Every tablet manufacturing line loses some material between the first weighing and the final sealed carton. That's expected — the question that matters isn't whether yield loss happens, it's whether the loss at each stage is predictable, within range, and accounted for. When someone asks "why is our yield low this quarter," the honest answer is almost never one single cause. It's usually a combination of small, stage-specific losses that compound.
This walks through where loss typically occurs at each of the four major stages, in the order material actually moves through the plant, so you can compare where your own numbers are falling out of range.
1. Granulation
Where the earliest and most variable losses occur
Granulation tends to have the widest swing in loss percentage between batches, because it involves more transfers, more equipment surfaces, and — for wet granulation — a drying step that changes the weight of the batch for reasons that have nothing to do with lost material.
| Loss point | Why it happens |
|---|---|
| Blender/mixer hold-up | Powder or wet mass sticking to blades, walls, and baffles; increases with batch density and moisture content |
| Screening and milling | Fines lost to dust extraction during dry milling, or granule retained in the mill screen and housing |
| Drying loss | This is a moisture loss, not a material loss — but if it isn't separated out in the reconciliation, it looks like a shortfall |
| Transfer between vessels | Material left in transfer chutes, IBCs, or drums when moving between blending, granulation, and drying equipment |
| Sampling | In-process samples pulled for LOD, particle size, or blend uniformity testing, often not weighed back into the loss column |
The practical fix here is less about reducing loss and more about separating moisture loss from material loss before anyone panics about a "low" granulation yield. A batch can show a 4–5% weight reduction that's entirely explained by drying and nothing else.
Check granulation loss against its expected range:
2. Compression
Where in-process rejects do most of the damage
Compression loss is usually more visible than granulation loss because it happens in real time, on the compression floor, and rejects are typically weighed or counted as they're removed. The risk here isn't hidden loss — it's under-recording it in the moment because the line is running and rejects get swept aside rather than logged.
| Loss point | Why it happens |
|---|---|
| Start-up and shutdown rejects | Tablets produced before the machine stabilizes on weight, thickness, and hardness, and again during shutdown as the feed runs low |
| In-process quality rejects | Tablets pulled for being outside weight, hardness, thickness, or friability limits during routine IPQC checks |
| Hopper residue | Granule remaining in the feed hopper and feed frame at the end of the run, which varies by machine design and granule flow properties |
| Sampling for testing | Tablets pulled for dissolution, hardness, friability, or retention samples |
| Punch/die adhesion | Material sticking to punch faces, especially with formulations prone to sticking, requiring periodic cleaning that removes tablets from the count |
| Dust collector loss | Fine powder pulled off tablet edges by the press's dust extraction system during compression, particularly with softer granules or higher compression speeds — small per-batch but consistent, and easy to overlook since it never appears as a visible reject |
A compression stage that's consistently losing more than its historical average is worth investigating for punch wear, granule flow issues, or a lubricant problem before assuming it's just "normal variation."
Check compression loss against its expected range:
3. Coating
Where percentage losses look small but absolute losses can be large
Coating loss is often reported as a small percentage, which can be misleading — on a large batch, even a 1–2% loss represents a meaningful number of tablets. Coating is also where sampling for in-process weight gain checks happens most frequently, because weight gain is checked repeatedly through the run rather than once at the end.
| Loss point | Why it happens |
|---|---|
| Pan loading/unloading | Tablets remaining in the pan, scoop, or transfer chute during loading and discharge |
| In-process weight checks | Samples pulled at intervals throughout the run to track weight gain — frequent, small withdrawals that add up |
| Spray-related loss | Overspray and fines generated during the coating process itself, particularly with aggressive spray rates or poor atomization |
| Core tablet attrition | Chipping or breakage of uncoated cores during tumbling, more common with fragile or high-friability cores |
| Exhaust/dedusting | Fine coated dust removed by the pan's dedusting or exhaust system |
If coating loss is consistently higher than expected, it's worth checking spray rate against pan speed and inlet air volume before assuming the core tablets themselves are the problem — the two are frequently connected.
Check coating loss against its expected range:
4. Packing
Where losses are the easiest to explain and the easiest to under-record
Packing loss should, in theory, be the most straightforward stage to reconcile, because rejects are usually visible and machine-generated rather than process-driven. In practice, this is often where reconciliation gaps get discovered — not created — because it's the last stage before the final count is compared against the original batch size.
| Loss point | Why it happens |
|---|---|
| Line rejects | Blister strips or bottles rejected for print defects, seal failure, or missing tablets, removed automatically or manually during the run |
| Line clearance and changeover | Partial strips or containers set aside during startup and changeover between batches |
| Retention and QC samples | Finished packs pulled for stability, retention, or release testing — legitimate but frequently the last thing added back into the reconciliation |
| Machine jams and startup waste | Blister strips or cartons damaged during machine start-up or jam clearance |
Because packing losses are usually well-documented on the line, a packing-stage reconciliation gap is more often a paperwork problem — a reject count that wasn't transferred from the line log to the batch record — than an actual physical loss.
Check packing loss against its expected range:
Putting it together
A useful way to think about total yield loss is that it's rarely one stage's fault. Granulation loses a little to hold-up and dust, compression loses a little to rejects and hopper residue, coating loses a little to sampling and overspray, and packing loses a little to line rejects. Individually, each is small and explainable. Reviewed stage by stage instead of only as one final percentage, patterns become visible — a specific stage trending worse over several batches usually points to a specific, fixable cause rather than random variation.
Tracking loss by stage, batch after batch, is what turns a reconciliation number from a pass/fail check into a genuinely useful diagnostic tool. If a batch actually fails reconciliation, the next step is a structured investigation — see How to Investigate a Batch Reconciliation Failure for the process to follow.