Three published sources describe what a tonne of RCN becomes. They are mutually incompatible, and this chapter says so rather than presenting a tidy balance that hides the disagreement.

The balance does not close

ReadingShellCNSLKernelTestaRejectsLoss
A50%20–25%, inside the 50% shell25–30% incl. testa(in kernel)
B~70%not stated~20% (derived by subtraction)~3%~3%~4%
C22–24% saleable packed kernel

Reading A gives 25–30% kernel. Reading B leaves about 20%. The production figure used throughout this guide is 22–24%. The shell fraction differs between A and B by twenty percentage points of the whole nut, and no arithmetic reconciles them.

Part of the divergence is measurement point rather than disagreement. Shell weighed at the shelling station is wet with condensate, gelled CNSL and adhering kernel fragments. Shell weighed dry after a week in a yard is a different number. Neither source states which it did.

So build outward from the number you trust and label the rest:

Fraction% of RCN as receivedBasisMeasurement point
Saleable packed kernel, all grades22–24Production KORWeighbridge at packing, after final conditioning
Shell50–70Published range, unresolvedShelling station, wet, with CNSL and adhering fragments
Testa~3PublishedPeeler cyclone / testa collection
Rejected kernel~3PublishedAfter revision at the shelling sort
Moisture and process lossbalanceDerived by differenceNot directly weighed anywhere

Weigh your own. ComCashew’s instruction is the right one and almost nobody follows it: “It is especially important to record weight loss of the kernel throughout the process.” Four weighing points — intake, post-shelling kernel, post-peeling kernel, packed kernel — plus shell and testa out, logged daily, give you a balance for your own nuts inside one season.

Switch between the three readings. The kernel fraction barely moves; the shell fraction moves by twenty points of the whole nut.

Saleable packed kernelShellTestaRejected kernelResidual / unresolved
Fractions of 100 kg of raw cashew nuts
Fraction% of RCNMeasurement point
Saleable packed kernel22–24Weighbridge at packing, after final conditioning
Shell50–70Shelling station — wet, with CNSL and adhering fragments
Testa3Peeler cyclone
Rejected kernel3After revision at the shelling sort
Residual / unresolved0–22Not weighed anywhere — derived by difference
The shell fraction differs between the two published readings by twenty percentage points of the whole nut, and no arithmetic reconciles them. Part of that is measurement point, not disagreement: shell weighed wet at the shelling station is a different number from shell weighed dry after a week in a yard. Weigh your own — four points (intake, post-shelling, post-peeling, packed) give you a balance that is true for your plant inside one season.

One separation that must stay rigid

Quality-check KOR is a laboratory cutting test used to price a lot of RCN, in pounds per 80 kg bag. ComCashew works the example: an outturn of 49 lbs per 80 kg bag means 49 lbs — 22.2 kg — of kernel usable by the manufacturer from 80 kg of nuts, which is about 27.75%.

Production KOR is kilograms of final packed saleable kernel per 100 kg of RCN actually processed: 22–24 kg.

These are two different measures and one must never be compared to the other. Buyers price on the first. Your bank is repaid out of the second.

Where yield is lost, stage by stage

StageMechanismPublished figureMass or grade?
StorageKOR decline over timeNo published curve existsBoth, unquantified
ShellingBrokens; uncut nuts; fragments left in shellBroken under 3%, 3–5% claimed; uncut under 5%Mostly grade
DryingMoisture removal; scorchingTo 3–4% moistureMass (water) and grade (colour)
PeelingBreakage; testa removed4–8% breakage; testa ~3% of RCNGrade, plus real mass as testa
GradingMachine-induced breakage“Consistently experienced problems”Grade
PackingFinal moisture adjustment; QC and detector rejectsTo 3–5%Mass (water)

Most of what a cashew plant loses is grade, not mass. A broken kernel is still in the bag, worth roughly 16% less. A scorched kernel is still in the bag, on a different price line.

That is why plants with identical KOR figures can post very different results, and why the whole-to-split and white-to-scorched mix belongs on the daily report next to tonnage. A tonnage-only report will show a good month in which you lost money.

The energy side, and the surplus nobody sells

Shell is the reliable by-product. Factories typically consume only 5–25% of the shell they produce for their own thermal needs — a surplus of four to twenty times — at a calorific value around 5,000 kcal/kg (about 20.9 MJ/kg), with briquettes near 4,800 kcal/kg.

Burn what the boiler needs and sell the rest. A plant that treats shell purely as waste to be disposed of is paying to remove a fuel it could sell, and the disposal cost appears in the accounts while the foregone revenue does not.

Measure energy per tonne processed as a standing metric. Without it, boiler efficiency is invisible, and a boiler quietly losing efficiency looks exactly like a raw-material price increase in the monthly accounts.