DIY LEDs - The write-up
- Thread starter Soundwave
- Start date
- Status
Constant current on the output, correct. The buckpacks are 85% efficient though, so to put out a constant 1.0amps of 22.2 volts on the output side, requires 1.18 amps on the input side of the buckpacks. The power supply for the buckpacks is providing 24V, so it is 1.18amps at 24 volts, which is the numbers I calculated above.
Am I missing something still?
tiffrobbins
New member
Let me run this by you and just make sure I'm getting this right in my head. (I'm a Newb with LEDs).
If I wanted to make this type of a fixture, but only half the size - I would run 2 lines of 6 white LEDS to their own 1000mA buckpucks and then both to a 1 24V power supply and I would run 2 lines of 6 blue LEDs to their own 700mA buckpucks and then both to a 24V power supply. Since I'd be running only half the lights and half the buckpucks, would it be wise to use 12V power supply or should I stick with the 24V? (I'm assuming there's a good reason to split the colors up to different power sources)
Also, lets say I want to get crazy and throw in some red for grow lights. Would I then need a third power supply or could it be tied into one of the existing supplies? or should I just shut up and keep it simple?
Thanks!
If I wanted to make this type of a fixture, but only half the size - I would run 2 lines of 6 white LEDS to their own 1000mA buckpucks and then both to a 1 24V power supply and I would run 2 lines of 6 blue LEDs to their own 700mA buckpucks and then both to a 24V power supply. Since I'd be running only half the lights and half the buckpucks, would it be wise to use 12V power supply or should I stick with the 24V? (I'm assuming there's a good reason to split the colors up to different power sources)
Also, lets say I want to get crazy and throw in some red for grow lights. Would I then need a third power supply or could it be tied into one of the existing supplies? or should I just shut up and keep it simple?
Thanks!
Soundwave
You might want to measure the current of the LEDs, I bet your not driving the LEDs at the current you think. I used Buck Pucks originally in my design but found they would not drive a string of 6 LEDs at the stated current. If you read the fine print on the data sheet the input margin is 2 V for the 350 mA Buck Puck with a side note that this increases with higher current units. So if you have a string of LEDs wih a foward voltage of 3.7 V total voltage drop is 22.2 V add > 2 V margin and you are over the rated 24 V supply and not droping the 3.7 foward voltage on the LEDs. If you look at the specifications for output current there is a footnote that states "Measured with single emitter; output drops slightly with additional series junctions to limit maximum power dissipation" . I suggest you put a .1 Ohm resistor in series with the LEDS and measure the voltage across it. It should read 100 mV if you are driving it at 1 A.
The kicker is the power dissipation..... at such high current any voltage difference between what you are dropping across the LEDs and Power supply must be kept low. Remember P = IV so if you have a 2 Volt difference at 1 A you are dropping 2 Watts of power on whatever IC is in that Buck Puck.
You might want to measure the current of the LEDs, I bet your not driving the LEDs at the current you think. I used Buck Pucks originally in my design but found they would not drive a string of 6 LEDs at the stated current. If you read the fine print on the data sheet the input margin is 2 V for the 350 mA Buck Puck with a side note that this increases with higher current units. So if you have a string of LEDs wih a foward voltage of 3.7 V total voltage drop is 22.2 V add > 2 V margin and you are over the rated 24 V supply and not droping the 3.7 foward voltage on the LEDs. If you look at the specifications for output current there is a footnote that states "Measured with single emitter; output drops slightly with additional series junctions to limit maximum power dissipation" . I suggest you put a .1 Ohm resistor in series with the LEDS and measure the voltage across it. It should read 100 mV if you are driving it at 1 A.
The kicker is the power dissipation..... at such high current any voltage difference between what you are dropping across the LEDs and Power supply must be kept low. Remember P = IV so if you have a 2 Volt difference at 1 A you are dropping 2 Watts of power on whatever IC is in that Buck Puck.
Keith Elliott
New member
Lynxvs: Can you tell use what you are using instead of the Buck Pucks please?
kcress
New member
Soundwave; You can use an ammeter and just multiply by the voltage.
However you want to take that meter,(changing its leads to voltage so you don't blow its innards out), and measure the voltage across the string of LEDs. This will give you the actual power across the string. Just the 24V only tells you the power you are using in the whole system including the BuckPucks, not the actual LED power.
Doahh; Yes you can use an analog voltage to control a MOSFET but there is sooo much more to it than just that. If you just try that you will fail at it. It is also a horrible method as any power not across the LEDs will just be wasted across the MOSFET causing it to heat up dramatically and probably fail.(soon)
Also your "how many LEDs?" question can not be answered as it depends on may other aspects of the configuration.
