A machine result can be true and still not describe the line
Each machine in a fruit line can pass its own test while the connected line still disappoints. The crusher may produce a good mash, the pulper may recover a good pulp, and the thermal section may hold its own conditions, yet the assembled route may lose product, change texture or create a bottleneck at a handoff. The missing evidence is usually not a machine result. It is the connection between results.
A connected trial treats the fruit lot as one campaign. The same material is followed through preparation, recovery, finishing, holding and treatment, with samples taken at each boundary. That does not make the trial automatically representative of commercial production, but it makes the result attributable. When the final product changes, the team can see where the change first appeared.
Give the campaign one identity
Start with a campaign record that stays with the material. Identify the fruit family and variety, the growing or purchase lot, maturity, size range, reception date, holding conditions and any preparation already completed. Give the campaign a unique batch code and give every sample its own code that includes the machine, the stream and the time.
The codes should be simple enough to write on a sample label by hand and specific enough to prevent two samples being confused later. A sample called “pulp after pulper” is weaker than a sample called “campaign 26-10-A, pulper outlet, 10:42, sample 04”. The second can be connected to an inlet sample, a reject sample and a final product sample without relying on memory.
Record the inlet at every machine
The material entering a machine is not the same as the material that entered the plant. Whole fruit becomes trimmed fruit, mash, juice, pulp or concentrate. Record the condition at each inlet: temperature, visible particle or fibre condition, solids where relevant, and the measurement method. Note any addition of water, enzyme, acid or other process aid, because it changes the mass balance and the interpretation of the next result.
Use one measurement basis consistently at each boundary. Raw-fruit kilograms per hour, liquid litres per hour, pulp kilograms per hour and accepted packages per hour are different quantities. The FAO's general juice chapter warns that laboratory results and commercial scale-up are not the same thing; a trial that does not state its scale, duration and interruptions leaves the reader unable to judge how much of the result can be transferred.
Follow the product and the reject streams
Every separation creates at least two streams. Stones, peel, seeds, press cake, screen rejects and retentate are not incidental waste; they are evidence about what the machine actually did. Sample them at the same time as the accepted product and record their condition and destination. A product sample without a reject sample can hide a separation problem.
The FAO chapter is explicit that juice recovery is a compromise between yield and quality: thorough comminution can extract more, but it can also extract undesirable components from seed, skin and core. That is why a trial record should not reduce success to a single recovery number. It should show the wanted product, the unwanted fraction and the quality result together, using matched samples from the same run.
Write down the handoff, not only the machine
Between two machines there is usually a buffer, a pump, a pipe, a valve, a gravity drop or a holding tank. Record the working level, the transfer method, the residence time and the product condition at that handoff. If the line stops, note what happens to the material in the buffer and whether it can be restarted without losing the sample identity.
Interface responsibility should be recorded too. Who supplied the pump, the connecting pipe, the control interlock and the sampling point? Which party confirmed that the passage is suitable for the product's particles and viscosity? A machine can be mechanically connected and still be a poor process fit. The handoff record is where that difference becomes visible.
Preserve the sample conditions
A sample is only useful if its condition at the time of collection is known. Record the container, fill level, headspace, temperature, holding time and any preservation step before laboratory analysis. For pulp and puree, texture and viscosity can change with temperature and time; for juice, air contact and holding can change colour and flavour. The final sample should be stored and transported under conditions that are stated in the record.
Where a sample is split for different tests, give each portion its own sub-code and keep the chain of custody. The person who takes the sample, the person who receives it and the method used should be traceable. This is ordinary quality-assurance discipline rather than a claim about a particular laboratory. It prevents a later discussion from depending on an unlabelled container.
Compare the trial with the commercial case
A short trial can show that a route is capable of producing a sample, but it may not show sustained output, changeover behaviour or cleaning downtime. Record the run duration, interruptions, start-up and shutdown, and the cleaning or flush steps between products. If the commercial case includes several fruit campaigns, the trial should identify which representative cases were covered and which remain open.
The acceptance decision should list deviations and unresolved items with an owner. A successful final sample does not erase a missing reject sample, an unrecorded handoff or a treatment condition that was not challenged. The line is accepted when the evidence covers the interfaces as well as the individual machines, and when the remaining uncertainty is written down rather than assumed away.
Sources and scope
- FAO, Principles and practices of small- and medium-scale fruit juice processing, Chapter 6
Unit operations, material-condition tables and a caution about laboratory results versus commercial scale-up. Used for qualitative trial-design principles.
- UNIDO, Small-scale fruit and vegetable processing and products
Production-method and quality-assurance context for small-scale fruit processing. No throughput, cost or financial return is inferred from this source.
Sources are cited for technical context. They do not certify a machine, a capacity or a process guarantee for a particular project. External pages were reviewed on 28 October 2026.