Why does the machine say it cannot dispense an item when the column is clearly full?

Because a full column and a working column are not the same thing. The three usual causes are product-to-coil mismatch, where the pack is too wide, too deep, too tall or too soft for the coil pitch and diameter it was loaded into and it wedges instead of dropping; a worn or unsuitable spring end that no longer releases the last pack cleanly; and the machine having no means of confirming that the item fell, so a failure is recorded as a successful vend and the stock count drifts away from what is physically in the machine. All three are configuration and verification problems rather than bad luck. Fit is decided before the build by checking each product's pack dimensions against the coil type, pitch and tray height; mechanical release is decided by matching the coil form to the pack shape and by cycling every motor and coil under load before the machine ships; and confirmation is decided at quotation, by specifying an elevator build with a drop sensor or locker cells with per-cell weight sensing.

Does the machine know when a product failed to drop, or does it just charge the user?

The KioskForce transaction record carries a Refunded state for exactly this case: the selected product failed to dispense after payment, so the money is returned to the customer automatically and the order is recorded against a receipt ID that can be searched later. That is the commercial half of the answer. The physical half depends on the build: an elevator machine carries a drop sensor that detects the item arriving in the delivery bin, and locker cells can be specified with weight sensors that verify a real weight change rather than a door that opened and closed. A plain coil position without either device knows that the motor turned, not that the pack left the coil. Which verification method your machine carries is a decision made at quotation, so ask for it explicitly rather than assuming it is fitted.

Our PPE machine is always out of the popular glove size. How is that prevented?

By treating every size as its own item with its own stock count, its own par level and its own low-stock alert, and by allocating facings according to measured consumption rather than giving each size one position because there are six sizes. A single glove line with six sizes behind one product record will always appear to be in stock while the two sizes most of the workforce wears are empty. Load each size as a separate selection, let the machine record consumption per size for a few weeks, then reallocate positions to match what the site actually draws — the common sizes usually need two or three positions each and the rare sizes need less than one machine's worth of shelf. Out-of-stock trays are disabled automatically so the worker is not offered a selection that cannot be filled, and low-stock alerts from the cloud dashboard are what should trigger replenishment, rather than someone noticing an empty column.

What can dispense verification not fix?

It cannot fix a wrong planogram — it reports the failure faster and refunds the user, but a pack that does not suit the position it was loaded into will keep failing until the position is changed. Weight sensing verifies a weight change, not identity, so two sizes sharing one cell cannot be told apart by weight alone. A drop sensor confirms the item reached the delivery bin, not that it arrived undamaged or that the right person took it. And no sensor prevents a stockout caused by replenishment frequency: if the site draws more than the machine holds between visits, the answer is more positions, more facings on the fast movers or a more frequent replenishment round, not more instrumentation. Verification tells you the truth sooner; it does not change the physical layout or the logistics.

Dispense Reliability and Stockout Prevention

Q: Do you check that our actual products fit the machine before it is built?

Yes, and it is the single most valuable hour in the whole project. Send the pack dimensions and weight for every item you intend to vend, in the packaging it will arrive in, and the layout is worked position by position against coil type, coil pitch and tray height before anything is cut. You can do a first pass yourself with the planogram configurator, which lets you set trays and coils per tray, change the coil option per position, adjust coil pitch per position and set tray heights, then send the resulting layout with your enquiry. Items that will not dispense reliably from a coil in supplier packaging are identified at this stage, and the honest options are a different coil form, a locker cell, or re-boxing the item — all of which are cheaper to decide now than after the machine has been crated and shipped.

Q: What testing is done before the machine leaves the factory?

Every completed machine is powered on and operated under load before it leaves the assembly bay, with every motor and spiral coil cycled through full rotation, payment peripherals verified with test transactions, touchscreen and keypad checked, connectivity registered against the server, and door interlocks and earth continuity confirmed. A separate load test runs a weighted coil rack to verify motor torque and spiral coil performance under a full product load rather than an empty one. Every machine is identified by serial number and its test results are recorded against that serial, which is also what lets a part be identified correctly later. Factory testing proves the mechanism; it cannot prove fit for your specific packs, which is why the planogram is checked before the build and why every selection should be test-dispensed with real product at commissioning.

Most failed vends have three causes — the product does not fit the coil it was loaded into, the coil or spring end no longer holds and releases the pack cleanly, or the machine has no way to confirm the item actually fell — and each one is removed at a different stage of the project. Fit is settled before the machine is built, by checking each product’s pack size against the coil type, coil pitch and tray height it will sit in. Mechanical release is settled by choosing the coil form for the pack shape rather than filling every position with the default, and by cycling every motor and coil under load before the machine ships. Confirmation is a configuration decision: an elevator build uses a drop sensor, and locker cells can be specified with per-cell weight sensing that verifies the dispense and corrects the stock count. Stockouts are a separate problem — a machine that never jams still fails the worker if the popular glove size runs out, which is a matter of holding each size as its own tracked item with its own par level and alert.

