The Device Keeps Its Own Parts Record: Parts and Service History, Liquid Contact and a 14-Day Window

Published 2026-11-07 · LuckyMDM Blog

The one-line version: liquid ingress is not an event, it is a state that surfaces late. A unit that boots, lights up and connects proves exactly one thing — that the circuit was continuous during those 30 seconds. It does not prove the absence of an electrochemical process already running under the shield. Two things make the difference: the hardware record the device keeps about its own replaced parts, and a 14-day observation window with two sampled rounds.

"It boots" is a weak test for a wet unit

One boot proves one conductive moment

The intake test almost everyone runs is a power-on: it reaches the home screen, the panel lights up, touch responds, so the unit is booked in as normal stock. The test is cheap, which is why it is universal, but it proves far less than it appears to. It confirms continuity across board, panel and digitiser for the duration of a button press.

Ingress damage is a curve, not a point

What damages a device after liquid contact is not the water, it is what the water carried. Dissolved salts stay on board surfaces and connector pins, and wherever two dissimilar metals sit in the presence of moisture and a voltage, a small galvanic cell forms and starts eating the joint. The process accumulates. On day one there may be no symptom at all; by day ten touch drifts at the edge of the panel; by week three the camera has fogged internally; by week six the cell has swollen.

Which is why two contradictory statements can both be true

"The unit boots" and "the unit fails inspection in eight weeks" describe different moments on the same curve. Intake sees the first point; the return sees the third. Treating the first point as the whole curve converts a timing problem into a condition problem, and the cost lands months later in a fault report that nobody can attribute.

The mechanism: why the device keeps a record it never announces

Layer one: the parts record is a pairing result, not a self-test

From iOS 15.2, Settings > General > About carries a Parts and Service History section. What it lists is not "a fault was detected" but "the pairing check between this component and the logic board returned this result". After a battery, display or camera is replaced, the system compares the new component against the pairing data held for that unit. Verified components read as genuine; unverified ones read as unknown. Because the source is pairing rather than diagnostics, a fully working unit can show an unknown part, and a unit with a real fault can show a genuine one.

Layer two: "unknown part" means "cannot verify", not "broken"

This is the most commonly misread label. Unknown part means the system cannot confirm the component is genuine. Functionally everything may be fine: the panel displays, touch tracks, True Tone works. Its ledger meaning is narrower and more useful — the hardware has been through a channel Apple cannot vouch for, so neither residual value nor warranty status can be estimated as an untouched unit.

Layer three: corrosion rate depends on residue and powered time

Two units that took on liquid on the same day can end up far apart. One took clean water, was powered down immediately and dried for three days. The other took a sugary drink, kept running for a fortnight, and never came apart. The first had most of its corrosion chain interrupted; the second still has residue reacting on its connectors. That is the reason "liquid contact" cannot be a yes/no flag in the ledger: the timestamp, whether the unit was powered down, and whether it was cleaned all change the slope of the curve.

The failure condition

The whole scheme depends on the ingress being noticed at all. If a unit took on liquid, shows no external trace and the customer says nothing, the observation window never starts. So the real function of the window is not to wait for failure — it is to create a declaration window. Write the rule as "any new functional fault within 14 days of intake is handled as ingress or repair", and the customer has a reason to be accurate at handover.

Three data sources, three different questions

SourceQuestion it answersHow it is producedCan it be fakedFreshness
Function telemetryDoes it work right nowDevice self-check, returned on queryNo, but only reflects nowMinutes to hours
Cosmetic gradingHas it dropped a resale tierControlled photos plus human gradingNo, but depends on the setupEvent-driven, needs re-capture
Parts recordWhat has been replacedComponent-to-board pairing checkNo, follows the hardwareUpdates on replacement, survives erase
Liquid contact flagHas it been wetIndicator colour plus customer declarationIndicator cannot be altered, declaration can be omittedOne-off event

The third row deserves attention: configuration can be erased, hardware records cannot. Many operators wipe a unit at intake and treat that as a reset to a clean state. The parts record is untouched by that, because it describes the hardware rather than the settings.

The liquid contact indicator: proves contact, not damage

The indicator sits inside the SIM tray slot. In its normal state it is white or silver; after contact with liquid it turns red. Sliding the tray out and shining a light into the slot is a visual check that needs no tools and no disassembly.

Its conclusion is narrow. A colour change proves the unit met liquid; it does not prove how much, and it does not prove whether damage resulted. In the other direction, sustained high humidity — a warehouse in a wet season, a charging spot next to a bathroom — can also turn it, which is a false positive. The correct use is as a trigger rather than a verdict: a colour change starts the observation window, and no colour change is not proof of dryness, because the indicator only reacts where liquid actually reached it.

