A burnt or unusually dry note during sample testing is easy to label and difficult to diagnose. The same observation can appear when the heating interface is temporarily under-supplied, when applied power is too aggressive for the configured resistance and liquid, when the puff interval is shorter than recovery time, or when one unit has an assembly or electrical fault.
For a hardware buyer, “it tastes burnt” is therefore an observation—not a root cause and not a batch disposition. The useful question is narrower: under which documented configuration and test sequence did the endpoint appear, and can the team reproduce it without changing several variables at once?
This guide describes a controlled fault-isolation method for filled disposable AIO samples. It does not define a universal wattage, viscosity limit, puff count, safe-use threshold, or emissions conclusion. Acceptance criteria belong to the specific product program and must be agreed before testing.
Start by preserving the first failed unit
Stop the sequence when the defined endpoint occurs. Repeated activations made only to “see whether it clears” can change liquid distribution, heating-surface condition, battery state, and the evidence the investigation needs.
Identify the unit and record:
- device model, revision, lot, and serial or sample number;
- formulation and fill-lot identity;
- target and actual fill mass, headspace approach, and capping time;
- conditioning duration, temperature, and orientation;
- activation mode, indicated battery state, and adjustable setting if present;
- puff number, duration, interval, and controlled flow or volume setting;
- the exact observation and whether it recovered after a defined pause.
Photograph visible changes and retain the unit. A replacement sample may help test a hypothesis, but it cannot describe what happened inside the original unit. This evidence should be captured before a qualification specification change or supplier escalation is warranted.
Lock the configuration before reproducing the complaint
A useful reproduction keeps the device, liquid, preparation, and machine sequence identifiable. “Same product” is not precise enough if samples came from different device revisions, fill lots, storage routes, or charge states.
Prepare matched controls from the same hardware and formulation lots where possible. Use the same fill method, cap delay, conditioning path, orientation, and starting battery condition. If the device has variable output or preheat, record the selected state rather than assuming the default remained active.
Resistance also matters. Two samples with nominally identical labels can draw different power if their actual resistance or control behavior differs. Measure only with a method suitable for the device, and do not open or probe an integrated assembly in a way that changes it. Where direct measurement is not practical, preserve supplier release data and compare device behavior using the buyer’s approved functional fixture.
Teams that need a broader baseline can use the ILEVA vape hardware sample testing protocol.
Define a puff sequence another team can repeat
A count without a sequence is not a method. Twenty puffs at long recovery intervals do not challenge liquid replenishment in the same way as twenty closely spaced activations. Record at least:
- puff volume or controlled flow;
- puff duration;
- interval between puffs;
- number of puffs per block;
- rest period between blocks;
- device orientation;
- preheat or button actions;
- stop rules and observation points.
CORESTA guidance emphasizes that no single machine-vaping regime represents every pattern of human behavior. For product qualification, choose a sequence that answers the program’s stated question, then document it. A standard machine condition can improve repeatability, but it does not automatically become a universal acceptance envelope for every AIO and liquid combination.
Track consumption, not only activation count
Activation count is convenient, but it does not prove that every puff consumed the same amount of liquid or delivered the same energy. Power, puff duration, device control, formulation, temperature, and liquid supply can all change consumption.
Weighing the configured sample at defined checkpoints can provide a practical mass-loss trend when the balance, handling, cleaning, and stabilization method are controlled. The project must define whether its instrument resolution and handling method are capable of detecting the change it wants to interpret. Treat mass loss as a process indicator, not a complete chemical analysis.
A sudden change in mass-loss rate, output, sound, temperature, or recovery behavior may help locate the transition before the final sensory complaint.
Hypothetical comparison: why the same puff count is not the same failure
The following illustration is fictional and is provided only to show the method. It is not measured data from an ILEVA unit or any supplier sample.
| Sample | Mass-loss pattern | Recovery behavior | Disposition-relevant signal |
|---|---|---|---|
| A | Steady trend to the recorded endpoint | Observation clears after the predefined rest | Test a repeatable supply-recovery interaction |
| B | Consumption declines before the endpoint | Observation persists after the same rest | Investigate supply, preparation, or irreversible change without assigning a cause yet |
| C | Stable trend until an abrupt interruption | Recovery follows electrical behavior rather than rest alone | Investigate control, contact, resistance, or unit-specific assembly evidence |
All three samples may reach a dry note near the same activation count. The shared count does not establish a shared cause.
Run one-variable comparisons
The first comparison should test the strongest hypothesis with the smallest controlled change.
Recovery interval
Repeat the sequence with a longer, predefined interval while keeping the device, liquid, power state, and puff duration unchanged. A repeatable improvement supports a supply-recovery interaction. It does not prove that viscosity alone caused the problem.
