Hi there. I am trying to power 600 WS2812B RGB leds with one 10A 5V power supply. The leds are placed evenly in a rectangular shape. I am only planning on having these very dim, but I can always pick up one or two more 10A 5V power supply to distribute power, since the math says it will use more power than the one 10a 5v power supply I currently have.
I am planning on distributing power every 50 leds, as I currently have 300 leds setup with power coming from one end, and I can see the colors getting orange and having some weird behavior already, if I use colors that aren't red, blue, or green as it draws more power. I am planning on adding the other 300 leds here soon.
I drew up a crappy diagram on paper on what it will look like, and the other diagram is what I think I should be doing for the end product. From how I have interpreted the neopixel uberguide about powering leds, that is how I thought I should do it, but figured I could ask here to confirm before I ruin the psu or leds.
My questions:
If I plan to distribute power every 50 leds from a central PSU, do all the grounds need to be connected together?
Should I plan on picking up other power supplys and distribute power differently / don't connect the grounds instead of the way I have planned out?
If I go the route that I drew up in my diagram with the central PSU powering every 50 leds... Will my power from the PSU be able to traverse ~5-15ft without losing voltage using small gauged wire (22-26), or will I get some kind of weird undefined behavior from this?
The PSU I have has the small terminal +/- ends, so it would be a tight fit for all of the wires as well.
P.S. I am a total noob so please inform and clear up any misconceptions I may have. Thank you!!
yes, grounds must always be connected. But also take ground lines directly to the central PSU. Otherwise you will still get voltage drop and incorrect colours. Remember the current has to flow in a circuit, and the circuit is only as good as it's weakest link, which would be the ground lines, looking at your schematic.
connect the grounds but use each PSU to power separate sections of LEDs. Don't connect the 5V lines from different PSUs.
thin wires will cause voltage drop and incorrect colours. Use thicker wires. Connect 5V to both ends of each strip to reduce voltage drop further.
Awesome, this clears everything I needed to know about, so thank you for that! One last question... do you have any recommendations on the size of the wire I should use?
Google for "AWG max current" to find a table showing max currents for different wire guages and choose something perhaps half as high again as you need. Some parts of your wiring will need less than others, so thin wires may be good enough for those. You could also run pairs of thin wires to double their capacity.
Can you expand upon basic circuit theory when applying resistors and capacitors? (...maybe specific to the layout I have explained please ) I was going to follow best practices based on other guides, but I think I need to invest some time to read more into it to figure out what I really need. I want to get a nice, reputable, and safe lab power supply to power these 600 leds so I have a central PSU and not use a 10a 5v PSU or multiple PSUs.
Resistors: maybe only one needed, ~500R, near to data in between Arduino and first strip. If data wires between strips are longer than a few hundred mm, you may need similar resistors on those lines too. Always run a ground line next to the data line to protect the data from getting corrupted.
Capacitors: a 470uF for each section of 50 LEDs should be ok, near the -/+ inputs or outputs to the strip.
Power supply: not sure what you mean by "lab". A bench power supply capable of this would be hugely expensive and a waste. You do not need to control voltage or current, which is what bench PSU do. You just need fixed voltage and sufficient current.
league_addict:
Awesome, this clears everything I needed to know about, so thank you for that! One last question... do you have any recommendations on the size of the wire I should use?
I am a relative newbie to using the WS2812 LEDs, so hold your snipes. Constructive criticism is appreciated however.
I currently have 300 LEDs (12 Volt WS2812 strings) powered by a 15 Amp supply based on the calculated rating of 60mA for each LED at full-power white. What I got when I measured the actual current draw was half that. (I suspect that this is from using 12V LEDs as opposed to 5V LEDs.)
Every different LED technology is different, so measure your current draw on a sample of the strings or strips you plan to use. Too much capacity (Amps) is never a problem.
I am using wire designed for LED installations, which is a repurposed speaker wire. 18 gauge is sufficient for my project, and at the end of a 300 LED series of strings, I am only seeing a one-volt drop. (12.5V at the PSU, and 11.5V at each end of the LEDs, measured in the middle of the strings).
SteveMann:
I currently have 300 LEDs (12 Volt WS2812 strings) powered by a 15 Amp supply based on the calculated rating of 60mA for each LED at full-power white. What I got when I measured the actual current draw was half that. (I suspect that this is from using 12V LEDs as opposed to 5V LEDs.)
Well, for 12 V LEDs, three per individual string, it is clearly 60 mA per three LEDs, isn't it?
SteveMann:
18 gauge is sufficient for my project, and at the end of a 300 LED series of strings, I am only seeing a one-volt drop. 12.5V at the PSU, and 11.5V at the end of the LEDs.
Which is presumably to say that you are powering the LED strip itself at least at both ends?
Paul__B:
Well, for 12 V LEDs, three per individual string, it is clearly 60 mA per three LEDs, isn't it?
Which is presumably to say that you are powering the LED strip itself at least at both ends?
I measured again using a single RGB LED to double-check my measurements. With the brightness set to 255, I measured:
Red=18.28mA
Blue=17.48mA
Green=10.94mA
White=18.21mA
I have no explanation why the white current draw is the same as the individual colors. Could the fastLED library be doing some kind of load leveling??
Post your code, fast led does have methods to limit current draw in the software, depending on options you set in setup.
Also, you may be getting voltage drop along the strip or from the powers supply when you go full white, you would need to measure the voltage at the strip while it is full whits to see.
Qdeathstar:
Post your code, fast led does have methods to limit current draw in the software, depending on options you set in setup.
Also, you may be getting voltage drop along the strip or from the powers supply when you go full white, you would need to measure the voltage at the strip while it is full whits to see.
I am going to double-check all my measurements today, time allowing. I am mysterified as to why my current draw is so low. If you see something in my sketch that accounts for this, please tell me.
You should really post your code in each topic you make so we don’t have to cross reference threads...
The “brightness” setting is set to 55 which will reduce the current given to the Leds globally....
The “green” color is only one led, while the white color is three Leds. To achieve the same level of brightness between “white” and “green” you need to drive each of the rgb Leds at about a third the current the green led used to make the “green” color.
Qdeathstar:
You should really post your code in each topic you make so we don’t have to cross reference threads...
The “brightness” setting is set to 55 which will reduce the current given to the Leds globally....
The “green” color is only one led, while the white color is three Leds. To achieve the same level of brightness between “white” and “green” you need to drive each of the rgb Leds at about a third the current the green led used to make the “green” color.
You didn't read my table above the sketch. I set the brightness to various values when I made my measurements. I know how RGB LEDs work. What I don't understand is how white draws about the same current as each individual color at all brightness settings.
Maybe you don't. And in that case probably I don't know either. It could be that the library is causing this effect. Or perhaps it could be something in the design of the chip itself...
