Your setup, sounds like you’re seeing some usual power drop issues.
A couple of points:
-
Voltage drops happen when you pass a current through a length of wire - any wire. The amount of drop depends only on the size of the wire, the length, and the total current. In electronics, this simple fact come from the physics of Ohm’s Law. I mention this basic rule because it actually shows up everywhere in wiring LEDs (we just choose to ignore it many times).
-
When you hook up multiple wires to a power supply for power injection points, each individual wire will have it’s own voltage drop by the time it gets to it’s part of the LED strip. If you stick to one size of wire (16AWG in your case), the drop will depend on the wire length and the amount of current flowing in that particular wire.
For your test of 5V when Off and 3.5V when On, you’re seeing a bunch of effects rolled into one as you likely have all the LEDs on at the same time (I mean it’s supposed to be BRIGHT!). What you’re seeing is the net effect of all the different drops from all your injection points trying to power the whole strip. I know you want the “whole thing to work”, but a better test is to measure the voltage drop for each injection wire separately.
The way to do that is to setup WLED to display just the section you think each wire will supply (out of 450) and light only that section. Starting with the endmost injection point, set WLED to light only the last 150 (LEDs #300-#449) and measure both the current through that wire and the voltage across it at the LEDs. Do the same thing for middle injection wire (LEDs, #150-#299) and the start (LED’s #0-#149).
My guess is you’ll get 3 similar current readings, although the 3rd will likely be about 0.45A higher than the 1st. Your voltage drops will likely be worse for the 3rd again and best for the 1st (try it and post your results). At this point you’re likely going “Huh, what’s going on!”
What you’re seeing is all the result of 1. above, it’s just we typically ignore a couple of wires in our usual setups. We know about each of the injection wires (we put in place after all), but there’s a set of wires that connects all the LEDs in the strip together as well. That’s either a PCB or individual soldered wires and they have their own voltage drops based on the current they have to supply to each LED. And to complicate matters more, each LED has a “standby current” ot uses when it’s “off” That current is pretty small (~1mA) but over 450 LEDs thats 0.45A, not huge but still appreciable.
So what’s the point of all this? The net result is you want to follow 2 general rules when doing power injection.
(1) - Keep the injection wire lengths as short and as even as you can get them. Move the PS if you can so its closer to the middle of the strip. Voltage drop depends on wire length, so keep the wires as short as you can.
(2) - Place your injection points symmetrically across the whole strip so you’re using the “strip wires” evenly. That means 3 evenly spaced injection points don’t go at the start, middle and end. They go at 1/4, 1/2, 3/4 along the strip. The advantage of that spacing is each point becomes responsible for supplying current to the same# of LEDS to the “right” and “left” of the injection point using the strip wires.
A last note, the brightness limiter in WLED doesn’t actual measure any current for your LEDs. It’s just a calculation based on what you tell it about your setup. Software as a safety feature for your power supply is dubious at best. In general, it’s a rough guideline and you’re better off actually measuring the current in your setup to tune the real current per LED or just disabling it altogether.
You’ll do much better managing your setup’s power with the path you’re on - measure it in the real world and adapt your design accordingly.