Elevator and Floor Access: The Test Plan a Buyer Should Demand
An evaluator should demand an elevator and floor access test plan before approving an integrated access release. A product demonstration, an interface list, or a verbal assurance that the elevator team will cooperate does not prove that a safe sequence will survive normal exceptions and network failures. The plan must turn the intended journey into repeatable scenarios with stated inputs, expected events, degraded behavior, named owners, and retained evidence.
Direct answer. In brief, this article covers elevator and floor access test plan from elevator and floor access test plan to elevator-and-floor-access. It states what to confirm about elevator and floor access test plan on the dated product page where a figure is involved. An elevator and floor access test plan shows what each component must assume about the others. elevator and floor access test plan is the interface that must be satisfied. Then commissioning starts on elevator and floor access test plan.
On elevator and floor access test plan, this article uses elevator and floor access test plan as the definition. It treats elevator and floor access test plan as a working definition specific to this market and product, rather than a general one.
- Map the interface — write down what hands over between lock, door and system at elevator and floor access test plan
- Test the handover — record the sequence that proves elevator-and-floor-access works end to end
- Keep the record — store the commissioning evidence with the file for approval boundary
- Treat this as the elevator and floor access test plan sequence, with elevator and floor access as the reference.
This document is a buyer-side process and timeline. It applies to a system integrator coordinating an access-control controller, credential reader, elevator or landing-door equipment, network services, and the client’s operating organization. It covers evaluation through acceptance and handover without assuming a particular product architecture.
Approval boundary: what the plan must settle
Begin with the operating outcome, not the equipment list. The plan should identify the authorized person or role, the point of presentation, the authorized destination, the service that receives the request, the service that performs the release, and the evidence produced at each boundary. It must also distinguish a call request, a dispatch decision, an arrival indication, a release command, and a reported door state. The product family this applies to is the smart door locks.
The plan should cover the intended use cases and their exclusions. For example, it may cover staff access, visitor movement under escort, service personnel, after-hours access, and a designated lobby. It should say whether freight cars, parking transitions, tenant changeovers, emergency operation, and mechanical elevator maintenance are outside the tested sequence.
For a hotel with separate guest and staff paths, use Elevator and Floor Access for a Hotel Property to separate property-system decisions from this execution plan. Do not let a broad property requirement substitute for a testable interface contract.
The approval boundary is reached only when the buyer can answer: what is sent, what happens when it is missing, who owns the clock, where the records live, what occurs during an outage, and which party owns the fix.
Baseline packet required before Gate 0
The buyer should issue a baseline request before any joint demonstration. Each artifact needs a named owner, an approval status, and a revision reference. Verbal descriptions are working notes until the responsible party signs or records acceptance.
| Baseline artifact | Minimum content | Accountable signatory |
|---|---|---|
| Approved use-case matrix | Person or role, credential, origin, destination, time condition, normal result, and degraded result | Client operating owner |
| Site and door schedule | Elevator or landing identity, controlled door, controller, reader, network path, and boundary devices | System integrator |
| Interface responsibility matrix | Sending party, receiving party, message, response, timeout behavior, and repair owner | Integration lead |
| Operating-state policy | Centralized versus standalone state, unavailable-service behavior, manual mode, and reset behavior | Design authority |
| Clock and evidence plan | Authoritative time source, event identifier, correlation method, record owner, and evidence format | Security or operations lead |
| Approval schedule | Fire, egress, privacy, elevator-service, and site-access questions with recorded responses | Client designated approver |
If a proposed SKU contributes a reader, controller, lock, or interface behavior, confirm on the dated product page for the specific SKU and attach that dated evidence to the baseline. A catalogue statement, roadmap entry, or verbal capability claim is not an observed test result.
Use the Office Smart Lock Site Survey and Smart Lock Retrofit Site Survey to capture physical door conditions, but do not treat either survey as evidence that the elevator interface works.
CONFIRM-WITH-LOCAL-FIRE-EGRESS-OWNER: Confirm which fire-signal, emergency-release, and egress authority approves the installed sequence, the required safe state, and any exception before acceptance testing begins.
