Quick Answer
A reliable 50-100 can snus vending mechanism should be designed from the measured package envelope, not from a catalogue photo or nominal can diameter. The mechanism must separate exactly one can, prove that it passed through the delivery path, prevent a customer from reaching stock, preserve SKU identity during refill, and recover safely when the result is uncertain.
For a recent Norway buyer brief, the requested capacity sounded modest: approximately 50 to 100 cans, individual dispensing, touchscreen selection, age verification, payment, and remote stock monitoring. In engineering terms, capacity is the easy line. The difficult line is “individual dispensing.” Round, shallow cans can bridge, slide in pairs, hang on lid edges, change friction with labels or condensation, and roll into a sensor in ways that create a false success.
This guide focuses on that mechanical and control problem. It does not change the legal conclusion for Norway. The Norwegian Directorate of Health says consumer self-service of tobacco is prohibited, including self-service machine sales, apart from the described prepaid-card or equivalent retrieval route. The exact sales model, employee supervision, concealed display, registration, and authorization method must be approved before the mechanism is placed in service.
Begin With a Product Envelope, Not One Sample
One perfect can from the buyer’s desk is not a design basis. Ask for production samples from every intended SKU, supplier, packaging revision, and storage condition. Measure enough units to find variation, then document the smallest and largest credible package rather than copying the nominal dimensions from a marketing sheet.
| Product input | What to measure or observe | Why it changes the mechanism |
|---|---|---|
| Body and lid | Minimum and maximum diameter, total height, lid overhang, rim or seam profile | Controls lane clearance, gate depth, and hang-up risk |
| Mass | Lightest and heaviest finished can | Changes gravity feed, motor load, impact, and sensor response |
| Surface | Label stock, wrap seam, coating, embossing, friction, loose edges | Changes sliding and stacking behavior |
| Condition | Cold, warm, dry, damp, slightly dented, lid not fully seated | Reveals conditions that can turn a passing lane into a jam |
| Commercial identity | SKU, flavor or variant, lot, date, supplier, approved storage | Prevents a mechanically compatible but unauthorized substitution |
Define the validated product envelope as the dimensional and physical range the mechanism is allowed to vend. If a future can falls outside that envelope, it requires assessment and perhaps a new insert, gate, or test. “It looks almost the same” is not configuration control.
Capacity Is a Lane Decision
A request for 50-100 cans is incomplete until the buyer states the number of SKUs and the minimum stock per SKU. A cabinet holding 100 identical cans is mechanically simpler than one holding ten SKUs with ten cans each. More lanes consume width, structure, motors, sensors, wiring, service access, and unusable clearance.
A practical first estimate is:
Usable capacity = sum of validated lane capacities – service reserve – test or quarantine positions.
Do not count every geometric void as saleable capacity. The last can may sit outside the gate’s controlled range. A lane may need a follower, low-stock flag, or loading clearance. The delivery elevator may block one position during service. The buyer should receive both nominal capacity and validated usable capacity by approved can family.
SKU mix also drives refill labor. Fixed lanes make identity clear but leave unused capacity when one product sells faster. Adjustable lanes improve flexibility but introduce setup error. Removable cartridges can speed refill and reduce hand contact, yet each cartridge needs positive machine and SKU identification. There is no free flexibility; every adjustment needs a controlled setting and a verification method.
Compare Mechanisms Against the Actual Can
| Mechanism | Strength | Typical risk with round cans | Best use |
|---|---|---|---|
| Gravity column with escapement | Compact, low power, simple refill | Double release, rim hang-up, stack pressure, friction variation | Consistent can family and sufficient vertical height |
| Motorized pusher lane | Controlled stroke and clear lane ownership | Can tilts, pusher rides over product, residual pressure after a jam | Horizontal shelves with stable guides |
| Short belt lane | Gentle transport and adjustable motion profile | Rolling, belt contamination, slip, product arriving in pairs | Mixed package surfaces when sensors can separate motion |
| Indexed carousel | Positive positions and good security | More parts, larger footprint, difficult refill reconciliation | Higher-value or tightly controlled selections |
| Lane plus delivery elevator | Reduces drop and centralizes customer delivery | Transfer alignment, elevator occupancy, two-stage jam recovery | Cabinets with several levels or protected delivery |
For many compact designs, a gravity or pusher lane can work, but only after the gate geometry is tuned around the real lid and rim. A round can should not rely on a spiral pitch chosen for snack bags. If the can can rotate into an unstable attitude, the guide surfaces should control that motion before the release point.
The mechanism should also avoid scraping, opening, denting, or obscuring required packaging information. Contact surfaces need suitable materials, rounded transitions, and access for cleaning. Tobacco may not be a food-contact application, but dust, damaged labels, and debris can still degrade reliability and product presentation.
Singulation Means Exactly One Can
A single-vend gate has to perform two jobs at once: release the first can and restrain the second. In a gravity stack, the weight of every can above acts on that boundary. If the first can moves only halfway, the second can wedge behind it. If the gate opens too far or rebounds, two cans may escape.
Specify the sequence in physical terms:
- The lane confirms that the gate is home and that the delivery path is clear.
- The controller reserves one saleable unit from the selected lane.
- The restraining feature captures the second can before the release opening is available.
- The actuator completes a controlled stroke rather than an unverified timed pulse.
- A passage sensor sees one credible product event inside the permitted time window.
- The gate returns home and the mechanism proves its home state.
- The customer bin independently confirms arrival or collection according to the design.
One beam break is weak evidence. A vibrating can can trigger the same beam twice, two cans touching can look like one long event, and a hand or service tool can create a false signal. Combine actuator position, lane state, timed passage, and delivery-bin evidence. The combination should distinguish success, no-vend, double-vend suspicion, and unknown outcome.
Use a Dispense Evidence Matrix
This evidence matrix is an OBO engineering framework for the project team. It forces the software result to match what the mechanism actually proved.
| Observed evidence | System state | Customer and inventory action |
|---|---|---|
| Gate stroke complete, one passage event, bin arrival confirmed | Confirmed vend | Commit one unit and complete payment |
| Actuator moved, no passage, lane still reports product | Probable jam or no-vend | Stop lane, do not silently retry, follow reversal rule |
| Passage event, no bin arrival | Product in delivery path | Lock path, retain transaction context, request service |
| Long or repeated passage event | Double-vend suspicion | Block further sales, flag count variance and incident |
| Bin event without an authorized stroke | Tamper or sensor fault | Lock machine or affected path and investigate |
| Power loss during motion | Unknown outcome | Recover state before accepting another sale |
Do not label an uncertain outcome as success because payment was approved, or as failure because the cloud did not receive a callback. The local controller should preserve the event through reboot and reconcile it when power or connectivity returns.
Design the Delivery Path Against Fishing and Roll-Back
The customer should reach the delivered can, not the storage lane. Use a baffled or interlocked bin that prevents a hand, hook, or flexible strip from reaching the release point. The inner door and outer collection door should not expose a continuous path at the same time. If an elevator is used, it needs a defined safe position before the customer door unlocks.
Round cans introduce a roll-back problem. A shallow slope or door impact can send the product away from the collection point. Shape the bin so the can settles where it can be seen and reached, while water, dust, and small debris do not collect around sensors. Test with the lightest and heaviest approved cans and with the machine slightly out of level within the installation tolerance.
Service access needs equal care. A technician clearing a jam should not release stack pressure unexpectedly or bypass the inventory record. The service mode must disable customer authorization, identify which lane was opened, and require reconciliation of removed, returned, damaged, and test-vended cans.
Make Refill Error Difficult
A beautiful dispensing mechanism can still sell the wrong product if the refill workflow is loose. Label each lane with an internal ID, not promotional graphics visible to the customer. The dashboard, service screen, route list, and physical lane should use the same identifier. Scan or confirm the SKU when the lane is assigned and again when stock is loaded.
Record starting count, added quantity, removed quantity, damaged units, test vends, and ending physical count. Use FEFO or the operator’s approved date rule where relevant. If two cans share dimensions but differ in product identity, mechanical fit must not authorize substitution.
Low-stock detection can be estimated from confirmed vends, measured by a lane sensor, checked by weight, or reconciled through a cartridge count. Each method has uncertainty. The dashboard should show whether a number is calculated, sensed, or physically verified. Avoid displaying false precision.
Connect Authorization, Payment, and Motion Carefully
The mechanism receives permission to release one unit; it should not decide whether the customer is legally eligible to buy. In a regulated project, the approved sales and age-control workflow produces a short-lived, one-item authorization. The payment system produces a separate financial authorization. The local controller then checks that the selected lane, product, gate, sensors, and delivery path are ready.
A robust transaction uses explicit states: authorized, reserved, payment authorized, dispensing, passage confirmed, delivered, committed, or recovery required. A timeout does not automatically move the transaction to the next state. Repeated API calls should be idempotent, so a retry cannot produce a second can.
When a jam occurs after payment authorization, follow the payment provider’s supported reversal or refund process. Do not improvise by opening another lane unless the operator has approved that substitution and the customer has explicitly accepted it. For Norway, the supervised legal workflow described in the related compliance guide remains the controlling gate.
