COB vs SMD LED Strip: What Changes Inside a Wine Rack

September 7, 2026

Everything that makes a length of LED tape a light source is mounted or deposited onto one narrow strip of flexible circuit board. On surface-mount device tape, what goes onto that board is finished components: each emitter is a molded package with its own lens material, its own phosphor coating, and its own solder pads, manufactured, tested, and sorted before it ever reached the strip. Chip-on-board tape carries no packages at all.



On chip-on-board tape, bare semiconductor dies are bonded directly onto the flexible circuit in a dense line, and one continuous phosphor layer is applied over the entire array. There is no molded body around any individual die and no stretch of bare board between them. The emitting surface is a line. On surface-mount tape it is a row of separated points at a fixed pitch.

QUICK ANSWER: Chip-on-board tape belongs where the strip sits close to bottles or glass and the housing is shallow. Surface-mount tape holds the advantage on color range, binning precision, and per-package output. Neither is the default.

That is a manufacturing difference before it is a lighting difference. Component makers treat the two as separate process families at the component level. ams OSRAM publishes handling and reflow-soldering guidance for its surface-mount emitters, and Lumileds publishes a chip-on-board primer describing die attach directly onto a substrate. Neither document covers flexible tape, but the two processes they describe are the two that produce the two kinds of strip. One process ends with parts on a strip. The other ends with an array that is the strip.



Everything below is downstream of that. How close the strip can be mounted, how shallow a housing it fits, where its heat exits, what color behavior is available, and what a failure leaves behind all trace back to whether the emitters arrived packaged or bare.

How Close Each One Can Sit to a Bottle

An emitting surface reproduces its own structure in anything reflective in front of it, and how far back that surface has to sit before its structure stops being legible is a property of the surface. A continuous phosphor line has no internal structure to reproduce, so what returns from a bottle shoulder or a glass rack end is a band. A row of discrete packages has structure at a fixed pitch, and closing the distance between the strip and the glass makes that pitch more legible.


Inside a rack, that sets a working distance for each format. Chip-on-board tape can run along the underside of a display-row lip a few inches above the bottles in front of it and still read as a band at that range. Surface-mount tape at the same range depends on everything downstream of the emitter to do the work, and it can get there.



Where the strip is set back behind a valance, or concealed entirely so only its light is seen, the format difference stops mattering at all. Proximity is the variable that decides whether it matters. Cellar racking happens to force proximity more often than most interiors do, because the members are thin and the reveals are shallow, which is why the question comes up here at all.

A surface-mount package stands above the board by its own body height. The emitting face of a chip-on-board array sits at close to board level, because the only thing above the die is a phosphor coating. That governs the minimum internal clearance an extruded aluminum profile has to provide before its lens can close over the strip.

Racking leaves lighting very little room. A shelf lip, a display-row reveal, or the narrow face of a rack upright is often the only place a run can go, and the profile has to sit flush inside it. A chip-on-board strip fits profiles that a surface-mount strip cannot physically close over without the lens bearing on the package faces. This is a fit question, not a blending one: it decides whether the run can be installed in the reveal the millwork already has.


The trade reverses as soon as depth is available. A shallower profile holds less aluminum, and thermal margin follows metal mass, so specifying the thinnest extrusion a chip-on-board strip allows gives up margin covered in the next section. Surface-mount tape is also made across a wider spread of board widths, voltages, and pitch options, so where the reveal is generous, the package height costs nothing and the broader catalog is available.



Bend behavior belongs to the same fit question. Both formats bend along the length of the board, and neither is built to bend across its width, so a return around a rack corner is made as a joint or as a change of profile, and the minimum bend radius is published against the individual series. Claims about which format curves more tightly are easy to find in both directions, which is a fair sign the answer is a data-sheet figure for a given series.

Junction temperature governs both light output and useful life on any LED, a point the US Department of Energy's solid-state lighting program has made in its thermal management guidance for years. What differs between the two formats is the chain of interfaces between the die and the board carrying it.


On a surface-mount emitter, heat leaves the die into the package's own substrate, crosses the package's thermal pad, crosses the solder joint, and only then reaches the copper of the strip. Each interface adds resistance, and the sum of them appears on the package datasheet as a junction-to-solder-point thermal resistance. On a chip-on-board strip, the die is attached to the board substrate directly. There is no package body in the path and no package-to-board solder interface, and the die-attach area across a dense array spreads that contact over a wider footprint of board than a row of separated pads does.

Fewer interfaces do not settle the question. A chip-on-board array packs far more emitting area into each linear inch, so at a given drive level the heat produced per foot of run can be higher even though each unit of it has a shorter path out. A strip with a short thermal path mounted against wood, with no metal behind it, is still a strip running warm.



Surface-mount has one real engineering advantage here. The package carries a published thermal resistance figure a designer can size a housing against, while a chip-on-board array is characterized at the strip level, which gives less to calculate with and more to test.