Responsibility map and controller ownership
A shared project can still have one owner for each decision. The buyer should reject a test plan that assigns only broad labels such as “vendor” or “building team.” The interface owner must be able to say which company receives a message, which company decides, and which company changes the relevant configuration.
| Decision | Accountable owner | Consulted party | Objective evidence | Reject when |
|---|---|---|---|---|
| Credential and floor authorization | Access-control owner | Identity and client operations | Decision record with reason code | Two systems can grant the same floor independently |
| Elevator call and dispatch | Elevator-service owner | Access-control integrator | Ordered request and response trace | Dispatch is described only as “the elevator handles it” |
| Landing-door release | Access-control integrator | Door-equipment owner | Command and door-state records | A request is treated as proof of release |
| Door-state interpretation | Controller owner | Door-equipment owner | Controller event and source state | Open, closed, and fault states are conflated |
| Degraded-mode decision | Client design authority | Operations, elevator and security teams | Approved state table and test result | “Fail locked” is used without approval |
| Record retention and disclosure | Client data owner | Security and privacy reviewers | Access matrix and retention decision | Test operators can read all event data without restriction |
| Emergency and service-mode behavior | Local designated authority | Elevator and egress owners | Recorded approval and test procedure | Normal integration logic is assumed to govern emergencies |
This decision table is also the starting point for the owner action register. Each unresolved boundary, missing response, disputed timeout, or absent evidence item receives one accountable party, a recorded action and a verification artifact. Shared ownership without a named decision maker is not acceptable.
The controller holds the controlled door state and makes the access decision; the reader reports the credential presented at the boundary. The plan must preserve that distinction when an elevator system, mobile credential, or supervisory service is added.
Stage-gate timeline for the evaluation
The timeline advances by evidence, not by a promised deployment date. A gate remains open when an exit artifact is missing, disputed, or supported only by a demonstration.
flowchart TD
A[Gate 0 Approve the use case] --> B[Gate 1 Freeze the interface contract]
B --> C[Gate 2 Build the test bench]
C --> D[Gate 3 Connect the live environment]
D --> E[Gate 4 Inject failure conditions]
E --> F[Gate 5 Reconcile evidence and ownership]
F --> G[Gate 6 Accept and hand over]
| Gate | Required work | Exit artifact |
|---|---|---|
| Gate 0 | Select use cases, define exclusions, nominate owners, and identify local approvals | Signed baseline and use-case matrix |
| Gate 1 | Define messages, state ownership, time source, reason codes, and outage behavior | Versioned interface contract and state table |
| Gate 2 | Reproduce the sequence with representative controllers, readers, and test identities | Bench script with expected observations |
| Gate 3 | Map live equipment, service endpoints, credentials, and physical door states | Site configuration record |
| Gate 4 | Run normal, degraded, race, restart, and tamper scenarios | Executed test set with raw evidence |
| Gate 5 | Reconcile records, investigate discrepancies, and close or assign defects | Review record and owner action register |
| Gate 6 | Approve the configured system and transfer operational responsibility | Acceptance dossier and handover package |
No gate should pass merely because the floor call worked once. The buyer should see the request, response, decision, release, door report, and resulting record aligned in the same test run.
Interface contract and message semantics
Elevator call sequencing begins with a contract for meaning, not merely connectivity. The parties must define what each message represents, when it is considered received, which response states are possible, and what a timeout means. A missing response must not be silently converted into a denial, and a denial must not be represented by an empty message.
| Field or element | Buyer requirement | Test observation |
|---|---|---|
| Message type | Distinct identifiers for request, acceptance, dispatch, release, denial, and fault | Receiver records the correct type |
| Request identifier | Unique value traceable across both systems | The same value appears in correlated records |
| Identity reference | Minimum reference needed to evaluate authorization | No unrelated personal fields appear |
| Origin and destination | Explicit source zone and authorized destination floor | The released landing matches the approved mapping |
| Requested action | Clear instruction such as call, dispatch, or release | Receiver performs only the authorized action |
| Issue time and expiry | Buyer-defined validity window and authoritative clock | Expired requests are not accepted |
| Response state | Accepted, rejected, unavailable, pending, or another documented state | Reason codes are stable and distinguishable |
| Device and source identity | Endpoint, controller, and application identity sufficient for diagnosis | Events identify their source |
| Error detail | Actionable reason without exposing restricted data | Operator can route the fault to its owner |
Where a service boundary is involved, document whether the requirement concerns a local API or cloud API, who operates it, and what remains available when the endpoint cannot be reached. The contract must also cover duplicate delivery, delayed delivery, out-of-order delivery, and replay of a request whose validity period has ended.
