SLK-9936S Door Thickness Compatibility Guide
The published specification for the SLK-9936S hotel smart door lock states a compatible door thickness of 36–100 mm. That single line decides more of a hotel lock project than most buyers expect. It gates which model goes on which door, how many line items end up in the RFQ, and whether a retrofit program runs as one standard order or splits into populations. Everything else in this guide exists to answer two questions engineering-style: does each of your doors actually sit inside 36–100 mm, and what exactly do you do about the ones that do not?
Not "hope the installer fudges it." Not "the datasheet says it covers almost everything." A measured answer, per door, recorded on paper you can send to a supplier.
1. The Compatibility Rule, Stated Once
Direct answer: the SLK-9936S is published for doors 36 mm to 100 mm thick. Doors measuring inside that band pass the first fit gate; doors measuring at the band's edges (roughly 35–37 mm or 99–101 mm) must be re-measured under a strict protocol before anything is ordered; doors outside the band are not automatic disqualifications for the program — they are a confirm-per-model exception that goes to the supplier with the actual measurement, before quotation.
Three consequences follow, and they govern the rest of this guide:
- Compatibility is a per-door fact, never a per-property assumption. A building does not fit a lock. A population of measured doors fits a lock.
- The envelope is a published specification, not a promise about your door. Panel construction, edge condition, warp, and hardware geometry are your side of the test; the 36–100 mm line is the manufacturer's side. Both must meet.
- Beyond the envelope, silence is the correct default. No supplier owes you a fit answer for a door nobody measured. No buyer should accept a "probably fine" answer for a door that was measured.
2. Why One Number on the Datasheet Controls the Whole Project
Door thickness is the first spec every serious hotel-lock buyer is told to check because the lock physically brackets the door. The through-body fastening, the case, and the cylinder all span the door's thickness dimension. The SLK-9936S block publishes a 5-pin cylinder and a 3-point latch; both are sized to engage correctly within the stated 36–100 mm band. Miss the band and the failure modes are mechanical and predictable:
| Symptom at the door | What the wrong thickness usually produces | Where it shows up first |
|---|---|---|
| Door thinner than the band | Loose bracketing, case rocking, cylinder end-play, sloppy latch throw | Fitting stage, then guest-visible rattle |
| Door thicker than the band | Fasteners that do not reach, forced alignment, latch not clearing the strike plate | Fitting stage, immediate — the lock will not seat |
| Borderline thickness with variance | Intermittent read or latch drag caused by mounting strain, not electronics | Weeks after opening, blamed on the wrong subsystem |
The last row is the expensive one. A lock installed under strain on a 35 mm door will pass its own RFID test — the Mifare 1 reader at 13.56 MHz with its published 0–5 cm read range does not care how happy the mounting is — and then misbehave in the field where everyone misdiagnoses it as a reader problem. Door-fit errors launder themselves into electrical complaints. This is why the survey protocol in Section 5 is worth the labor, and why the audit/event memory (up to 1,000 records per the published block) is more useful when the mounting is honest enough to make the records trustworthy.
3. Reading the 36–100 mm Line Like an Engineer
The published line commits to a thickness range. Careful buyers learn to notice what it does not say, because those gaps become supplier questions rather than assumptions.
| What the published line covers | What it does not cover — and how to handle it |
|---|---|
| Door thickness between 36 mm and 100 mm | Door width, height, and edge/stile geometry. A thick rail at the lock edge still has to accept the case and cylinder bore; verify stile width per door during the survey. |
| The nominal flat band | Variance within one door. Warped or bowed doors can span several millimetres between top and bottom. The protocol measures each door at multiple heights for exactly this reason. |
| A door that is a uniform slab across the lock region | Sandwiched, rebated, or hollow sections at the lock edge where a caliper on the face reads something the mounting cannot use. Note construction type in the register. |
| Mechanical fit only | Everything else about the assembly — frame condition, strike alignment, existing bore sizes. Those are installation-scope items, handled in the sibling guides; this document owns thickness. |
One more engineering note: thickness is measured through the door, not across the whole door-plus-frame-and-gap. It sounds trivial. Surveyors new to the job get it wrong, and a register contaminated with "gross opening" readings will misroute entire floors into the wrong decision bucket.
