LED Strip in Furniture: Voltage Drop, Driver Sizing, Groove Dimensions, and R9 on Cherry

The commission that starts this conversation is almost always a display piece: a walnut bookcase with a lit recess, a cherry bar cabinet, a floating shelf the client wants to "glow." The hard part is that every guide you find on 12V versus 24V strip is written by someone selling strip, and it stops at the wall. A Houzz thread titled "Recessing under cabinet LED strips?" has a woodworker asking for "the size and type of router bit" for a recessed channel; the only reply is a photo. Nailer Guy's guide says you "can rout a small groove in your wood to perfectly house the LED strip" and gives no dimension. So here is the version I wish had existed, in the order the decisions actually get made: strip, wire, driver, groove, and then what the light does to the wood.

The Strip Is a Spec Sheet, Not a Roll of Tape

Three numbers on the strip decide everything downstream: width, watts per foot, and lumens per foot. LEDYi Lighting lists the common single-row PCB widths as 8, 10 and 12 mm, and width has to match the channel, so pick profile and strip together. Watts per foot drives the driver and the wire. Lumens per foot drives whether the piece looks right: HitLights' application guide puts accent lighting at 100–350 lm/ft and under-cabinet at 175–525 lm/ft, and for furniture I aim at the low end of the accent range. A lit recess that reads as a light fixture has failed as furniture.

Two products bracket the market. Waveform Lighting's CENTRIC HOME strip is $139 per 16.4 ft reel, 5.5 W/ft, 450 lm/ft, CRI 95+ with R9 80+, in 2700K or 3000K. Armacost's RibbonFlex Pro 24V COB strip is $54.99 per 16.4 ft at Super Bright LEDs in its 165 lm/ft version: 1.2 W/ft, 8 mm wide, CRI 90+. One is a bright, high-fidelity strip that needs a real driver; the other is dim, dense, cheap and easy to power.

Voltage Drop Is Measured at the Far End, Not on the Reel

The rule of thumb, as Super Bright LEDs states it, is that 12V strip runs about 16.4 ft (5 m) before noticeable brightness loss and 24V about 32.8 ft (10 m). Super Bright LEDs prefers 24V for new installations, and so do I, but the reel-length figure is a manufacturer maximum, not a promise about brightness at the far end.

Lightstec's bench test is the number I plan around. A single-fed 12V strip read 12.54 V at the start, 11.5 V at 5 m (an 8.29% drop) and 9.92 V at 10 m (20.89%). The 24V strip read 23.8 V, then 21.7 V at 5 m (8.82%) and 18.8 V at 10 m (21.01%). Lightstec's point is that a 24V strip puts 6 LEDs in series per segment against 3 on 12V, so each LED bears roughly half the drop. QuinLED is more conservative still: 24V single-feed segments of about 2.5 m (8.2 ft), feed both ends at 5 m, feed start, middle and end to reach 10 m, and never put multiple supplies on one strip. A bookcase recess is rarely longer than 8 ft; a run around a bar or headboard is where you plan a second feed point before assembly.

Feed Wire Is Arithmetic You Can Do at the Bench

The wire from driver to strip gets sized by what's in the drawer, and that is where I see the most avoidable loss. Armacost's voltage-drop calculator sets the limit: if the drop in the feed exceeds 0.75 V on a 12V system or 1.5 V on 24V, increase the gauge or shorten the run. HyperPhysics publishes copper resistance per 1,000 ft: 16 AWG at 4.016 Ω, 18 AWG at 6.385 Ω, 20 AWG at 10.15 Ω, 22 AWG at 16.14 Ω.

Work it once. A 10 ft feed is 20 ft of conductor round trip; at 18 AWG that's 0.128 Ω. Eight feet of 5.5 W/ft strip is 44 W, which draws 3.67 A at 12V, so the feed drops 0.47 V, inside Armacost's limit. At 24V the same load draws 1.83 A and drops 0.23 V. Swap to 22 AWG at 12V and you're at 1.18 V, over the limit; at 24V on 22 AWG it's 0.59 V, still under 1.5 V. That is the whole 12V-versus-24V argument in one line: 24V lets you run thinner wire through a smaller chase, and a smaller chase is a woodworking problem solved.

Driver Headroom: Divide by 0.8, Stay Under Class 2

Inspired LED states the industry standard plainly: size the driver so the connected load doesn't exceed 80% of its rating. Total strip watts divided by 0.8, rounded up to a standard size. Their example: 20 ft of 24V strip at 3.0 W/ft is 60 W, 60 ÷ 0.8 is 75 W, so you buy a 96 or 100 W driver. Waveform's calculator uses 120% of draw instead; both land near 20–25%. Oversizing is typically safe, HitLights notes, because a constant-voltage strip draws only what it needs. Voltage must match. Wattage is a floor.

