It's kind of fun to see the desk get more and more crowded as the project goes on.

@nigeljames6017

Жыл бұрын

A Canadian rats nest ?

My suggestion - Use the WLED firmware on an ESP8266 (or ESP32). I easily control strings over 1,500 bulbs long, and use multiple lines of strings out from my controller chips. It works very well, even with four different strings of lights running different patterns simultaneously. And that's on trees with 1,000 lights each (big trees). It's really easy to do, and I also use 74hct125 chips as drivers. Cheap and easy. Loved how crowded the bench got. Forgot to drink the beer!

Thank you!!! I think this is the first time someone actually mada an effort to elaborate how the information is passed to the LEDs to give us viewers an understanding of how we can "abuse" this for advanced use cases.

@pileofstuff

Жыл бұрын

My explanation glossed over a lot of the finer details, but I'm glad it was useful as an introduction to the topic.

If it did mess up, you could have a single transistor driver using the data pin and the 5V for it. Since it did not, look at sandwich boxes and the cheap ' tupperware ' types for outdoor projects. Another channel I follow uses them alot with great success when they can be air tight ( and thus water tight ).

I thought Neopixel was the name for Adafruit's boards (and associated library), but the LEDs themselves are still 2811/12/13. By the way, max string length at 60hz is 1024 LEDs :)

No LEDs were harmed in the creation of this video ! 😄

“The only way to discover the limits of the possible is to go beyond them into the impossible.” Arthur C. Clarke aka Mr Stuff

More LEDs? Another! MORE! I'm impressed at how effective a method this is. Thank you for confirming that it works.

In my limited electronics experience, the addressable leds are just glorified shift registers, and yeah nothing should break if done parallel... the driving circuitry might not have enough fanout power tho. I remember seeing 20 somewhere, but AVR chips are always conservative with their numbers (I assume you are using an AVR based controller) EDIT: Should've watched more before commenting, you did say you had an Arduino Nano for the test bench (as I assumed)

Definitely good to know, and also not take everything you hear about hobbie electronics at face value. Thanks!

Now this what I like tinkering with electronics and beer!

Funny, I tried the exact same thing a few weeks ago and was pleasantly surprised to see it worked

If you were worried about overloading the data output, i suppose you could have used a buffer Maybe for next year, run a 3 wire bus up the trunk with thick wire for current, adding a connector every so often, then just plug in strings to radiate If a string has a problem, it would be easy to unplug and repair ;) I can understand not paralleling for something like a scrolling message (unless you want duplicates), but this is fine, oh wait, you could cover your desk with scrolling signs :D

What a great experiment that will come in handy because I was looking for some sort of easy lighting for my bar and pass thru to the dining room. I'd like to give this a shot and see how long it holds out. I had to listen again to the programing portion of the video and I think I got what I needed. Thanks Kent, for the fun little useful project.

holy crap that wide strip of leds, if it would be tighter (no gaps) it could make easy large led screens, self-lit

There is no addressing in WS2812's they just strip the first 24bits off the data stream and output the remaining data bits. Looking at the data sheet the input current is +/- 1uA so i can’t see a problem with putting them in parallel. Even if the Arduino /ESP8266 could only output 5mA it still could power quite a few. The only issue i could think of is the capacitance of the wiring slowing down the edge, but again this depends on the output impedance of whatever is driving the data line. The only draw back is that all the parallel LED's will display the same. Can i recommend that folk should read the data sheets themselves for devices they are going to use, they are there, so you can work out if it will work or not.

I've done plenty of testing with ws2812b and ws2811. Neither of those need level shifters on a 3.3 volt microcontroller like the almighty esp32. Just make sure to get a strip made in 2020s. The older revision 1.3 was prone to glitching but the new revision works beautifully on 3.3 volt signals. You can tell which version you have by observing the pwm frequency. If the led flickers at 400 Hz it's rev 1.3. If it flickers at 2 kHz it's rev 1.4. You can tell them apart by moving your eyes left and right vigorously.

I also read on the internet that WS2812 LEDs draw about 4x as much current as they actually do. I think most people writing about them don't have a clue (or their experience is from fakes/clones, not genuine WorldSemi chips). I've certainly run them in parallel before with no issues.

@SpeccyMan

Жыл бұрын

You read it on the internet, eh? 🤔

WHADHAFAK ? HAHAHA . What a contraption ! Nice ! Smokey !

WOW, all my life I was into electronics and made a very successful career out of it. And I can tell you, I never knew there were such "addressable LEDs" before seeing this. WOW, being retired now sure stops the latest knowledge and knowing about such parts... Thumbs Up!

@pileofstuff

Жыл бұрын

I only discovered the a few years ago. They're a lot easier to deal with than trying to control a discrete matrix

@chuckholmes2075

Жыл бұрын

really? what rock were you hiding under?

I only get reliable results with 3.3v controllers and "neopixels" with fast level shifters (3.3->5v) and following "neopixel uberguide" from Adafruit to the letter. So this make me thinking -- with proper fast buffer/Schmitt triggers ICs sky is a limit for paralleling strings of neopixels.

To my mind the only thing that should matter is the trigger voltage of the input pin on the address IC. Stack as many in parallel as you want as long as you have enough tickle on the address/data pin to flip it.

@pileofstuff

Жыл бұрын

the capacitance of the cabling could impact the data as well (softening the leading edge of the square wave), if it gets too far out of whack. And if the input impedance was low enough it has the potential to drag the voltage down too low.

Definetely time to shift into Christmas mode/mood. Great idea of these in parrallel.

Kent, you always seem to amaze me, when I think nah it’s not gonna happen, you seem to make it happen, great job mate, keep up the great work mate 🤙🏼🇦🇺

Thank You! Very useful video showing what is possible ‼️👍😊

Thank you! It was an interesting experiment, and good for a laugh on a dreary Friday.

Controllers vary greatly in what they can *source* to the first pixel. A related, but different test is to fan out the data output from a WS2812 to a bunch of parallel data inputs. There's at least one commercial addressable LED strip display kit on the market that feeds the data output from controller to a 'stem' strip whose output feeds to a hub of 6 or 8 strips with parallel inputs that radiate out like a flower or an exploding firework, if you want to look for some kind of validation from manufacturers. Or lots of us could have told you 8 or even 12 parallel inputs is unlikely to be a problem with such short data and power connections. But there are so many interacting factors that can contribute to a real world setup where the data integrity breaks down somewhere, and your test setup is practically a best case environment in comparison. There's NO question that sometimes people run into a problem with glitching LEDs somewhere in a large setup at first light, and one or more of a dozen different 'solutions' ends up making the problem go away.

If you use a logic level shifter and split the input and then run the 5 volt output to each You shouldn't get an issue with loss of signal. Not quite sure if that's how that is supposed to be used lol.

Cool, you've disproved that theory. :)

I'm sure you must know about it but, just in case... Have you played with WLED yet? - I think you'll like it.

Short wires and 5V supply no issue, otherwise all you need is any buffer, like a 74 HC4050 to give you 6 buffered outputs to use. Long wires would cause issues, but under 10 feet of wire total to the first led no issue, length after that so long as each string works the rest do not care, just so long as power supply can handle current draw.

@pileofstuff

Жыл бұрын

I was also concerned that the cable length would be an issue when I took it outside and deployed it on the tree. But it worked, so I'm happy. I do have a big box of "various Lengths of wire" ... I suppose I could repeat the experiment with random mis-matched wires.

@SeanBZA

Жыл бұрын

@@pileofstuff You really have to go far to make it stop, but before that you would have the odd corrupted colour, and the odd wrong flashing, before it stops working.

yep they say you shouldnt but ive done this many times. multiple sets in parallel and they do seem to stay addressed. if you look at data sheet it says that each chip has a signal repeater or amplifier. knowing this is really useful and i do wonder if i can make an led tree/branch.

