the software first).

Then I connected the LED's directly to the 24v PSU and it lit up faintly.

I then went on to try to measure the watt usage threw a watt meter and it showed that the power draw was negative 0.1W. I'm guessing this has to do with the inaccuracy of my watt meter.

I want a power draw as close to 21W as possible, it will probably be less than this but now there is almost nothing.

 

I know that I could group series and parallel differently, like you would do with batteries for solar/wind, but it would require a ton of soldering since I have 5 rows of 37 LED's on my aluminium heat sink.

My plan is to split each row into pairs of 17 LED's and then connect these in parallel to the dimmer channels. I'm letting the connectors "hold hands" by stacking them on top of each other so + and - make contact.

 

How can I push more current threw to my LED's or am I asking the wrong question? Any help is received with thanks

 

 

 

Cheers from Scotland

[evilc66]

You are going about this a little wrong.

 

First, the numbers you are getting from your meter with the diode test mean nothing. That's just the breakdown voltage that the LED needs to start passing current across the die. That has nothing to do with the operating voltage.

 

Second, the controller you are using is a constant voltage PWM driver. You can make it work, but it's not ideal. To make this work, you will first have to measure the output voltage of one of the channels to see what you have to work with. Most likely it will be slightly under the voltage from the power supply due to losses from the FETs internal to the controller.

 

Once you have that voltage, then you can figure out how many LEDs you can connect to each channel in series. Lets start with the first LED, and use 3.4v as the average LED voltage. Under ideal circumstances (being the channel voltage is 24v), you would be able to run 7 of these LEDs per channel. You most likely won't be able to, so we will say 6 LEDs is practical.

 

Now, if you were to connect 6 LEDs and turn the controller on, you would have some very dead LEDs very quickly. The reason is that the 6 LEDs would be splitting the channel voltage equally. Lets assume that the output is actually 22v (so about a 2v loss to the controller). 22v / 6 LEDs = 3.666v per LED. That voltage would put the current at the LED close, if not over 1000mA, which will kill those LEDs quickly. To fix this, you need to add a resistor that eats up the difference between the LEDs and the controller. At 3.4v, the LED string would come to 20.4v, leaving a difference of 1.6v. We are going to assume that at 3.4v the LED is drawing the 600mA that you stated (not having a datasheet makes this all guess work). A little Ohms Law application and we need a 2.67ohm resistor (1.6v / 0.6A = 2.66666666ohms). This resistor will have to be rated for 1W or higher (1.6v * 0.6A = 0.96W). This of course is assuming that everything that has been stated is true. You may want to err on the side of caution and increase the resistor value to 3ohms so that the current is lower (~533mA), and under the ratings that were given.

 

Apply the same thinking to the other LED.

 

All of this hard work is just another reason that a constant current LED driver should have been considered. All you would have had to do is make sure that the LED string voltage fell within the output range of the driver and hook it up. The driver keeps the LED current solid (this is important), and varies the voltage as required without having to figure out resistor values and channel voltages.