The buyer should require written response semantics before joining equipment to a network. Connectivity alone cannot establish that a floor selection was authorized, that a car reached the intended landing, or that a door was released.
Nominal sequence from credential to landing door
The nominal test should begin with a clean system state and use a named test identity with deliberately limited floor rights. The test operator records the expected result before execution. The sequence below is a control pattern that the approved architecture may adapt, but it should not collapse distinct decisions into a single success flag.
| Step | Action | Required observation | Evidence |
|---|---|---|---|
| 1 | Present the credential at an approved origin | Reader reports the credential; controller resolves identity and permitted floors | Presentation and authorization events |
| 2 | Controller creates the access request | Request carries identity reference, origin, destination, identifier, and validity | Controller request record |
| 3 | Elevator-side service evaluates the call | Service acknowledges, denies, or identifies the request as unavailable | Ordered response record |
| 4 | Dispatch selects the required landing | Destination mapping and dispatch decision are explicit | Dispatch trace |
| 5 | Car movement or arrival is observed | An arrival event is used only if the interface exposes one | Movement or arrival source record |
| 6 | Floor access release is requested | Designated release equipment receives the authorized command | Release command and target identity |
| 7 | Door state is reported separately | Open or closed state is not inferred from the release command | Door-position or equipment event |
| 8 | Both systems close the transaction | Request, response, decision, release, and door report share a traceable identifier | Reconciled event sequence |
| 9 | Operator view updates | Display matches the source records and shows any unresolved state | Operator export or screenshot with context |
| 10 | Test identity is removed or expired | The credential no longer grants the test destination | Revocation and repeat-attempt evidence |
A successful floor access release must be distinguishable from a successful elevator call. It must also be distinguishable from an open door. This separation prevents a controller or elevator service from reporting success when the user was transported to the wrong floor or when the authorized release never occurred.
Run the nominal sequence from each relevant origin, including normal occupied periods, after-hours conditions, and a service-person scenario. The buyer chooses the conditions that can produce different authorization or dispatch decisions; it should not copy one successful route and call it complete coverage.
Failure, degraded mode, and race tests
Degraded mode operation is part of the elevator and floor access test plan, not an appendix. Before each test, approve the required state, the maximum tolerable uncertainty, the observable result, and the party responsible for recovery. A missing response must remain distinguishable from a denied request.
| Condition | Test stimulus | Required observable behavior | Evidence and owner |
|---|---|---|---|
| Link unavailable before request | Disconnect the approved integration path | The system does not fabricate a release; the user receives the approved unavailable result | Link fault, endpoint response, repair owner |
| Link lost after authorization | Break the path between authorization and dispatch | The pending state is visible and does not become an undocumented grant | Correlated request and fault record |
| Access server unavailable | Stop the relevant server or service | The selected centralized or standalone behavior is applied as approved | Configuration evidence and outcome |
| Controller unavailable | Isolate or restart the controller | No stale authorization is accepted after the approved expiry | Restart trace and repeated attempt |
| Duplicate credential presentation | Repeat the same valid request | The system follows the documented duplicate policy without unintended extra dispatch | Pair of request records |
| Simultaneous requests | Present two identities at different controlled points | Each request remains correlated to its own destination and decision | Concurrent traces |
| Out-of-order response | Delay an earlier response until after a later one | The receiver does not apply a stale response to the wrong request | Ordered source timestamps |
| Stale floor grant | Restore a previously valid grant after its validity window | The grant is rejected or refreshed only through the approved path | Expiry and policy trace |
| Door already open or obstructed | Open or obstruct the controlled door | The system reports the physical state separately from the command | Door-position and command records |
| Service or manual mode | Place the elevator or landing in an approved service state | Normal release logic does not override the service authority without approval | Mode record and operator trace |
For a centralized design, determine which decisions live on the server and what the controller can still enforce when the server is unreachable. For a standalone design, identify what remains on the device, how its rules are updated, and how updates are audited. The buyer should not accept a design that says only “the door fails locked”; that is an approval question, not an engineering default.
