Are LED strips not subject to the laws of physics?

[The Magic Rabbit]

So here I am, designing some lighting using LED strips off eBay... It doesn't seem to take long for those currents to add up! No problem, thought I, as I have a power supply for a PC hanging around I could use and that says it can deliver 20Amp on the 12V rail and 18Amp on the 5V rail, and it even has a 5V line for the control chips that is not switched. Perfect!

Then I start looking at wire... and I'm thinking either my maths is wrong, or there's no way those LED strips can work!

The 12V strip is 5m and I calculated at 6 Amp ... but to get that voltage over 5m without massive voltage drop and associated heat would need quite a chunky wire, so how could it possibly work with the thin tape-like circuit they use!?

Then I start calculating the RGB stuff and it just gets silly! There's 300 of those at 60mA each... That's 18Amp over 5m ... That cable is thicker than the stuff you wire cookers in with... How can that possibly ever work on that flexible circuit tape?!

So am I barking up the wrong tree, or is it simply that the stuff doesn't work as advertised and has to be powered in short lengths with very fat cable to it?

Sent from my XT1039 using Tapatalk

[The Magic Rabbit]

1. No they're not, the PWM is often varied but there has to be a significant difference in duty cycle to get different shades.

2. Every time you want white, or off-white. It clearly depends on the colour you want to produce.

3. 60mA is LOW for these kind of LEDs and its not uncommon for them to be overdriven to more than double that on some applications. I have a few LEDs that use 2.8Amp (but not in this project)

Sent from my XT1039 using Tapatalk

[David_AVD]

Almost all 12V LED strips have the LEDs arranged in sections of 3.  That means only 100 sections (not 300) over the 5 metres.

So, 100 x 0.06A per comes out to 6A for the 5M length on full white.   In reality (and especially if you only power from one end) voltage drop in the strip will see the actual current draw less than that.

60mA per LED is 20mA per colour and this is the max I've ever seen with strips.  Some others are rated closer to 15mA per colour per section.

The 150 LEDs/metre strip figures come in at approx half of the 300 LEDs/metre stuff.

[The Magic Rabbit]

The RGB strip is 5V, not 12V

Sent from my XT1039 using Tapatalk

[David_AVD]

Well, you did say:

The 12V strip is 5m and I calculated at 6 Amp

[The Magic Rabbit]

Sorry, I should not have assumed. The RGB strip is comprised of individually addressable  WS2812b LEDs.

Even with the current reduced, surely voltage drop over that length of wafer-thin circuit is going to be very significant?

Sent from my XT1039 using Tapatalk

[David_AVD]

Yes, there is some voltage drop along the length of the strip.  You do need to power the strips from both ends.

Have you actually measured the current yet?

[sleemanj]

The tape is a bus, the leds (or series strings thereof) are across said bus.  Any point on the bus only carries the current for the (series sets of) leds after said point on the bus.

In other words, your length of tape might require 6A, but the tape is not passing 6A all the way from one end to the other, only that first section (and your initial wires) are passing the full 6A and then it drops off linearly along the tape to 0A after the last LEDs.

The individual die on these strips will only be passing about 20mA each, these are not high power leds (and the tape is certainly not a heatsink for one!).

Yes the intensity will taper off the further down the tape you go, but as you know human vision and light intensity is a funny old business and it will probably have to taper considerably before you actually notice it.

Powering the strip from both ends is not uncommon if necessary.

[The Magic Rabbit]

The RGB strip hasn't arrived yet. The 12V white strips claim 6Amp and are drawing 2Amp, which I suspect is due to voltage drop. The strip also gets pretty warm.

Powering the strip from both ends will cause the drop to be most significant towards the middle wouldn't it? Could I run a 12V rail alongside it and "tap off" at intervals of say, 0.5m, or would that be worse?

Still going to need some hefty cable though wouldn't I?

Sent from my XT1039 using Tapatalk

[The Magic Rabbit]

Obviously I'd need a5V rail for the RGB

Sent from my XT1039 using Tapatalk

[David_AVD]

Most of the claimed specs for LED strip are complete bullshit (or gross copy/paste errors).  Measure it once you have it and go from there.

[sleemanj]

>  which I suspect is due to voltage drop

Time for you to break out your multimeter and confirm your hypothesis then.  Ohm's law to the rescue.

For starters, look at what the configuration of the LEDs and resistors on your strip is, what resistor is installed, what current do you calculate should pass.

[mikeselectricstuff]

For 5V strips,  to light any significant proportion of the LEDs you need to feed power in from chunkier cable at regular intervals, e.g. for 48 Leds/metre, about every 2m

Of course controllable LEDs are often used for effects, and if, for example, you're showing a small 'blob' moving along the length of the strip you may well be able to get away with feeding a 5m length form one end. 
 
Another factor is that at least for the APA102 (SPI/Dotstar) I've tested, the driver does NOT deliver constant current as it's well inside its dropout region. Therefore LED brightness is moch more affected by voltage drop than it ought to be if it was a proper constant-current driver.

You can get higher quality non-controllable strips with constant-current regulators built in. 