Ah, good ol' NeoPixels. (Adafruit sells the WS2812 LEDs as NeoPixels.)

nice good fun led, lots of mess (still cleaner than my desk) and a christmas tree excellant 🙂

what I do in this case, I use diff line transcievers like the sn75176 and the controller stays inside the house, and the led groups are just sitting on the bus with their own diff line receivers. Its pretty much like dmx but the actual data is ws2812 data stream instead.

I run two strips in parallel for my stair lights. 236 LEDs on the long side. Works great for me. They are mostly off and run at a relatively low brightness. I run them from an ESP32 D1 Mini running WLED.

they actually sell repeater/spliters for this i think. but a cool thing to do that is not just paralleling would be to starburst off of one led node in the main string a few pixels so it forms a partial star or snowflake , so it acts like a ornament. i not sure i will be doing all my outdoor pixel stuff this year not alot of time this year , and not sure if my 16foot x 4 foot 100x20 WS2812 grid held up in storage.

@monad_tcp

Жыл бұрын

its just a small fan-out problem, just but a buffer/latch chip in the data line to increase the possible sink current

So my WiFi Pixel design uses a 2n7000 mosfet to implement a level shifter for the 3v ESP module to drive the pixel signals at 5v like they want. I have to assume the D1 has a similar level shifter as the pixels really do not like 3v signals all that well. It may work OK 3v with short distance to the first pixel and then after that, the signal is "regenerated" (amplified?) by the WS2811 controller to proper 5v. I would have loved to see if the WiFi Pixel board could drive all those pixels- it is something I've never tried. Your next step is to use the FREE Xlights to run the pixels on the tree for more patterns...

@waylandsmith

Жыл бұрын

I haven't seen a level shifter on an ESP32 board before (but that would be handy!) and it doesn't look like the D1 does either. I just threw a cheap SN54AHCT125 on my project and that worked great, but I haven't tried to run multiple stands in parallel.

What you are doing is called "fan out" the specs on your microcontroller will tell you how many TTL loads (or min impedance, max current, max capacitance) any output can drive. Then you look up the specs on your LEDs (you only need to consider the first LED on each strip for the data line). The WS2812B LEDs have a typical TTL load and most controllers can handle about a 10 TTL device fan out before you overstress/blow out your microcontroller pin.

4:36 fan out in CMOS for a 50mA pin at 5V, probably like 50 in parallel ?

You could build your own jumbotron.

hey what is that green ruler like square thing on your desk that you build the thing on top of

@pileofstuff

Жыл бұрын

It's called a"self healing cutting mat". They're commonly available at craft and sewing stores. Or(of course) online (pileofstuff.ca/r/a6mea)

That is SO cool !! how did you get Santa to come and do the wiring ??......cheese.

Brilliant!

The chip is the black thing inside the led itself, the capacitors outside are actually useless. I received few boards with some capacitors desoldered or completely missing and it still works.

I would be interested to see if that still worked with a data signal going through 15+ feet of wire.

@pileofstuff

Жыл бұрын

There's about 20 feet between the outdoor controller and the tree.

you can use buffer IC to run parallel LEDS.

Running WS2812 from a 3.3V controller can go wrong, according to specs they need 3.5V on their Din pin to recognize it as a "High" signal. It may work (as evidenced in your video), but it can also cause all sort of strangen behaviour. So, to be on the safe side, you should add a level shifter in between to get the 3.3V data signals up to 5V levels. (Or, alternatively, drive the LEDs from a 4.5V supply.)

@szaszafaja

Жыл бұрын

simpler hack is to use an extra first led and lift its ground up with a diode.

@WoodmanFFM

Жыл бұрын

@@szaszafaja Lifting the ground won't work - the LED would see even less amplitude on the data pin than before. If you want to use a diode you need to lower the Vcc of the first LED with it. Or just use an NPN transistor and two resistors to cleanly and reliably shift the signal into the required range. Cheaper than "sacrificing" an LED and uses less power as well...