CONFIRM-WITH-LOCAL-FIRE-EGRESS-OWNER: Confirm the required behavior for each unavailable, restart, service-mode, and emergency condition against the approved installation design.
Clock, event, and record reconciliation
The access event audit log must be reconciled with the source equipment record. An audit log records an event with a time and source; event memory is a buffer a lock keeps when it cannot reach the server. The two serve different purposes, and neither proves delivery of an event that was never received.
One organization should own the authoritative clock for correlation. The test record must show how each endpoint’s time is checked, how drift is handled, and what buyer-approved tolerance applies. When timestamps conflict, the reconciliation method must identify the original source rather than silently rewriting the event.
| Evidence artifact | What it must prove | Reconciliation check |
|---|---|---|
| Source device record | The credential, request, command, or door state originated at the stated endpoint | Device identity and local time |
| Controller decision record | Why the controller accepted, denied, delayed, or failed the request | Reason code and request identifier |
| Elevator or service record | How the call, dispatch, or release was handled | Transaction identifier and source sequence |
| Door-state record | The controlled door’s separately reported state | Target identity and event source |
| Consolidated audit record | How the system presented the combined event | Ordering, deduplication, and correlation |
| Operator evidence | What the user or operator was told or shown | Export linked to the same test run |
| Test record | Which precondition, action, and expected result were used | Script revision and witness sign-off |
A screenshot without source data is supporting evidence, not a complete record. The acceptance pack should preserve raw exports where available, the query or export method, the relevant configuration, and a reference to the corresponding test scenario.
CONFIRM-WITH-LOCAL-DATA-PRIVACY-OWNER: Confirm the approved purpose, operator access, retention, disclosure, and regional privacy conditions for call, credential, and door-event records before test data contains personal information.
Fire, egress, privacy, and security boundary checks
Keep ordinary access logic separate from approved emergency behavior. A request-to-exit signal reports that an exit device was operated from the escape side; it is not a release command. A door position contact reports open or closed, not the cause. The buyer should determine whether either signal belongs to this interface and require the approved interpretation rather than inferring one.
CONFIRM-WITH-LOCAL-FIRE-EGRESS-OWNER: Confirm the approved source, destination, authority, and reset behavior for every fire or emergency signal included in the elevator and landing-door interface.
Test data should use named test identities and only the personal fields required by the scenario. Operators should receive access according to their assigned function, while exports, screen views, and test reports should avoid exposing unrelated credential or movement data. Review these controls with the client’s privacy owner before populating a production-like dataset.
CONFIRM-WITH-LOCAL-SECURITY-APPROVER: Confirm the approved authentication, authorization, retention, and disclosure rules for integration service identities, test credentials, and event logs.
| Security check | Test method | Pass evidence |
|---|---|---|
| Replayed request | Re-submit a captured request after its approved validity period | No new release occurs from the replay |
| Altered destination | Change the destination in an intercepted test request | The system rejects or safely resolves the alteration |
| Unauthorized floor policy | Attempt a floor outside the test identity’s rights | A policy reason code identifies the denial |
| Interface privilege change | Use a restricted test role to change floor policy | The change is denied and recorded |
| Service interruption | Make an endpoint unavailable during the request | The approved unavailable state occurs without a false release |
| Log disclosure | Review operator views and test exports | Restricted personal data is absent or masked as approved |
CONFIRM-WITH-LOCAL-ELEVATOR-AUTHORITY: Confirm which local authority or service authority must review connected elevator equipment and what evidence it requires before operational handover.
Running the test room and preserving evidence
Integration acceptance criteria should be written before the test session and linked to each executable script. The test lead starts from a documented configuration, confirms the test identity’s rights, executes one scenario, records the result, and restores the baseline before moving to the next scenario.
| Run phase | Required action | Record |
|---|---|---|
| Prepare | Freeze configuration, assign the script, verify time, and name the witnesses | Run header and baseline reference |
| Execute | Perform the script without skipping intermediate observations | Ordered source events and operator observations |
| Reconcile | Compare request, response, decision, release, and door state | Correlation worksheet |
| Recover | Restore the approved baseline and confirm that no test grant remains | Reset and cleanup evidence |
| Close | Classify discrepancies, assign owners, and link the raw evidence | Test case disposition |
Use a consistent evidence index rather than free-form filenames. Each item should identify the scenario, environment, endpoint, source record, tester, and relationship to the acceptance criterion. Preserve raw logs before filtering or reformatting them for presentation.
