Managing voltage drop and "pinking" in high-density pixel and Neon Flex runs

If the far end of your pixel strip or Neon Flex run turns pink, dims, or flickers while the near end looks perfect, you are almost certainly looking at voltage drop. It is the most common commissioning problem in high-density pixel installs, usually appearing the first time the whole run is driven to full white. The fix is not guesswork: it is one line of arithmetic, a wire table, and power injection in the right places.

What "pinking" is and why it happens

Every LED strip carries power along two thin copper rails printed on the strip itself. Copper resists current flow, and that resistance costs voltage: the further along the strip the current travels, the less voltage is left for the pixels there, like a long hose where the far sprinkler barely dribbles despite full pressure at the tap.

The colour shifts pink because of LED physics. Blue and green LED dies have a forward voltage of roughly 2.7 to 3.3 V, while red needs only about 1.8 to 2.2 V. As the rail sags, blue starves first, then green; red keeps going the longest, and white minus blue and green is pink, then red. Sag further and the pixel IC loses enough headroom that data timing destabilises, adding flicker or frozen pixels. On 12 V and 24 V strips with internal constant-current regulators, colour holds until the regulator drops out: a later, more abrupt failure, same math.

A second, separate cause of erratic flicker is failing to connect the negatives (GND) between independent power supplies; more on this below, because it is the most common reason "we injected power and it still flickers".

The one formula you need: V = I × R

Voltage drop is Ohm's law, nothing more: V_drop = I × R, where I = P ÷ V_supply and R = resistance per metre × cable length × 2. The ×2 is the round trip: current flows out on the positive conductor and back on the negative; both legs count.

Drop scales with current, and the current needed for a given wattage shrinks as voltage rises. Combined, percentage loss scales with 1 ÷ V²: doubling the system voltage quarters the percentage drop for the same power and cable. This is why 5 V hits the wall first, every time.

Worked example: 5 m of 60 px/m RGB strip at 0.3 W per pixel

Total load: 5 m × 60 px/m = 300 pixels, and 300 × 0.3 W = 90 W at full white. Feed it through 3 metres of AWG 14 (2.08 mm²) two-core cable: round trip 6 m at 8.3 mΩ/m gives R ≈ 0.050 Ω. The same 90 W then behaves very differently by voltage:

The 5 V system has lost almost a volt before the first pixel even sees power. Many 5 V pixel ICs specify a supply window of roughly 4.5 to 5.5 V (check your strip's datasheet), and colour shift is visible well before the lower limit, so this run is marginal at pixel one, with 18 A × 0.90 V = 16 W burned as heat in the feeder. The 12 V and 24 V systems have barely noticed. Planning target: keep total drop under about 5% of the rail, PSU terminal to last pixel (0.25 V at 5 V, 0.6 V at 12 V, 1.2 V at 24 V).

Drop along the strip itself adds to the feeder figure. Strip PCB rails are much thinner than feed cable, and their resistance varies widely with copper weight and is rarely published, which is why the multimeter check below matters. One rule helps: because current falls off as pixels peel away along a run, far-end drop is half what a lumped-load calculation predicts. That factor of two helps, but it does not rescue an 18 A run on thin copper.

The advantage of higher voltage compounds: 24 V is roughly 23 times more tolerant of the same cable than 5 V (the ratio squared), and 12 V roughly 6 times. To hold the 5% budget at 18 A over that 3 m feeder you would need about AWG 8 (8.4 mm², welding-cable territory), or, far more sensibly, a supply next to the strip and multiple injection points. This is why 5 V systems are wired with short, fat, frequent feeds while higher voltages can run leaner. Treat that as physics, not a free pass: real-world 24 V tape still commonly needs injection around every 10 metres, so higher voltage stretches the interval rather than removing the planning. Within ENTTEC's range the choice is 5 V or 12 V pixel strip, and the long-run answer is the constant-current 12 V 8PXB series, which holds colour over long runs by design instead of leaning on rail headroom. Settle tape voltage before the cable order, not after, then plan injection from the numbers regardless.

Wire gauge reference

Figures are solid copper at 20 °C; stranded cable reads a few percent higher, warm copper higher again (about +0.4% per °C). Multiply by round-trip length, not one-way distance.