12:30 *Liquid electrical tape!* It's exactly what you need to seal joins like this.

Sending dmx(e131) over wifi, Doesn't work well if you want fast framerates and large # of leds. I've attempted running an small-mid size xmas light show over wireless to my controller and it didn't work well. Once I switched to wired ethernet, it worked without issues.

Not sure about anyone else, but I have this overwhelming urge to go solder something. Something that blinks. 💡💡💡

@pileofstuff

Жыл бұрын

do it!

"Gaming ARGB" PC components do this all the time, they connect 4-5 in parallel, manufacturer provide splitter cable.

there is no addressing and no "stripping" of anything, it's a serial stream of bits. In the WS2812A/B case (as well as many others), the first chip will receive 24 bits and if it receives another 24 bits before the latch period, will simply pass along it's current 24 bits to the data out and receive the next incoming 24 bits. After the latch period, each chip "applies" the received values as pwm signals to each of it's r,g,b "channels" / leds. You cannot "address" any of the leds, you can only spit out as many bits as you want leds on.

I heard that after about 600-800 LEDs you could get input lag and thats why they suggest only 600-800 LEDs per data channel. So with your 14 strips in parallel with only 16LEDs per strip, you are no where near close to the max number of LEDs. I would love to see 100 LEDs per strip and see how many strips you could do in parallel.

@pileofstuff

Жыл бұрын

That would be a fun experiment if I could get someone to fund it.

maybe the world is not telling the truth about what we can do great vid love it,.//.,

when i was playiong with them i created 4 instenses (if thats the right expression) in the code so had 4 outputs, but i like the way u did it so will have to try that :)

@pileofstuff

Жыл бұрын

Ive seen that done as well, just to keep the code easier to understand

Parallel strings with only 8 LEDs used in each string is one thing, but can it maintain data integrity with longer strings?

@pileofstuff

Жыл бұрын

Each LED repeats the data stream, so there isn't any degradation if it's 5 0r 500 LEDs.(as long as the power wiring is capable of passing the cu rent for every LED)

@danman32

Жыл бұрын

@@pileofstuff there can be degeneration down the line. All it takes is a few lost or late bits, even if the LEDs are close together.

@waylandsmith

Жыл бұрын

@@danman32 I'm curious, is there a reason why bits would be more likely to get lost later in the packet than near the beginning?

The address space is infinite because it's based on a sequence of data, not an address

@pileofstuff

Жыл бұрын

yeah, the limiting factor is more the propagation time versus the datarate/frequency

try to do parallell with several 5m strips, that's when you'll have a problem.

They actually don't have a "address" in theory you could connect a infinite long strip but then the update rate of the last would be infinite slow

Instad of taping when you cant heatshrink things use liquid electrical tape. Paint on and its waterproof

I know that voice.. enjoy your other channel but this is nice.. " hands doing stuff with commentary"

@pileofstuff

Жыл бұрын

Other channel?

Eh, there are no addresses in the protocol? The protocol is a preamble, and then shift in N bits. It's a large shift register. Once you stop shifting it will be latched.

You can tell the second one that was added is not as bright is the back one

I think you have to take a good look at the data sheet again because the WS2812 does not work like you think they work.

That poor Nano should be toast. Is it really powering those LEDs?

@pileofstuff

Жыл бұрын

I've abused my Nanos way worse than this. They take quiet a lot to kill.

@tinygriffy

Жыл бұрын

The power lines of the LEDs are connected directly to the bench power supply.. just the data lines are in parallel on a pin on the ESP. It is kind of hard to see given the monstrosity of this experiment (:P) .. maybe at 11:00 if you look at the breadboard top left you can see it..

@pete3897

Жыл бұрын

No

Mate can u please teach the whole code 😉 its c++ right?

