Null Pixels and why Wire Size does matter

davrus

Silent elf
Joined
Jan 9, 2014
Messages
168
Location
Whitianga
I wanted to test, to see when I would need a Null Pixel for my situation. It is commonly said that for distances greater than 5 metres, you will need a null pixel.


My conclusions are that the Ground cable size is very important. At a larger distance, the voltage drop caused the data signal "return" to fail, as the ground was being used mostly by pixel power return. Use a bigger cable for the ground, and suddenly you can increase the distance. At 6 metres using standard wires, the lights fail. If you double the ground cable, then 6 metres works fine. Use thicker cable for positive and ground, and then you can get to 9 metres, without a null pixel.



I used a basic test bed, and intended to vary the cable type and length to see what I would discover.
PixLite 16 (12V) -> pigtail -> Watt's Up meter -> test wire -> pigtail -> pixel string -> digital voltage meter


The pixel string for most of my tests was not a simple standard string. It is 5 x one metre string of 20 pixels, with pigtails on each end, connected together giving 100 pixels over about 6 metres. So my test bed had the signal flowing through 12 pigtails, in total.


They are all Ray Wu black 4 core pigtails [SIZE=78%](presumably 18 AWG)[/SIZE][/size][SIZE=78%].[/SIZE]
The wire is Ray Wu's 4 core RGBB cable (presumably 18 AWG).
I also used some 2 core 12V garden cable - 12 AWG.


First test, using 4 core RGBB - 6 metres - when white (255) the lights were flasing, not steady. Far end voltage was 11.3 V when all off, 6.9V when full white. Peak Amps 3.2A


Second test using 4 core RGBB for signal and positive - 6 m - plus 1 core 12AWG for ground, 8 m. Everything worked as normal. [/size][SIZE=78%]Far end voltage was 11.3 V when all off, 6.9V when full white. Peak amps 5.2A[/SIZE]


Third test, using a single core of the 4 core RGBB cable (for data) 9 metres, plus 2 cores of the 12AWG, 8m. Everything worked as normal. Far end voltage was 11.6V when all off, and 7.2V when full white. Peak amps = 5.4A


Fourth test, same as third but data cable was 6 metres. Far end voltage was the same, Peak Amps the same.


Fifth test using 4 core RGBB, with Black & Blue joined (ground), 6 metres, everything normal, Far end voltage 11.4V for off, and 5.54V for all white. Peak amps = 4.8 A


Last test using a "normal" pixel string ( 100 pixels on 10 metres), 4 core RGBB (not using blue), 6 metres, and as expected, lights flashing on full white.. I watched this one closely. As the white level increased, the pixels got brighter, up to about 188 (far voltage showing 5.7V). As the level continued to increase, more and more pixels started to flash. And as more pixels flashed, the end voltage increased so that at 255, the voltage was 8V, and the pixels were really flashing erratically.
 

David_AVD

Grandpa Elf
Community project designer
Generous elf
Joined
Jun 12, 2010
Messages
4,681
Location
Victoria Point (Brisbane)
Yep, as you've found voltage drop plays a huge role in keeping pixels happy. The ground wire doubly so due to the data needing a solid reference.
 

DrNeutron

Just starting in this crazy hobby
Joined
Oct 19, 2012
Messages
228
Location
Glace Bay
Would the same rules apply for WS2801 pixels that have a separate 4th wire for clock?
Great write up BTW Davrus, very informative.
Thanks.
 

David_AVD

Grandpa Elf
Community project designer
Generous elf
Joined
Jun 12, 2010
Messages
4,681
Location
Victoria Point (Brisbane)
In general the same applies to all pixel types. Clocked pixels can be slowed down however and this can help in marginal situations.
 

fasteddy

I have C.L.A.P
Global moderator
Joined
Apr 26, 2010
Messages
6,648
Location
Albion Park NSW
You generally know you have an issue when you start to see flickering when you use white of a combination of 2 colours due to the increased load on the shared common wire so always best to test in all situations including dimming, and mixed colours.

The other thing to note for Australian climates is that what may seem to work well on the bench in testing during winter may in fact have some issues when used in your display in summer due to the increased ambient temperature which in turn will increase the cable resistance. Also there is added electrical noise generated by your whole display running which could also affect your data signal.
 

gerry

Senior elf
Joined
Dec 19, 2012
Messages
855
Location
Surrey Hills,Melbourne
I recently saw a post/discussion (unfortunately too late for me this year) that said that if you have an extra wire , then use it to double the ground connection.

I assume that this means that if you were using a 4 wire 4/20 cable , to use two wires together for the ground connection (rather than cut and tape off the extra wire ?)

Is this correct in all cases ?
 

fasteddy

I have C.L.A.P
Global moderator
Joined
Apr 26, 2010
Messages
6,648
Location
Albion Park NSW
gerry said:
I recently saw a post/discussion (unfortunately too late for me this year) that said that if you have an extra wire , then use it to double the ground connection.

I assume that this means that if you were using a 4 wire 4/20 cable , to use two wires together for the ground connection (rather than cut and tape off the extra wire ?)

Is this correct in all cases ?

This is correct because by using the additional unused wire as a return ground you then 1/2 the resistance path for the return ground cable, so dont leave the 4th cable unused, it has a good use in helping reduce voltage drop.

So if using 4 core cable with 2811/2812b or any other 3 wire pixel then use the 4th wire as a shared ground wire
 

gerry

Senior elf
Joined
Dec 19, 2012
Messages
855
Location
Surrey Hills,Melbourne
Fasteddy said:
This is correct because by using the additional unused wire as a return common you then 1/2 the resistance path for the return common cable, so dont leave the 4th cable unused, it has a good use in helping reduce voltage drop.

So if using 4 core cable with 2811/2812b or any other 3 wire pixel then use the 4th wire as a shared common return

Thanks Eddy. I knew to double the wires each for the power injection runs (you told me to do that in my first post), but missed this item in the million odd posts that I read. Will do so henceforth.
 

logandc99

Dedicated elf
Joined
Jul 13, 2012
Messages
1,153
This is great info so thanks to all. I am about to make up some cables for my pixels using some 4 core cable and was wondering which of the wires to double up but now my question has been answered. :D
 
Top