A change to controller configuration, network policy, credential rights, interface mapping, or equipment state can invalidate affected results. The owner must identify the regression subset and rerun it. The buyer should not permit a failed scenario to disappear because a later demonstration succeeded.
Defect classification, ownership, and retest
Defects should be classified by observed impact, not by which company encountered them. A release to the wrong floor or acceptance of an unauthorized floor is blocking because the authorization outcome is wrong. A missing reason code or incomplete operator view may be major but not safety-blocking, depending on the approved workflow.
| Class | Example observation | Required treatment |
|---|---|---|
| Blocking | Wrong destination, unauthorized release, untraceable request, or behavior outside the approved boundary | Stop the affected scenario and gate; escalate to the accountable interface owner |
| Major | Incorrect timeout state, missing source record, duplicated event, or unexplained difference between interfaces | Assign a corrective action and require a focused retest |
| Minor | Ambiguous wording, incomplete field description, or nonessential presentation defect | Record in the owner action register with the agreed verification evidence |
| Observation | Environment limitation or unapproved use case outside the baseline | Confirm scope and decide whether another gate is required |
The owner action register records the symptom, expected result, actual evidence, root cause, corrective action, accountable party, affected scenarios, and retest reference. A status update without new evidence is not closure. The retest must reproduce the original conditions closely enough to show that the cause, not merely the symptom, has been addressed.
Where a defect touches emergency behavior, the relevant local approver must review both the corrective action and the retest before the gate can close.
Acceptance dossier and operational handover
The final dossier should let an independent reviewer reproduce the decision without relying on the people who ran the demonstration. It must distinguish a claimed capability, a configured behavior, an observed result, and an approved operating rule.
| Dossier component | Acceptance question | Signatory |
|---|---|---|
| Approved baseline | Does it match the use cases, exclusions, equipment, and owners being accepted? | Client operating owner |
| Interface contract | Does it define messages, states, time, expiry, and failure behavior? | Integration lead |
| Nominal test record | Does every approved journey show the request through the door-state report? | Test lead |
| Degraded test record | Does every interruption produce the approved visible outcome? | Design authority |
| Reconciliation worksheet | Can both source systems be tied to the same transaction? | Operations or security reviewer |
| Local approval record | Are emergency, elevator-service, privacy, and security questions answered by their owners? | Relevant designated authorities |
| Defect and risk record | Is every unresolved issue visible with an owner and verification method? | Project sponsor |
| Handover package | Can the operating team detect, diagnose, and recover from the tested failures? | Operations owner |
For a portfolio, adapt the gates through Elevator and Floor Access for a Multifamily or Office Portfolio. Use Elevator and Floor Access Cost and Effort Drivers Without Quoting Prices to document effort-bearing interface choices, and complete the handover with Documenting Elevator and Floor Access for Handover and Audit.
It is worth pairing this with smart locks for office buildings. That is the same question elevator and floor access test plan has to answer here.Approval should be an explicit decision. Any unresolved issue must state its operational effect, accountable owner, verification method, and the condition under which it may be revisited. A silent deferral is not a recorded risk decision.
Final buyer questions before releasing the gate
- Which exact event requests an elevator call, and which event authorizes a floor access release?
- Does the interface expose an arrival indication, or would arrival be inferred?
- Who owns the authoritative clock and the correlation identifier?
- What remains available when the server, controller, network path, or service endpoint is unavailable?
- Which failures fail visibly, and which states require local approval?
- Can every decision be reconciled across source equipment, controller, elevator service, and audit record?
- Which test identity, floor, door, and service mode did each result cover?
- Who owns every open item in the owner action register?
- What evidence will the operating team receive at handover?
- Which change requires a full regression set rather than one repeated demonstration?
Require the supplier and incumbent service parties to return the completed test plan, named owner matrix, executable scripts, and evidence index for review. Do not release the final gate on a roadmap statement or a single successful demonstration. The buyer’s acceptance package is the commercial control: reproducible results, clear responsibilities, and a documented route to correct anything that fails.
The commercial step that follows is to request a quote.