4. Survey Methodology: Populations, Not Properties
The unit of hotel-lock compatibility is the door population: a group of doors sharing construction, age, and likely geometry. A single property routinely contains three to five populations, and a competent survey names them before it counts them.
| Project situation | Typical populations you will find | Survey coverage rule |
|---|---|---|
| New construction, one door schedule | One engineered population | Measure a validated sample per door type; full census rarely needed, but keep the sample documented |
| Renovation, rooms completed in different years | Two to four populations split by fit-out phase | Full census of every room door; samples fail here because phase boundaries are invisible from the room number |
| Like-for-like lock replacement on legacy doors | One aging population plus its outliers | Full census; legacy doors wander, and the outliers are the program risk |
| Mixed estate (main building + annexes, resort wings) | One population per building vintage | At minimum, census each building's first and last room and every visually different door; realistically, census all |
| Heritage or listed-building doors | A population you cannot change | Full census, thickness and stile width both, because this is where confirm-per-model decisions decide the architecture |
Method rules that separate a usable survey from a decorative one:
- Measure doors, not room types. Two rooms identical on the plan can carry doors 6 mm apart. The register's primary key is the room/door ID.
- One surveyor, one shift, one tool set per building — or a documented cross-check between surveyors. Inter-observer bias is real at the band edges.
- Photograph the outlier, not the average. Every door that reads at the edges or looks like a different construction gets a photo filed against its register row.
- Close the loop before the RFQ. A survey with 100% of doors recorded and 0% of edge cases re-checked is a survey nobody did.
5. The Measurement Protocol
This is the part most projects skip. It takes ninety seconds a door and it is the difference between a defensible register and a guess with decimals.
Tools. A vernier caliper is the primary instrument. A steel rule or rigid tape is the fallback and must be flagged in the register as a lower-confidence reading. Folding rules and loose tapes deform; on a 40 mm door they will eat 1–2 mm without complaining.
Procedure, per door — record every value, never a range:
- Measure with the door closed, at the lock edge (the edge the hardware sits on), never across the top or bottom edge where stiles and rails differ.
- Take readings at three heights on the lock edge: about 300 mm from the bottom, mid-height, and about 300 mm from the top. Thickness on a healthy door is constant; a spread of more than ~1 mm across heights is a warp or construction note, not a number to average away.
- If the door has moulded panels or a rebated edge, take the reading at the exact hardware region — where the case and cylinder will actually sit.
- Record the minimum and maximum readings per door plus the tool used. The decision in Section 7 runs on the extremes; averaging hides a bow right past the edge of the band.
- Any reading that lands in the borderline zones (see Section 7) gets a second person, second tool, fresh measurement on a different day. Two disagreeing borderline readings are a re-survey trigger, not a rounding opportunity.
- Paint and lacquer count as door. Do not scrape readings "to bare wood" unless the survey brief explicitly covers stripped dimensions; you order locks for the door that exists.
| Protocol element | Pass condition | Fail condition → action |
|---|---|---|
| Position | Closed door, lock edge, hardware region | Reading taken on face or top edge → void and redo |
| Coverage | Three heights per door | One "obvious" reading → void and redo |
| Consistency | Spread across heights ≤ ~1 mm | Spread larger → flag warp, add construction note + photo |
| Borderline re-check | Two independent readings agree within ~1 mm | Readings disagree → escalate to verify in writing-per-model with both values disclosed |
| Traceability | Tool, date, surveyor initials recorded | Register row without provenance → row is advisory, not quotable |
6. Recording the Survey: The Door Register
The register is the deliverable. It travels to the supplier, it justifies the model split, and after installation it is the document that answers a warranty argument. A workable row template — fill it in per door:
| Room / door ID | Building / floor | Reading min (mm) | Reading max (mm) | Tool (caliper/rule) | Stile width noted? | Construction note | Photo filed? | Decision bucket |
|---|---|---|---|---|---|---|---|---|
| EXAMPLE — 214 | Block B / Fl 2 | 42 | 42.5 | Caliper | Yes | Standard solid core | No | In band |
| EXAMPLE — 307 | Block B / Fl 3 | 35 | 36.5 | Caliper | Yes | Original 1990s door, slight bow | Yes | Borderline low — re-check |
| EXAMPLE — GF-11 | Annexe | 101 | 102 | Rule | Pending | Reclaimed heavy panel | Yes | Out of band — confirm per model |
The decision bucket column is filled by the rule in the next section, mechanically. Surveyors who let judgment creep into bucketing produce registers where the bad news has been smoothed into "in band." It should not be smoothed. It should be bucketed.
