Start-up problems are among the most expensive faults to diagnose, because they often appear intermittently and disappear as soon as an instrument is connected. A display that fails to initialise one time in fifty is usually a sequencing or inrush problem rather than a defective panel.

This article covers the order in which a display system’s rails should come up, how inrush causes failures, and how to verify the behaviour before it reaches production.

The rails in a display system and their order

A display-based product usually has several supply rails: the main logic supply, the interface or transceiver supply, the panel bias and timing supply, and the backlight driver input. In many designs the panel has its own internal regulators, so only the logic and backlight rails appear at the connector – but the internal order still matters.

The conventional order is logic first, then interface, then panel bias, then backlight. Backlight last is deliberate: it prevents a bright flash before the panel has valid data, and it keeps the highest current load from coinciding with the inrush of the other rails.

Why sequence matters to the panel

Dimming behaviour interacts with the same rails, and the flicker-free dimming article covers the combination of drive method and supply behaviour. Two mechanisms are at work. The first is electrical: voltage present on an input while its supply is still rising can forward-bias internal structures in a way that increases current draw, which in a marginal design damages the panel over time rather than failing immediately.

The second is logical. Many panels latch their configuration on reset. If the reset is released while the supply is still below the panel’s threshold, the panel can latch into a state that produces a blank or wrongly timed image until the next power cycle. That is the classic “works after a second power cycle” symptom.

Inrush: causes and measurement

Cable resistance is the variable most often ignored. A short bench cable limits inrush more than the production harness does, so a design that passes on the bench can exceed the source’s current limit in the product. Measure with the production cable at its production length, and if the harness has not been finalised, measure with a deliberately pessimistic length so the result covers the worst case.

Inrush is the short current pulse drawn while input capacitance charges. Its amplitude is set by the input capacitance and the total series resistance of the supply path – including connector contact resistance and cable resistance – and its duration is short, which is why a bench supply often hides it.

Measure it with a current probe and an oscilloscope, not with a multimeter, and measure at the source rather than at the load. Record the peak, the duration and the conditions: supply voltage, cable length, temperature, and whether other loads start simultaneously.

Limiting inrush without breaking start-up

The standard methods are a series resistor bypassed after start-up, a controlled slew-rate driver, or simply a supply with adequate current capability. Each has a cost.

A series resistor reduces the initial pulse but also reduces the voltage available during the critical start-up window, which can cause exactly the latch problem described above. A slew-rate limit is more elegant but must be checked against the panel’s own power-on timing requirements – too slow can be as wrong as too fast.

Hot-plug and rapid power cycling

Connector design matters here too. Contacts that make and break in a different order – power before ground, for example – produce a start-up that the panel never experiences on a bench. Where hot-plug is a real use case, specify the mating sequence, verify it, and make sure the connector’s own design supports the order you need.

Where the harness carries both power and signal, the routing advice in FPC cable design applies equally to discrete wiring. Hot-plugging a display into a live system is worst-case in both directions: the display’s capacitance draws a pulse from the supply, and any residual charge in the display meets a source that may already be at a different potential.

Rapid power cycling is the other stress. If the supply collapses and recovers before the panel’s internal circuits have fully discharged, the second start-up begins from a partially charged state and the internal reset may not complete. Specify a minimum off time in the design and, where the application can cycle power quickly – for example a key-switched vehicle installation – verify the behaviour explicitly.

Brown-out and under-voltage behaviour

Temperature changes the timing of start-up. Internal regulators and reference circuits settle more slowly when cold, so a delay that is comfortable at 25 ¡ãC can be marginal at -20 ¡ãC. If the product is specified across a wide temperature range, the start-up margins should be verified at both ends rather than at room temperature with an assumption about the rest.

Temperature changes the timing of start-up. Internal regulators and reference circuits settle more slowly when cold, so a delay that is comfortable at 25 ¡ãC can be marginal at -20 ¡ãC. If the product is specified across a wide temperature range, the start-up margins should be verified at both ends rather than at room temperature with an assumption about the rest.

Real supplies sag. The panel’s behaviour during a sag determines whether the product recovers automatically or needs a power cycle.

Define the threshold at which the panel is expected to reset, and the minimum duration of a sag that must be tolerated without a reset. These two numbers belong in the requirements, because choosing them is easier than discovering them.

Symptoms of incorrect sequencing

It is worth recording the lab conditions alongside each symptom you encounter. Sequencing faults are notorious for appearing only with a particular supply, a particular cable length or a particular production lot of the panel, and a note that a fault was seen once with a bench supply is of little use when it appears in the field with a different one.

It is worth recording the lab conditions alongside each symptom you encounter. Sequencing faults are notorious for appearing only with a particular supply, a particular cable length or a particular production lot of the panel, and a note that a fault was seen once with a bench supply is of little use when it appears in the field with a different one.

Symptom	Likely sequencing cause
Blank screen, works after a power cycle	Reset released before the supply reached threshold
Correct image but unstable, settles after seconds	Interface running before the panel’s internal supply stabilised
Bright flash at start-up	Backlight enabled before valid data
Supply shuts down or current-limits at start	Inrush exceeding the source capability
Fails only when several units start together	Aggregate inrush on a shared supply
Fails only at low temperature	Slower internal start-up colliding with a fixed delay

Verifying the sequence on the bench

Capture all rails on one oscilloscope screen with a common trigger, then repeat at the temperature extremes and with the supply at both ends of its tolerance. Two further cases are worth running: several units starting on the same supply, and a rapid off-on cycle within the shortest interval the product could experience.

These four captures will find almost every sequencing defect, and they take an afternoon rather than a redesign cycle.

Documenting limits for production test

Write the limits into the test specification with their conditions attached: minimum supply voltage, temperature at which the limit applies, and the off time used. A limit without conditions drifts as the production line changes equipment, and the test gradually stops catching the problem it was added for.

Write the limits into the test specification with their conditions attached: minimum supply voltage, temperature at which the limit applies, and the off time used. A limit without conditions drifts as the production line changes equipment, and the test gradually stops catching the problem it was added for.

Once the behaviour is understood, two numbers belong in the production test: the minimum supply voltage at which start-up must succeed, and the minimum off time before a restart is guaranteed to work. Both are simple measurements, and both catch the marginal units that would otherwise reach the field.

Start-up checklist

Item	What to verify
Rail order	Logic, interface, panel bias, backlight – and reset timing relative to each
Inrush	Measured peak and duration at the source, with cable and temperature recorded
Supply capability	Margin with all loads starting simultaneously
Brown-out	Threshold and sag tolerance defined and tested
Cycling	Minimum off time verified at temperature extremes
Production test	Minimum start voltage and minimum off time included

If a display product in your programme intermittently fails to start, capture the rail sequence at the temperature extreme before replacing any parts – the capture usually shows the cause in the first few milliseconds. The driver side of the same design matters too: the backlight driver’s own inrush and settling behaviour interact with the sequence described here, so the two should be specified together.

Frequently asked questions

Why does a display work on the second power cycle?

Usually because the panel latched its configuration before the supply stabilised. Adding a proper reset delay after the rail has settled normally resolves it.

Is inrush a problem if the bench supply handles it?

Yes. Bench supplies are usually far more capable than the production power source. Measure with the real supply and the real cable.

Do all panels need the same sequencing?

No. The order is broadly similar, but the required delays and the minimum off time are panel-specific and should come from the panel documentation or the supplier.