A practical guide from the engineers at Moetion Technologies

IoT products are everywhere. Factory floors, in hospitals, on kitchen countertops, and often worn by people. Many of them need a user interface and more often than not, that means a small display. But selecting one is usually far more difficult than people expect.

Here’s the problem: once you drop below 5 inches, the familiar rules stop applying. Standardized interfaces disappear. Options go end-of-life with little warning. And getting your graphics to look good on the actual device is a constant battle.

Why does this happen, and, most importantly, what can you do about it? This guide walks you through how our team thinks about display selection for embedded systems, from first principles to final choice.

Why Small Displays Are So Difficult

The display industry is driven by volume: smartphones, tablets, monitors, and TVs. Manufacturers invest heavily in standardization and long product lifecycles for those categories because the margins justify it.

Small displays (the kind used in thermostats, ovens, and medical devices) don’t get the same treatment. Each application is unique, volumes are lower, and resolution requirements are modest by consumer standards. In addition, most manufacturing happens overseas using chipsets with minimal or no documentation, making integration and debugging far harder than it should be. The result is a fragmented market where:

• Interfaces are rarely standardized across manufacturers
• Products go end-of-life the moment a key customer stops ordering

Embedded systems add another layer of complexity. Many products use microcontrollers with limited processing power and I/O pins. Even a great display at a good price can be a dead end if your processor can’t support it. Some teams add a co-processor dedicated to the HMI, but that means maintaining two codebases and introduces serious headaches around OTA updates.

The good news: most of these pitfalls are avoidable with early, deliberate planning.

Step 1: Start With What You Actually Need to Show

The most common mistake we see? Teams start by searching for displays without first defining what those displays need to show. This almost always leads to over-specifying on some dimensions and under-specifying on others.

Before you open a single datasheet, answer these questions:

• Does the device need full video, simple animations, or largely static graphics?
• Will it display photos or high-contrast imagery, or will the HMI be simpler: icons, numbers, status indicators?
• Could black and white work, or is full color necessary?
• How detailed do the graphics actually need to be?

We recommend pulling together sample graphics and storyboarding the key screens early in the design cycle. This single step will do more to narrow your options and protect your budget more than almost anything else.

💡  Our Recommendation Find an evaluation kit for a display close to what you think you need and upload test graphics at the lowest resolution you think you can get away with. People consistently overestimate the resolution they require, and starting low gives you room to move up if needed, rather than painting yourself into a corner.

Step 2: Understand Your Interface Options

The interface between your display and microcontroller is one of the most consequential decisions you’ll make. Unlike larger consumer displays, there’s no universal standard here, and the wrong choice can force a costly redesign later.

Here’s how to think about the main options:

High-Speed Interfaces

RGB / Parallel

RGB sends red, green, and blue pixel data across 24–30 pins, plus additional pins for clock and sync. It’s typically the lowest-cost option for color displays, but it demands a lot of I/O and can be electrically noisy. Displays under 3″ often use a reduced pin-count “parallel” version. Sometimes, you can reduce pin going by tying pins together to turn an 8x8x8 RBG input into a 6x6x6 or even 5x6x5, but this sacrifices color depth.

Best for: cost-sensitive applications where you have sufficient I/O.

MIPI-DSI

A proprietary interface controlled by the MIPI Alliance, DSI interleaves graphical data with backlight, touch, and other signals. The display parses all of this in real time, which means significant CPU overhead on your microcontroller. That said, if you need cell phone-class resolution and already have a powerful processor, DSI opens up a wide range of high-volume, competitively priced panels.

Best for: high-resolution applications with a powerful microcontroller already in the design.

LVDS

A word of caution: “LVDS” is used loosely in the display industry and can refer to 3–4 different interface types. When applied to displays, it typically refers to a variation of FPD-Link (pioneered by National Semiconductor / Texas Instruments). It serializes RGB data into 3–4 differential pairs, enabling longer cable runs and fewer pins than RGB. Backlight, touch, and other signals must be handled separately.

Best for: applications with longer cable runs or pin-constrained microcontrollers that need high-speed data.

Figure 1: 6 and 8 bit versions of LVDS

Low-Speed Interfaces

SPI

SPI is a popular choice for prototyping and for displays with an onboard controller (for example, the NHD-4.3-800480FT-CSXP-CTP from Newhaven). Rather than handling every pixel yourself, you upload an image or animation, and the display takes care of rendering. The tradeoff: these displays are more expensive, less common, and generally not suited for video or fast animations. Black-and-white and trans-reflective LCDs are natural fits for SPI due to their lower data throughput requirements.

Best for: rapid prototyping or simpler HMIs where display-side processing offload is worth the premium.

UART, 4-Wire, and 2-Wire

These low-speed serial interfaces are common on budget microcontrollers. For text-only, black-and-white, or very low-speed color displays, they’re an excellent choice: simple to implement, widely supported, and low overhead.

