Basic setup diagram, with power injection example

[Kotche]

Hey ACL team

I'm finding that even reading through various material (inc the 101 Manual), posts, watching videos, etc. that I'm not yet sure I have a handle on some of the basics. This diagram is what I think I understand so far, so I'm looking for some validation here.

My diagram depicts a basic setup of:
- Computer or Pi
- Pixel Controller
- Multiple strings of 12V pixels
- Power supply
- Fuse line, for power injection

My assumptions here are:
- the Power Supply is used to power the Pixel Controller
- you can run multiple strings of 12v Pixels (~100 pixels max) off the Pixel Controller, without power injection
- for 12V Pixel strings longer than 100 pixels (up to 300 pixels in my example) requires power injection
- the power injection occurs at pixels #150 and #250, as it supplies power to the 50 pixels either side of the injection points
- I have a 15A fuse to manage the power injections for the longer string
- I can run multiple power injections through the same fuse so long as I don't exceed what the fuse can handle


Am I more or less on the right track here? What other considerations do I need to factor in? Thanks in advance for your input!

 

[thewanderingpine]

So this is a really good video for Power Injection:
View: https://www.youtube.com/watch?v=eR3QbzjpZy8

One key thing -- think of the power as flowing both ways. And that you want any given pixel to be no more than 100 pixels from a power distribution point. So to simplify your diagram - you can power inject just once at node 200.

(If you're looking at 5V, then you shrink those numbers)

Another good reference: http://spikerlights.com/calcpower.aspx

The bill porter video linked above has some comments on fuse selection (I think your 15A is too big, but I'm in no way an expert)

 

[Pavle]

Nice diagram, what did you use to put that together?

best thing you can do when starting out is get the items you need (for the controller I’d suggest you start with a PiHat, before going for a more comprehensive solution), and work on getting a single string of lights blinking. It will really help cement the required knowledge to begin planning a larger display. It will also help you work out if its for you.

Lots of fun and plenty of support around to get you through almost every idea you could come up with.

 

[Kotche]

So this is a really good video for Power Injection:
View: https://www.youtube.com/watch?v=eR3QbzjpZy8

One key thing -- think of the power as flowing both ways. And that you want any given pixel to be no more than 100 pixels from a power distribution point. So to simplify your diagram - you can power inject just once at node 200.

(If you're looking at 5V, then you shrink those numbers)

Another good reference: http://spikerlights.com/calcpower.aspx

The bill porter video linked above has some comments on fuse selection (I think your 15A is too big, but I'm in no way an expert)

Thanks - I just finished watching that video a little earlier today.

Per my diagram, I'd understood that power injection for 12V pixels would cover up to 100 pixels, hence 50 pixels in either direction from point of injection, though you seem to suggest that it would cover 100 pixels in either direction? So I'm happy to be corrected!

Nice diagram, what did you use to put that together?

I used Lucidchart.

 

[Kotche]

So to update my original diagram, I would now have one power injection for the string of 300 pixels; power injection at pixel #200, supplying power to the 100 pixels either side of the injection point. The power for the first 100 pixels is supplied from the Pixel Controller:

 

[Croydon Lights]

As you are running 12V pixel you can get away with adding the first injection at pixel #200 and have only one injection point for the 300 pixels and I would use a 5 amp fuse which would then need a smaller size cable than if you run a 15amp fuse as the cable needs to be rated above current of the fuse or else the cable will melt before the fuse

 

[i13]

The diagram looks correct although I would suggest showing a separate wire for the positive and negative. If you're fusing the positive, it would be sensible to cut the positive wire in the pixel string between pixels 100 and 101. This ensures that if there is a short, all of its current will either flow through the injection fuse or the controller's fuse. It would prevent the two fuses from being able to share the extra load.

 

[Kotche]

Diagram updated in light of further feedback, showing +v and -v lines, with the break in the +v line for the 300 pixel string:



Other considerations not depicted:
- length of wire from Controller to strings (how long can the wire be before voltage drop?)
- AWG of wire from Fuse to String A (what would be required given the 5A fuse and the length of String A?)
- power supply specs
- type of controller

Anything else others think would be beneficial to add?

 

[Notenoughlights]

If you run the whole lot from one power supply, I wouldn't see the point in breaking positive when injecting, that only comes into play when you inject power with a different power supply than what the lights are being fed with.

Edit; nevermind, I see the fuse now.

 

[Kotche]

If you run the whole lot from one power supply, I wouldn't see the point in breaking positive when injecting, that only comes into play when you inject power with a different power supply than what the lights are being fed with.

Edit; nevermind, I see the fuse now.

Fuse is there for additional precaution / safety in this setup, I think. Can't quite recall the rationale behind including it now, but I think that was it ...

 

[Notenoughlights]

Fuse is there for additional precaution / safety in this setup, I think. Can't quite recall the rationale behind including it now ...

I don't usually use fuses, then again I shouldn't be saying that too loudly The "Fs" had me confused till I read the fine print above it.

