Which Pixels?

[DaveRNZ]

So I've got my RPi, and a Hanson Pi Hat which arrived a week ago. I'm looking at Ray Wu's store on Ali Express, getting very confused by all the options. I'm wanting to build around 5x Candy Canes using some coreflute (or similar) with 100x pixels per cane. These will go up on the fence so be really visible.

I know that I want:

• Bullet Style
• Intelligent (in case I want to re-use one day)

What I'm not sure on is:

• Resistor vs Regulated? I don't think it matters for my application. In reality I have almost unlimited access to PC power supplies so I'm not worried about having to inject etc.
• Water rating? I'd probably go for IP68 because why not

Other than that, the only differences appear to be, wire colour, connector, distance between pixels. Is this correct?

 

[Skymaster]

• Resistor vs Regulated? I don't think it matters for my application. In reality I have almost unlimited access to PC power supplies so I'm not worried about having to inject etc.
• Water rating? I'd probably go for IP68 because why not
• 8mm vs 12mm? I thought 8mm was the standard but I'm struggling to find any

Resistor/Regulated - for 12V, both will do the same thing; Regulated will give you a bit more headroom when it comes to voltage drop. However, you're talking small runs and you're not afraid of PI, so 6 of 1, half a dozen of the other. Resistor will be slightly cheaper and have less things to go wring. If you're going 5V, then there is no resistor or regulator - as the pixels run on 5V natively.
Water Rating - IP68 is relatively standard. If it's "real" IP68, it means it's submersible. IP65 is good enough for really hard jets of water from all directions. But that's if you trust the Chinese vendors to build to the exact IP specs.
LEDS - Bullets are an 8mm LED in an sleeve/epoxy, for a total of 12mm diameter. One and the same in this case.

The wiring is important too - you'd want to get 18awg, rather than the standard 20/22awg - it's very minimal increase in cost, for less voltage drop.

 

[DaveRNZ]

Thanks for the prompt reply. This is what I was looking at, specifically because I can't see any reason to go with the more expensive ones from what I can tell. I need white to match the prop. https://www.aliexpress.com/item/100...store_pc_allProduct.8148356.52.69583c60gjvhoE

 

[Mark_M]

Resistor is fine in most cases.

I like expensive regulator for the colour accuracy, but no one else notices that!

 

[LA0]

I'm going to deviate slightly from the resistor vs regulated and say that I like 5v personally. I can run more pixels off of less power supplies, and 5v are more efficient with energy usage, and they run cool as a result; they are also cheaper.

The only thing you really need to get your head around is power injection, but it's pretty easy to understand, and once you know it you won't forget it.
As insane as it might sound.. I actually like doing power injection at this point lol.

I have some pixel strips that run on 12v, but if I were new and starting out, I would've used 5v for everything if I knew then what I do now.

IMHO.

 

[merryoncherry]

People who bought pixels 2 or more years ago are going to tell you something different than I will, as the designs and most common chips have changed!

If you go order 12V pixels today (pixels made this year or last year), odds are the batch of regulated pixels are going to be noticeably brighter and use at least twice as much power as the batch of resistor pixels. You'll run 100 regulated pixels per controller port without PI, and 200-300 resistor pixels per port without PI, so plan accordingly. (Of course, with PI, you can do all 500 on one port, still 40FPS, but you'll spend more time and money messing with this than the $150 for 500 pixels. Then again, maybe you like spending time on the hobby, and that'll be more fun for you.)

You will not notice any color differences with any of them, nor will you notice any difference at the start / end of the strings, because they have constant current drivers in them. If you care about color fidelity, you might turn down the red and blue slightly for a warmer feel, and set the "gamma" to 2.6.

I only have 1 data point on 5V pixels, without PI they pinked slightly at 100 pixels, but I just adjusted it in xLights and nobody noticed.

For comparison purposes:

 

[TerryK]

I look more at the distance between the pixels and the power source. The greater the distance the more voltage drop one sees generally which then makes the regulated pixels more advantageous. As already mentioned, regulated pixels are more immune to voltage drop. Also as mentioned, 5 Volt pixels are the most economical and power efficient. Next are the 12 Volt resistor followed by 12 Volt regulated pixels.

The different IP ratings are mostly seen with strip, not string. I'm sure there's some bullet pixels out there that are not IP68 but I've yet to see them. Wire color and connector are personal preference. xConnect seems to be the most popular or seems to be becoming the most popular.

Pixel spacing is pretty much standard at 4 inch although it seems different manufacturers measure that from slightly different points. Most coro props work fairly well with 4 inch pixel spacing. Some with a smaller spacing and some where a jump is necessary and one needs to implement a lengthening splice.

I might mention pixel types, WS2811 or GS8208. The GS8208 pixels are more efficient than the WS2811s but do not see much use from individuals. Also more expensive than 12 Volt regulated pixels and somewhat a bit more difficult to obtain.

 

[merryoncherry]

No disrespect @TerryK but this is all outdated. This was probably true of pixels of years past but it isn't true now. Maybe the GS8208 or another chip is the reason (I can't see any markings on the chip in the pixel). The fundamental assumption that resistors are more sensitive to voltage would appear to be wrong because of the constant current drivers (see test results below), and the regulators draw more current, so they drop the voltage more, which makes them far more sensitive. The only advantage to regulated pixels is that they are brighter. The brightness and color of today's resistor pixels is more uniform across more pixels running farther from the controller than for the regulated pixels.