Strips take constant-voltage drivers; Inspired LED is explicit that constant-current drivers are for modules and downlights, not strips. And a Class 2 supply under UL 1310, per Super Bright LEDs' guide, is capped at 60 V DC, 5 A and 100 VA, which by my arithmetic is 60 W at 12V and just under 100 W at 24V. That is why 96 W and 100 W are the largest common 24V Class 2 drivers. Cost is not the obstacle: TRC Electronics lists a Mean Well LPV-60-24 (24 V, 60 W, UL 1310 Class 2, non-dimmable) at $16.05. Dimming is where the money goes, with Waveform's flicker-free dimmable supply at $249 against $129 non-dimmable, so I quote it as a line item and let the client decide.

The Groove: Half-Inch Bit, Three-Eighths Deep

This is what the retailer guides never say and the channel manufacturers do. Diode LED's CHROMAPATH Slim Recessed spec sheet instructs: route with a straight 1/2 in bit, track depth 3/8 in. The profile is 16.4 mm (0.65 in) wide across its flange with a 10.8 mm inner width, roughly 3/8 in tall, frosted polycarbonate cover, listed to UL 2108 with compatible tape. Lumenstar's LS-LL15, a 25 × 7 mm profile, calls for a 3/4 × 1/4 in (18 × 6 mm) cut and takes 8–10 mm strip; its LL10 is 15 × 6 mm and needs a 16 × 6 mm (5/8 × 1/4 in) groove. Klus's MICRO-ALU is 15.2 × 6 mm with an 11.2 mm gluing width for one 8 mm strip, rated 7.3 W/ft max.

Put those together and the recessed profiles cluster at 6–9.5 mm deep and 13–18 mm wide. In 3/4 in stock, a 3/8 in groove leaves 3/8 in of wood; a 1/4 in groove leaves 1/2 in. A 1/2 in straight bit cuts the Slim Recessed slot in one pass; the 3/4 in cut is a larger bit or two passes. The distinction that matters at the bench is flange versus flush. The Slim Recessed (0.65 in), Diode LED's standard Recessed (1.03 in) and the LL15 (25 mm) sit proud over the groove edge, so a slightly rough edge hides. MICRO-ALU and LL10 are flush and demand a groove cut to the profile's width, which in my shop is a CNC job.

Depth also decides whether you see dots. The Hook Up bought every Amazon diffuser channel and landed on one rule: the LEDs must be farther from the diffuser than they are from each other. At 30 LEDs/m every channel showed hotspots; at 144 LEDs/m the deep channels were clean. Shallow channels cost about 22% brightness on average, deep ones 36%. No recessed profile above is deep enough to diffuse a 60 LEDs/m strip (a 16.7 mm pitch), so for a dot-free line in a shallow groove the answer is a COB strip like Armacost's at 320 LEDs/m, or a channel-and-tape pairing the maker rates as diffused.

Finally, why the groove holds aluminum rather than bare strip. SuperLightingLED warns that wood's poor heat dissipation can overheat white strips and yellow the phosphor, and recommends aluminum profiles on wood. HitLights puts typical strip temperatures at 30–55 °C, about 30 °C over ambient; Waveform claims a lower 20 °C rise for CENTRIC HOME. Neither is an independent test, so I treat the channel as cheap insurance on a client's piece. Stick the strip to the channel, not the finish; LEDSupply is blunt that 3M backing won't hold on glossy or oily surfaces.

What R9 Does to Cherry

The last decision is the one clients notice and can't name. CRI averages eight pastel samples; R9 is the saturated-red sample left out of the calculation. Bees Lighting explains that white LEDs are naturally strong in blue and yellow and weak in long-wavelength red, so without proper red phosphors "cherry cabinets, mahogany furniture, and leather" look flat or muddy. Portfolio Lighting describes low R9 flattening cedar to a generic brown indistinguishable from pine or spruce. GRE Alpha gives the trap in numbers: a CRI 92 product can have single-digit R9 while a CRI 88 product posts R9 70.

That is what the Waveform strip's $139 buys: CRI 95+ with R9 80+, published. Armacost's COB strip is CRI 90+, and I could find no R9 figure on its pages. Color temperature is the cheaper lever. LEDVANCE describes 2700K as emphasizing warm tones such as wood, beige and gold, with 3000K cleaner and slightly cooler. For cherry and walnut I default to 2700K; for maple or ash I'll show 3000K.

I've come to believe that a lit piece of furniture is a small electrical system that happens to be made of wood, and it deserves the same intent as the joinery: 24V strip inside aluminum, a Class 2 driver loaded to 80%, feed wire sized by the arithmetic, a groove cut to the profile's published dimension, and a strip whose R9 is printed on the box. Cut the test groove in scrap before you cut the cherry.