What “unable to dispense” costs you

The complaint that ends a machine’s credibility on site is not a breakdown. It is a full machine that will not hand over an item. The worker comes for gloves, the column is visibly stocked, the screen refuses, and after the second or third time they stop trying and go back to the store cupboard — which is the arrangement the machine was bought to replace.

Two things go wrong at once when that happens. The user loses trust, and the stock record silently becomes fiction: a position that failed to vend but reported success shows product going out that is still sitting in the coil, so the replenishment run is planned against numbers that are wrong in both directions.

This page sets out why vends fail, what removes each cause, and what to ask any supplier before you order. It applies equally to a snack machine and to a PPE or tool machine, but the consequence is different — a missed chocolate bar is an annoyance, and a missed pair of safety glasses is a person working without them.

The three causes of a failed vend

Cause	What the user sees	Where it is actually fixed
Product-to-coil mismatch — pack too wide, too deep, too tall or too soft for the coil pitch and diameter it sits in	The pack rotates forward, tilts, wedges against the neighbouring coil or the glass, and stays put	Before the build, in the planogram: coil type, coil pitch and tray height chosen per product
Mechanical release — a spring end that no longer holds and releases the pack cleanly, or a coil form that never suited the pack shape	Intermittent failure, usually worst on the last one or two packs in the column	Coil form chosen for the pack shape at order; worn coils replaced as a wear part in service
No confirmation the item fell	Nothing at all — that is the problem. The transaction is recorded as a success	At quotation, by specifying a verification method: elevator with drop sensor, or locker cells with per-cell weight sensing

The third cause is the one buyers most often leave out of the specification, and it is the one that converts a mechanical fault into a data fault. A machine that cannot detect a failed vend will keep selling from a jammed position all week and report a healthy sales pattern while doing it.

Fit is decided before the machine is built

Most dispense failures are designed in months before anyone loads the machine. The fix is unglamorous: measure the packs.

For every item you intend to vend, record the following in the packaging it will actually arrive in, not the manufacturer’s bare-product dimensions:

Field	Why it decides the position
Width, depth and height of the pack	Sets the coil diameter and pitch it can occupy and the tray height above it
Weight	Affects motor torque and how the pack behaves as the column empties
Rigidity — box, blister, bag or soft pack	Soft and deformable packs are the highest-risk items in any coil machine
Orientation on the coil	Many packs vend reliably one way round and jam the other
Consumption rate	Decides how many positions the item needs, not just which one

You can do the first pass yourself. The planogram configurator lets you set the number of trays and coils per tray, change the coil option per position, edit coil pitch per position, and set tray heights — then export the layout and send it with your enquiry. A layout attached to an enquiry turns a generic quotation into a specific one, and it surfaces the awkward items while they are still cheap to solve.

Items that will not dispense reliably from a coil in their supplier packaging have three honest answers: a different coil form, a locker cell instead of a coil, or re-boxing the item into a pack that vends. All three are decisions to take before manufacture.

Coil type and coil pitch, chosen per product

Filling every position with the default coil is the most common cause of a machine that jams on a minority of its lines. The published coil options exist because pack shapes differ:

Option	Suited to
Option A — small items, coil diameter around 3–4cm	Compact products needing tighter spacing
Option B — can form factors	Cylindrical items such as aerosol or drink cans; the mechanism lifts the product slightly for a smooth release
Option C — thin and wide items	Flat, wide packs that need a shim to elevate them for a proper release
Option E — default	The standard configuration, suitable for most common products
Option G — long stick shapes	Elongated items such as chisels or pens, held securely along their length

Pitch matters as much as form. Too tight and the packs press against each other and bind; too loose and you lose capacity for no reliability gain. Pitch is set per position in the planogram, so a single tray can carry a tight pitch for small boxes and a wider one for bulky packs.

For items too wide for a single coil, adjacent coils can be coupled to turn in sync so the pack is pushed by both — see motor coupling. Coupling follows configuration rules (a motor already assigned to a selection cannot be coupled, and the couple code must not clash with the admin PIN), so set it up at commissioning rather than mid-shift.

Where the pack is genuinely unsuited to any coil — a helmet, a harness, a boxed instrument, an item that must come back — the correct answer is a locker cell rather than a cleverer coil. That is covered on industrial vending and, for returnable equipment, in the hybrid dispense-and-return arrangement described on the weight sensor cell page.

Confirming the item actually fell

Verification is a specification decision, and machines differ. Ask for it explicitly.

