Questions about Amperage in an LED circuit

[UF8FF]

Hello everybody,

I'm playing around with some LED strips and have a few questions. As I've searched for a good solution for my project, I've decided that I'll need 3-4 5m strips of 5v LEDs and I'll be powering them with a 300W PSU and a DigQuad Controller since it has fuses included already. But, while doing all this research, I have built up some questions about amperage that I need some assistance answering. (And probably just the basics of how electricity moves, to be honest here).

Details about the strips:
5m strips with 30 LEDs/meter. At full RGB brightness they consume 7.5A per strip.

I want to inject power every 5m to avoid voltage drop and keep color accuracy. Here are some questions. (Please note that I am not accounting for voltage drop over distance in these images, AWG of the wire, or the draw from the little MCU that will run the data):

Q1. If I make a number of t-junctions, does the main conductor still carry the full amperage of all the LEDs included in the circuit of that conductor? Or does the main wire have less load on it? (like in the top image)

Q2. In image two, is it correct to assume that using two "home runs" that also act as parallel runs would distribute the load and allow me to use a higher gauge wire? Again, are the T junctions helping me in terms of load? or do they only help distribute 5v power?

Q3. Does the current on a conductor dissipate as electrons are consumed? Meaning, if I have a conductor with 4 T junctions, will there be less current as the main conductor get further away from the supply source because of the LEDs 'using' the power?




Apologies if these questions are super elementary, but it's been driving me crazy!


Edit: just realizing I didn’t take into account the current that will be going through the strips themselves, as well

[strawberry]

https://www.inchcalculator.com/wire-gauge-size-and-resistance-calculator/
both directions so 10m
ohms law wire voltage drop V(volts)=I(amps)*R(ohms)

[Martin Miranda]

don't forget that the current on the negative side is equal. so you should use the same thickness of wire.

[mariush]

I'd suggest getting a bunch of  dc-dc converters , to do 12v to 5v conversion  ...
For example these can do up to 3A at very high efficiency : https://www.ebay.com/itm/223048616138
So you could use 3 of these on each rgb strip, injecting 5v at every 33% point. You'll have 9A of current available for a peak 7.5A so the converters won't be overloaded.
You could also use thinner cables over the length of the whole strip because you may not care that much about voltage drop, the dc-dc converter will produce 5v with 12v or 11.5v

You say a 300w power supply but keep in mind a lot of ATX power supplies have a limited 'budget' for 3.3v and 5v ...  usually 100-120w for both combined, and around 15-20A on each 3.3v or 5v
With the dc-dc converters you could power your strips from 12v of your computer power supply saving you from needing separate power supply. 

[Zero999]

Use higher voltage LED strip. 5m of 5V LED strip is just asking for trouble. Change it for 12V, or even 24V LED strip, which will use much less current.

5V LED strip is horribly inefficient. The LEDs have a typical forward voltage of around 3V, which means 2V is being dropped inside the resistors, built-in to the strip, so ⅓ of the power is wasted, before it even gets to the LEDs. 12V strip consists of 3 LEDs in series, giving a total forward voltage of 9V, so ¼ of the power is wasted. A higher voltage also means lower currents, so less voltage will be dropped, over the length of the strip, giving a higher efficiency and more uniform brightness.

[tooki]

Use higher voltage LED strip. 5m of 5V LED strip is just asking for trouble. Change it for 12V, or even 24V LED strip, which will use much less current.

5V LED strip is horribly inefficient. The LEDs have a typical forward voltage of around 3V, which means 2V is being dropped inside the resistors, built-in to the strip, so ⅓ of the power is wasted, before it even gets to the LEDs. 12V strip consists of 3 LEDs in series, giving a total forward voltage of 9V, so ¼ of the power is wasted. A higher voltage also means lower currents, so less voltage will be dropped, over the length of the strip, giving a higher efficiency and more uniform brightness.

The DigQuad controller appears to be a controller for addressable LEDs, so none of the above applies, nor is there really any alternative to 5V. I have yet to see an addressable LED designed for higher voltages, alas. So the only option, really, is to feed power in via fat wires at regular intervals.

[Ian.M]

You *could* use individual buck converter modules right next to each strip, and distribute the power at a higher voltage on thinner wires. However to prevent possible damage if one strip looses power, you should run the data line (and clock line if present) through a dual supply level translating buffer that can tolerate partial power-down, with one side powered from the controller (or previous strip if daisy-chained) and the other from the buck module's 5V output at the strip.  e.g. use 74LVC1T45 for a single signal or 74LVC2T45 for two signals, both with their DIR pin tied low so they act as unidirectional buffers B => A.

I wouldn't recommend multiple buck modules feeding a single  strip, as any slight difference in their output voltage will result in one providing most of the current and probably overloading, so you'll still need heavy wire for both 5V and 0V  between the buck module and each power injection point on a single strip.

[mariush]

The DigQuad controller appears to be a controller for addressable LEDs, so none of the above applies, nor is there really any alternative to 5V. I have yet to see an addressable LED designed for higher voltages, alas. So the only option, really, is to feed power in via fat wires at regular intervals.

