5v or 12v

[chicken]

Looking at a bunch of RGB stuff this year. I have some 5v pixels laying around from last year to play with but I see Ray has 12v ones now. I like the idea of being able to get longer runs without having to do a midspan power but how much extra power am I going to waste in each of the pixels with them having to drop the voltage down from 12v vs 5v. Just looking to get some ideas on what others are doing. Thanks

 

[random]

I'll preface by saying I'm a n00b, these opinions are based on the research I've done since becoming interested in RGB pixels.
From what I have seen, the 5V strings can be one of any number of control ICs but the 12V ones are TM1804 IC which are constant voltage drive. What this means is a larger variation in brightness from the start to the end of the string as a result of voltage drop. If you have 12V at the start of a string and 8V at the end of the string the lights at the end will be a little dimmer. Maybe enough to notice, maybe not.
I decided to go with constant current drivers for that reason, ie WS2801. These ICs have a voltage tolerance of 3.3V-5.5V so you need to reinject power more often but your lights stay consistent. Nothing to stop you using TM1804 and injecting power again whenever required, the 3.3V lower limit of WS2801 is a hard limit (below that, stops working) whereas TM1804 is a soft limit (no reinjection just means dimmer lights but they still work). So TM1804 is a better choice if you are running a long string. I am looking at running a matrix so reinjection is no problem.
From memory the increase in power consumption is of the order of 15-20% for the 12V strings as a result of having to drop more voltage at each pixel.

 

[fasteddy]

I'll preface by saying I'm a n00b, these opinions are based on the research I've done since becoming interested in RGB pixels.

You description is pretty much spot on, so for a noob your doing really well in understanding some of the differences. thanks for taking the time to reply to this thread, I must have missed this one as no one had responded previously.

 

[Superman]

Just looking to get some ideas on what others are doing. Thanks

High Chicken,
I use a mix of both 12v RGB Strip and 5v Pixels (ws2801) as Random said the ws2801's do need power injection and I find that this needs to be done at the start and the end of a string of 50.

 

[CustomTB]

Anyone have good diagram on how the power injection works?

 

[fasteddy]

This may be some help for you
http://auschristmaslighting.com/forums/index.php?action=dlattach;topic=1211.0;attach=1200

 

[zeph]

From memory the increase in power consumption is of the order of 15-20% for the 12V strings as a result of having to drop more voltage at each pixel.

From another n00b - something here is not yet making sense to me - just because there's some missing piece of info, I'm sure.

If one were driving a roughly 2 - 2.5v LED, the resistors (and to some degree the chip's transistors) need to drop the difference between that and the power supply as waste heat. At 5V, that would make a single LED about 40-50% efficient (2 - 2.5v/5v). At 12 V a single LED would only be about 15-20% (2 - 2.5v / 12v) efficient. However at 12v, one could put multiple LEDs (per color) on the circuit; 3 LED's would make it about 50-65% efficient (6-7.5v / 12v). At 4 LED's it could be more efficient still, but would also lose some of the headroom for compensating for voltage drop. (However, this would mean three - or four - RGB LED's per controllable node, not one.)

This would imply that 12v power was either far less efficient (singled LED), or a bit more efficient (3 LED), than 5v power. So what am I missing?

(IF the control chips were doing per-pixel DC-DC voltage conversion that would be a game changer, but I see no sign of that)

 

[fasteddy]

From memory the increase in power consumption is of the order of 15-20% for the 12V strings as a result of having to drop more voltage at each pixel.

From another n00b - something here is not yet making sense to me - just because there's some missing piece of info, I'm sure.

If one were driving a roughly 2 - 2.5v LED, the resistors (and to some degree the chip's transistors) need to drop the difference between that and the power supply as waste heat. At 5V, that would make a single LED about 40-50% efficient (2 - 2.5v/5v). At 12 V a single LED would only be about 15-20% (2 - 2.5v / 12v) efficient. However at 12v, one could put multiple LEDs (per color) on the circuit; 3 LED's would make it about 50-65% efficient (6-7.5v / 12v). At 4 LED's it could be more efficient still, but would also lose some of the headroom for compensating for voltage drop. (However, this would mean three - or four - RGB LED's per controllable node, not one.)

This would imply that 12v power was either far less efficient (singled LED), or a bit more efficient (3 LED), than 5v power. So what am I missing?

(IF the control chips were doing per-pixel DC-DC voltage conversion that would be a game changer, but I see no sign of that)

What you find is generally that the strips and modules are mostly 12vdc and use 3 x 5050 LEDs in series for each section, there are also some with 1 or 2 LEDs per section at 5vdc
The strings mainly come in 5vdc due to the fact that they are only driving 1 LED and this is the most common and efficient design. The 12vdc variations are a special design to overcome the voltage drop issues of using 5vdc and allow for longer string lengths but also come with a cost of more energy wasted. The 12vdc pixel IC controls with constant voltage as apposed to constant current for the 5vdc and constant current is better for driving LEDs as LEDs are current devices and is generally the preferred method for controlling pixels. The other thing is the 12vdc pixels are driven at around 10ma instead of 20ma for the 5vdc variations so some intensity differences may be seen. So with the strings if you use 12vdc there are some trade offs, in the end most wont notice or care for these trade offs as long as the pixel works and does what it supposed to do.