When puffed pellets vary between runs, teams often change flour and fryer settings together, mixing two causes. A clearer approach is staged trials: groups within one lot test only conditioning, geometry and downstream expansion, while raw-material comparison is a separate stage. The method below suits your own line and does not replace specifications or the contract.
1. Separate the two kinds of trial
A same-lot group draws one lot, mixes and splits it into sub-samples, codes each and records draw date, bag number and shelf life. It answers one question: with material held constant, what changes when conditioning, geometry or expansion conditions change?
Raw material comparison is separate. Each candidate uses its own independent lot, replicated in parallel under the same formula, equipment and operating conditions; sub-samples are taken within each material and never mixed across materials. A same-lot group cannot answer this question.
Measure the basics first: moisture, particle size or bulk density, and reducing sugar if relevant. Three separately prepared and measured replicates are a parallel determination, not evidence that the sampling represents the whole lot; a value reflects only the samples tested and cannot explain every difference between groups.
Stabilise a baseline first: repeat one lot under fixed conditioning, geometry and equipment until within-group scatter is acceptable, then change the next variable. No single order is prescribed and reverse order is not the point; change one variable at a time and record the rest.
Several specimens taken from one run describe variation within that run; they do not replace independently repeated process trials. With only one lot per candidate, material and lot effects remain confounded. That can be an initial screen, but assessing consistent supply requires multiple independent lots, planned repeat runs, and records of storage and opening history.
2. Distinguish three water steps: mixing water, drying after extrusion, equilibrating dried pellets
Mixing water is added to flour in the mixing step; drying after extrusion removes water from formed pellets; equilibrating dried pellets adjusts moisture before expansion. The three differ in record fields and cannot replace one another. This article mainly evaluates the moisture of dried pellets before expansion; it does not imply that spraying water onto dried pellets is enough for production, and any addition or conditioning must follow a validated process and internal food safety management.
What matters is total moisture and its evenness, not only how much water went in. State the calculation basis (wet or dry) and measure moisture at several points before and after conditioning by the same method; without a stated basis the addition cannot be recalculated at scale.
| Field | Record | How | Why |
|---|---|---|---|
| Group code | Grouping sheet match | Spreadsheet or paper | Links results |
| Moisture before | Value, sampling points | Meter or oven, method stated | Basis for calculating water addition |
| Target, basis | Wet or dry, calculation | Calculation sheet | Scale-up check |
| Water added | Amount, spraying, duration | Weighing, settings | Checks moisture distribution |
| Mixing, tempering | Time, speed, room conditions | Display and measured | Within-group scatter |
| Moisture after | Values, point count | As above | Checks the agreed moisture target |
If moisture varies widely within a group, check mixing, tempering and sampling before touching expansion settings.
3. Keep the same geometry when comparing conditioning
Thickness, diameter or width, length, single-piece weight and surface condition all affect heat transfer and expansion. When comparing conditioning, use the same die and geometry and change only moisture; otherwise size effects are read into the conditioning conclusion.
Study geometry as its own variable: take a random set of pieces per group, measure thickness and diameter with calipers, weigh them, and record mean, range, piece count, operator and date. Size can come from the die, cutter, conveying or drying; hold it fixed or log it group by group so it is not read as a material problem.
4. Log actual expansion conditions, not assumed temperatures
Frying, hot air, microwave and pressure-release expansion transfer heat differently, but the record logic is the same: equipment number, medium, setpoint and measured value, residence time or line speed, load per batch and batch interval. You set temperature and time for your equipment and product; no recommended values are given here, because the same setting behaves differently on different equipment and loads.
Record appearance after expansion (colour, blistering, rupture, hollow centres), volume or size change, moisture content, and the number of assessors and how texture was judged. Use one machine and one operator per trial day where possible.
Changing one variable at a time is useful for initial troubleshooting, but it does not establish that the variables act independently. Moisture effects may depend on pellet thickness or expansion conditions. Before scale-up, verify the intended combination and, where needed, use a multifactor design to investigate interactions (NIST guidance on one-factor-at-a-time trials).
5. Reading order and the next step
- Check repeatability within a group; wide scatter points to sampling, weighing or equipment stability.
- Read conditioning in the lot groups: fix material and geometry, change moisture, look for a trend rather than a single point.
- Read geometry: fix material and conditioning, change only size and observe its effect under those fixed conditions.
- Read expansion last, adjusting on fixed conditions to find a window worth scaling up.
- Read the raw material comparison separately: independent lots, same formula and equipment, replicated. Repeatable differences suggest checking material and process fit; a single run does not establish a cause.
Keep result differences separate from missing records: if a key field such as moisture after conditioning is missing, that group is only a lead. Before scaling up, put trial conditions, lot numbers and equipment settings on one page, with research and development confirming use and trial conditions, production confirming equipment fit, quality defining inspection and release, and sales putting grade, specification version and delivery requirements into one order record.
If you are running a similar trial, send your grouping, conditioning record fields, pellet size range and equipment type, and we will review the variable set-up and reading order against your format.