Method	What it confirms	Notes
Motor rotation only	That the motor turned	The baseline behaviour of a plain coil position. It does not know whether the pack left the coil
Drop sensor (elevator builds)	That the item reached the delivery bin	On machines running with the elevator, the multi-select shopping cart is disabled so the drop sensor functions correctly — the machine switches to instant purchase, one selection at a time
Per-cell weight sensing (locker cells)	That the cell weight changed by the expected amount, on dispense and on return	Optional and factory-fitted per cell; it also corrects the stock count and verifies that a returned item is physically present, not just that the door closed
Vision-verified pick (robotic builds)	That the item was grasped, with a retry if the grasp fails	The mechanism used on robotic kiosks, where the arm picks rather than the coil drops

Weight sensors are integrated during manufacture and cannot be retrofitted in the field, so the cells that need them are chosen when the order is placed. That is a reason to decide which items are worth verifying before you buy, not after.

What happens when a vend fails anyway

Even a well-configured machine will occasionally fail to dispense. The behaviour that matters is what happens next:

The customer is not left out of pocket. The order report carries a Refunded state — the selected product failed to dispense after payment, so the money was returned automatically. Every transaction generates a receipt ID, so a specific order, its items and its payment events can be traced afterwards.
Empty positions stop being offered. Trays that are out of stock are disabled automatically to prevent failed vending, and stay disabled until inventory is restocked and the Refill key is pressed.
The failure is visible remotely. Transaction states, stock levels and low-stock alerts are visible in the cloud dashboard rather than only on the machine, so a position failing at one site does not have to be discovered by a person standing in front of it.

One configuration note worth reading before you enable it: the optional automatic refill setting marks the machine as refilled as soon as an administrator enters the admin portal. It saves a step for teams that always refill fully, and it overstates stock for teams that open the door for other reasons. Choose deliberately.

The other kind of failure: the silent stockout

A machine that never jams still fails the worker if it is empty of the only size that fits them. This is the most common end-user complaint about PPE vending, and it is not a mechanical problem at all — it is a stock-modelling problem.

Hold every size as its own item. One glove line covering six sizes behind a single product record will report healthy stock while the two sizes most of the workforce wears are gone. Each size needs its own selection, its own count, its own par level and its own alert.

Allocate facings by consumption, not by symmetry. Giving each of six sizes one position is tidy and wrong. Let the machine record real consumption for a few weeks, then reallocate: the common sizes usually justify two or three positions each, and the rare sizes may not justify a permanent position at all — a small buffer in the crib or a locker cell serves them better than a coil that turns over twice a year.

Set par levels against replenishment frequency, not against capacity. The question is not how much the column holds, it is how much the site draws between replenishment visits. A position that empties in four days on a weekly round is under-provisioned no matter how full it looks on day one.

Let the alert trigger the round. Low-stock thresholds in the cloud dashboard exist so that replenishment is driven by data rather than by someone noticing. On PPE and MRO sites, consumption data can be pushed onward to a procurement or ERP system so the reorder happens without a manual step — the integration pattern described on MRO vending and ServiceNow integration.

Watch for the substitution signal. If a size runs out and consumption of the size next to it spikes, you are not seeing demand — you are seeing people taking what is left. That pattern in the transaction data is a stockout that has already cost you something, and it is one of the clearest arguments for size-level tracking. The commercial side of that argument is on PPE cost control.

What is tested before the machine ships

Fit is your data; the mechanism is ours. Every completed machine undergoes full powered-on testing before it leaves the assembly bay, covering:

every motor and spiral coil cycled through full rotation under load
payment peripherals verified with test transactions
touchscreen response, display rendering and keypad input
connectivity — network registration, server handshake, remote management heartbeat
door interlock switches and earth continuity

A separate load test operates a weighted coil rack to verify motor torque and spiral coil performance under a full product load rather than an empty one, and a full-system burn-in runs the machine under load with all motors, coils, sensors, payment modules and control logic verified. Every machine is identified by serial number and its test results are recorded against that serial — the same record that later lets a replacement part be identified correctly, as described on spare parts and support. More detail on the QA stations is on the manufacturing facilities page.

Factory testing proves the mechanism works. It cannot prove that your packs vend, because the packs are not there.

Commissioning: prove it with real product

The last chance to catch a fit problem cheaply is the day the machine is filled for the first time.

Load real product, not samples. Vend the actual packs, in the actual packaging, from the actual positions.
Test-dispense every selection. The selection settings module includes a Test control that simulates the dispensing process for validation and troubleshooting — run it on every position, not a sample of them.
Test the near-empty condition. Most failures appear on the last one or two packs in a column, not on a full one. Run the column down.
Run the sizes and shapes most likely to jam first. Soft packs, wide packs and tall packs earn the attention; rigid boxes rarely surprise anyone.
Use the factory mode test where the whole machine needs proving. It cycles the rotation of all coils and the opening of all locker cells, and exists for distributors and workshops commissioning a machine before handover.
Record what you changed. A position that needed a different coil or a different orientation is information for the next machine and for the next planogram.