 

[TerryK]

Other considerations not depicted:
- length of wire from Controller to strings (how long can the wire be before voltage drop?)
If you are asking how long a wire must be before one sees voltage drop, 0 feet actually. However, for short distances voltage drop in most situations is so small it can be ignored. If you are asking how long a wire can be before voltage drop becomes a problem, it depends upon the wire AWG and current through the wire. In this case with pixels, the number of pixels and their drive level (which I think you have not yet mentioned).

- AWG of wire from Fuse to String A (what would be required given the 5A fuse and the length of String A?)
The wire AWG size would need to be sized to safely carry 5 Amp of current and then adjusted larger in size (a lower AWG number) to give the voltage drop desired or necessary. There is a problem here though. Two hundred pixels assuming WS2811s at 100% White (0.055 Amp each) would be 11 Amp.

- power supply specs
Assuming WS2811s at 100% White, (500 * 0.055), 27.5 Amp, an approximate 330 Watt. If you would drop the drive level a bit a Meanwell RSP-320-12 would carry them fairly well as a minimum. If you would want some headroom in the supply capacity then a Meanwell RSP-500-12. A Meanwell LRS-350-12 would work similar to the RSP-320 model.

- type of controller
Definitely an opinion decision. I like the Falcon F4.

A point to keep in mind; the first pixel of each string sees the voltage drop from the supply to the controller and then individually from the controller to that particular string. Essentially, wire size and length from the supply to the controller is just as important as from the controller to the string and in this case, the supply through the fuse to the String A trailing section.
Also keep in mind that voltage drop is seen in both the V+ and the V-/GND wires. So voltage drop calculations involving wire resitivity values typically need doubled.

 

[Kotche]

That is really helpful, thank you, @TerryK.

I have some follow up questions for you and/or the wider community:
- what is meant by "drive level"?
- If I'm housing the Power Supply and Controller in the same enclosure, what wire AWG would be recommended through to each string, if I'm running pixels at 40% brightness:
-- Strings B and C will be up to 3-4 meters away from the Controller
-- String A will be within 2 meters away from the Controller / Power Supply
- Is it a good rule of thumb to match the lead wire AWG size with the AWG size of the pixel string?
- Falcon controllers seem out of stock. I could work an Advatek controller into my model?

 

[Pavle]

Is it a good rule of thumb to match the lead wire AWG size with the AWG size of the pixel string?

Generally speaking, all of the pixel strings coming out of China indicate they have 18AWG wire, however real world results indicate it can be as bad as 26AWG (much smaller than it should be). In my personal experience, I use security cable for all of my additional wiring 6 Core 14/0.20 ( with 3 cores used for 0V, 2 cores for +ve and 1 core for data) and it would typically be in your best interests to use a higher gauge for all of your runs between controllers, power supplies etc. It's common to see people using 14AWG

Most people here will advise you to build your system to work at 100% and there is plenty of valid reasons to support this advice. For most of us, it's probably going to be an extra power supply or two at most and with such a small investment relative to other costs in this hobby it's probably the last place you want to cheap out.

 

[i13]

Drive level: The maximum brightness setting that you've chosen for your pixels. You may decide that 100% brightness is too bright. You mentioned a drive level of 40%.

Recommended AWG between the power source and each string: It depends on the distance and current draw but it would often be too thin if you match it to the wire AWG between the pixels. You'll want it to be thick enough to safely power the pixels at 100% brightness and also thick enough that there's a tolerable amount of voltage drop within it. Using an example of 100 pixels placed two metres from the power supply:
Reference table: https://www.engineeringtoolbox.com/wire-gauges-d_419.html
At 100% brightness, each pixel may draw 0.0555 amps. This is a worst-case-scenario number but it is better to be safe than sorry. As you're powering 100 pixels, that is 5.55 amps. 16 AWG wire would be safe to use unless it has 43 or more cores. Now we need to check how much voltage drop will occur. The total length of wire is four metres because the current needs to travel both directions in order to complete the circuit. Multiplying the resistance by the current gives you the voltage drop as per Ohm's law. The resistance is four metres multiplied by 0.013 ohms per metre = 0.052 Ohms. Multiplying this by 5.55 amps gives you a loss of 0.2886 volts which is excellent. At 40% brightness, you'll only lose 40% of this voltage which is 0.11544 volts (even better). I have now determined that 16 AWG is sufficient unless it has more than 43 cores. This would also be fine at four metres away from the controller. you've doubled the distance and therefore doubled the voltage drop. This means that you'll have a loss of 0.23088 volts at 40% brightness which is still acceptable. I like to overengineer and run my pixels at 100% brightness. You could start the calculation with 40% of 0.0555 amps. This would allow you to use thinner cable but it's best to use a fuse that blows with less current than the cable's maximum and be prepared to blow the fuse when you misconfigure something and run the pixels too bright. Another technique is to double the power cables which halves the voltage drop. Be careful not to break one of them because this may overload the remaining one.

Remember that your injection point is powering 200 pixels, not 100.

There are plenty of options other than Falcon. Advatek controllers will do the job although I have no experience with them myself. Their main limitation is that they can only support one pixel signal type at a time. I expect this to be a non-issue because most pixels use the same WS2811 signal nowdays.