Pixels in parallel yet still addressable

Notenoughlights

400,000+ twinkly lights
Joined
Oct 16, 2016
Messages
649
Location
Christchurch, New Zealand
The ws2811 chip can do both speeds (taking pin7 high puts it in slow mode)

The motivation always seems to be a cost saving exercise (remember the old 3 wire to to 2 wire lights) so whatever they have done would have to be cheaper than just running a return data wire or running them in a zigzag pattern.

Extra circuitry on the first pixels should be visible.
Can you take a close up photo of the first pixel on either the second or third string.

I don't believe the ws2811 is reprogrammable outside the factory and imagine how messy it would be from a QC point of view ensuring you solder a different 1st pixel for each string to shift or ignore the data of the the preceding string(s)

There's no obvious circutry on any of the pixels, they all look similar - all the same size physically etc. A photo won't show much as they're potted in white frosted resin and you can't see anything, and sorry I don't want to destroy the resin as I want to still use this light set.
 

Grozzy

Oh great, the idiot's back
Joined
Aug 16, 2018
Messages
104
There's no obvious circutry on any of the pixels, they all look similar - all the same size physically etc. A photo won't show much as they're potted in white frosted resin and you can't see anything, and sorry I don't want to destroy the resin as I want to still use this light set.
Yeah Nah don't be cutting into the resin.
If there was any extra circuitry it would be clearly visible with the pixel being bulkier than the others
 

Grozzy

Oh great, the idiot's back
Joined
Aug 16, 2018
Messages
104
Can you take a close up of the pixel anyway.
Do they look like a single 5050 profile size or 3210 rectangle.
 

Notenoughlights

400,000+ twinkly lights
Joined
Oct 16, 2016
Messages
649
Location
Christchurch, New Zealand
@Grozzy pictures as requested;
Pixel next to a DC barrel jack for size comparison, these must be the 3210 rectangles as they are way too small to be a 5050 formfactor
View: https://i.imgur.com/qBvdDlU.png


First pixel in line; top and bottom view
View: https://i.imgur.com/IcGQDTg.png
+
View: https://i.imgur.com/fsn4Qwv.png


First pixel in line from another string top and bottom view
View: https://i.imgur.com/bmaB5Gi.png
+
View: https://i.imgur.com/KYK0fcW.png


Second pixel down the line of the top 2, top and bottom view images
View: https://i.imgur.com/bLZpJxB.png
+ https://i.imgur.com/htQ6rAx.png

Second pixel down the line from the other string top and bottom view; https://i.imgur.com/IwbXoMX.png + https://i.imgur.com/GbRAzUm.png
 

Grozzy

Oh great, the idiot's back
Joined
Aug 16, 2018
Messages
104
So potentially something like this. But still a 2811 chip which doesn't explain how they get the other strings to be different. Anything that happens in the data stream would effect each string equally 😖
 

Notenoughlights

400,000+ twinkly lights
Joined
Oct 16, 2016
Messages
649
Location
Christchurch, New Zealand
So potentially something like this. But still a 2811 chip which doesn't explain how they get the other strings to be different. Anything that happens in the data stream would effect each string equally 😖
Exactly
It's as if string 1 takes pixel data 1-13
String 2 takes pixel data 14-26
String 3 27 - 39
and so on. I would expect them all to mirror each other being that only 3 wires come from the PCB at the start of each string.
There will no doubt be some of these returned to Bunnings with faulty strings that don't do what they are supposed to if the first pixel in each string is doing some sort of preprocessing magic. Part of me wanted to think DMX but the controller was fine driving standard ws2811 5v pixel nodes.

Here's another one doing the same trickery, this one has mobile control https://www.aliexpress.com/item/1005003251499675.html
 

Grozzy

Oh great, the idiot's back
Joined
Aug 16, 2018
Messages
104
So I came across a chip that seems to have some sort of built in test sequence.
I wonder if that's all the controller does.

Internal control patterns
GS8208 gets into RGB test mode right after power on. If the device does not receive the
external display data for a long time, the chip will go into the built-in display patterns
mode. Pattern on a total of six categories of 32 series, including integral color jump,
integral color gradient, meteors, waves, colorful gradient, color jump of water circulation.
It is 10 minutes about, and the image refresh frequency is 100hz.
Under the internal display mode, the first chip’s SDI/SDI2 needs to be connected to
GND.

 

Notenoughlights

400,000+ twinkly lights
Joined
Oct 16, 2016
Messages
649
Location
Christchurch, New Zealand
Doesn't sound right. I powered them on without any data feed and nothing happened. I assume "long time" is a few seconds, these were on for about 10 with no change. The controller doesn't just give a test sequence, it puts 195 pixels worth of data out (Maybe more, haven't tested this) (As tested with a controller running WLED attached) if I set 13 pixels to be on, the first string lights up, 14 pixels, the first string and the first pixel of the next string light up. Remembering these 2 strings share the same data feed directly from the controller and are effectively in parallel with each other.
 

Grozzy

Oh great, the idiot's back
Joined
Aug 16, 2018
Messages
104
Doesn't sound right. I powered them on without any data feed and nothing happened. I assume "long time" is a few seconds, these were on for about 10 with no change. The controller doesn't just give a test sequence, it puts 195 pixels worth of data out (Maybe more, haven't tested this) (As tested with a controller running WLED attached) if I set 13 pixels to be on, the first string lights up, 14 pixels, the first string and the first pixel of the next string light up. Remembering these 2 strings share the same data feed directly from the controller and are effectively in parallel with each other.
OK can you confirm I read that right.
Even with WLED pixel 14 lights up Pixel 1 on string 2.
The test mode hypothesis was that the controller was just taking Data In to Low or some combination of H/L to trigger the test sequence and excuse the chinglish cut and paste from the data sheet 🤣 but powering them on with no data should have done something if they were this type of chip.
 

Notenoughlights

400,000+ twinkly lights
Joined
Oct 16, 2016
Messages
649
Location
Christchurch, New Zealand
OK can you confirm I read that right.
Even with WLED pixel 14 lights up Pixel 1 on string 2. - Yes
The test mode hypothesis was that the controller was just taking Data In to Low or some combination of H/L to trigger the test sequence and excuse the chinglish cut and paste from the data sheet 🤣 but powering them on with no data should have done something if they were this type of chip. They were powered on with just + and - connected, I lifted the Data line off the PCB. Not even a flicker from the leds.

Text added in quote to answer questions - see quote.
 

AAH

I love blinky lights :)
Community project designer
Joined
Dec 27, 2010
Messages
4,190
Location
Eaglehawk
Is it possible to see what's up in the ring around the top? The simplest solution would be to have the smarts integrated up there rather than in each string.
 

Notenoughlights

400,000+ twinkly lights
Joined
Oct 16, 2016
Messages
649
Location
Christchurch, New Zealand
The ring has 3 seperated tracks, no components. All datas are connected, all positives are connected, all negatives are connected. Imagine twisting all the wires together, that's what's in there. I can take a photo when I am home, but there's nothing to see.
 
  • Like
Reactions: AAH

Notenoughlights

400,000+ twinkly lights
Joined
Oct 16, 2016
Messages
649
Location
Christchurch, New Zealand
That's what everyone with something to hide says ;)
Righto i'll get a photo when I get home :p Expect a round white PCB with 20 outputs all linked together. No electrical components top or bottom layer, a controller with 4 wires marked + - D L and a pair of wires going to a plug for the star as well as the 5v + - feed coming into the tree ( for some reason it doesn't go through the controller)
L and - go to the star as as they are controlled by a mosfet to switch the star off when the power button is pressed.
 
Top