Some 12V verse 5V musings... with some math...

nutz4lights

Full time elf
Joined
Dec 12, 2012
Messages
305
Location
Melbourne, Florida
I had some more time so I pulled out my 93 pixel string (don't ask) of 12V square-type pixels with 10cm spacing and made the same measurements outlined above. Please see the picture below... Some commentary... it appears that these are 0.020A per color LED elements. I know Eddy's 12V versus 5V sticky says that most or some of the 12V pixel lights use 10mA, so I was surprised to see a current draw of 0.020A on these. The bottom two sets of numbers are a bit off, but the measurements are the measurements... and I believe the calculations are correct.

[attachimg=1]

I'm just going to throw this out there... but based on what I'm seeing from these measurements, I am very tempted to go with 12V lighting. I have read the 12V versus 5V thread and I can see the following reasons for not using 12V lighting:

1. 12V will use and waste a lot more power than 5V requiring more power supplies
2. 12V will dissipate a lot more heat than 5V due to the wasted energy
3. 12V will output less light than 5V due to running at lower current

#3 is just plain not true for the setup I'm looking at. The colors are running between 13mA and 15mA in color mode for the 5V and spot-on at 20mA in color mode for the 12V. If anything, for my setup, the 12V lights should output more light than the 5V lights in my possession. #2 there's no getting away from, that is definitely going to happen. I let the 12V string run for a good hour tonight and I didn't feel the injection molded section getting warm, which is good, but there's no arguing that #2 will happen. #1 is a bit trickier... There's really no arguing that 5V will require a lower wattage power supply and the math below agrees with that... but I also don't like the thought of under-driving the pixels producing lower current ratings...

- Twelve outputs, 100 pixels per output, 12V draws 4.0A per output running white which means 48A / 80% = 60A x 12V = 720W worth of power (I think I'd be ok with two 350W power supplies) Twelve outputs, 100 pixels per output, 5V draws 1.2A per output running white which means 14.4A / 80% = 18A x 5V = 90W worth of power.

I dunno... the thought of just running one cable to a pixel string per output of the P12R is appealing (no power injection). The cost difference between a single 5V power supply ($15) and two 350W power supplies ($40) is not really all that bad...

If the lights put out less light, that would be a huge negative... more power supplies just can't be the show-stopper here considering how cheap they are, right?

Thanks for reading,
 

Attachments

  • 12V 93 square pixel.jpg
    12V 93 square pixel.jpg
    58.2 KB · Views: 115

fasteddy

I have C.L.A.P
Global moderator
Joined
Apr 26, 2010
Messages
6,648
Location
Albion Park NSW
Beefer said:
agreed - power injecting is just a pain...

Its a pain but a requirement of low voltage lighting which you cant get around.

Now as far as the strings nutz4lights has, then i did post a warning on that as well
http://auschristmaslighting.com/forums/index.php/topic,2653.msg23122.html#msg23122

Im sure they would be running fairly warm if thats the case and is why you have no chance at running these without injecting at every 50 pixels, you would have to treat these the same as you would treat 5vdc strings but with more than double the power consumption than a 5vdc string. Do you have a link to the ones you got.
 

David_AVD

Grandpa Elf
Community project designer
Generous elf
Joined
Jun 12, 2010
Messages
4,681
Location
Victoria Point (Brisbane)
With regard to the wire gauges; don't take the claimed numbers as accurate. Use a micrometer to measure the strand diameter, count the strands and key it into a wire calculator (like this) to get the real size. :)
 

nutz4lights

Full time elf
Joined
Dec 12, 2012
Messages
305
Location
Melbourne, Florida
ɟɐsʇǝppʎ said:
Its a pain but a requirement of low voltage lighting which you cant get around.

Now as far as the strings nutz4lights has, then i did post a warning on that as well
http://auschristmaslighting.com/forums/index.php/topic,2653.msg23122.html#msg23122

Im sure they would be running fairly warm if thats the case and is why you have no chance at running these without injecting at every 50 pixels, you would have to treat these the same as you would treat 5vdc strings but with more than double the power consumption than a 5vdc string. Do you have a link to the ones you got.
Eddy,

I'm definitely not trying to stir the pot here because your information and opinion is worth a heck of a lot more around here than mine is. All I can do is present what I am seeing. I don't know the Lithgow lights folks, but what I can tell you is that my lights are not running hot. My garage (Florida stays warm year round) is right around 25-30C right now, so maybe my results are skewed but they don't feel warm to me at all. Now, what I can do is take the 5V string and the 12V string to work and put a thermocouple on both of them all day while I work and look for temperature rise. I work as a research engineer for a company that builds electronics, so I have access to some pretty nice measuring equipment.