Color Range and Color Consistency

This is where surface-mount offers capability chip-on-board has not matched. A package is a container, and a container has room in it. One surface-mount package can hold a red, a green, and a blue die together, or add a fourth white die, or hold two white dies at different correlated color temperatures so a single run can be tuned across its range. Larger packages can carry a control IC alongside the dies, which is what makes per-emitter addressing possible at all. Every one of those configurations exists because there is a package.


Binning is the second half of it. Packaged emitters are tested and sorted after manufacture, which is what lets a specifier ask for a narrower color-consistency bin and actually receive one. A chip-on-board array is covered by a single poured phosphor layer applied at strip level, and batch-to-batch color consistency is a known weak point of the format.



For a collector adding racking in phases, or a restaurant extending a display wall into an adjoining room, that is not an abstract concern. It is the difference between one white across a room and two.

What a Failure Does to Each Format

On surface-mount tape, a failed emitter is one component going dark in a row of them. The segment around it keeps working, the affected part is identifiable by eye, and the defect sits at a specific point on the board. On chip-on-board tape there is no discrete component to identify. A fault inside the array shows up as a dark length of the continuous line, and the phosphor layer covering the dies means the emitting surface cannot be worked on piece by piece.



In practice that pushes chip-on-board service toward replacing a whole length between cut points. That is a design-time question more than a repair-time one. A service route to every run belongs in the design, and so does treating the strip as a replaceable assembly from the start instead of a permanent part of the millwork. It also compounds with binning, because replacing a length of chip-on-board tape years later means matching a poured phosphor layer from a different production batch.

TIP: Order spare footage from the same production run when the strip is first specified, and store it with the project file. A replacement length pulled from a later batch can read as a visibly different white beside the original.

Neither format is something a cellar owner should open a channel to investigate. A dark section, a flickering section, and a section that has drifted in color are all diagnostics for the installer or a licensed electrician, and all three can originate well upstream of the strip itself.

Claims That Do Not Follow from the Construction

Marketing copy routinely asserts that chip-on-board is brighter, more efficient, and longer-lived than surface-mount, and the reverse is just as easy to find asserted on at least two of those three. Some pages manage both inside one article. None of the three is a property of the format. Brightness, efficacy, and useful life each depend on the specific series, its drive current, and its thermal design, which is why a buyer should require a datasheet or a measured output before accepting any of them.


Two things are safe to say. A single surface-mount package can be built to a higher output than any comparable length of chip-on-board line, because concentrating output at one point is what a package is for. And output per foot of tape is set by the chosen series, the current it runs at, and where its heat goes, none of which is disclosed by the three letters on the label.


The first of those holds inside the tape comparison and nowhere outside it. In spotlights and downlights the point-source role is held by packaged chip-on-board emitters, single high-flux modules built on ceramic or metal-core substrates, and Lumileds lists its LUXEON CoB Core Range for exactly those fixtures. Clustered high-power surface-mount emitters filled that role before packaged chip-on-board became common, and the scope of this article is chip-on-board tape rather than chip-on-board emitters.



A specifier who treats the format as shorthand for performance is relying on a letter code. The useful comparison is between two named strip series, with their published data, mocked up in the housing and at the distance the installation will actually use.

Choosing Between Them for a Specific Run

The choice belongs to the individual run, and a single cellar can carry both without anyone reading the difference in construction. Where the emitting surface will be seen at close range or returned by glass, and where the millwork gives a shallow reveal, chip-on-board fits the geometry the room already has. Where the run is set back, concealed, or asked to do something a package enables, tunable white across a display zone, addressable segments, or a bin held to a stated tolerance across phases, surface-mount is the format that can do it.



What settles it is never the label. It is the strip series, the housing it will sit in, the distance to the nearest reflective surface, the color tolerance the project has to hold across its life, and a sample of both viewed in the actual reveal before the racking goes up.

Frequently Asked Questions

  • What do the numbers in an SMD part designation mean?

    They are the package footprint in tenths of a millimeter. A 5050 emitter measures 5.0 by 5.0 millimeters across its package, a 3528 measures 3.5 by 2.8, and a 2835 measures 2.8 by 3.5. The convention describes the outside of the package; it says nothing about the die inside, which is why two emitters sharing a designation can hold different chips and produce different output. Chip-on-board strips have no equivalent designation, because there is no package to measure, so their data sheets identify a strip series.

  • Does chip-on-board come in tunable white or addressable versions?

    Yes, though the catalog is narrower. Strip manufacturers now list chip-on-board products in tunable-white configurations and in addressable versions where a control IC drives defined segments of the continuous line. The distinction is that the addressable unit is a zone of the phosphor line fixed by the strip series rather than an individual emitter, so the smallest controllable increment is a manufacturing decision made before anyone specifies the product.

  • Are the two cut and joined the same way?

    Cut spacing and connector type belong to the individual strip series and should be read off its data sheet before a layout is drawn. Physically, both formats expose copper solder pads at each marked cut point, and on chip-on-board tape those pads sit in a deliberate gap in the phosphor line, which is the one place the continuous surface is interrupted. Solderless clamp connectors engage those pads. Cutting, joining, and terminating are installer work.