Bench-Test the Mechanism Before Building the Cabinet
Start with an adjustable test rig that reproduces one lane, the gate, relevant stack height, sensors, and delivery transition. This is faster to modify than a finished cabinet. Use real cans from the buyer, including dimensional extremes and credible packaging defects.
Challenge variables one at a time, then in combinations:
- minimum and maximum can dimensions and mass;
- full, half-full, two-can, and last-can lane conditions;
- label seams facing different directions;
- cold, warm, dry, and permitted damp-surface conditions;
- slightly tilted installation and realistic vibration;
- rapid repeated sales and long idle periods;
- low supply voltage, reboot, communication loss, and actuator timeout;
- blocked outlet, restrained product, empty lane, and deliberately introduced double-release risk;
- normal refill, wrong SKU, overfill, poor stacking, and service-door interruption.
The protocol should define the number of cycles and confidence required for the project. There is no honest universal cycle count. Wrong-SKU and double-vend events may have zero tolerance, while an agreed recoverable jam rate is a commercial and operational decision. Record every failure with the can, lane setting, event trace, video or photo where useful, cause, correction, and retest.
Move From Bench Proof to Integrated Acceptance
| Stage | What is proven | Do not advance until |
|---|---|---|
| Product characterization | Validated envelope and SKU families | Worst-case samples and storage conditions are available |
| Single-lane bench rig | Singulation, gate force, sensors, basic recovery | Failure modes are understood and repeatable |
| Multi-lane module | Lane spacing, wiring, cross-lane effects, refill | SKU mapping and service access remain controlled |
| Cabinet integration | Delivery path, anti-fishing, doors, power, thermal environment | Installation and tamper challenges pass |
| Software and payment integration | One-item token, reservation, idempotency, reversal, logs | Successful and failed transactions reconcile |
| Supervised pilot | Real refill, customer collection, support, stock variance | Open risks have owners and approved limits |
The prototype should include a configuration record: lane width or insert, gate revision, actuator parameters, sensor positions, firmware, approved can families, and test report. Without that baseline, a later factory adjustment can quietly invalidate the passing result.
Acceptance Questions for the Supplier
- Which measured can envelope and approved SKUs does the mechanism support?
- What is the validated usable capacity for each proposed SKU mix?
- How does the gate restrain the second can while releasing the first?
- Which independent signals distinguish confirmed vend, no-vend, double-vend suspicion, and unknown outcome?
- How does the machine recover after power loss during motion?
- Can a software retry, delayed callback, or duplicated command release another can?
- How are the customer bin and stock path protected against fishing?
- How are lane identity, SKU, count, date, and service removals reconciled?
- Which parts are adjustable or replaceable when can dimensions change?
- What cleaning, wear inspection, spare parts, and preventive-maintenance intervals are required?
- What bench, cabinet, integrated, transport, and pilot tests are included in the quotation?
- What evidence will the buyer receive before quantity production?
Norway Compliance Still Comes First
Mechanical reliability does not authorize unattended tobacco sales. Norway’s official guidance says tobacco products may not be visibly displayed, self-service sales are prohibited, and the prepaid retrieval exception has neutral-design conditions. It also describes registration and written internal-control duties for sales locations. A transparent front, branded product menu, or direct customer payment-to-dispense flow shown in a generic machine photo should not be copied into a Norway specification.
Use the Norwegian Directorate of Health tobacco sales guidance and the underlying law and regulation as starting points. Have the Norwegian operator obtain advice on the exact location, supervision, sales authorization, product presentation, and token flow before the mechanical prototype is commissioned.
Related Custom Vending Guides
- Norway snus vending compliance and BankID guide
- Tobacco vending machine technology overview
- Vending machine testing before mass production
- Security for high-value products
- Inventory shrinkage and reconciliation controls
- Custom vending machine development cost
- Custom vending software integration checklist
- Custom vending compliance matrix
Send Real Cans Before Asking for the Final Quote
For a 50-100 can project, OBO needs the intended SKU list, multiple samples from each can family, measured variation if available, storage conditions, target capacity by SKU, refill interval, cabinet limits, supervision and authorization workflow, payment route, desired sensors, security level, and acceptance criteria. We can then propose the lane and delivery architecture, build a mechanism rig, document failure modes, and price the integrated prototype from evidence rather than guesswork.
A proven can mechanism is only one budget line. The Norway snus prototype quotation guide shows how to price mechanical proof, integration, cabinet build, acceptance, and pilot support separately.
A delivery sensor becomes useful only when its evidence is reconciled with authorization, payment, channel stock, and the customer outcome. Use this snus vending inventory and dispense-reconciliation guide to connect 50-100-can mechanism data with the remote dashboard.