another 5v vs 12v question

[jeffcrouse]

5V LED strip has always been my default for some reason, but I have a big build coming up with very long strips (40 strips, 200 pixels each), and it seems that using 12V with my PixLite MKII Long Range Controller (with Long Range Isolated Receivers) would solve SO many problems. If I use 5V, I would need to inject power every 50 pixels, whereas, with 12V, it seems like I can run power and data at one end of a 200 pixel strip and I'd be fine...

I've read many discussions...

• Choosing a Pixel Voltage: 5V vs 12V
• What Digital LED chip to choose
• 12v vs 5v LEDs
• 5V or 12V WS2811 LED for my project?

Despite all of these, I still can't wrap my head around why anyone would still choose 5V except power consumption. It seems that 12V "smart" strip is newer, so there aren't as many options, so maybe that accounts for some of it. Also, I've read that 12V strip is only available in "sets of 3" configuration, but WS2815 seems to work just like other 5V strip, so I don't think that's relevant any more.

Am I missing some other important factor as to why 5V is still so popular? I'm nervous to make the jump to 12V, but it seems like the thing to do.

Also, ALITOVE 12V WS2815 should work fine with the Advatek controller, right?

 

[AAH]

I don't know anything about the WS2815 but I know enough about physics, electronics and marketing to be very dubious about an ad that claims "WS2815 LED strip has no voltage drop ". A strip that defies the laws of physics makes me doubt some of the other stuff on the page. 1 thing that has me puzzled is the claim of 7.2W/m for 30Led/m. This works out to 0.24W/Led. The WS2815 data sheet says that the led current is 15mA/channel which equates to 0.54W/Led. The 0.24W/Led is actually what it works out to for 5V. The 72W/strip seems to be based on the 0.24W/Led. My calculations show it as 162W/strip.
As far as I know every variety of pixel/strip uses the 3 leds in parallel except for 12V GS8208. Instead of the 55mA/45mA @12V per pixel it uses 18mA @12V per pixel. It does this at the expense of wasting some power when there is differing brightness levels between the RGB channels.
Both of these led/strip types will work with Advatek.
5V strip and nodes are more efficient for single led resolution than 12V WS281x. 12V 3 led/pixel is more efficient than 12V 1led/pixel. The WS2815 pixels need to get rid of a fair bit of heat in a small package so I wouldn't be a fan. I would also try some before going bulk and actually take some measurements. Having a theoretical 72W compared to an actual 162W could be VERY bad.

 

[TANSA]

In your question of 5V or 12V the WS2811 LED's are actually only 5V LED's.
So it comes down to preference and a little electrical knowledge.
I am a sparky of 30+ years and the main difference to me is I like 5V but as everything there are pros and cons.
5V you need to inject power more often but also is less affected by water/moisture and actually uses less power.
12V you do not have to power inject as often but at each pixel it has to drop the voltage that goes to each LED.
To make it more complicated each colour (Red, Green, Blue) requires a different voltage.
But this site has a good reference to the WS2811 pixels so you can make up your own mind.
It is all about choices and there is no one choice that is better than the other but it does limit or send you into other choices that may be different.
https://www.wiredwatts.com/learn-regulators
I use a Pixlite controller/s and a DMX controller, always read the manuals to see which standard/s they support and how

Happy christmas lighting
(PS it can get very frustrating but hang in there)

 

[jeffcrouse]

Thanks for the responses

@AAH -- I will definitely do my due diligence and test voltage drop. I understand that they are using marketing speak when they say there is none. I'm hoping that I can get 180 pixels (3m) of strip without visually noticeable voltage drop.

@SAALTFAM - My lights will be indoors and I will have plenty of power available, so the tradeoff of not having to inject power as frequently is very appealing to me.

 

[TANSA]

A little secret you can run power to both ends (even if it a small wire) and you will probably never notice the very minor difference in the middle.
a very little known fact I got from working on airport runway/taxi lights the human eye has a hard time noticing a 5% drop in luminosity good eyes can see about 10% down to about 5%

 

[Benslights]

been powering both ends for years it is nothing new and spoken about at very mini i have ever been to.

 

[tooms]

5V pixels always seem to be cheaper

 

[TerryK]

I made voltage/current tests on a WS2811 5 volt string and a WS2815 12 volt strip. The results were...




Edit Apr. 29, 2020 WS2815 ma/Node and mW/Node decreased by factor of 3

Comments: First, current draw with all pixels off was not possible so while it is assumed that current draw drops near zero for black, it remains essentially unknown. The controller driving the string/strip is also a BTF Lighting device and colors are assumed to be full on but it too was not possible to verify. Measurements were also taken at the string/strip ends so calculations for single nodes must be considered a 'rough' average. Measurements were made with a Fluke 79. I also ran twice through the test sequence. Values changed a bit I am assuming due to heating of the devices. The value change though was so slight that the first test sequence numbers were used.
World Semi indicates the WS2815 Min/Max supply range as 9.5 to 13.5 V. If string to string operation is fairly consistent and distance between strips is kept short, it is reasonable to conclude that 2 strings could be daisy-chained with a single power injection point. Four possibly 5 strips if power is injected at both ends before any IC in the chain drops out of its supply range. The World Semi WS2811 datasheet has that device's absolute maximum supply rating as 6.0 to 7.0 volt. This a rather atypical for a 5 volt device and World Semi may have increased the rating to compensate for the VDD pin series resistor. Be that as it may, it makes determining how many strings can be daisy-chained before voltage drop takes a chip outside of its ratings more difficult. It is also rather well known that visible effects may not be seen even though a node is operating outside of its manufacturer's specifications.
Regarding wattage requirements, for the WS2811 it is easily seen that current /wattage increase as a second and the third LED is turned on. Obviously the LEDS are operating parallel. The WS2815 however maintains a rather consistent 0.42 0.14 watt dissipation when any LED or LED combination is on. This then indicates that there is a rather clever design in the LED drive implemented with the WS2815. The datasheet does not indicate where/how the LEDs are wired but it appears they are in series. A last power comment, at white, the WS2811 uses slightly less than half of the WS2815. No longer accurate, depends upon activated color(s).

I would at this point agree with AAH that the WS2815 is similar to the GS8208 albeit the power calculations I arrived at differ somewhat. Mark_M also has a rather nice video analyzing the WS2811 12 volt devices.

 

[algerdes]

I'm curious as to why you could not get current draw without lights on. The chips themselves are always on and using. What method did you use to get the readings? Inline or induction?

 

[TerryK]

Measurements were made inline with a Fluke 79; 10 Amp or 40 mA input whichever worked best. The controller I used (a BTF Lighting 'Colorful X2') has a fair number of color pattern states but not an all colors off. When I posted the chart my assumption was that users would be more interested in what a string is doing full on (white) to better plan power requirements. Essentially I put little thought into checking the colors off current draw.
There was also interest in the WS2811 12 Volt device measurements. I did not have a string so I ordered one. When I get it I'll run tests on it as well and double back to pull color off measurements from the WS2815B strip and WS2811 5 Volt string.

 

[algerdes]

Thanks for that information. Looking forward to any further results.

 

[TerryK]

Correction to previous post...

I had a factor of 3 error in the Excel formula used to calculate mA/Node which also rippled into the mW/Node. Root cause was testing a 150 node strip and incorrectly making relative calculations from carrying WS2811 number of nodes to the WS2815 section.