Choose a tea bag packaging machine by fixing the finished bag first — shape, closure, filter material and fill weight — because that one specification removes most machine types from the list before you open a datasheet. Then match the filling system to how your tea flows, size the output from good bags per hour over your real running hours, and buy against an acceptance test run on your own tea rather than against a rated speed.
Fix the finished bag before you look at any machine, because shape, closure and filter material each point to a different machine architecture, and a machine built for one bag type is not converted to another. The table gives the four cases and what to confirm for each.
| Finished product | Machine type to evaluate | Key point to confirm |
|---|---|---|
| Flat, single-chamber tea bag | Single-chamber filter-bag machine | Bag dimensions, heat-sealable filter material, and whether string and tag are applied on the machine. IMA’s EC24 range is this configuration. |
| Double-chamber tea bag | Dedicated double-chamber forming and closing machine | The exact closure method and the filter paper it requires. Some machines, such as IMA’s MD20, run non-heat-sealable paper and close the bag with knots. |
| Pyramid tea bag | Pyramid-bag forming and sealing machine | Compatibility with the chosen filter material and bag dimensions. Ultrasonic machines are available; some, such as FUSO’s FP-100S, also produce rectangular bags. |
| Already-made tea bags requiring outer packaging | Envelope wrapper, flow wrapper, counting system, or cartoner | Whether you need individual wrapping, grouped packs, retail cartons, or a combination. |
Fill weight filters the list next. Of the machines in the table, the single-chamber EC24 is rated to about 2.2 g of tea per bag and the double-chamber MD20 to 4 g, so a fill weight above 2.2 g has already ruled out the first row before dosing or speed are considered. A pyramid machine is specified by edge length instead — 50 to 80 mm on the FP-100S — and the fill weight follows from the volume the bag closes around.
Two points in the table read differently from what buyers expect. On a pyramid machine the string and tag are usually not applied by the machine at all: the FP-100S runs pre-tagged filter material and carries no string or tag unit, so for pyramids the question moves to the filter-material supplier.
An outer envelope is often an option on the bag machine itself rather than a separate wrapper. The EC24 Multipla discharges bags already in heat-sealed envelopes, and the MD20 runs either naked bags (no envelope) or enveloped ones, so the last row of the table applies when bags arrive naked from elsewhere, or when the machine you have chosen lacks the option and a flow wrapper or envelope wrapper has to follow it.
The filling system is chosen by how the tea flows and by what a wrong weight costs, not by the bag format: the same pyramid machine is sold with a volumetric slide doser as standard and a load-cell weigher or an auger as options, and the wrong choice is what produces dust, broken leaf and giveaway on a machine that is otherwise right. Three cases cover most teas.
A volumetric doser — a cup or slide that fills a fixed volume — is the right system for CTC grades, fannings and any blend that pours the same way every time, because it is the fastest and simplest option and its accuracy is limited only by how much the tea’s bulk density varies. Volumetric dosing fills volume, not weight, so a denser lot goes into the bag heavier by the same proportion.
Before you accept a volumetric system, measure bulk density across several lots of each blend. If the spread is wider than your weight tolerance, either set the target on the densest lot and accept the giveaway on the others, or move to weigh-based dosing.
Whole leaf, flowers and fruit pieces need weigh-based dosing on a gentle feed, because the volume per gram changes from one bag to the next and a volumetric cup cannot hold a weight. A weigh cell settles each dose by weight; the feed into it — a vibratory tray or a rotating hopper built for non-free-flowing product — has to move the blend without crushing leaf or compacting it into a bridge above the outlet.
Blends also separate in the hopper: heavy fruit pieces sink, light petals rise, and the first and last bags of a run come out with different recipes. Ask how the machine keeps the hopper level and agitated, because a weigh cell confirms weight and nothing else.
Matcha, finely ground herbs and dust grades need an auger filler with an agitated hopper, because aerated powder floods through a volumetric cup and settled powder packs solid in it. An auger meters by screw rotation while the agitator holds the powder at a steady density at the screw inlet, which is what makes the dose repeatable.
Powder also decides the filter material and the seal. Fine particles pass through an open mesh, and powder that settles in the seal area produces a bag that leaks at the seam, so the doser has to place the powder clear of the seal zone.
