WS2801 LED 25meter

Other Hardware LEDs and Multiplexing
1

Hi guys,

I'm doing a small project at home with 5meter LED strips ws2801 & 5050smd / 60 per meter. The total length of the project would be 25 meter so 5 strips in total and thus 1500 LEDs. The controller would be a raspberry in my case (sorry for that:)) using it's SPI. The idea is to use the 74AHCT126 for level shifting. The data & clk would be connected from one strip to another so from data-bus point of view it would be one large 25 meter strip. For power however I would inject roughly every 5 meter. The main goal is actually to have constant and dimmable light. The fact that every LED is 'addressable' would be an extra. So I'm not after anything fast but it should at least be able "to snake". For ex, the effects shown here WS2801 Raspberry - YouTube is something that should at least be possible.

Not sure if this is the right place to ask but I was hoping maybe someone did similar projects with said components? My question is for most if this will work; is a 25m data-line not to long in terms of interference etc? Thanks!

2

You may look into FastLED and a ESP8266 this gives you so many more options
as well as on a ESP32 you can get direct webaccess to controll from a Smartphone

the Raspberry is not the best option on your guess

3

The data line isn't 25 meters long, the signal gets re-generated by each WS2801 chip so the longest data line will usually be the connection form the control board to the first WS2801.

One change I would recommend would be to connect the power and ground more often than every 5 meters. At full brightness WHITE an RGB LED will need around 60mA, 5 meters (300) would need 18 amps, that is a lot of current for the small conductors on the LED strips. In your particular application, it might be better to use 12 volt strips, that will reduce the current requirements to about 1/3 of what is needed on a 5 volt strip, with the disadvantage that the LEDs can only be controlled in groups of 3.

4

bmbbsr:
You may look into FastLED and a ESP8266 this gives you so many more options
as well as on a ESP32 you can get direct webaccess to controll from a Smartphone

the Raspberry is not the best option on your guess

Thank for the tip, but the raspberry has more options wrt programming it (read: easier). Afaik the SPI functionality is comparable with that of an Arduino or ESP. Maybe it's missing some "realtimeness" but for the application that I'm building this is not required. So the question is not perse about the raspberry it was more on the other components and general setup.

5

david_2018:
The data line isn't 25 meters long, the signal gets re-generated by each WS2801 chip so the longest data line will usually be the connection form the control board to the first WS2801.

True. So maybe loss of signal quality won't be an issue indeed. But what about transmission delay? Suppose I turn all leds on at the same time? Afaik the last led needs to "wait" until all other leds received there value. So, which frequency do I need to run at minimum to have the delay between first and last led (lets say the last 20 leds) as small as possible? Will such a delay be noticeable with the human eye? or are 1500 leds not that "impressive" and will any frequency do?

david_2018:
One change I would recommend would be to connect the power and ground more often than every 5 meters. At full brightness WHITE an RGB LED will need around 60mA, 5 meters (300) would need 18 amps, that is a lot of current for the small conductors on the LED strips. In your particular application, it might be better to use 12 volt strips, that will reduce the current requirements to about 1/3 of what is needed on a 5 volt strip, with the disadvantage that the LEDs can only be controlled in groups of 3.

With 5v the current is really getting high indeed, that was my first concern too. However, I was thinking that if they sell a 5meter led strip, it must be possible to operate it at 100% no? To be on the safe side I could inject more frequently but then I would need to interrupt the strip which will probably create 'dark spots'. The idea was to power each section either at the beginning/end or in corners. In corners the next section would be powered from the other side, so in a corner only 1 section gets powered, so the other section can "overlap" eliminating dark spots. Also, while it should be able to operate at 100% it probably will never be used this way; it's meant to be for terrace lighting. As its direct-lighting it's only goal is to give a bit of light just to give a bit of ambiance :slight_smile:

I also searched alternatives but there seems to be no (of the shelve!) 12v or 24v alternative that offers equivalent functionality: each LED addressable that is. There are also "neon led tubes", but the led strips fits nicely into a special "ground" led profile, I can make 90° corners, the strips are ok price wise, they can be cut each 15cm or so, and it's also easily accessible and broken sections can easily be replaced as well...

6

JohnSPI:
But what about transmission delay? Suppose I turn all leds on at the same time? AFAIK the last LED needs to "wait" until all other LEDs received there value.

They do indeed. The thing is that all LEDs needs to "wait" until the data stops for a set delay, so they all update simultaneously.

JohnSPI:
I was thinking that if they sell a 5meter led strip, it must be possible to operate it at 100% no?

Whatever makes you think one thing follows the other?

These strips are primarily intended for chasing displays, not for illumination. And the strip length is primarily going to be a manufacturing convenience. When you buy a spool of lace or ribbon, you are not expected to use all on one dress or wrap one gift!

JohnSPI:
To be on the safe side I could inject more frequently but then I would need to interrupt the strip which will probably create 'dark spots'.

How would it do that? If you do not inject power every metre or so you are likely to get dark (or in fact, "hot" red) areas.

JohnSPI:
Also, while it should be able to operate at 100% it probably will never be used this way; it's meant to be for terrace lighting. As its direct-lighting it's only goal is to give a bit of light just to give a bit of ambience :slight_smile:

One problem is that dimming is performed by PWM, If the PWM happens to synchronise (and it may do this on and off over time) then the current draw when the PWM cycles coincide will approach the maximum. Guess what happens then?

JohnSPI:
I also searched alternatives but there seems to be no (of the shelve!) 12v or 24v alternative that offers equivalent functionality: each LED addressable that is.

Not in strips, but in these apparently.

The thing is, they are somewhat useless. With only one LED per colour, the WS2811 is still driving the same current but absorbing the difference between the LED voltage and 12 V instead of 5 V. So you are using exactly the same current, but doing it at 12 V.

Admittedly, you will see much less "red shift" because the constant current control will then tolerate a loss of almost seven Volts rather than 2 Volts. Whether the wiring will tolerate it is another thing. :astonished: