The question that decides whether a honey packing room makes money is rarely about the fastest machine. It is honey filling machine capacity batch matching: whether the machine in the room finishes the lot you actually have, in the shift you actually have, without giving away more honey in overfill than the machine saved in wages.
This is written for packers and importers specifying equipment for export. It sets out how to convert a batch in kilograms into a machine specification, why viscosity changes every number in the brochure, what the net quantity rules allow you to give away, and how much each machine class really costs per batch once labour and cleaning are counted.
Start from the batch, not from the brochure
A batch is the quantity of one lot that has to move through the line as one unit: one origin, one moisture reading, one lot code, usually one jar size. Its size is set by what arrives from the apiary or the drum, not by what the machine would like. Honey is a traded agricultural commodity and production volumes are published by country in the FAO FAOSTAT crops and livestock products dataset, which is worth checking against any volume assumption in a business plan before you size a line around it.
Convert the batch into jars first. A 5,000 kg lot packed into 500 g jars is 10,000 jars, before losses. That single number sets everything: at 1,200 jars per hour it is more than one full shift; at 4,000 jars per hour it is under three hours. Nothing else in the specification matters until that division is done, and it is the division most often skipped.
Then count the SKUs that have to run through the same machine. A room packing six formats from one filler spends more hours changing over than filling. Where SKU count is high, a lower peak rate with fast changeover beats a high peak rate with slow changeover, and this is the point most buyers get backwards.
Machine classes and the batch each one suits
Throughput figures below are rates published by equipment makers for honey-service machines. They are not measured across manufacturers on a common test, and every figure is qualified by fill volume, viscosity and operator count. Read them as brackets and then require a run-off test on your own honey before accepting delivery: rates vary by source and model — ask the supplier for a measured test report.
| Machine class | Published throughput | Published accuracy | Batch it suits |
|---|---|---|---|
| Tabletop or single-head piston, semi-automatic | Manual pacing, tens to low hundreds of jars/hour | Varies by source or model — ask the supplier for a measured test report | Up to roughly 500 kg per lot; sample runs; gift and minipack channels |
| Two to four head semi-automatic piston filler | About 1,000–1,200 jars/hour published for four-head piston units | ±0.5% to ±1% commonly published | 500–3,000 kg per lot |
| Inline piston or lobe filler, four to twelve heads | About 1,200–4,000 jars/hour published for honey service | ±0.5% to ±1% commonly published | 3,000–20,000 kg per lot |
| Full servo-driven multi-head line with capping and labelling | Up to about 4,000–7,000 jars/hour for 40–1,000 ml ranges; one maker quotes up to 5,000 jars/hour | ±1% published; tighter figures quoted on some units | Above 20,000 kg per lot, or continuous packing |
The practical reading: most regional packers are buying into the second or third row, and the fourth row only pays for itself when the line runs most days. A multi-head line that runs one day a week is a very expensive way to buy capacity you use at 20%.
Viscosity: why every printed rate needs your own number
Honey’s resistance to flow is not a constant. Figures published by pump manufacturer Viking Pump give about 14,000 centipoise at 20 degrees Celsius against about 1,300 centipoise at 40 degrees Celsius, roughly an order of magnitude across that twenty-degree band. Specific gravity sits around 1.4. Any throughput figure quoted without the product temperature and moisture attached to it is not a specification, it is a wish.
| Product temperature (°C) | Approximate viscosity (cP) | Effect at the nozzle | Action in the room |
|---|---|---|---|
| 20 | 14,000 | Slow draw, long dwell, stringing at cut-off | Avoid filling at this temperature unless the format is wide-mouth |
| 30–35 | Between the two figures above — varies by source or model, ask for a measured test report | Usable on most piston heads | The common working band for honey that is not too coarse |
| 40 | 1,300 | Free flow, short cycle, minimal stringing | Fastest filling; hold here and do not go higher |
| Above 40 | Falls further | Quality risk exceeds the speed gain | Not recommended: HMF rises and diastase falls with heat |
The heating method matters as much as the number. Jacketed hoppers and heated product circuits using warm water keep viscosity steady through the whole run. Direct elements create hot spots near the wall, and hot spots are exactly what the Codex freshness limits punish: the standard caps HMF at 40 mg per kg and requires diastase of at least 8 Schade units. Warming the room or the water, and logging the temperature once an hour, costs almost nothing and is the cheapest quality control in the building.
Accuracy, giveaway and the net quantity rules
In the United States, net quantity is checked against NIST Handbook 133, administered by the NIST Office of Weights and Measures. Two conditions decide pass or fail. The average net content of the packages in the sample must be at least the labelled quantity, and no individual package may fall below the Maximum Allowable Variation for that label weight. Published examples put the MAV for a 454 g label at 9 g, and describe MAV as running from about 10% of label weight for weights under 36 g to about 2% for packages above roughly 24.67 kg, where a 25 kg package carries a 0.5 kg MAV.