Since I am new to these pixel lights, I should point out what my frame of reference is on "warm". I have been running a mix of incandescent and LED lights for the last 8 years. I started out with the 7W per bulb incandescent C9 bulbs (YOUCH those are HOT!). I moved to the 0.3W per bulb C7 and C9 LED bulbs and then up to the 1W per bulb LED C9 bulbs when they came out (5 LEDs per bulb). Those 1W bulbs used to get a little warm because they were the retrofit bulbs that screwed into a normal incandescent C9 stringer (120V up to the bulb which was then rectifying and dropping voltage to what was required by the LEDs). All of my other LEDs were 120V up to the stringer and ran cool (7W per 60 count C6 bulbs is around 0.12W per bulb).

You had asked about the pixels I am using, here are two points of reference//comparison:

12V:
http://www.aliexpress.com/store/product/DC12V-input-WS2811-pixel-node-100pcs-a-string-IP68-rated-with-all-black-wires/701799_762876648.html

5V:
http://www.aliexpress.com/store/product/WS2811-LED-pixel-node-DC5V-input-full-color-RGB-string-50pcs-a-string-IP68-rated-in/701799_677496152.html

Those are both 10cm spaced. I also have those 20cm spaced bulbs that I'm not too happy with because of the excessive voltage drop. What I can tell you about the above two lights is that:

1. The 12V feels just like the 5V in regards to temperature. Neither are hot, they are both room temperature, which feels like 25-30C in my garage. Neither are certainly "cool".

2. The 12V is drawing more than double the power (over triple actually: 11.5W versus 3.5W) than the 5V light. For my megatree, I would need two P12R units, each with dual 350W power supplies. That is $80 worth of power supplies. If I did 5V, I would need a single 200W power supply with each P12R, so $30, a savings of $50.

3. There is no way I would inject power into the 12V based on what I'm seeing. The 93 count string (I had to cut 7 off because there was a bad pixel) looks just as good as the 10cm spaced 5V pixel string. For 100, it would truly be a single connector system, which would be very convenient for me.

4. One item we haven't discussed is cost... The 12V pixels cost more ($0.40 a pixel compared to $0.28 a pixel for 5V). My megatree, which is going to be 48 strings of 50 pixels, would end up costing $290 more with the 12V compared to the 5V...

I spent the weekend writing this post and reconsidering my position and I think I will go with the 5V after all. The only thing that has me irritated at this point is that, the 12V pixels are the only ones that are actually pumping 20mA through the pixels... even my 10cm 5V pixel strings are only 12-13mA... you made reference to this in a post somewhere... but shouldn't a 20mA burning pixel run brighter than a 13mA pixel?

Take care,
 

fasteddy

I have C.L.A.P
Global moderator
Joined
Apr 26, 2010
Messages
6,648
Location
Albion Park NSW
I know your not trying to stir the pot and im only trying to provide answers for you. :D

What has me mystified is that the 12vdc string is drawing so much more current but your seeing none of that energy being disapated as heat. So when working this out based on maths the energy needs to go somehwere

If we work on Rays specs (Which may be wrong or different)

12v @ 0.3 watts = 25 mA (30mA Rays spec) per node

5v @ 0.3 watts = 60 mA (60mA Rays Spec) per node

So for 100 pixels at 5vdc = 6 amps

Then for 12vdc with 100 pixles = 2.5 amps

So based on that something is not right, either rays specs or there could possibly be an internal short either on a pixel or wiring going into the pixel. Especially that its drawing so much more but not showing itself in dissapated heat. I presume you only have the one string of 12vdc you are testing. Can you cut the string in half and then test both halves to see if the current is the same.

It just doesnt seem to match what should be going on, so im wondering if your results are being skewed by a small short on one of the pixels.
 

nutz4lights

Full time elf
Joined
Dec 12, 2012
Messages
305
Location
Melbourne, Florida
I'm glad we both understand each others intentions!

I can definitely cut the 12V string in half... I had already done just that in my process of finding and fixing the dead pixel. I had re-soldered the two strings back together but can separate them again. I will get to that one night this week.

Ray specs both the 12V and the 5V at 300mcd per pixel... and what is odd about that is that I feel like the light output should be lower for the 12V based on the spec'd current of 30mA compared to 60mA with the 5V string. As I said earlier though... the 12V are definitely comparable in brightness and even-ness of color to the 5V.

I will cut and measure that 12V stringer and get back to the forum.
 

fasteddy

I have C.L.A.P
Global moderator
Joined
Apr 26, 2010
Messages
6,648
Location
Albion Park NSW
You will find it hard to notice a difference between an LED running at 10mA and 20mA due the way the human eye see light which is not linear like an LED. The only real way to see the differences is with a lux light metre as the total overall output will be less.
 

Beefer

Apprentice elf
Joined
Feb 19, 2012
Messages
55
ɟɐsʇǝppʎ said:
Beefer said:
agreed - power injecting is just a pain...

Its a pain but a requirement of low voltage lighting which you cant get around.


You got that right - I have just started testing my real-world setup which includes a bunch of 3m 4-core extension cables, only to find that now I don't get sufficient voltage to all 100pixels (at 12v).

Before I was plugging the pixels straight into the E682 controller and it was powering the whole string fine - but as soon as I have added the extension cables (adding resistance) I don't quite have enough power - as soon as I inject power from the other end, all is good again....
 

nutz4lights

Full time elf
Joined
Dec 12, 2012
Messages
305
Location
Melbourne, Florida
ɟɐsʇǝppʎ said:
So based on that something is not right, either rays specs or there could possibly be an internal short either on a pixel or wiring going into the pixel. Especially that its drawing so much more but not showing itself in dissapated heat. I presume you only have the one string of 12vdc you are testing. Can you cut the string in half and then test both halves to see if the current is the same.

It just doesnt seem to match what should be going on, so im wondering if your results are being skewed by a small short on one of the pixels.
Ok, gonna split up my latest test data into two posts... first one answering the question you posted above. I took the string of 93 of the 12V pixels and cut it into two equal 46 pixel strings (that poor one pixel looked so lonely!). I ran my test routine as I've been doing above, but only measured the current draw on each string (no voltage data tonight). They were both equal (within 0.01A) and these were the values:

White: 2.40A (52mA per pixel)
Red: 0.91A (20mA per pixel)
Green: 0.91A (20mA per pixel)
Blue: 0.91A (20mA per pixel)

The values for the 93 count string were:

White: 4.00A (43mA per pixel)
Red: 1.83A (20mA per pixel)
Green: 1.84A (20mA per pixel)
Blue: 1.85A (20mA per pixel)

So it appears that, aside from white, the pixels are behaving as if they are 20mA per red, green, and blue element. I ran the 12V and 5V together today at work (all day on my desk)... which will lead to my next post...
 

nutz4lights

Full time elf
Joined
Dec 12, 2012
Messages
305
Location
Melbourne, Florida
As I started to say above... I ran the 12V 93 count string and the 5V 50 count string on my desk today at work. That was the first time I had hooked up one power supply to one rail of the P12R and a second supply to the other rail. Pretty cool.... so I brought these into to work so that I could hook up a two channel thermocouple meter, one to each string. I don't know how common thermocouple meters are with those that are not into science & engineering, but they are basically very fancy thermometers. I ran the P12R in test mode since I can't have my personal PC at work... I hooked up a thermocouple to the underside of the 12V and 5V square pixels (the underside is nice and flat). I connected the thermocouple to pixel #2 of each string, figuring that the voltage was the highest at the beginning of the string and hence the voltage dropped by the resistor would be the highest (higher heat generation). I couldn't connect to pixel #1 because of how close it was to the P12R...

So, with all of that said... the room temperature was 23C in my office... both pixel strings started off at 23C. Within the first hour, the 12V string was at 34C and the 5V string was at 26C and that is where they stayed all day long (roughly 8 hours). The P12R test routine pretty much goes through red, green, and blue, with very little white thrown in... so I don't doubt that those temperature values would rise up, but that is pretty much in line with what I expected. 34C is not that hot which is why I didn't feel any hotness with my finger thermometer test in the garage over the weekend.

Anyhoo, just another data point... I wish I could bring the thermocouple setup home and measure it after letting the string sit on white for an hour... but even that would not be a realistic test because who uses white for an hour in their display? I sure as heck don't... mostly colors...
 

fasteddy

I have C.L.A.P
Global moderator
Joined
Apr 26, 2010
Messages
6,648
Location
Albion Park NSW
The info you provided now proves that these 12vdc LEDs are running closer to the 20mA than the 10mA and the reason you are seeing such high current readings, Rays specs are incorrect for that product and should in fact be approx 0.72 watts each node instead of the rated 30mA

So not all 12vdc strings are alike and is something that people must be aware of. The 12vdc you have are very inneficient wasting more than 1/2 its energy in just heat dissapated through the resistor.

It seems like the square type you showed can suffer from this because your not the only one who has seen similar results using the square pixel. I havent come across this yet with the normal bullet 8mm type pixel strings.

In the end this design offers no real benifits because it will still suffer from voltage drop due to the higher levels of current it requires to run the pixels.
 

nutz4lights

Full time elf
Joined
Dec 12, 2012
Messages
305
Location
Melbourne, Florida
Turns out... I have a major correction to make. What I have been calling the 20cm spaced (8") pixel strings above are actually 25cm spaced (10") pixel spacing... UGGGHHHHH.... I don't remember changing that in my quote session with Ray Wu but I must have because it is in the quote... I think I was very nervous about how the twig tree would wrap with those... it turns out, the largest branch row to branch row spacing is only 17.5cm (7") and the majority of them are 15cm (6") or less. I think on my next batch of lights for the remaining 11 trees, now that I know what the voltage drop will do, I will order 15cm (6") spacing and should experience voltage drop and current draw much closer to the 10cm (4")...
 
Top