12v or 5v - Current draws compared

[AussiePhil]

This is not quite a like for like but it is head for head and a comparison of two 100cnt strings
Parameters
For the record we capture the LED current and the Pixel current as two separate numbers
All current numbers unless stated are for 100% white.

12v Pixels

• Unregulated, resistive bullet pixels
• 20CWG wiring
• 2811 protocol
• these were a simple purchase of pixels from a pixel supplier from China
• testing revealed these are 27mA per LED
• 29.8mA per pixel

5v Pixels

• 2811 based square nodes
• 18CWG wiring
• These came from Ray Wu in Nov
• testing had 47mA per LED
• 47.8mA per pixel

Both strings feed from each end with 1M pi leads to minimise any voltage drops for PI leads

So first measurement was to check current when just powered up and when blacked out.
This is known as the quiescent current.... and gives the power used when the lights are powered up and doing nothing

	5V	12V
Quiescent	70mA	200mA
Black out	80mA	280mA

So lets draw that out to a 2000 pixel megatree with 20 x 100cnt strings
At 5v
0.08 x 20 = 1.6A x5v = 8W
at 12v
.28 x 20 = 5.6A = 67W

So what Current do we see for 100 pixels at 100%

	5v	12v
Amps - 100 cnt draw - 100%	4.24	2.87
Power in watts	21.2	34.44

So we see less current on the wires which helps with voltage drop meaning we do have longer runs between injection points
BUT
we also see 63% greater power consumption for the same number of pixels

Lets again use the 2000 pixel megatree with 20 x 100 cnt strings

12v
20 x 34w = 680w.
5v
20 x 21w = 420w

Now a couple of observations

• The 12v resistive pixels got warmer but not where you may suspect, see the thermal image below but it appears to be the LED itself
• Using a light meter the 5v pixels were brighter at the same % level
• Reducing the light reading on the 5v pixel to match the the 12v pixel seen the 100cnt current drop to 3.2 Amps and the drive % to 73%
• This would see the power required for the tree drop to 306w
• This actually fits with an old statement that 12v pixels need twice the power

Video of the ramp from 0 to 100%

View: https://vimeo.com/670152673


Open to questions and further test requests
Cheers
Phil

 

[ezellner]

To me, this is important especially for those just trying to decide what type of pixels to purchase. My home is almost 90 years old. Exterior power outlets were not a thing back then. It is extremely difficult to run a new line from the breaker box due to slab and crawl space construction. I installed one 120 V 20 Amp circuit to power my entire show. The show was designed around this limit which demanded use of 5 V pixels to get all the props I planned. The amount of extra effort for power injection was trivial in the overall scope of things. Thanks for the all the time you put into the comparison, Phil.

 

[TerryK]

... Now a couple of observations

• The 12v resistive pixels got warmer but not where you may suspect, see the thermal image below but it appears to be the LED itself
• Using a light meter the 5v pixels were brighter at the same % level
• Reducing the light reading on the 5v pixel to match the the 12v pixel seen the 100cnt current drop to 3.2 Amps and the drive % to 73%
• This would see the power required for the tree drop to 306w
• This actually fits with an old statement that 12v pixels need twice the power

...

First, what type of light meter? Are you taking an actual measurement or just making a relative comparison?

"light meter the 5v pixels were brighter at the same % level" Expected in my thinking as you were using 12 Volt Resistor pixels. The internal pixel arrangement of components and their values change the 'rules' a bit.

"This actually fits with an old statement that 12v pixels need twice the power" Politely, I do see this statement as being very accurate in my opinion. Rather power being a calculated value would depend upon the voltage drop from the source to the level required/needed by the internal components of the pixel and current draw at the component voltage. So, a ratio of 5 to 12, approximately 2.4 times as much. Baring in mind that voltage drops along the length of a string, it is also somewhat a 'moving target'.

 

[MichaelF5]

Brilliant thankyou!

do you have more pixels to test the limits of injection points?

 

[AussiePhil]

First, what type of light meter? Are you taking an actual measurement or just making a relative comparison?

It was an actual electronic Lux meter, not calibrated so the values are good for relative comparison and general illumination values. I tried to control the readings and conduct them in a consistent manner. For the record I cut a short length of 40mm pvc that i used to block out external light, the pixels were positioned directly up with the Leds at the same distance from the reading head, the head neatly covers the top of the pipe completing the sealed setup.
It's not perfect but is far better than just a visual look by eye.
lets say the 12 measured 270 lux and the 5v was 330, i then would down the 5v brightness/intesity in vixen to 73% to get the same 270 reading.

