Device Repeatability & Reproducibility

Overview

Repeatability and reproducibility (R&R) testing answers a fundamental question: does your product perform consistently? Are two cartridges from the same batch interchangeable, or is there significant unit-to-unit variation? This is the foundation of any quality control program. Without R&R data, you are guessing — hoping that the cartridge a consumer pulls off the shelf performs the same as the one you tested in the lab last month.

The multi-channel Universal Vaping Machine makes R&R testing fast and practical by running multiple identical products under identical conditions simultaneously. Instead of testing one cartridge at a time and hoping that ambient conditions, operator technique, and session timing don't introduce confounding variables, the UVM tests 4 or 10 products in parallel — same puff profile, same power, same environment, same moment in time. The result is clean, comparable data that isolates the variable you actually care about: the product itself.

Definitions

• Repeatability — the consistency of results when the same product is tested multiple times under the same conditions on the same machine. If you test the same cartridge three times and get the same vapor density curve each time, the measurement system has good repeatability.
• Reproducibility — the consistency of results across different machines, operators, or test sessions. If two different labs running two different UVM systems produce equivalent data on the same product, the method has good reproducibility.
• Machine-side control — the UVM's programmable puff profiles, precise power control, and integrated sensors minimize machine-side variability. Every puff is executed to the same duration, flow rate, and wattage, and every measurement is captured by calibrated sensors with known precision. This isolates the product-to-product variation you actually care about, rather than conflating it with measurement noise.

Why It Matters

R&R testing is relevant across the entire cannabis vapor supply chain, but different stakeholders use it differently:

• Cannabis producers and MSOs — incoming hardware inspection. Does this batch of cartridges from your supplier meet your quality standards? Are they consistent enough to fill and sell under your brand? R&R testing turns "we think the hardware is fine" into "we have data showing the batch is within specification."
• Oil formulation providers — does a given formulation perform consistently across multiple cartridge units? If the same oil produces wildly different vapor output in different cartridges, the problem may be the hardware, the fill process, or a formulation-hardware interaction. R&R testing helps isolate the root cause.
• Hardware middlemen and distributors — verify batch quality before reselling to brands. Provide statistical evidence that a shipment is good (or bad). A distributor who can show R&R data on every batch they ship builds trust and differentiates on quality.
• Regulatory submissions — demonstrate that your product performs within defined tolerances. Regulatory bodies expect quantitative evidence of consistency, not anecdotal claims. R&R data is the standard way to provide it.

Testing Methodology

A well-designed R&R study follows a structured protocol that controls for as many variables as possible while allowing the product-level variation to surface clearly:

1. Select N products from the same batch. A minimum of 4 samples can be tested in a single run on the 4-channel UVM. For statistically meaningful results, 10 or more samples are ideal — achievable in a single run on a 10-channel system, or across 3 sequential runs on the 4-channel system.
2. Connect all products to the UVM. Use the same puff parameters, the same power level, and the same environmental conditions across all channels. The multi-channel architecture ensures that all samples experience identical test conditions within a run.
3. Run a standardized test regimen. A typical R&R protocol might specify 50 puffs at 55 mL puff volume with a 30-second interval between puffs. The specific parameters should match your product's intended use case and any applicable standards.
4. Collect per-puff data on all channels. The UVM automatically logs vapor density, pressure drop, coil resistance, and other metrics on every puff for every channel. No additional setup is required.
5. Calculate batch statistics. For each metric, compute the mean, standard deviation, and coefficient of variation (CV) across all samples. CV — the standard deviation divided by the mean, expressed as a percentage — is the most useful single number for comparing consistency across metrics with different units and scales.
6. Identify outliers. Products falling outside 2 standard deviations from the batch mean warrant investigation. These may indicate manufacturing defects such as bad seals, inconsistent wicking, or coil assembly issues.