@pileofstuff

Жыл бұрын

Yes, Arduino is generally programmed in a slightly modified C++ There is a ton of good beginner level tutorials on the Arduino site docs.arduino.cc/tutorials/

@bishalscreation

Жыл бұрын

@@pileofstuff ok thanks 🙏💐

I'm curious about the current being pulled with all those parallel LEDs.

@tinygriffy

Жыл бұрын

the power lines of each and every ws2812 pixel is always in parallel ! (on a 5V strip) PS: and it is roughly 55mA per pixel (datasheet) for the brightest white.. so 55 times the number of LEDs is the maximum current in milliamps

@jyvben1520

Жыл бұрын

he has powerline connections at several points in the string, so it does not depend on the led string power lines too much.

@kyleallred984

Жыл бұрын

@@jyvben1520 I was referring to the breadboard testing. Not the lights on the tree.

@pileofstuff

Жыл бұрын

It peaked around 350ma. If I set them to white, it would bave been about 3x that amount. And if i had more than 1 LED lit at a time (per string) it would have been a *lot* higher.

you can in theory solve the bogging down of the signals by building a driver and power supply.

@pileofstuff

Жыл бұрын

If my experiment showed any problems, that was my plan.

All the parallel issues can be solved with a buffer, dunno where you heard all that but it's just not applicable if you know what you're doing

Literally every (non-USB) ARGB fan hub sold for gaming PCs connects the data lines in parallel, so every fan gets the same pattern just like your first experiment. Edit: They usually have data logic buffers to amplify the signal.

Do you know that what you are point to at 1:25 is just an alim decoupling capacitor ?? The controller is inside the LED it self. That's why it has 6 pins instead of 4 like a classical RGB 😛

@pileofstuff

Жыл бұрын

Yes, and I believe I mentioned that the controller is inside the LED package more than once during the video.

@skuizhopatt5318

Жыл бұрын

@@pileofstuffThankls for the answer ! Fair enough then :) But why did you point at it in the first place ?? Must admit I didn't watch all along, I know well enough the basics of that WS-like modules... But I beleive what you are "abusing" here, is more your alim than the MCU or LED strips... (I may have missed a point btw). I mean, as the data line draw virtually NO current, it's not very astounding you can put MUCH of them in parallel. It's not a critism, i'msure plenty of viewers will find it interesting ! Years ago when I started to play with MCU... YT existed already, but was not full of this kind of content, and that's cool, thanks for the newbies :) (Note to the newbies : go to IRC, though !) My point is you can have mislead a couple of them here...

@pileofstuff

Жыл бұрын

@@skuizhopatt5318 During editing sometimes the voice over and video can get a bit out of sync. What I was pointing at a second ago may not be exactly what I am talking about at that exact instant.

@skuizhopatt5318

Жыл бұрын

@@pileofstuff unsure if you should insisted, here, but... 🤷‍♂

Are you drinking Fernet?

Online electrical and programming forums can be very helpful, but i find they're always VERY by the book kind of crowd and they lack creativity actual implementation.

you probably have deuteranopia

@pileofstuff

Жыл бұрын

Yup. What did I mess up this time?

@csepes

Жыл бұрын

@@pileofstuff you called the brown wire green. People often don't know about it for a long time. A friend of mine learned about deuteranopia only at university. We once drank a light lager and he was imbued with its beautiful "green" color 😁

@pileofstuff

Жыл бұрын

@@csepes Ahh. After all these years, i should know better than to try and identify things by colour.

🤍

I've been using 2812'2 (ARGB's) for a good decade now and yes I have ran across a few issues with parallel circuits. most notably the cheap SP105e and other controllers from amazon. I've since graduated from Arduino's (too complex and limited) I'm using the PizelBlaze (From Ben Hencke) and have had no issues (much easier to program and upload programs to it - plus it's WiFi). there is a PWM multiplexer module that you can add to the signal and theoretically it can be boosted indefinitely. (never bought one or used one)