"light meter the 5v pixels were brighter at the same % level" Expected in my thinking as you were using 12 Volt Resistor pixels. The internal pixel arrangement of components and their values change the 'rules' a bit.

both styles use 2811 led controller but the 12v pixels have components that are limiting the max white current level to 27mA, the 5v ones were 42mA
the 2811 data sheet shows 18.5mA at 5v, this seems to fall off as Vdd drops over a string ... future test to confirm.
All of this is variables and actually can be estimated based on the loop resistance and the input currents to the circuit.

"This actually fits with an old statement that 12v pixels need twice the power" Politely, I do see this statement as being very accurate in my opinion. Rather power being a calculated value would depend upon the voltage drop from the source to the level required/needed by the internal components of the pixel and current draw at the component voltage. So, a ratio of 5 to 12, approximately 2.4 times as much. Baring in mind that voltage drops along the length of a string, it is also somewhat a 'moving target'.

As above the moving target can be calculated, the practical comes in around 2x with a bunch of variables in play it would seem.

Cheers
Phil

 

[MichaelF5]

that reminds me... It'd be interesting to see you push the limits of the Volt regs... even just on a single pixel, from vague recollection they were rated for max volt 25 or 30?

 

[AussiePhil]

Brilliant thankyou!

do you have more pixels to test the limits of injection points?

I don't have more 12v pixels at this time and unlikely to buy more 2811 12v ones. The limits though can be fairly accurately estimated based on the minimum operating voltage both to display white and to also get consistent intensity levels.
I did do an improptu test to confirm minimum operating voltage using the electronic load to simulate additional pixels.
The pixels stopped being white around 3v which is around the nominal Vf for a blue LED. As the Vf for blue can be as high as 3.4v a safe Min voltage would be 3.5v.
What I observed fits with the guideline "no more than 100 pixels to power" for 12v. for 18cwg this may well stretch to 150 pixels and of course if you limiting the intensity levels then it will be higher.

Cheers

 

[AussiePhil]

Update Info:
Counted the strands and measured the strand size
Took multiple readings with the digital calipers and got around 0.14mm diameter
There were 21 strands in the bundle,
This gives a calculated real AWG of 22, the measured resistance equates to a 24AWG wire
When i count i bundle to 10 strands each so 21 strands in the image



I've spotted a couple vendors now start to quote strands and strand size in listings

 

[TerryK]

What is the wire type in the graphic? Obviously not bare copper. Tinned copper? Aluminum?

If vendors are beginning to list strands and strand size then wire type is just as important.

 

[AussiePhil]

What is the wire type in the graphic? Obviously not bare copper. Tinned copper? Aluminum?

I can't tell by looking at it but there is no copper colour in the cut ends and it measured resistance is close to the published numbers for aluminium wire of that sizing, so expect it to be aluminium.

If vendors are beginning to list strands and strand size then wire type is just as important.

The vendor I bought some M16 pigtails from was listing type (copper) and mmsq at least, another also listed the strands.

We should be asking this info from the vendors as I agree it's important.

 

[scamper]

these tests just prove the theory. As all led's are the same voltage essentially, it is natural that a 12v string is going to use more power as you need to dump the excess voltage which cannot be done without some kind of power draw /loss

 

[OldMarty]

For the record we capture the LED current and the Pixel current as two separate numbers
All current numbers unless stated are for 100% white.

12v Pixels

• Unregulated, resistive bullet pixels
• 20CWG wiring
• 2811 protocol
• these were a simple purchase of pixels from a pixel supplier from China
• testing revealed these are 27mA per LED
• 29.8mA per pixel - Incorrect, 3x (R.G.B) 27mA = 81mA per PIXEL (at full white)

5v Pixels

• 2811 based square nodes
• 18CWG wiring
• These came from Ray Wu in Nov
• testing had 47mA per LED
• 47.8mA per pixel - Incorrect, 3x (R.G.B) 47mA = 141mA per PIXEL (at full white)

Which means all your calculations thereafter are all the wrong figures.....

Just saying...

 

[merryoncherry]

@OldMarty - No. ~20mA per channel is top of the mark.

There are new 12V designs that are much better, remains to be seen what the pace of adoption will be.

 

[Skymaster]

@OldMarty - No. ~20mA per channel is top of the mark.

This is correct.

The WS2811 data sheet previously specified 18.5mA per colour for a total of 55.5mA per pixel (plus quiescent current) and the later revisions dropped this to 16.5mA per colour for a total of 49.5mA (+QC) per pixel.

WS2811 chips are constant current drivers so are designed to supply exactly this current to each LED segment, irrespective of the voltage drop across the particular colour LED segment.

12V Resistor WS2811 pixels generally have resistors that are too high to even reach these constant current values, which is reflected in the testing above.

Dimming is achieved using PWM where the LED is full on then full off for different time periods. The longer on vs off, the brighter it is. This means the instantaneous current is the full current of the LED but the average is lower.