Specifying a 0.40-Inch 1440x1080 Near-Eye Module for a VR Headset

Key takeaways

1. A 1440x1080 image on a 0.40-inch diagonal works out to roughly 4,500 pixels per inch, which is the headroom that lets a 30-degree field of view be filled without obvious pixel structure.
2. Resolution alone is not sharpness: the perceived image is set by angular pixel pitch, so a 1440-pixel width spread over 30 degrees gives about 0.02 degrees per pixel, close to the central acuity limit of a normal eye.
3. MIPI carries the pixel stream as a high-speed serial link, so lane count, lane rate and host PHY settings have to be agreed before the board layout is frozen rather than patched afterwards.
4. An emissive OLED pixel needs no backlight, which removes the diffuser and the light-guide stack, but it also means every interface in front of the panel must be judged as a reflection surface and a transmission loss.
5. Colour coverage is a gamut statement rather than a calibration statement, so 90% DCI-P3 tells you what the panel can reach, not what reaches the eye after the eyepiece and the compositor have had their say.

Head-mounted displays invert the usual display problem. On a desk monitor you can make the panel bigger to gain resolution; in a headset the panel sits a few centimetres from the eye, so size has to shrink while the pixel count stays high. That pressure is what makes 0.40-inch class near-eye modules interesting to VR integrators: the optics can be short, the front of the headset stays light, and the image still resolves detail in the middle of the field of view.

The module described here is a 0.40-inch near-eye optical module with 1440x1080 resolution, a 30-degree field of view, 90% DCI-P3 colour coverage and a MIPI interface, positioned for AR/VR glasses and head-mounted displays. Each of those numbers is a constraint on something else - the eyepiece focal length, the frame buffer, the host PHY, the colour pipeline - and none of them can be evaluated in isolation from the optical stack it will be bonded into.

This article looks at why this size, resolution and format fit a VR headset, then works through what to verify before the module is written into a bill of materials: how the field of view is defined, what the MIPI link implies for the host, which colour and alignment assumptions need owners, and which failures tend to appear only after the first worn prototype.

Why 0.40 inches and 1440x1080 land where VR optics want them

Panel diagonal and image angle are locked together by the eyepiece. For a fixed field of view, a smaller panel needs a shorter focal length, which pulls the optic closer to the eye and shrinks the distance from the front of the headset to the face. That is the mechanical argument for a 0.40-inch emitter in a VR viewer: the whole optical train gets shorter, and shorter optics are easier to keep light and stiff enough that the image does not swim when the wearer moves.

Resolution then sets angular pitch. A 1440-pixel width spread across a 30-degree field lands near 0.02 degrees per pixel, which is in the same region as the resolving power of a healthy eye at the centre of gaze. Below that threshold, extra pixels mostly buy margin: headroom for lens distortion correction, for the resampling that correction performs, and for content rendered at a higher internal resolution than the panel can show. Above the panel's native pitch, no amount of rendering restores structure that the emitter cannot present.

Colour and emission behaviour come next. An OLED pixel is its own light source, so black is the absence of emission rather than a shutter closing, and there is no backlight assembly to route around in a compact viewer. ^[3] The stated 90% DCI-P3 coverage describes how much of that wide-gamut space the panel can address; it says nothing yet about how the headset's compositor maps and clamps content into it, which is a separate design task with its own measurement plan.

What the pixel density figure does and does not cover

Pixel density is simply the number of pixels along a stated span, most often quoted per inch of diagonal. ^[2] It is a useful first filter because it tells you how much structure the emitter can show, but it is not a statement about the image the wearer sees, because the optic magnifies the panel and the eye samples the result at an angle.

Two panels with the same density can look different once bonded, depending on how much of each pixel cell actually emits, how the sub-pixels are arranged, and how much light the eyepiece scatters between neighbouring cells. Ask for the active area drawing and the sub-pixel layout rather than a density figure alone; the layout decides how visible the pixel grid is on a bright field, and on a near-eye display the pixel grid is often the first artefact users notice.

Establish what the 30-degree field of view actually measures

A field of view quoted without an axis is ambiguous: diagonal, horizontal and vertical values for the same image differ substantially at a 4:3 aspect ratio. For a 1440x1080 panel, a 30-degree diagonal figure corresponds to a noticeably smaller horizontal angle, and that changes the angular pitch you get for your pixel count as well as the eyepiece prescription.

