Maker Pro
Maker Pro

Using ohms law to calculate how many devices can be used with power supply

Rob From Pluto

Feb 9, 2015
14
Joined
Feb 9, 2015
Messages
14
Hi,

Me and some friends are making some LEDs for fun, we came up with a list of To Do's and split them amongst us. My To Do happens to be my weakest point- working out the power part. I think this is probably easy for people who know ohms law, i've been reading a ton but this part is confusing to me-

We are using high power LEDs- 40W RGBW (led engin brand). They are actually four LEDs in one package so we can mix them to make any color (using an LED driver). Red is powered with 8v, the others are powered with 12v.

I have it working, very bright and everything works as it should. I'm powering it with a computer ATX power supply using a constant current controller.

What i'm trying to work out is how many of these LEDs i can power using this power supply. The power supply is rated at 700 watts. My first thought was to simply divide 700 by 40watts (since the LEDs are rated at 40 watts) to get 17.5 (call it 17), meaning theoretically i should be able run 17 of these rgbw LEDs (and figure in 80% for safety).

Is this right? Someone told me that i should be doing it differently- by using the amps drawn- i.e. each channel (there are 4 channels, each color is called a channel) is held at a constant 600ma, so 600 * 4 channels (reg, green, blue and white) = 2.4 amps. The power supply's specs written on the side say it can supply 52 amps at +12v, and 52A/2.4A = 21, meaning i should be able to run 21 of these RGBW LEDs.

Which do you think is the right way to do it?

Thanks:)

Rob
 

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
If you are using a constant current of 600mA for each of the 4 LEDs you are probably not using the entire 40W capability of the LED.

What is the forward voltage of the 4 LEDs at 600mA? That will tell you how much power you are actually drawing.

It will also tell you how many you can put in series when driven by 12V max.

Bob
 

Rob From Pluto

Feb 9, 2015
14
Joined
Feb 9, 2015
Messages
14
Ok, that makes sense to me now, i was wondering what am i missing. But, I'm sending 12v to the controller, it's a constant current controller so i think it varies the voltage going to each led to make sure the current for each stays constant (600ma); how would i find a Vf for it when it changes?

I was hoping to see if i was figuring it right so i at least have a ballpark, but of the two ways i figured it out, i think maybe they are both close? You're right i think- since the controller is keeping current at 600ma, it's not max- each led (of the 4) can run at i believe 1.2a, so the numbers i was getting would be different.

Anyway, were the ways i figured it right, or it won't work without knowing the Vf?

Thanks again
 
Last edited:

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
You could measure the Vf by simply putting a voltmeter across the LED when it is operating. It is normally around 2.5V for red and 3.6V for green and blue. But I suspect your 40W LEDs are actually 3 LEDs in series, so they would be more like 7.5V and 10.8 for the others.

Then the current draw will actually depend on how the controllers work. If they are linear regulators, they would draw the same current as they are providing to the LED. If they are buck converters, it is more complex, they will draw a little more current, but for less than 100% of the time dependent on the Vf of the LEDs.

Bob
 

Rob From Pluto

Feb 9, 2015
14
Joined
Feb 9, 2015
Messages
14
I forgot that the Vf is listed in the specs- it's 8v for red and 12v for the other three.

So 12v * 3 (12v on each of three channels) =36v plus 8v (for red channel) = 44v

44v times .600 (total Vf for all four LEDs, times current) gives power = 26.4 watts. I think this makes sense for ohms law?

If that's right then i think i would still use the power supply's rated 700w and divide by 26.4 (and figure in 80% for safety)? Or would it be better to find it using current (600mA times four to get 2.4A and divide into max amps rated for the supply)?

I hope that's right, i haven't used ohms law in quite awhile, i was surprised i was able to remember some of that without cheating (google).

Thanks
 

Rob From Pluto

Feb 9, 2015
14
Joined
Feb 9, 2015
Messages
14
I get 26.514 for the first way, and 21.6 for the latter, that's not taking into account 80% for safety. I don't know why each method gives different results, unless i didn't do something right, but after checking everything, seems right.
 

ramussons

Jun 10, 2014
462
Joined
Jun 10, 2014
Messages
462
Since you are going to control Each LED package independently for colour mixing, they are all going to be with their current controllers Across the 12 V supply - Not in Series.

" ....Someone told me that i should be doing it differently- by using the amps drawn- i.e. each channel (there are 4 channels, each color is called a channel) is held at a constant 600ma, so 600 * 4 channels (reg, green, blue and white) = 2.4 amps. The power supply's specs written on the side say it can supply 52 amps at +12v, and 52A/2.4A = 21, meaning i should be able to run 21 of these RGBW LEDs..."

I would agree with this in principle.