However you want to take that meter,(changing its leads to voltage so you don't blow its innards out), and measure the voltage across the string of LEDs. This will give you the actual power across the string. Just the 24V only tells you the power you are using in the whole system including the BuckPucks, not the actual LED power.
Doahh; Yes you can use an analog voltage to control a MOSFET but there is sooo much more to it than just that. If you just try that you will fail at it. It is also a horrible method as any power not across the LEDs will just be wasted across the MOSFET causing it to heat up dramatically and probably fail.(soon)
Also your "how many LEDs?" question can not be answered as it depends on may other aspects of the configuration.
EBOLII
New member
Sorry for the linked image but I have exhausted my search for information
The site is calling the center circuit board a FET but I have been unable to find any information. Is anyone able to explain or point me to right location
Just trying to get a grasp on the DIY process....I am planning for a large build [72"Length x 24"Width x 30"Deep] so I still have design flaws/questions
The site is calling the center circuit board a FET but I have been unable to find any information. Is anyone able to explain or point me to right location
Just trying to get a grasp on the DIY process....I am planning for a large build [72"Length x 24"Width x 30"Deep] so I still have design flaws/questions
ReefEnabler
Premium Member
if its an FET its probably what they're using to dim the LEDs since they allow you to control the amount of current that passes through.
I've got no experience with one like that though.
I've got no experience with one like that though.
stugray
Premium Member
EBOLII,
The FET Can control the current through a LED in an analog fashion ( known as the linear region of the device ), but it is very inefficient & wastes a lot of heat.
The better way to do it is using "pulse width modulation" or PWM.
That is where you turn the FET fully ON & fully OFF very fast.
When a FET is fully ON it wastes very little power.
By varying the width of the ON vs OFF times ( pulse width ), you can vary the apparent intensity of the light.
As long as the frequency of the pulses is high enough, the human eye cannot see the blinking effect from the LED.
I suspect the circuit boards in the pic above are the PWM circuit boards. The FETs can be seen on each corner of the boards.
This method is what I use and I have a working prototype that can be programmed via Enet.
Stu
The FET Can control the current through a LED in an analog fashion ( known as the linear region of the device ), but it is very inefficient & wastes a lot of heat.
The better way to do it is using "pulse width modulation" or PWM.
That is where you turn the FET fully ON & fully OFF very fast.
When a FET is fully ON it wastes very little power.
By varying the width of the ON vs OFF times ( pulse width ), you can vary the apparent intensity of the light.
As long as the frequency of the pulses is high enough, the human eye cannot see the blinking effect from the LED.
I suspect the circuit boards in the pic above are the PWM circuit boards. The FETs can be seen on each corner of the boards.
This method is what I use and I have a working prototype that can be programmed via Enet.
Stu
I made it more complicated then it probably needs to be but.... I used a constant current source LED driver. It's basically a very small surface mount IC that requires few external componants to operate. It provides the LED a constant current source and has a PWM dimming pin as well as a shutdown pin which allows it to be turned off and on. This by all means is not for the average DIY it requires the fabraction of a PCB as well as some good soldering skills.
If you look around you can find lots of LED drivers of varying types (Buch, Boost, etc.) If you are looking for a dimming function make sure it has a pin for that.
I would go with a IC that is designed to be a LED driver. Trust me even with this you are going to find problems when you hook everything up. Unless you really know what you are doing I would never attempt to build your own driver.
I know this looks like the simplist thing in the world to do but there are a lot of details that ,unless you know about them , are going to cause you lots of problems. I have been working on a design for almost a year and still do not feel comfortable with it. Everytime I think i have a handle on it something comes up that I didn't expect. Now I know the reason Solaris had so many problems with the design......
This is an example on what I mean when I say the devils in the details..... I designed my first prototype Light pretty much like the one in this thread only I was using Luxeon III Stars. I bought the LEDs, Buckpucks etc. and put it all together. I measured the current and found that the Buckpuck was not driving the LEDs at the rated current.
Last night I broke out the Buck pucks that I had to remeasure the output only this time I had no problem driving the LEDs at the rated current.....
I didn't understand the results for a while but it turns out the LEDs I was using last night are not the same that I used on the first test. The forward voltage of the Luxeon stars are 3.9 V and the Luxeon Rebels that I am currently using have a forward voltage of 3.4 V. Doesn't seem much does it but it's enough for it not to work......
- Status
Similar threads