The data wires can remain 5v, but you can cut / break the 5v wires at various locations on the strip  (for example on a 5m strip, if you use 3 dc-dc converters you could have one voltage insertion at start of the strip, then at 1.75m , then again at 3.5m
All three segments will receive 5v because you set the output of each regulator to 5v, and as the segment doesn't go over the maximum current output of the dc-dc converter (3A in the ones I linked to) you won't see too much brightness variations.

[tooki]

The DigQuad controller appears to be a controller for addressable LEDs, so none of the above applies, nor is there really any alternative to 5V. I have yet to see an addressable LED designed for higher voltages, alas. So the only option, really, is to feed power in via fat wires at regular intervals.

The data wires can remain 5v, but you can cut / break the 5v wires at various locations on the strip  (for example on a 5m strip, if you use 3 dc-dc converters you could have one voltage insertion at start of the strip, then at 1.75m , then again at 3.5m
All three segments will receive 5v because you set the output of each regulator to 5v, and as the segment doesn't go over the maximum current output of the dc-dc converter (3A in the ones I linked to) you won't see too much brightness variations.

Why are you responding to me? I was responding to the suggestion to use higher voltage LED strips, nothing more.

Ian.M addressed both possible topologies for using multiple buck converters.

Regardless of how many power domains you have (1 big one or multiple smaller ones), it's still wise to feed a strip frequently from its power domain, because RGB LEDs change color balance as voltage drops, because of the different Vf of the three LED colors, resulting in blue going darker sooner than green, and green sooner than red. So in the end, in severe cases, the end of the strip is not only dimmer, but takes on a pink hue.

[Ian.M]

@Mariush,
The problem with that is what happens if one segment powers up slower than the rest, or something causes transient loss of supply to one segment?  Applying signals before power can cause latchup and destruction of the first LED in the strip's integrated controller hence my suggestion to use  dual supply buffers that tristate if only one side is powered.   You could get away with using series resistors in the data (and clock) lines before the first LED of each strip, and dual Schottky diodes (e.g. BAT54S) to clamp the signals to the (local) supply rails at the input(s) of the first LED in each strip.

[UF8FF]

I'd suggest getting a bunch of  dc-dc converters , to do 12v to 5v conversion  ...
For example these can do up to 3A at very high efficiency : https://www.ebay.com/itm/223048616138
So you could use 3 of these on each rgb strip, injecting 5v at every 33% point. You'll have 9A of current available for a peak 7.5A so the converters won't be overloaded.
You could also use thinner cables over the length of the whole strip because you may not care that much about voltage drop, the dc-dc converter will produce 5v with 12v or 11.5v

You say a 300w power supply but keep in mind a lot of ATX power supplies have a limited 'budget' for 3.3v and 5v ...  usually 100-120w for both combined, and around 15-20A on each 3.3v or 5v
With the dc-dc converters you could power your strips from 12v of your computer power supply saving you from needing separate power supply.

Thank you for your thoughts! Just for some additional info I am using a Meanwell  PSU (LRS-350-5), not an ATX! So I have 60a out of the entire supply -- I believe it's something like 20amps per terminal. I will double check the spec sheet.

As for the idea of the buck-converters: I have seen people use that same idea before! Your approach would absolutely guarantee that I won't suffer any voltage drop, but in the final project I am expecting to limit the current to about 75% at most and in some testing with the strip I have here, the brightness and color accuracy seems ok so far. This is close on my mind if I am unhappy with the results though or want to boost brightness.

Use higher voltage LED strip. 5m of 5V LED strip is just asking for trouble. Change it for 12V, or even 24V LED strip, which will use much less current.

5V LED strip is horribly inefficient. The LEDs have a typical forward voltage of around 3V, which means 2V is being dropped inside the resistors, built-in to the strip, so ⅓ of the power is wasted, before it even gets to the LEDs. 12V strip consists of 3 LEDs in series, giving a total forward voltage of 9V, so ¼ of the power is wasted. A higher voltage also means lower currents, so less voltage will be dropped, over the length of the strip, giving a higher efficiency and more uniform brightness.

Thanks for chiming in. You know, I have tossed the idea around about using 12v strips but ultimately decided not to because I wanted to use less power day-to-day (as I don't plan to have these running at 100% brightness all the time). But you make a great point about the power being wasted. I'm robbing Peter to pay Paul, I guess. I have only purchased one strip and I do plan to put these strips in different areas of the house. Maybe for the bigger rooms the 12v will be a better idea. I did also want to stick with 5v since it can be argued it's 'safer' than a 12v circuit. But starting a house fire isn't all that safe, either .  I'll have to do a bit more tinkering. Thanks again for your insight.


Any thoughts on the theoretical questions I posted? I'm hoping to make sure I'm understanding how current works in a circuit. Also, as I edited in my first post, I did forget to include the current running through the LED strips themselves, so those should share half the load.

[Zero999]

I'm currently looking into using it to light my walk-in wardrobe. It's extra low voltage, so very safe, being self-adhesive, it doesn't need lots of holes drilling and it spreads light in all directions, rather than in a narrow beam, but it does have disadvantages. Low voltages mean high currents and the fact that at least 25% of the power is wasted in the current limiting resistors, make it unsuitable for high power applications.

To summarise: LED strip is great, but I wouldn't recommend it for high power levels.