PRECISION ROAST / SOURCES / RAW MATERIALS
The receiving sample should represent the input batch, and the result should show not only the removed defects, but also the loss of suitable grain.
Before roasting, you can divide the flow by size and visible differences, remove some heavy, light or metal impurities by separate nodes, select the sequence according to the known composition of the lot and the purpose of the product, assess the quality by material balance and repeated manual check of all fractions.
Mark the item closed only when there is a document, measurement, observation or designated responsible person.
select a representative sample with preserved identity
describe defects with photographs and physical samples
assign a sign: size, color, density, metal or light impurity
coordinate the order of nodes and number of passes
weigh the entrance, the fitting exit and every break
Check pass and false selection manually
to establish purification and traceability between batches
Each stage preserves the boundary between the revealed fact, the editorial method and what still needs to be confirmed.
A sample from the top layer of the bag or a pre-selected handful may not represent the batch. Describe the method of sampling, combining and dividing the sample, save the batch number and the original control sample. The supplier of the machine must receive the material without prior improvement.
The sieve sees the size, the optics see the available external difference, the magnet sees the ferromagnetic material, and the destoner uses the different flow behaviors. One defect may not have a stable external feature, and these constraints are fixed before the test.
Pre-removal of large impurities or sizing alignment can stabilize the next step, but adds output streams, heights and purification.Compare multiple sequences on a single material and don't skip the product again without a separate result label.
Weigh the original sample, the main product, all the fractions and the technical residue. Manual check of the received flow shows the pass, and the analysis of the cutoff shows the loss of good grain. The result without the sum of the flows does not allow you to estimate the useful yield.
Record the recipe, the tempo, the surrounding conditions, the number of passes, the air settings and the shelf life criteria. To commission, prepare a new closed sample to test reproducibility, rather than repeating the work on a familiar sample.
Short answers retain the same limitations as the main material.
Not always. It depends on the purpose and the heterogeneity of the batch. Compare the sequences on one sample and evaluate whether calibration improves the useful result of the next step.
Only if a particular system consistently sees them as a predetermined difference, and the test confirms this. For density differences, a separate destoner is usually considered; metal inclusions require their own control.
There's no reason to think it's universal. Origin, processing, humidity and appearance change, so the recipe is tested on each significantly different batch group.
It's identified, weighed, checked and sent according to an approved procedure, so you can't bring it back to the main stream before analysis, otherwise the material balance and traceability are lost.
Compare optical and mechanical models according to the approved product stage.