[The Magic Rabbit]

I can only do this for the 12V LEDs at the moment...

I measured 1.5 ohm along the 5m strip (3 ohm both ways).

LEDs have 390 ohm resistor for current limiting.

There are 300 LED in 100 segments.

So, this 6 Amp chain should be drawing 3 Amp... That's still more than the 2 Amp measured.

?

Sent from my XT1039 using Tapatalk

[The Magic Rabbit]

12V one end, 10.2V the other.


Sent from my XT1039 using Tapatalk

[David_AVD]

Not all LED strip are created equal.  Some may use thicker copper or wider tracks, etc.

Is this 12V strip the white one?  If so and there is one resistor per 3 LEDs, that's 20mA per section so 2A total is reasonable.

[The Magic Rabbit]

Yes, the 12V strip is white, with a 390 ohm resistor, which I made to be 30mA per segment. At the far end though, it would be less.

So in theory at least, the 12V strips can take almost 24V to get the LEDs up to spec...

The idea was to have white strips for an actual light source, and RGB strips so I could have scrolling letters in the display (5x5 font).

Sent from my XT1039 using Tapatalk

[tszaboo]

You should consider that the average current through the strip is half of the maximum current, since it branching off every now and then. Also, powering it from both ends will give you 1/4 of the total current. I would suggest to measure the resistance of the strip GND, between both ends. I'll do the same, I have a "dumb" RGB strip.
I also noticed, that with the supplied controller, the duty cycle was never 100%.
I started calculating at some point, that it would be good practice, to create a DC-DC stepdown converter, very small size, maybe even integrated inductor in some SO8 package, with preset voltage output, and minimal passives. That would reduce cabling cost and make addressable LED strips easier to install. And place this DC-DC on each section. Like connect 12/24V on one input, you are done, voltage drop doesnt matter. But solution cost was too much.

mod: I see someone aready measured it.

[The Magic Rabbit]

As to whether it's reasonable - does it seem reasonable that a 12V 6A 72W 5m strip is only actually 12V 3A 36W Max and only actually running at 2A/24W?

Sent from my XT1039 using Tapatalk

[rich]

Am I missing something here? 12V / 390R = 30.7mA, but that hasn't allowed for the forward drop of the 3 LEDs in the segment.

2A for 100 segments gives a neat 20mA and a rating of 4.8watts/meter.
If yours was advertised as 12V 6A 72W 5m, (14.4 watts/meter) you clearly don't have that and has nothing to do with volt drops down the length. Your worse case voltage was 10.2V, which is only a 15% drop.

Is it reasonable? Not if it was advertised and payed for as 14.4W/m. This ebay seller seems to offer (4.8W/m, 12W/m and 14.4W/m) which adds credibility to  your strip being 4.8 W/m.

http://www.ebay.co.uk/itm/12V-5630-5050-3528-SMD-Flexible-LED-Strips-Strip-Light-Waterproof-/161230174915?var=460280855066&hash=item258a113ac3:m:md8T8dmWUnXl-lBFHuOgk2A

[The Magic Rabbit]

They are SMD5630 from this listing:
http://pages.ebay.com/link/?nav=item.view&id=171576163813&alt=web

I'm challenging the seller over this, as I've also noticed their picture shows 470ohm resistor, so couldn't possibly offer the output claimed in their table.

Am I correct to think if I want the output the LEDs claim (looking at several different data sheets for them, 60mA is ideal) I would need to double the voltage?

Sent from my XT1039 using Tapatalk

[mikeselectricstuff]

They are SMD5630 from this listing:
http://pages.ebay.com/link/?nav=item.view&id=171576163813&alt=web

I'm challenging the seller over this, as I've also noticed their picture shows 470ohm resistor, so couldn't possibly offer the output claimed in their table.

Am I correct to think if I want the output the LEDs claim (looking at several different data sheets for them, 60mA is ideal) I would need to double the voltage?

Sent from my XT1039 using Tapatalk

"470" means 47 ohms (47x10^0) not 47. Assuming 3V droip per LED, current = 3/47=63mA = 0.756W, so their figure of 75W/5m is about right, but unless power tracks are reinforced, voltage drop will reduce this
.
That's a lot of heat & I suspect these won;t last well if run at full power, and the sticky will almost certainly come unstuck.
 

[Towger]

Oh, the joys of Chinese Watts.  :-) If you want to run them at full power (not recommended at all) you will need to feed them at 14+ volts every meter or so.
The WB2812s are nice, from memory they will run quite happily down to not much above their forward voltage drop.

[The Magic Rabbit]

So I just learnt how to read resistance on a surface mount component... Cool!

Maximum is 150mA... I wouldn't want to run them at that, since that's going to shorten their life and require expensive heat sinking.

If I assume the cheapest brand SMD5630 then they would be about 50mA at 3.4V (crap compared to brands like CREE) which would need a 39 ohm resistor at 12V with three on series if my maths is right at last.

The RGB chain is going to have to wait until it arrives to determine what it's actually running at.

Still, that should be 5A per strip, so they clearly need regular tapping to a 12V feed.


Sent from my XT1039 using Tapatalk