Applied power or control setting
Where the device is designed for an adjustable setting, compare only approved settings. Calculate or verify the electrical condition using actual configuration data rather than a generic voltage label. Improvement at a lower setting may show that the tested energy demand exceeded the current supply behavior; it does not establish a universally correct setting.
Fill and conditioning route
Compare matched samples prepared through the approved route and a deliberately bounded alternative only when the project allows it. Differences may point toward fill placement, trapped air, headspace, cap timing, or conditioning. Do not change formulation and preparation simultaneously.
Device unit or lot
If one unit fails while matched controls do not, preserve it for focused inspection. Check observable activation, airflow, charge behavior, external damage, and lot identity before declaring a design failure.
An isolated observation and a recurring pattern are not equivalent. Do not invent a universal number of units. Classify recurrence against the buyer’s approved sampling plan, randomized sample selection, defect catalogue, and escalation rule. If that plan does not define how this endpoint is treated, the method needs revision before the observation can support a lot decision.
Interpret patterns without overclaiming
A reversible endpoint after a controlled rest suggests that liquid replenishment or thermal recovery belongs in the next hypothesis. A persistent endpoint with abnormal electrical behavior shifts attention toward control, contact, resistance, or battery performance.
A complaint paired with restricted airflow belongs in the separate clog or leakage investigation described in Why Vape Cartridges Clog.
A sudden one-unit deviation may justify inspection of assembly or component variation. These patterns narrow the next test; they do not prove an internal mechanism by themselves.
Taste observations are especially vulnerable to expectation and carryover. If sensory evaluation is part of the program, define panel qualification, blinding where practical, sample order, palate reset, and the rule for confirming an endpoint. Sensory recovery is a hypothesis signal, not evidence of chemical recovery, emissions performance, or product safety. If a program includes laboratory emissions work, that requires a fit-for-purpose analytical method and should not be replaced by informal tasting.
Turn the result into a bounded decision
Close the investigation with one of four states:
- Accept within the tested envelope. The configuration met the predefined endpoint under the recorded sequence.
- Revise the method. The sequence, conditioning, measurement, sampling, or stop rule was not controlled well enough to support a product decision.
- Hold for investigation. The failure is reproducible or lot-linked, but the responsible variable is unresolved. Preserve affected samples and keep the qualification decision open under the buyer’s controlled process.
- Requalify the interface. A change to power, resistance, liquid, fill route, component, or device revision requires a new bounded comparison. Record the proposed change, identify affected decisions or orders, route any specification/addendum question to its owner, and notify the responsible supplier contact according to the buyer’s process.
Do not erase an adverse result by averaging it with additional good units after the stop rule was reached. Record retests separately and state why they were performed. For a wider endurance decision, use the separate disposable AIO battery-endurance qualification method.
A compact qualification worksheet
Before release, capture the following as a controlled checklist or worksheet—not as narrative notes.
| Field | Test record |
|---|---|
| Device identity | Model, revision, lot, serial/sample number |
| Liquid identity | Formulation and fill lot |
| Preparation | Fill mass, headspace approach, cap delay, conditioning, orientation |
| Electrical state | Battery state, approved setting, available resistance/control evidence |
| Puff sequence | Flow or volume, duration, interval, block size, rest, stop rule |
| First endpoint | Puff/block, exact observation, associated sound/output/temperature/airflow change |
| Consumption evidence | Checkpoint masses or other project-approved proxy and method capability |
| One-variable comparison | Variable changed, variables held, result |
| Pattern classification | Isolated or recurring under the approved sampling/escalation rule |
| Decision state | Accept envelope, revise method, hold, or requalify interface |
| Owner and next evidence | Responsible person, due action, evidence required before release |
For programs moving from qualification into controlled production, the disposable AIO pilot-run readiness guide provides the next operational boundary.
Final takeaway
A burnt-taste observation during AIO qualification becomes useful only when the team preserves the unit, freezes the configuration, and reproduces the event under a controlled sequence. Power, resistance, puff timing, liquid supply, preparation, and assembly should enter the record before a qualification specification change or supplier escalation is warranted.
When those variables remain identifiable, the complaint becomes a testable boundary. The team can decide whether to accept the configured envelope, revise the method, hold the sample population, or requalify the hardware–liquid interface with evidence instead of guesswork.
Teams selecting samples for this kind of configuration-specific work can review ILEVA empty AIO hardware for configuration-specific qualification.
References and scope
- CORESTA Guide No. 22: Technical Guide for Setting Vaping Regimes for E-Vapour Devices
- ISO 20768 overview
- Dry puff conditions for electronic cigarettes
- Effects of Coil Power on E-Cigarette Aerosol
External nicotine e-liquid literature is used only to support test-design principles. It is not evidence of performance for oil-filled ILEVA AIO products.