Running the window: 14 days, two sampled rounds

LuckyMDM (Sichuan Starlight Network LLC) keeps the three-tier parts conclusion, the liquid contact flag and the observation expiry date as three separate fields on the device record, and a unit inside its window is excluded from rentable allocation — it can be held and inspected, but it cannot be assigned to a subscriber.

Round one on day 1: capture the variables

Three things on intake day: pull the tray and read the indicator colour, open the parts record, and ask the customer about ingress or prior repair. Store all three together, because that combination is the baseline. Run a five-point function check the same day — touch, camera, speaker, microphone, Face ID — recording each as pass or fail rather than a single "normal".

Round two on day 14: read the slope

Repeat the same five checks and compare them item by item. Identical results close the window and the unit is graded normally. A newly failing item moves the unit to the ingress path: out of the rentable pool, into testing or into resale. The value of the pair is not either result — it is the difference. A unit that passed all five on day 1 and fails touch on day 14 has produced evidence no visual inspection could match.

A device-management service like LuckyMDM, working with rental and subscription fleets, treats this as a scheduling problem rather than a sorting problem: the unit is not moved to a separate shelf and forgotten, it is given a start date, an expiry date and two comparable readings.

Why 14 days rather than 30

It is a compromise. The early symptoms — touch drift, internal fogging — mostly appear within the first two weeks, while shelf occupancy and capital turnover make a full month expensive. A unit that shows nothing new in 14 days is much less likely to fail afterwards, but the probability does not reach zero, so expiry means "can be listed" rather than "risk cleared".

The arithmetic on 500 returned units

Take 500 units through intake with 3 percent showing ingress or repair traces — 15 units. The difference between the two policies is when those 15 are identified.

Without a window, all 15 are booked as normal and listed. Experience suggests roughly 60 percent of them show symptoms within 60 days: 9 units. At 400 USD per repair, plus 80 USD for a swap shipment and 3 days out of service at 30 USD a day, that is 570 USD per unit and 5,130 USD across the nine — before counting the subscriber experience.

With a 14-day window, around 9 of the 15 surface at the second round and go straight to testing or resale, while the remaining 6 are listed on schedule. The added cost is labour: 10 minutes per unit per round, 15 units, two rounds — 300 minutes, about 5 hours. Five hours of inspection against 5,130 USD of repair and downtime is a trade that holds at any fleet size.

The quieter benefit is the refusal basis. Without a window, a unit that fails two weeks after handover has no clean attribution and the operator absorbs it. With a window, the handover sheet states that new functional faults within 14 days are handled as ingress or repair, and the boundary is drawn at the moment of intake rather than argued about later.

Misconceptions and boundaries

Misconception one: if it boots, it did not get wet.

A boot proves one conductive moment. Ingress damage is delayed, so the test that matters is the indicator colour and the customer declaration, not the power-on.

Misconception two: an unknown part means the unit is unusable.

The label speaks to verification, not function. Its meaning is residual value and warranty exposure. Conflating the two both rejects good units and overpays for units that should be discounted.

Misconception three: one erase returns the unit to a clean state.

Erase clears configuration and data. The parts record follows the hardware and survives the wipe, a restore, and a backup recovery.

Boundary one: this does not fit brand-new stock. An unactivated unit has no replacement history and no ingress exposure, so a window only burns shelf time. It applies to trade-ins, refurbished intake and customer returns.

Boundary two: an indicator change is not a rejection. High humidity can turn the indicator without any ingress, so the decision has to combine the colour with two rounds of function results. Colour alone is a trigger, and only colour plus a new functional fault within the window forms a complete case.

FAQ

Which components appear in Parts and Service History?

Battery, display and camera. Those three are the most frequently replaced and the most consequential for residual value; other components are not covered by the section.

Are genuine and unknown the only two states?

No. Newer iOS versions distinguish a brand-new genuine part from a used genuine part pulled from another device. That middle tier sits between the two on resale value and should not be merged with unknown parts.

Can a unit inside its window be rented out?

Better not. The point of the window is to keep ungraded units out of the rentable pool. If it must go out, grade it on the ingress path at a lower tier and price the risk in.

What if the customer denies liquid contact?

Rely on the handover sheet. Recording the indicator colour, the parts record and the five function results at intake, with a signature, is what makes the 14-day rule enforceable. Without the sheet, no mechanism survives contact with a dispute.

Does the risk end when the window closes?

No. It drops, it does not disappear. Expiry means the unit can be listed; the ingress flag and both rounds stay in the record and are reviewed again at resale grading.

Checklist

Five checks decide whether intake actually controls ingress and part replacement:

All Articles