  • Is one format more durable in physical handling?

    The encapsulated surface gives chip-on-board an edge on mechanical exposure. Surface-mount packages stand proud of the board with edges and corners that can be knocked or snagged while racking is being fitted around them, while the poured phosphor layer presents one covered face with nothing standing above it. Better electrostatic tolerance is sometimes claimed for chip-on-board on the same grounds, though that is a series-level specification and belongs on the data sheet.

  • What should a lighting submittal ask for so two strips can be compared fairly?

    Ask for an IES LM-79 test report on the finished strip. LM-79 covers absolute photometric and electrical testing of a complete solid-state lighting product, which is the level a fair comparison has to be made at. LM-80 data, by contrast, measures lumen maintenance of an LED package, so it maps cleanly onto a discrete surface-mount emitter and far less cleanly onto a chip-on-board array with no separable package to test.

  • Can chip-on-board and surface-mount tape be used in the same cellar?

    Yes, and it is common where different runs have different jobs. The thing to control is white-point agreement between them, since the two come out of separate binning processes. Ask the supplier to state color-consistency tolerance in SDCM steps, the standard deviation of color matching, for both products at the same nominal color temperature, and confirm both figures are that manufacturer's stated tolerance for those exact series.

Compare both formats in the actual reveal before the racking is built — the housing depth, the mounting distance, and the color tolerance decide it more surely than the label does. Cachet Wine Cellars designs and builds custom wine cellars from Irvine, California, with a lifetime warranty on every installation. Call (949) 569 7857.

< Older Post Newer Post >

Recent Posts

Luxury wine cellar with wooden racks and shelves of assorted liquor bottles displayed on warm lighting.

Damp Rated Wine Cellar Light Fixtures: Which Zones Need Them

September 7, 2026
A cellar holds humidity almost no other room does. What dry, damp, and wet listings actually mean, what an IP code misses, and which zones need which.
Modern home hallway with a large glass wine rack and hardwood floors

What New Build and Remodel Each Decide for Cellar Lighting

September 7, 2026
Already framing walls, or working around finished ones? Learn what a new build lets you hide and route for cellar lighting that a remodel simply can't.
Rows of wine bottles stored horizontally on tall racks in a cellar or shop.

Wine Cellar Lighting Beam Angle Matters More than Lumens

September 7, 2026
Two cellar fixtures at identical lumen output can light a room completely differently. Beam angle is the variable that settles which one you get.
Modern dining room with glass wine wall, round table, and staircase in the background

Glamorous Glass Residential Wine Cellars in a Dining Area

May 8, 2026
Having residential wine cellars adds sophistication and elegance to the space and reflects your refined wine taste. It can serve as an icebreaker and create a visually stunning focal point.
Backlit glass wine cellar with bottles in a modern room, flanked by two armchairs

Modern Glass Wine Closet Wine Cellars for Your Wine Collection

May 8, 2026
Wine collectors put a lot of time and effort into developing their collections, and their wine storage solutions should reflect that same level of sophistication. Modern glass wine closets are a stylish and practical way to showcase your wine collection. With sleek designs and elegant finishes, these closet-style custo
Wooden wine rack and glass shelves in a narrow home wine cellar with bottles stored horizontally on the wall

A Modern Wine Cellar Inside a Chic, Farm-Style Designed Home – An Absolute Favorite!

May 8, 2026
Many Orange County homeowners are considering modern wine cellars as a beautiful and functional addition to their home. Their first function is to securely and effectively protect and store the precious wines within. But a secondary and equally important function is that the cellar is stylish, contemporary, and beautif
Glass-front wine cellar with wooden racks and bottles in a warm-lit room

A Walk Through One of the Gorgeous Glass Wine Cellar Doors in California

May 8, 2026
As a wine collector, you understand the importance of having a professionally designed wine cellar at home to keep your most valuable wines safe and in perfect condition. Reliable wine cellar components like glass wine cellar doors in California help maintain the quality of your wine collection.
Wine cellar with floor-to-ceiling racks and bottled spirits displayed on shelves

Residential Wine Rooms Built in Orange County: A Closer Look at This New Vintage Style Wine Cellar in Mission Viejo

May 8, 2026
Residential wine rooms have become increasingly popular among California homeowners. This style combines both functionality and aesthetics. It also pays homage to the region’s rich winemaking heritage.
Interior with wooden staircase, wine rack, and large mirror in a modern home foyer

Under the Stairs Wine Cellar Project Unveiled in Santa Ana, California

May 8, 2026
In modern home design, having a residential wine cellar under the stairs is a trendy and space-efficient way to create a unique and functional display for your wine collection.
Wall of wine bottles surrounding a lit shelf with boxed bottles and wine glasses in the center

Metal Wine Racks: An Ideal Racking System for a High-End Residential Space in California

May 8, 2026
To ensure your modern wine cellar is up to date with the latest trends, it may be beneficial to consider using sophisticated metal wine racks. These racks offer a sleek and stylish design that can add to the overall aesthetic of your cellar and keep your collection organized.
Show More