7. Sorting Results Into Decision Buckets
Every measured door lands in exactly one of five buckets. The actions differ, and the escalation rules are deliberately asymmetric: the burden of proof always points toward the exception path.
| Bucket | Register condition (min–max per door) | Action | Owner before RFQ |
|---|---|---|---|
| In band | Both extremes ≥ 36 and ≤ 100 | SLK-9936S proceeds on the published envelope | Buyer, documented in register |
| Borderline low | Any reading 35–<36 mm | Re-measure per Section 5; if confirmed, move to exception path | Second surveyor, then supplier |
| Borderline high | Any reading >100–≤101 mm | Re-measure per Section 5; if confirmed, move to exception path | Second surveyor, then supplier |
| Out of band | Confirmed reading <35 mm or >101 mm | Never quote against the published envelope; raise as a named exception with actual measurements | Supplier confirmation Confirm per model |
| Unmeasurable | Door geometry defeats a clean reading (rebates, compound profiles) | Escalate as an engineering exception; physical sample fitting decision only | Supplier + installer jointly |
Direct answer for the two borderline zones: a door that reads 35–37 mm or 99–101 mm is not "close enough" and not "failed." It is a re-measurement trigger. Half of borderline readings evaporate under a caliper, three heights, two surveyors. The half that survives is a real exception and belongs in Section 8, with numbers attached.
Follow one door from the caliper face to either an RFQ line or a written exception, and the bucketing rule becomes a path nobody has to interpret:
flowchart TD
A["Survey row closed · min and max recorded · three heights on the lock edge"] --> B{"Both extremes inside the published 36-100 mm band?"}
B -->|"yes"| C["In band · proceed on the published envelope"]
B -->|"no"| E{"Where does the reading sit?"}
E -->|"35-37 mm or 99-101 mm"| F["Borderline · re-measure trigger"]
E -->|"below 35 mm or above 101 mm"| G["Out of band · never quote blind"]
E -->|"rebates · compound profiles"| H["Unmeasurable · engineering exception"]
F --> I["Second surveyor · second tool · different day"]
I --> J{"Both readings agree?"}
J -->|"yes"| K["Confirmed exception · both raw values disclosed"]
J -->|"no"| L["Re-survey trigger · row is advisory not quotable"]
L --> F
C --> M["Register rows cited as evidence in the RFQ"]
K --> N["Supplier package · register extract + exception file + written question"]
G --> N
H --> N
H --> O["Physical sample fitting decision only · supplier and installer together"]
O --> N
N --> P["CONFIRM PER MODEL answer in writing before quotation"]
P --> Q["Separate line item or a deliberate reroute of those doors"]
M --> R["Program order line"]
Q --> R
R --> S["Pilot on the worst-borderline population before the program order"]
8. Beyond the Envelope: Confirm, Don't Assume
Doors confirmed outside 36–100 mm are where B2B discipline separates from catalogue optimism. The honest engineering position has four parts.
First, published facts stop at the edge. Within this catalog, the two thickness envelopes I can cite are the SLK-9936S at 36–100 mm and the SLK-9905A hidden-reader hotel lock at approximately 38–100 mm (both PDP-published). Note what that comparison already does for you: on a genuine 36.5 mm borderline door, the two published bands differ at the low end, and a door that is borderline on one model sits inside the other. Model selection is population arithmetic, not habit.
Second, every other path is a per-model question, not a per-catalog question. Thin doors below 36 mm, heavy doors above 100 mm, glass assemblies, exterior aluminium doors — the catalog contains categories for these uses (the site publishes 10 outdoor models and 3 glass models among its 79 models, and a 6-model cabinet line shows the factory thinks in door geometries, not just room types), but no fit claim transfers across categories by association. A thin legacy door does not become compatible with a hotel lock because the factory makes locks. It becomes a candidate for a model whose own published envelope covers it — confirmed per model, in writing, before quotation.
Third, exceptions change program shape, not just line items. An exception population forces decisions about credential policy (do the odd doors share cards with the standard population?), spares strategy (an entire extra model's worth of batteries and shells?), and software (the published multilingual management software and the standalone-or-centralized database modes both run per lock, so a mixed fleet is a mixed-configuration fleet). Buyers weighing system architectures should read Standalone vs Centralized Hotel Lock Systems: Which Architecture Fits Your Property as the companion decision.
Fourth, never let "adapter" become an assumption. Door-thickness adapter kits are a common accessory conversation in this product class. They are also a classic source of invented specifications. If your register produces out-of-band doors and someone offers adapters, the adapter's dimensional range and its effect on the latch and cylinder path become a written supplier confirmation item, tied to your actual measurements.
| Door population found | Published-catalog status | Required action |
|---|---|---|
| 36–100 mm (confirmed) | Inside SLK-9936S envelope | Proceed; cite register rows |
| ~38–100 mm (confirmed) | Inside both cited bands | Proceed; 9936S; 9905A available where hidden reader matters |
| 36–38 mm only | Inside 9936S band; below the 9905A cited band | Proceed with 9936S; do not assume 9905A fit on this sub-population |
| <36 mm or >100 mm (confirmed) | Outside both published envelopes | Confirm per model with measured extremes disclosed; expect a different model family or engineering exception |
| Glass, exterior aluminium | Different categories entirely (3 glass, 10 outdoor models published) | Route to those categories as their own survey; envelope numbers from the hotel line carry zero authority there |
9. Packaging the Survey for the Supplier
A register with no voice reaches the factory as noise. The supplier package is small and specific:
- Register extract, machine-readable — the Section 6 table, all rows, buckets filled.