Key Metrics for R&R

The UVM captures several independent metrics on every puff. Each tells you something different about product consistency:

• Pressure drop (Pa): mean and CV across units. This tells you about draw consistency — are all units equally easy (or hard) to pull on? A high CV in pressure drop suggests variability in airpath geometry, wick packing, or oil fill level.
• Vapor density (IR differential): mean and CV. This tells you about vapor output consistency — do all units produce the same amount of visible aerosol? Vapor density variation can stem from coil performance differences, oil wicking rates, or power delivery inconsistencies in the product's own electronics.
• Total puff count (if running to depletion): mean and CV. This tells you about product longevity consistency — do all units last roughly the same number of puffs? High variation here points to inconsistent fill volumes or variable oil consumption rates.
• Coil resistance (Ω): mean and range. This tells you about hardware manufacturing consistency. Coil resistance is set at the factory and should be tightly controlled. A wide range indicates poor manufacturing tolerances in the heating element.
• Mass per puff (if using gravimetric filters): mean and CV. This tells you about aerosol delivery consistency — how much material is actually deposited per puff. Combined with vapor density data, it gives a complete picture of aerosol output uniformity.

Multi-Channel Advantage

The 4-channel UVM tests 4 products simultaneously under truly identical conditions — same ambient temperature, same humidity, same puff timing, same session. This eliminates session-to-session variability that would confound sequential testing. When you test one product on Monday and another on Tuesday, differences in room temperature, barometric pressure, or even oil settling time can introduce noise that masquerades as product variation. Multi-channel parallel testing removes that noise entirely.

The 10-channel system scales this to 10 simultaneous products — approaching meaningful sample sizes in a single test run. With 10 data points collected under identical conditions, you can compute batch statistics with real confidence. For operations that need to test incoming hardware regularly, the 10-channel system transforms R&R testing from a periodic special study into a routine QC step that takes minutes, not days.

Example QA/QC Workflow

Here is a practical workflow for using R&R testing as an incoming quality gate for cannabis cartridge hardware:

1. Receive shipment of 1,000 cartridges from your hardware supplier.
2. Pull 10 samples randomly from different positions in the shipment (e.g., different boxes, different layers). Random sampling is critical — do not cherry-pick from the top of one box.
3. Run all 10 through the UVM using a 10-channel system in a single run, or across 3 sequential runs on a 4-channel system. Use your standard puff profile and power settings.
4. Calculate batch statistics for pressure drop, vapor output, puff count, and coil resistance. Generate the per-metric mean, standard deviation, and CV.
5. Compare against your acceptance criteria. Typical thresholds might include: CV < 15% for pressure drop and vapor density, no individual sample more than 2 standard deviations from the batch mean, and coil resistance within the supplier's stated tolerance.
6. Accept or reject the batch based on quantitative data. If the batch fails, you have the data to support a return or renegotiation with the supplier. If it passes, you have documentation for your quality records.

Interpreting Results

Low CV (< 10–15%) — a consistent batch. The hardware is well-manufactured, the fill process is uniform, and the products are likely to perform predictably in the field. This is your target range for production-quality hardware.

Moderate CV (15–20%) — acceptable for some applications, but worth monitoring. If this is a new supplier or a new hardware revision, investigate whether the variation is trending upward across shipments.

High CV (> 20–30%) — significant unit-to-unit variation. This level of inconsistency will be noticeable to end users. Investigate supplier quality, filling process controls, and storage conditions. A batch with high CV across multiple metrics likely has a systemic manufacturing issue.

Outliers — individual units that fall far outside the batch mean (beyond 2 standard deviations) may indicate manufacturing defects: bad seals, inconsistent wick material, misaligned coils, or improper fill volumes. Outliers should be examined physically when possible to identify the root cause. If outliers appear frequently across batches, the supplier's process control needs attention.

The QA/QC charts — showing mean ± SD with individual sample values and outlier highlighting — are generated directly from UVM data exports. These charts provide an at-a-glance summary of batch quality and are suitable for inclusion in quality reports, supplier scorecards, and regulatory filings.