Buffer/line driver ...
 
Notifications
Clear all

Buffer/line driver - null pixel without pixel

347 Posts
90 Users
0 Reactions
197.7 K Views
RobG
Posts: 1255
 RobG
Topic starter
(@robg)
Noble Member
Joined: 14 years ago

Yes, there's one series resistor on the input and I am still undecided about the one on the output.
I was trying to find a perfect solution for this, but with so many different pixel types, wires, and almost complete lack of documentation from pixel manufacturers, it's almost impossible to do so.


Reply
Posts: 366
(@technician)
Reputable Member
Joined: 12 years ago

I too was frustrated about the lack of electrical data on most pixel chips. Due to this I started doing some of my own evaluation on some WS2811 strings. In general this is what I found.

The input appears to be CMOS expecting a 5 V signal, with a threshold near 2V. Some overvoltage clamping is provided for voltage blow -2 and above +7 on the data. Input impedance is very high relative to the wire impedance. Input current indicates an impedance >10K, which would enable a fairly wide fan out for fireworks bursts for example where 1 pixel string then drives 6 or more directly off the end so a skyrocket for example could run up a tree and then burst at the top without needing fancy sequencing for all the arms as they could be in parallel.

In deciding on the resistor for driving high speed data transmission lines, there is a 2 step process. First is the hardest and relates to the available drive current of the driver IC. What effective impedance is this current? If equal to or greater than the impedance of the data line, then NO resistor is desired. If the impedance is less than the wire impedance, then a resistor to make up the difference is used. For example the driver if effectively 50 Ohm, and the wire is 110 Ohm, then a resistor near 60 ohm is recommended in series.

The reverse is true on the far end. Because pixels are low current draw, they are effectively higher impedance than the wire. The ringing on data wire is the result of this mismatch.

Here is the engineering problem for those trying to terminate this line. It has a DC offset. It does not swing above and below ground like an AC signal. If you are running 5V pixels, the solution is a virtual ground with the offset to 2.5 V. To do this a pair of resistors is used, but the formula for parallel resistors is used for the load. For example to terminate a 110 ohm wire, use 2 220 ohm resistors in series. One to 5V and one to ground. This will kill the ringing, but there is associated attenuation with this, so knowing the pixel threshold, and if the input has a hysterisis (different levels for detecting a zero and another for detecting a one) is still an unknown so far for the pixels. Need more engineering time to reverse engineer the pixel input thresholds and signal conditioning. When I get more time, I do plan on doing some data level shifting to find the pixel input threshold levels for typical pixesl as well as minimum signal amplitude.


Reply
RobG
Posts: 1255
 RobG
Topic starter
(@robg)
Noble Member
Joined: 14 years ago

I did a quick test and it looks good.

My setup was:
24V power
pixel controller -> 3p pigtail -> 3-4-3 T splitter -> n sections of 10ft 3p extension cables -> 3-4-3 T splitter
10W flood at the first splitter, 20W flood at the second
I used 3, 6, and then 10 10ft extensions, for a total of 30ft, 60ft, and 100ft.
My second flood was working as expected with or without the driver (that was surprising,) so I had to use my scope to verify quality of the signal.


Reply
tucsonlights
Posts: 667
(@tucsonlights)
Honorable Member
Joined: 15 years ago

That's good news. Let's get this going.


Reply
wirekat
Posts: 23
(@wirekat)
Eminent Member
Joined: 19 years ago

I'm in for 20.


Reply
Aurbo99
Posts: 2804
Moderator
(@aurbo99)
Member
Joined: 19 years ago

I'm in for 10


Reply
Posts: 23
(@savageelf)
Eminent Member
Joined: 12 years ago

Ok, rookie question...

In simplest terms, it appears that this board is spliced into a 3-core extension (?). Is that correct?


Reply
Posts: 23
(@savageelf)
Eminent Member
Joined: 12 years ago

I would be in for 10...


Reply
wirekat
Posts: 23
(@wirekat)
Eminent Member
Joined: 19 years ago
savageelf wrote:Ok, rookie question...

In simplest terms, it appears that this board is spliced into a 3-core extension (?). Is that correct?

Yes, the positive and negative connect through while the data runs through the board (powered by the + & -)


Reply
Posts: 1344
(@mrsmoofy)
Noble Member
Joined: 12 years ago
RobG wrote:I did a quick test and it looks good.

My setup was:
24V power
pixel controller -> 3p pigtail -> 3-4-3 T splitter -> n sections of 10ft 3p extension cables -> 3-4-3 T splitter
10W flood at the first splitter, 20W flood at the second
I used 3, 6, and then 10 10ft extensions, for a total of 30ft, 60ft, and 100ft.
My second flood was working as expected with or without the driver (that was surprising,) so I had to use my scope to verify quality of the signal.

RobG has exceeded their stored private messages quota and cannot accept further messages until they clear some space.

🙂


Reply
Page 6 / 35
Share: