I am trying to wire some LED strips but they are getting a lot hotter than expected. I have noticed that they are also drawing much more current than expected. I am going to show what I've measure first as the product details seems to be contradictory.
I have two lengths of 2m of 24v CCT/dual-white LED strip connected to a 150W 24v constant-voltage PSU (https://www.nationallighting.co.uk/indoor-lighting/led-strip-lighting/150w-24v-constant-voltage-led-driver-not-dimmable).
The strips are about 4.8m away (they will need to be about 10m once installed properly) and are wired using the standard flat RGB wire (4 x AWG22/0.32mm^2). Two of the wires are connected together for the positive side. The measurements:
- 2.8A coming from the power supply, for both strips, so this should be 1.4A for each strip
- 22.6v at the start of the strip
- 22.2v at the end of the strip.
The product listing has seeming contradictory info: https://www.ledison-led-lights.co.uk/strip-lights/led-strip-lights/54watt-led-strip-light-5m-dual-colour-10-8w-meter.html
- 10.8W/m in the image and specs
- 1200 lm/m according to the one of the images - using 1. this implies ~111lm/W
- 90 lm/W in the specs - different from 2.
- 800 lm(WW)/900 lm(CW) in the specs - implies a total of 1700lm/m, and assuming 10.8w would yield 160 lm/W.
I did try to ring the seller before purchasing to get clarification and ask for a spec PDF, but they carried on quoting the 1200lm and 10.8W, but didn't really address points 3 and 4 after I tried to explain my numbers. I ended up buying it anyway just to see what happened.
The main/intended outcome: how to reduce heat/increase lifespan
I have mounted the strip to aluminium profiles, but the profile is still getting too warm for my liking and I would like to run them at a lower temperature/current to increase their life. Ultimately I would be driving them using PWM from a custom STM32 based circuit, but I still want to know that at 100% light they are not getting too hot. I assume that PWM alone wouldn't reduce the temperature that much as the LEDs are still driven at the same current during the cycles.
Whilst the specs are contradictory, this is the first LED strip I have seen draw more current than specified, is this even possible or do leds just draw as much as they need immediately after turning on (so no thermal runway)?
I am fairly new to working with LED strips and am mostly a computer scientist with a bit of embedded software/electronics knowledge, but not really the electrical characterstics side of things.
I have learnt that too much current is ultimately what kills LEDs (and subsequently heat) as they care more for a fixed current close to their rated voltage than a fixed voltage at about the correct current (I need to find the link/forum post where I read this). Because of this, they can be driven at a lower voltage and actually have a higher efficiency when doing so.
Most questions online are about LED strips now drawing as much current as expected or being bright enough etc.
I was also expecting that the voltage drop after 4.8m of quite thin wire would have have reduced the voltage enough (and therefore the current?) enough to have reduced the heat output. I may not have a correct understanding of the relationship between current/voltage and resistance here and which one causes the other.
I have tried to look at multiple voltage drop calculators, which seem to be giving different values as well, all of which are higher than what I actually measured, so I'm not sure what knowledge I'm missing there:
I suspect that I will need to use a voltage step down converter and use the current limiting feature of that (I've got one of these https://www.ebay.co.uk/itm/DC-DC-BUCK-CONVERTER-STEP-DOWN-8-40V-TO-1-25-36V-8A-12A/173554594886).
The real question(s)?
I think that maybe I have a misunderstanding of how current limiting at a given voltage actually works and perhaps that is what the question should have been. I am also confused how using a buck converter to reduce the current is acceptable, and yet using a power supply that can only provide the same current is not OK.
I did also have a probably crazy idea that maybe the voltage drop could be beneficial if it reduces the current to the LEDs without needing to use a buck converter? I was thinking it would distribute the heat along the entire wire instead of needing another module that will need heatsinking, provided we don't go past the temperature/current ratings for the wire. It didn't seem to work in this case, and I suspect wouldn't at the 10m either or would be too unpredictable. Given that the voltage drop calculators didn't match my results, I may also lack some knowledge there as well.