- Population summary — count of doors per bucket, per building.
- Exception file — every borderline and out-of-band door with both raw readings, tool used, and photo reference.
- The written question — for each exception: model recommendation against these exact dimensions, and any constraint on cylinder, latch, fixing, or adapter for these doors.
- The context — credential and system mode you intend (standalone vs centralized), so the answer comes back as a coherent configuration.
That package is what a quotation request should look like when thickness risk exists — it is the same door-reality evidence the Hotel Smart Lock Project Procurement Checklist demands before its own sign-off lines. Which is most orders, honestly — because most hotel orders are retrofits, and retrofits are door archaeology.
10. An Illustrative Scenario: 88 Rooms, Three Populations
Consider a hypothetical 88-room coastal property — a 1990s main block of 64 rooms, a 2010 renovation floor of 16 rooms, and an 8-room annexe of reclaimed solid doors. A like-for-like replacement project assumed "one lock for all rooms" before the survey.
The caliper version found three populations. The renovation floor measured a uniform 42–43 mm — squarely in band. The main block measured 38–41 mm across most doors, but six doors read a 2 mm spread across heights (bow), and two of those dipped to 35.5 mm at one corner: borderline-low, re-checked with a second surveyor, confirmed. The annexe, which looked identical to the main block from the corridor, averaged 96 mm with three doors reading 101–102 mm: two borderline-high (confirmed out on re-measure) and one clean out-of-band.
The RFQ that followed had three line items, not one: 64+16 doors quoted against the published 36–100 mm envelope; the two confirmed 35.5 mm doors and the two confirmed 101+ mm annexe doors raised as named exceptions with raw readings disclosed; the remaining annexe door kept as an in-envelope line but photographed for install planning. The property's buyer got a worse-looking spreadsheet in week one and avoided a worst-case scenario in month three: a shipment fitted to a lie. The exceptions either resolve through a supplier-confirmed path or they route those five doors to different hardware deliberately — which is a program decision made with data instead of an installation failure made with hope.
11. Survey Failure Modes to Refuse
Five patterns break thickness projects, and each is cheaper to prevent than to repair:
The single-reading survey. One measurement per door is not a survey of doors, it is a survey of that caliper's placement. Three heights, extremes recorded. Non-negotiable.
The averaged register. Columns that store a mean instead of min and max delete the exact information the edge cases live in. Buckets run on extremes.
The sample extrapolated across renovation phases. A 10% sample is a tool for homogeneous populations. If the building was fit out in phases, the population boundary is invisible from the room numbers and the sample will silently miss it.
The verbal borderline pass. A door that re-measures at 35.5 mm does not "fit closely enough." Borderline confirmations become field failures, and field failures come back to blame the lock.
The self-certified out-of-band path. If a buyer quotes doors outside the published envelope anyway, the resulting failures are commercially indefensible. The confirm-per-model loop is the protection on both sides of the transaction.
12. The Compatibility Sheet, in One Table
| Question | Answer | Evidence needed |
|---|---|---|
| What door thickness does the SLK-9936S publish? | 36–100 mm | PDP specification line (cited, dated to this guide's reading) |
| What do we do with in-band doors? | Proceed; register rows as evidence | Min/max per door, three heights, tool recorded |
| What about 35–37 or 99–101 mm? | Re-measure twice, independent tool + surveyor | Both raw readings, photo, construction note |
| Confirmed below 36 or above 100? | Never quote blind; Confirm per model | Exception file sent with register extract |
| Other model envelopes? | Only two are cited here: 9936S 36–100 mm, 9905A ~38–100 mm | Anything else is a supplier question, not a blog claim |
| Glass or exterior doors? | Separate categories (3 glass, 10 outdoor models published), separate surveys | Zero envelope transfer from the hotel line |
Door thickness is the cheapest engineering gate in a hotel lock project and the one most often paid for after the fact. Measure in populations, record in extremes, bucket by rule, and let nothing outside the published band reach a purchase order without a written, per-model answer.
When your register is closed and the buckets are filled, the exceptions and the standard populations become two different conversations with the same supplier — start both through the quote request with your register attached, and pull a physical pilot on a real door from your worst-borderline population through order-samples before the program order. The lock-side companion — full buyer walkthrough of the SLK-9936S — and the door-population angles covered in the replacement-program stage-gate guide and the motel exterior-door resource pick up where the caliper leaves off.