ENTTEC's tape installation guides recommend 1.5 mm² / AWG 14 stranded cable as the minimum for all PSU-to-tape power connections. Check the cable's continuous current rating too: 18 A through AWG 18 is a fire question before it is a voltage question. Fuse each injected section per the strip and controller datasheet limits.

Power injection: what it is and where to inject

Power injection means feeding V+ and GND into the strip at additional points along its length, not just at the start, so no pixel is far from a low-resistance supply. The data line is untouched: it runs in one unbroken chain from the controller through every pixel.

• Calculate from the tape's actual wattage, not a generic rule of thumb: high-wattage tape draws more current and needs more aggressive injection.
• Start and end (dual injection) is the baseline for anything beyond a couple of metres at 5 V: it halves the worst-case distance and splits the current both ways, so worst-case drop falls to roughly a quarter, with the residual sag mid-run.
• Interval starting points by voltage: every 5 metres for 5 V, every 7 to 10 metres for 12 V, every 10 metres for 24 V. These are starting points, not guarantees; ENTTEC's per-strip installation guides give tested single and dual injection figures, based on a feeder of 3 m or shorter.
• Keep feeders short. If the PSU-to-injection cable would exceed 4 metres on 5 V or 8 metres on 12 V, place an additional supply near that point instead of stretching the cable.

Injection cures voltage drop only; data distance and controller channel limits are separate ceilings, covered in Pixel Strip: Maximum Length Factors.

Running data further than power

In large installs the controller cannot sit near every injection point. ENTTEC's Pixelator Mini MK2 with PLINK injectors separates the two problems: data travels long distances from the Pixelator to a PLINK at the fixture, and power is supplied locally at the PLINK. The 48 V PLINK variant shifts the drop math further still (48 V quarters the percentage drop again versus 24 V), useful when run length is fixed but injection access is limited.

One PSU or several: the wiring rules

With a single supply feeding several injection points, every feed shares V+ and GND by construction; size it with at least 10% current overhead above calculated full-white draw. Run each injection as its own wire pair back to the PSU terminals (a "star" layout) rather than daisy-chaining feeds. Once a run needs more current than one supply can deliver, or feeders get too long, three rules become critical:

• Common the grounds. Tie the 0 V / GND terminals of every supply on one data domain together, and to the controller's ground, keeping the strip's ground continuous end to end. Without this the data signal has no shared reference, causing intermittent corruption that looks exactly like voltage-drop flicker but does not respond to injection. If flicker persists after correct injection, suspect a missing ground tie first.
• Never join the positive outputs of separate supplies. Where two PSUs feed the same strip, cut the strip's V+ rail between the zones so each supply powers only its own section. Ordinary supplies with paralleled outputs will not share load evenly: the one set slightly higher sources most of the current, and back-fed current can trip protection or damage the lower supply. Some industrial supplies are designed for parallel operation with current-share terminals, but unless the datasheet says so, assume yours is not. GND stays common throughout; only V+ is sectioned.
• Match voltages: same nominal voltage on one strip; set adjustable supplies to matched output before connecting.

Choosing supplies with the right protections, headroom, and adjustable outputs is covered in Pixel: Power Supplies.

Verifying with a multimeter

Calculation gets you the design; measurement confirms it. The critical rule: measure under full load. An idle strip draws almost nothing and will read healthy everywhere even on a badly undersized feed.

• Set every pixel to full white (all channels at 100%) and leave it running: worst-case current.
• With the meter on DC volts, measure at the PSU output terminals first; this is your reference. Adjustable supplies can be trimmed up slightly to offset feeder drop, keeping the first pixel within datasheet range.
• Measure across V+ and GND at the strip's input pads, then the midpoint and far-end pads. Probe the pads, not connector shells, to exclude contact resistance. The PSU-to-strip-input difference is your feeder drop and should match the V = I × R prediction; if it is much larger, look for undersized cable or a poor crimp.
• Compare the far-end reading with the datasheet minimum. On a 5 V system, far-end readings sliding toward 4.5 V line up with visible pinking; if the far end sags below range at full white, add or move an injection point and re-measure.
• Re-measure warm: copper resistance rises with temperature, so a run that passes cold can pink after several minutes at full white.

Test a representative run before committing: ten minutes with a multimeter is cheaper than re-wiring injection points after the cladding goes on.

Related reading

• Pixel: Power Supplies
• Pixel Flickering Guide: Diagnose and Resolve Common SPI LED Issues
• Pixel Strip: Maximum Length Factors