Turn that into giveaway arithmetic for a honey room. If the target is 500 g and the filler averages 0.5% over — 2.5 g per jar — a 10,000-jar lot gives away 25 kg of honey. At 1% over, the same lot gives away 50 kg. Published accuracy figures for servo and piston honey fillers run around ±0.5% to ±1%, so moving from a loose filler to a tight one is often worth more per year than the difference between two machine classes, and it is the argument that wins internal approval.
A checkweigher is how you keep it there. Catching drift in the first twenty minutes of a run costs one jar; catching it at the end of the shift costs the lot. Servo-driven systems also allow the dose to be corrected from the touch panel without stopping, which is worth specifying if your jars vary in weight from batch to batch.
Hygiene sits alongside accuracy. The current good manufacturing practice rules in 21 CFR Part 110 set out what the room around the machine has to look like: cleanable surfaces, protected product, and controlled conditions. In practice that means sanitary tri-clamp connections rather than threaded pipe, product contact parts identified as 304 or 316 stainless with certificates, and a cleaning routine the operator can actually complete inside the shift.
Changeover, cleaning and the hidden hours
Published throughput measures steady-state filling. Real days are not steady-state. Count the hours that are not filling:
Warm-up. Product has to reach working temperature before the first acceptable dose comes off the head. On a cold morning with crystallised stock, this can be an hour.
Changeover between formats. Published figures for quick-change systems run from under a minute to a couple of minutes on some units, but that applies only where the change parts and a written sequence already exist. Without them it is thirty minutes and a mop.
Cleaning at the end of the lot. Honey leaves traces that crystallise and then attract moisture. Any part that cannot be stripped without tools will eventually be the part that is skipped.
Capping and labelling. A filler with no downstream capacity simply moves the queue to the next table. Rate the line at its slowest station, not the fastest.
The rule of thumb that survives contact with real rooms: take the published rate, halve it for the first month of operation, and size the batch around that. If the batch still fits the shift, the machine is right.
Season windows: northern and southern calendar
Packing capacity is bought against a harvest that arrives once and then disappears.
Northern hemisphere. Extraction generally runs May to August, so the heavy packing months are September through December, and the peak strain on an oversized filling set-up coincides with festive-format demand. Equipment must be installed and commissioned by late summer; a machine arriving in October has missed the year it was bought for.
Southern hemisphere. The same months are build-up. Extraction runs roughly November to March, which moves the packing peak into January to April. Ordering in mid-year is normal there, and holding a shorter, higher-rate busy season is often better served by one fast line than two slow ones.
For exporters the planning implication repeats: a Chinese-built line shipping in July supports the southern season, while anything wanted in Europe or North America for autumn has to leave the factory much earlier. Lead time is part of the capacity calculation, not a footnote to it.
What one 5,000 kg batch costs to pack on each machine class
The comparison below fixes the batch at 5,000 kg of honey into 10,000 jars of 500 g, and compares machine classes on the things that are measurable without a price list. Operator-hours come from published throughput divided into 10,000 jars at the stated rate, with no allowance for changeover or cleaning, so treat every figure as a best case and add your own realism to it. Capital cost is deliberately left out: it varies too widely by specification, and any figure quoted here would be meaningless next month.
| Machine class | Published rate (jars/h) | Best-case filling time for 10,000 jars (h) | Operators typically required | Where the cost actually lands |
|---|---|---|---|---|
| Single-head semi-automatic | Rate varies by source or model — ask the supplier for a measured test report | Over one shift; plan as a multi-day lot | 1–2 | Labour dominates; machine cost almost irrelevant |
| 4-head piston, semi-automatic | 1,000–1,200 (published four-head figure) | 8.3–10.0 | 1–2 | One operator per head deployment; fits a single long shift |
| Inline 4–12 head piston or lobe | 1,200–4,000 (published honey range) | 2.5–8.3 | 1–2 | Cleaning time and change parts, not wages |
| Full servo multi-head line | Up to 4,000–7,000 (published top of range) | 1.4–2.5 | 1 plus downstream attendants | Capital recovery and capping/labelling bottlenecks |
Spent at the right scale, the qualification is simply this: below about 3,000 kg per lot, labour is the cost and buying more machine buys nothing. Between roughly 3,000 and 20,000 kg, throughput and changeover decide it. Above that, the constraint leaves the filler entirely and moves to capping, labelling and how many pallets the loading bay can hold.
Two lines that fit the middle band are worth pricing against each other before deciding. The automatic and manual honey filling machine covers rooms that still want a manual override for small lots, and it is usually paired with a stainless steel honey storage tank holding the lot at working temperature. The wider plant equipment, presses, melters and insemination devices sits under beekeeping machinery, and mixed containers are loaded from there. Production context to set against any volume plan is available through the FAO pollination programme.