Best for: simple, low-cost HMIs with minimal display requirements.

Step 3: Choose Your Panel Technology

Panel technology determines how your display produces light and color, with significant impact on power consumption, thickness, visibility, and cost. There are five broad categories to know:

Backlit Color LCD (TN, IPS)

The most common choice for full-color embedded displays. A set of white LEDs illuminates the LCD from behind. IPS panels offer better viewing angles and color accuracy than TN; TN panels are faster and cheaper. Generally, the most cost-effective option for color applications.

Figure 2: NHD-4.3-800480FT-CSXP-CTP display from Newhaven

Reflective LCD

Uses ambient light instead of a backlight, making it ideal for outdoor applications or products where power consumption is a concern. Best for character-based or black-and-white displays. Some models also offer an optional backlight for low-light conditions. Displays such as the NHD-C12864WO-B1TFH#-M from Newhaven are examples of these.

Figure 3: NHD-C12864WO-B1TFH#-M from Newhaven

OLED (Self-Emitting)

Each pixel emits its own light, eliminating the need for a backlight entirely. This makes OLEDs the thinnest option available, and they deliver outstanding contrast ratios and color depth. They also tend to draw less power. The tradeoffs: higher cost and potential burn-in concerns for static UIs.

E-Ink and Segmented

Niche options best suited to specific use cases: e-ink for ultra-low-power applications with infrequent updates, segmented LCDs for fixed-format readouts (think digital clocks or simple meters). Both typically require significant volumes or highly specific power constraints to justify the added complexity.

Step 4: Narrow Your Selection

Display and microcontroller selection need to happen in parallel; a mismatch between the two is one of the most common (and most expensive) mistakes in embedded design. Here’s the process our team uses to move quickly and confidently:

1. Set a size range, not a fixed dimension

The small display market is too fragmented to lock in on a single size. Give yourself ±0.3‟–0.5‟ of flexibility. If you think you need 4″, search 3.5‟–4.5‟, dramatically expanding your viable options.

2. Match your interface to your refresh needs

• Full video (>10 FPS): RGB, LVDS, or MIPI-DSI, plus substantial RAM and ROM on the microcontroller
• Simple animations (1–10 FPS): SPI may work, but RGB, parallel, or LVDS is more reliable
• Largely static graphics: SPI or low-speed serial interfaces are often sufficient

3. Factor in cable distance

• Under 3″: most interfaces will work
• 3″–12″: evaluate signal integrity carefully for high-speed options
• Over 12″: use a serial interface. LVDS or MIPI-DSI for high speed; UART, I2C, or 2-wire for low-refresh displays

4. Audit your available I/O

Pin-limited microcontrollers should favor LVDS or MIPI for high-speed needs, or a serial option when speed is not critical. If your microcontroller lacks a built-in LVDS or MIPI driver, you may need to distribute functions across two or more chipsets, which affects both cost and complexity.

One thing worth noting: an RGB display may have a lower unit cost than LVDS, but if it forces you to add a co-processor to compensate for pin limitations, LVDS may actually be cheaper in total. In addition, if your display is 10” from your controller, an RGB display may require a costly harness with up to 40 pins vs an LVDS which may require only 10 pins.

5. Address environmental and integration requirements

• Outdoor use or ceiling-mounted devices: target >800 nits for direct sunlight; >500 nits for indirect. Many LCD manufacturers can upgrade backlights on existing panels for a premium, worth exploring before sourcing a different display
• Space-constrained or thin products: OLED eliminates the backlight and can make a meaningful difference
• High durability or chemical cleanability (wipe-down with solvents): factor in cover glass, which may affect your touch technology options

A Few Tips on Graphics and Memory

Once you’ve selected a display, how you develop and manage graphics matters more than most teams expect. Two principles to keep front of mind:

Develop graphics at your display’s native resolution — not higher.

Resizing images on the device is memory-intensive and produces poor results. If your display is 800×480 at 5×6×5 color depth, build your assets at exactly that spec. Don’t develop at 940-pixel resolution with 8×8×8 color depth and expect it to look good after downscaling.

Use rendered graphics instead of photos wherever possible.

Renderings give you direct control over shading, shadows, outlines, and resolution. This pays dividends at small display sizes where photos tend to lose critical detail. It also makes it easier to make edits to your images later.

Final Thoughts

Display selection for embedded systems isn’t a component decision — it’s a system decision. Interface type, panel technology, microcontroller capabilities, and software architecture are all tightly coupled. Getting one wrong can unravel the others.

The teams that navigate this successfully share one trait: they start the conversation early. Not at the industrial design phase, and certainly not at the supplier selection phase. Start at the very beginning, when requirements are still flexible and options are still open. If you’re working through an embedded display decision and want a second opinion, we’re happy to talk. Reach out to the team at Moetion Technologies.