"The greater the distance the more voltage drop one sees generally which then makes the regulated pixels more advantageous." This is false because the regulated pixels on the market use more power than the resistors, so they drop the voltage more at the same configuration. This is not hypothetical, I just tried it and took pictures for you :-D

I have for your comparison today:
Regulated 12V Bullets: Wired Watts regulators, delivered to me last month.
Old Resistor-style: 12V HolidayCoro strawberry bulbs, delivered to me last October. (I don't have any old bullets, these'll have to do.)
Modern Resistor: 12V Alling, delivered to me 2 weeks ago

This first picture is 100 regulators, 100 strawberry, and 200 resistors at 12ft from the controller . Works well, as one expects.


Actually, maybe NOT as one expects. You can't run 200 regulators or 200 strawberries even that close to the controller without yellowing. Very noticeable yellowing of 150 regulators.


Very noticeable yellowing for 150 strawberries also (just like it says in the ACL manual).


With 300 modern resistors, there is slight dimming but the color is actually pretty constant.


OK, so that proves that modern resistors exchange quantity for brightness, and are less susceptible to yellowing as the voltage drops, moreso than regulators.

But what about the claim that they're better when far from the controller? That was only 12 feet. I'll add a 20ft extension, first the regulators:


That didn't work. They flickered violently because they pulled the ground up too close to the data signal voltage (see Keith Westley's excellent video about this at VCS 2020). Adding a signal booster and they'll run at 32 feet, but they'll yellow.

I'm done discussing the regulators now. Running 100 close to the controller, without PI, is the max for full white.

The strawberries were quite usable at 32 feet. So I bumped it to 92 feet (this required a data signal booster mid span) and they do dim and yellow a little:

(Attachment limit hit... to be continued)

 

[merryoncherry]

OK... picking up where I left off, 100 resistor-style strawberries at 92 feet from the controller, no PI:


A closeup of that last bulb vs. the first:


OK, now for the most versatile pixel of the bunch, I'll put 200 of the modern resistor bullet pixels out at 92 feet from the controller, and those will be next to 100 at 12 feet from the controller, for a real serious comparison.


They are a smidge dimmer, but the color is good. Keep in mind that there are now 150 feet of wire between the controller and my last pixel (92 feet, plus 200x4" spacing is another 60 feet):


I tried HolidayCoro pixels from last year, Ray Wu, Alling (Paul Zhang), ETopKeji, and whatever Gilbert sent me, very similar result across the board except the HolidayCoros wasted more energy. I just got a box from RGBMan, I'll update if these turn out to be any different.

If @DaveRNZ lived in the US I think his pixel choice would be clear... he needs 500 pixels in strings of 100, and Amazon sells exactly that (Alling for xConnect, ETopKeji for Ray Wu connectors), delivered next day free shipping with no waste, only 20% over the cheapest price you could possibly find for a 10k bulk buy.

 

[merryoncherry]

RGB Man 12V resistor pixels, arrived this afternoon - same as the others. 2.2A for 100 (12ft from controller); 3.1A for 200 at 12 ft.

 

[LA0]

RGB Man 12V resistor pixels, arrived this afternoon - same as the others. 2.2A for 100 (12ft from controller); 3.1A for 200 at 12 ft.

Just so you know, RGB Man is just a reseller for Paul Zhang / Etop.
So I'd expect the results to be similar.

 

[Mark_M]

You can't run 200 regulators ... even that close to the controller without yellowing. Very noticeable yellowing of 150 regulators.

This is an odd one that got me last year.
In 2020 I had 200 regulated icicles. All at 100% and perfect white.
Last year I put them up and the last 30 were soo badly yellowing.

It read 10v at the end.
I found it to be oxidation in the connector which caused the issue.
Voltage didn't bump back up, but must have been something about current capacity which caused it.

All I can think of it current limiting through the corrosion. My testing video from 2019 certainly showed that regulated could work down to 7.3v at full brightness, and even down to 4.8v before turning off-white.

 

[merryoncherry]

@Mark_M I probably have different regulated icicles, but 100 on a 5A port was all it would do. Chaining on a second one caused all kinds of problems, up to the whole string going dark. (Though, in retrospect, it probably pulled the ground voltage too high for the signal to be read, and not any other reason.) But since the 100 icicles pull about 5.1A full white when close to the controller... something would have to give. I love those icicles though, I'll keep running them on one port per string since they're closest to SLRRs anyway.

 

[merryoncherry]

I got some Mattos 12V resistor pixels today. They don't look identical to the other resistors I got so far this season, as the boards are blue. But the power readings are very much in line with the others, and the light output brightness/color is the same.
100 string full white: 2.29A; receive black .19A
200 string full white: 3.09A; receive black .38A
Maybe the chips and LEDs and design parameters are the same...

Can't remember if I mentioned earlier, but while 2021-era HolidayCoro pixels used a lot more power off / black than the rest, I got a few in 2022 and they are now distinguishable from the others only by the fact that the spacing is 1/2" closer together and therefore the wire resistance is less and the full white power draw is more... about 2.5A for a string of 100 instead of 2.2-2.3A for all the others.