A better way of doing this for many more LED's would be to use Switched or Pulsed Current Feed. This will give the same intensity, but the average current drawn will be less. So the power supply will be able to feed more LED's.
 

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
You cannot run a constant current supply off 12V when the forward voltage is 12V. The supply will need a volt or two more than 12V to maintain the correct current. Can you link to the constant current driver specs? It should specify a maximum output voltage when run by 12V in.

And do you realize that you will need a constant current controller for each LED. I.e. 4 per device?

Bob
 

Rob From Pluto

Feb 9, 2015
14
Joined
Feb 9, 2015
Messages
14
Thanks for saying, and good catch- i was quoting the specs i found on the mouser page for the LED engn LED and didn't think about the full specs- the voltages that i said before are right but they are "max* voltages (8V for red and 12v for the others), the *min* Vf is 6V for red and 12V for the others.

So for G, B and W, i don't see a problem. I'm not sure why red works with no problems when the Vf range is 6V-8V and it's getting 12V. Obviously the controller is dissipating it somehow but it's a very inexpensive chinese driver that has no data on its in/out configuration, the only specs i have for it say that it works with 12-24vdc input. Maybe the red will burn out after awhile.

I know constant current controllers will vary voltage to maintain the current, maybe that's what's happening, would be good to know. Regardless, it's an exercise to learn, and that i did for sure, anything else that i can pick up from this project will help for the next.

And the other major thing i'd like to confirm is if i have the theory right for verifying how many of these LEDs i would be able to run on the pc power supply- i think a couple of people implied i have it right, but can anyone confirm?

Making this LED light has been a great project- i was surprised that LEDs could be so bright.
 

Rob From Pluto

Feb 9, 2015
14
Joined
Feb 9, 2015
Messages
14
I missed it again, sorry- the min Vf for red is 6v and the min Vf for the other colors is *9.3* ..Long day.. Anyway that explains why they work with 12v via constant current since the Vf has room to decrease, not sure i understand all of the technicalities, i hope that makes sense. The thing i'm still unsure of is why red works at all- since it's Vf range is 6v-8v and i'm giving it 12V (as far as i know!), shouldn't red burn out? Maybe it's just ok since the controller is feeding the LED 600mA when the LED is rated to 1A (per channel).

But the main point i'm still hoping to confirm- if i can calculate how many of these LED/driver combos i can run on this 12v ATX power supply using one of these solutions(neither includes 80% for safety):

**1**
Max Vf on each channel, 8V on red, 12V on the other three:

So 12v * 3 (12v on each of three channels) =36v plus 8v (for red channel) = 44v

44v times .600 (total Vf for all four LEDs, times current) gives power = 26.4 watts. I think this makes sense for ohms law?

Since the power supply is rated at 700w, divide it by 26.4 watts figured above to get 26.5, rounded down to 26, so the power supply could run 26 of these.

Or
**2**

Would it be better to find it using current- 600mA times four channels to get 2.4A and divide into max amps rated for the supply, which is 52A in this case, which comes to 21.6, rounded down, the power supply could run 21 of these.

Two different methods, two answers, so one method must be incorrect?

Thanks
 
Last edited:

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
A constant current supply provides exactly the voltage required to make the required current flow. If you measure the voltage across your red LEDs you will find that it is somewhere between 6 and 8V even though you are putting 12V into the driver.

Now, as I stated before, the max number of LEDs you could run depends on how the constant current driver works. Most likely (given the input voltage rating) they are switching regulators, which means that, theoretically, you could be pulling less current from the supply than is going through the LEDs. But with the GB & W LEDs needing up to 12V, this effect would be small.

But you can be safe be using the full current drawn by the LED, and the current capability at 12V of the power supply. As you calculated, it will run 21 of them no matter what.

You cannot use the power rating of the supply because it cannot supply all that power at 12V and because there will be wasted power that you are not taking into account.

Bob
 

Rob From Pluto

Feb 9, 2015
14
Joined
Feb 9, 2015
Messages
14
Thanks, everything you said makes sense. I do remember picking up somewhere that the ATX power supplies can't use all power on just one rail, i didn't think about it until reading what you wrote, that's very clear thanks for writing it. I wanted to be sure that i was calculating the right way for worst case scenario (disregarding 80% for safety), and it makes sense. I also read that different atx supplies distribute their power differently across their rails (i.e. from reading, i might or might not be able to power all 21 of these from connecting all of them to one set of yellow/black (12v) connectors cut off of a molex, probably i need to divide the 21 evenly between however many 12v "rails" are supplied. Also from reading, most only have one 12v rail, but Y's it to provide 3-4 separate lines, each of which are in turn Y'd to connect several molex/SATA power/etc.
 
Top