Treat this as the first question for the supplier, not a detail to be resolved during bring-up. If the optical design is ray-traced against a horizontal figure while the module is characterised diagonally, the first prototype will show a field that is either larger than expected and soft at the edges, or smaller and tighter than the mechanical envelope allows. Both outcomes cost an optical revision, and optical revisions are the slowest kind to absorb in a headset programme.

The productive habit is to convert every headline number into the downstream quantity it constrains, then check that quantity against your own budget.

A verification matrix to run before the optics are frozen

The MIPI link, lane mapping and the init sequence that trips people up

MIPI DSI is a high-speed serial interface for display modules, and it exists because a parallel RGB bus wide enough for a 1440x1080 stream at headset frame rates would need an impractical number of traces and matched lengths. Serialising the pixels over a small number of differential lanes keeps the flex and the board narrow, which matters when the display sits at the end of a moving hinge or a temple arm. ^[1]

The consequence is that bring-up happens in layers. The host PHY must lock and the lanes must be mapped in the order the panel expects, then the panel's init sequence has to run before any pixels are sent, then the timing parameters define the visible window and the blanking. A frozen image on a black screen with a locked PHY is almost always an init or lane-order problem rather than a broken panel, so keep the link-status register visible in the debug log from the first power-up.

Two habits save a spin. First, capture the vendor init sequence as data, not as code buried in a driver, so you can diff it against the next panel revision. Second, bring the link up at a conservative rate and raise it only after the image is stable, because marginal timing shows up as intermittent sparkle or a single corrupted column that is easy to misdiagnose as a mechanical defect in the flex.

Write the lane mapping and the init revision into the build notes; both change more often than the mechanical outline does.

An integration review sequence for a two-eye headset

1. Fix the optical geometry on paper first. Decide the eyepiece focal length and eye relief from the panel's active diagonal, then compute the resulting field of view and angular pitch before ordering any glass.
2. Lock the interface budget with the host team. Agree lane count, lane rate and PHY settings with the SoC or bridge documentation in hand, and reserve the debug pins you will need if the link does not lock.
3. Define colour ownership end to end. Name who measures the panel, who calibrates the bonded assembly and who owns the compositor tone mapping; a gap between those three roles shows up as a colour mismatch between the two eyes.
4. Design the mechanical datum around boresight. Two modules have to be aligned to each other, not just to their own housings, so define the datum that sets interpupillary geometry and the tolerance stack that keeps both images converged.
5. Plan the thermal path before the enclosure is drawn. An emissive panel with no backlight still dissipates power in proportion to the scene's average brightness, so decide how heat leaves the module stack while there is still room to route it.
6. Build one bonded coupon before a full prototype. A single panel bonded into a representative eyepiece will expose field curvature, ghost reflections and colour shift far more cheaply than a wearable assembly will.

What to settle with the module supplier before tooling

Start from the published numbers on www.displaymodule.com/products/0-40-near-eye-optical-module-1440x1080: a 0.40-inch near-eye optical module at 1440x1080 with a 30-degree field of view, 90% DCI-P3 colour coverage and a MIPI interface, intended for AR/VR glasses and head-mounted displays. Those figures are the baseline for the review, and the useful work is turning each one into a question that only the supplier can answer.

Ask for the field-of-view axis, the output outline with datums, the bond line and adhesive recommendation, the MIPI lane configuration and a reference init sequence with its revision number. Ask separately for the tolerance on the optical alignment, because two panels bonded into one headset multiply that tolerance into a convergence error the wearer will read as eye strain.

Compliance paperwork belongs in the same request. RoHS restricts the hazardous substances that may be present in electrical and electronic equipment, so the supplier's declaration should be filed with the drawings rather than chased later by whoever runs the shipment. ^[4] Confirm as well how panel revisions will be announced, how long a given revision stays available, and what the minimum order quantity is for a bonded assembly, since those three commercial facts decide whether the design can survive a production ramp without a second optical qualification.

Finally, agree the acceptance test. A near-eye module is judged through its optic, so the specification should name the test conditions: drive level, warm-up time, ambient temperature and the colour targets to be measured. Without that, the first shipment will be evaluated against whatever instrument happens to be on the bench, and both sides will be right.

Frequently asked questions