Two displays from the same specification can differ visibly when they sit side by side. The difference is not a defect: it is the accumulated effect of tolerances in the LEDs, the driver, the light guide and the assembly. Binning is the practice of controlling part of that variation, and understanding where the rest comes from is what makes an acceptance criterion enforceable.

This article covers the sources of unit-to-unit variation, what binning does and does not control, and how to set criteria for a production lot.

Where unit-to-unit variation comes from

The measurement side of the same problem is covered in brightness, contrast and uniformity measurement, which sets out the conditions that make two measurements comparable.

Five sources dominate. The LEDs themselves vary in brightness and colour. The driver’s output current varies between units and across channels. The light guide and films vary in transmission and diffusion. Assembly varies in the pressure and alignment of the optical stack. And the panel’s own transmission differs slightly from unit to unit.

Each source contributes a small amount, and their effects partly add and partly cancel. The result is a distribution rather than a value, which is why a single-unit measurement cannot describe a production lot.

Binning concepts without the jargon

LED manufacturers sort their output into bins by brightness and by colour. A panel maker can then choose LEDs from a narrow range of bins, which reduces variation at the cost of yield and price, or accept a wider range, which reduces cost and increases variation.

The relevant questions for a buyer are which bins the maker specifies, whether the specification applies to the LEDs as purchased or to the finished light output, and whether the maker guarantees a matching tolerance between the modules supplied against one order.

A statement about LED bins is not a statement about finished brightness. The optical stack sits between the LED and the viewer, and its own variation can exceed the benefit of tighter bins.

Driver tolerance and current matching

Backlight drivers deliver current to several channels or strings, and the matching between those channels affects uniformity within a single unit, while the absolute accuracy affects variation between units.

Two specifications matter: the channel-to-channel current matching, which determines whether one part of the screen is brighter than another, and the output tolerance across production, which determines whether two units are similar. Both should be stated separately, because a driver can be excellent at one and mediocre at the other.

Light guide, films and assembly effects

The light guide is often the largest contributor to uniformity within a unit. Its moulding tolerances, its surface texture and its coupling to the LEDs determine how evenly light spreads across the panel. The films above it control the distribution further.

Assembly then adds its own variation. The pressure of the frame against the stack, the compression of any gap filler and the alignment of the layers all shift the result slightly. Two units built from identical components can measure differently because of how they were assembled, which is why the assembly process belongs in the variation budget.

Acceptance criteria for a production lot

Lot acceptance is a sampling decision as much as an optical one; the rule set in AQL inspection determines how many units must be measured to make the statement.

It is worth setting the criteria from the installation rather than from the panel. Where units are shipped separately and never appear together, the relevant tolerance is the one the end user sees: a machine in one building compared with a machine in another, which is far looser than a side-by-side comparison. Quantifying that difference before writing the specification usually relaxes it substantially and removes cost.

Lot acceptance should be based on a distribution rather than a single value. A workable structure states a window for the lot’s average, a maximum spread within the sample, and a limit for the difference between any two units in the sample.

Criterion	What it controls
Lot mean within a window	Whether the batch as a whole is brighter or dimmer than intended
Standard deviation or range within the sample	How much units differ from each other
Maximum difference between any two sampled units	The worst pair a customer could receive
Colour difference within the sample	Whether a side-by-side installation will look matched
Uniformity within each unit	Whether an individual screen looks even

The last two criteria matter most in practice. Two units can both be within specification and still look mismatched when installed next to each other, because the criterion was applied to each unit separately rather than to the pair.

Matching requirements for multiple units

Where several displays appear together, write the requirement against the installation rather than the product. A control desk with four screens needs a matching tolerance between them; a fleet of machines in different buildings does not.

Matching is a similar discipline to the one used for display specification terms, and the same instruments and conditions should be used for both. The practical consequence is commercial: a tight matching requirement means the supplier has to select units for the order, which usually means a minimum quantity or a price premium. Writing the requirement only where it is needed keeps both cost and risk in proportion.

Line measurement approach

Full optical measurement of every unit is usually impractical. Three approaches work in production: measure every unit at a single point with a simple instrument and use that as a screening value; measure a sample fully and use it to characterise the lot; or measure every unit in the final assembly, in the state the customer will see.

The third is the most meaningful, because brightness measured on a bare module differs from brightness measured through bonded glass and in a housing. Where the product has a front glass, the line measurement should be taken after assembly or the screening value will not represent the shipped product.

Documentation and lot traceability

Record the lot identifier, the LED bins used, the driver batch and the measurement results for each production lot. That record supports three later activities: investigating a customer complaint about colour match, comparing a new lot against the previous one, and identifying which shipments might be affected if a supplier changes a component.

Without lot identification, a uniformity complaint can only be answered with a replacement, because there is no way to establish whether the issue is a one-off or a systematic shift.

Corrective options when a lot drifts

Where the drift follows a period of thermal or humidity stress rather than production, the explanation may lie in the environmental behaviour of the assembly; the schedule in reliability testing provides the conditions that make such an effect visible.

Whichever correction is applied, record the change and its scope. Adjusting driver current to bring a lot back into the window changes the optical behaviour of every unit in that lot, and a later investigation that does not know about the adjustment will misinterpret the data. A one-line note in the lot record is enough to prevent that.

When a lot measures outside the window, the options are ordered by cost. Screening within the lot and matching units for multi-unit orders is the fastest. Adjusting the driver current to bring the mean back into the window is next, provided the change does not alter colour. Rejecting the lot and requiring tighter bins is the most expensive and is justified when the drift indicates the supplier’s process has changed rather than normal variation.

The decision depends on whether the drift is a shift in the mean or an increase in spread: a shifted mean can be corrected; a wider spread means the process is less controlled and the specification needs reviewing.

Variation-control plan

Source	Control	Evidence
LED brightness and colour	Bin selection stated in the specification	Bin definition and lot record
Driver current	Channel matching and output tolerance	Driver specification and batch record
Light guide and films	Moulding and material tolerances	Supplier process control data
Assembly	Defined compression and alignment	Assembly instruction and first-article check
Final measurement	Screening or sampling in the shipped state	Lot measurement record

If a multi-unit installation is showing brightness mismatch, send us the installation layout and the measured difference – the answer usually lies in whether the requirement was written against each unit or against the set. For the measurement side, see the article on measuring brightness, contrast and uniformity.

Frequently asked questions

Why do two displays of the same model look different?

Because each is within its own specification window rather than identical. The difference is the accumulated tolerance of LEDs, driver, optics and assembly.

Does tighter LED binning guarantee a better match?

It helps, but the light guide, films and assembly can contribute more variation than the LEDs themselves. Ask what the finished module’s tolerance is, not only the bin.

How many units should be measured for lot acceptance?

Enough to characterise the distribution – commonly five to ten units per lot for a screening measurement, with fuller characterisation on the first lot and after any change.