This is also the argument for asking for a sample or pilot unit before a volume order — covered in the manufacturer vetting guide.

Where verification is the wrong answer

Sensors report the truth sooner. They do not change physics or logistics, and a page that implied otherwise would not survive contact with a maintenance team.

Verification does not fix a wrong planogram. A pack that does not suit its position will keep failing; the sensor just tells you faster and refunds the user. The fix is still to change the position.
Weight sensing verifies weight, not identity. Two sizes or two similar items sharing a cell cannot be distinguished by weight alone. Separate them, or accept that the record is cell-level rather than item-level.
A drop sensor confirms arrival in the bin, not condition. It does not know whether the pack arrived damaged, or whether the right person collected it.
Instrumentation does not fix replenishment frequency. If the site draws more than the machine holds between visits, the answer is more facings on the fast movers, more positions, or a more frequent round.
Some items should never be in a coil. Long, heavy, awkward or deformable items belong in locker cells or a shelf. Forcing them into a coil buys a reliability problem that no sensor removes.
Every added device is a part that can fail. Sensors are worth specifying where a failed vend has a real cost — PPE, controlled tooling, high-value consumables. Specifying them on every position of every machine is a maintenance liability bought for a benefit you may not need.

Questions to send any supplier

Copy these into your enquiry. The answers separate suppliers who have thought about dispense reliability from suppliers who have not.

Will you check the fit of our actual pack dimensions against coil type, pitch and tray height before the machine is built, and will you tell us which items you consider high-risk?
Which coil forms are available, and which one will each of our products sit in?
How does the machine confirm that an item was actually dispensed? Is that fitted on every position, or only on some?
If a vend fails after payment, what happens to the customer’s money, and where is that visible afterwards?
Does a failed vend correct the stock count, or does the count keep drifting until someone recounts by hand?
Are out-of-stock positions disabled automatically, and what has to happen before they are offered again?
Can each size or variant be held as a separate tracked item with its own par level and low-stock alert?
Can weight sensing be added to specific cells, and can it be added after delivery or only at manufacture?
What testing is done on motors and coils before the machine ships, and is the result recorded against the serial number?
Which parts are the wear items on the dispense mechanism, how are they identified, and how are they ordered?

How KioskForce answers each point

Point	Our answer
Fit checked before build	Yes — pack dimensions are worked position by position against coil type, pitch and tray height. Use the planogram configurator for a first pass and send the layout with your enquiry
Coil choice	Five published coil options covering small items, cans, thin and wide packs, the general default and long stick shapes; pitch set per position; adjacent coils can be coupled for oversized packs
Dispense confirmation	Elevator builds carry a drop sensor; locker cells can be specified with per-cell weight sensing that verifies dispense and return and corrects stock. Which method your build carries is confirmed at quotation
Failed vend after payment	Recorded as a Refunded order — the money is returned automatically and the order is traceable by receipt ID
Out-of-stock behaviour	Positions that are out of stock are disabled automatically until restocked and marked refilled
Size-level control	Each size is loaded as its own selection with its own stock count, par level and low-stock alert in the cloud dashboard
Pre-shipment testing	Every machine powered on and operated under load, every motor and coil cycled, plus a weighted-rack load test; results recorded against the serial number
On-site verification	Per-selection `Test` dispense in the admin interface and a factory mode test that cycles all coils and locker cells
Wear parts	Vend motors, coils and springs are treated as orderable wear items — see spare parts and support
Retrofit limits	Weight sensors are factory-integrated per cell and cannot be retrofitted in the field, so specify them with the order

Planogram configurator — build and export the layout before you request a quotation
Coil options — the five published coil forms and what each suits
Weight sensor cell tracking — per-cell verification of dispense and return
Motor coupling — coupling adjacent coils for oversized packs
Industrial vending — PPE, tool and MRO dispensing, including locker cells
PPE cost control — quota-based issue and what stockouts really cost
Tool crib modernisation — inventorying items and dimensions before automating a crib
Spare parts and support — wear parts on the dispense mechanism and how to order them
Manufacturing facilities — the QA stations and load testing referred to above
Vending machine user manual — selection settings, test dispense, refill and order states
How to vet a manufacturer — why a sample unit loaded with real product matters
FAQ — short-form answers, including dispense reliability

Get started

Send us the list of items you intend to vend with their pack dimensions, weights and expected consumption, and tell us which failures would actually cost you something. We will come back with a layout, the coil form proposed for each position, and a straight answer about which items we would not put in a coil at all.

Contact Us for More Information