Size the machine on good bags per hour over your real running hours, not on the rated bags per minute, because the rated figure assumes the machine never stops and rejects nothing, and no tea line runs that way. Work backwards from annual volume:
Required rated speed (bags/min) = annual good bags ÷ (running hours per year × 60 × line efficiency)
Line efficiency is the fraction of running time the machine actually spends producing good bags, and it is the input most buyers set too high. If you have a measured overall equipment effectiveness for a comparable line, use it; if not, plan on 60%, which is closer to what most plants achieve than the 85% usually quoted as world class.
Rated speed follows machine architecture far more than bag shape, which is the opposite of what most buyers assume. Among the machines in the section 1 table, the ultrasonic pyramid machine is rated at 80 bags/min, the knotted double-chamber machine at 200, the single-chamber heat-seal machine at 240 naked bags/min, and a single-chamber line forming joined pairs of bags at up to 2,000. Closure method and the number of bags formed per cycle set the speed; the shape does not.
Two rated speeds are comparable only when quoted for the same bag. Rated figures are normally given for the smallest bag and lightest fill, and a larger bag, a heavier dose or whole leaf that needs a slower feed each bring the real figure down. Ask for the rated speed at your bag size, your fill weight and your tea, and treat any figure without those conditions as unproven.
The filter material and the machine’s sealing method are one decision, not two: heat-seal paper works only on heated sealing jaws, ultrasonic materials only on an ultrasonic machine, and non-heat-seal paper only on a machine that closes the bag with a knot or a fold. Choose the material family with the machine, and prove the exact grade on the machine before delivery.
Heat-seal paper seals because it is not pure cellulose: it carries thermoplastic fibres that melt in the jaw and bond the two layers. Conventional grades use plastic fibres; plastic-free grades replace them with PLA, a biopolymer usually made from corn starch, which seals at different jaw settings. If a plastic-free bag is required on a heat-seal machine, name the PLA grade in the enquiry, because a machine proven on conventional paper is not proven on PLA paper.
Ultrasonic machines join the material with vibration rather than heat and accept nylon, non-woven fabrics and PLA-based materials made for ultrasonic sealing; they also fix the roll width — 120 to 180 mm on the FP-100S — so the filter supplier’s roll has to match the machine, not the other way round. Non-heat-seal paper carries no sealing fibre at all, which is why a knotted double-chamber machine can run it and a heat-seal machine cannot.
Buy the level of automation your labour cost and your outer pack justify, and no more, because the bag-forming head is the same whether the machine discharges naked bags into a tray or feeds a cartoner, and everything added above it removes an operator while adding a station that can stop the line.
The levels rise in steps. Naked bags with hand packing is the cheapest machine and the most labour: one operator on the machine plus the packers behind it. A machine-applied envelope removes the hand wrapping, a counting and cartoning attachment removes the hand counting, and a full line adds flow wrapping for a hermetic or modified-atmosphere pack, cartoning, case packing and palletizing.
Each step also costs speed. The same single-chamber machine is rated at 240 naked bags per minute and 200 with envelopes, and an integrated line runs at the speed of its slowest station and stops when any station stops, so a cartoner rated below the bag machine sets the line speed and every added station lowers the efficiency planned in section 3. A first-time buyer moving off hand work usually does best with envelope and counting built into the machine and hand cartoning behind it, adding secondary packaging once the bag machine runs at a known efficiency.
Inspection is specified with the automation level, not after it. A checkweigher after the bag machine catches dosing drift and a metal detector after the envelope catches metal that arrived in the tea, and each has to be quoted with its reject mechanism, because a detector that stops the line instead of rejecting the bag costs more output than the defects it finds.
Check the three things that are not on the datasheet: how long the machine takes to clean between blends, what the building has to supply before it will run, and where the parts and the technician come from when it stops. Each one is a cost that arrives after the invoice is paid.
Tea is a dry product, so cleaning is brush and vacuum rather than washdown, and the risk is carry-over: peppermint after chamomile, or a fruit blend after a plain black, leaves flavour and allergen traces in the hopper, the doser and the forming head. Ask for tool-free access to every surface the tea touches, stainless steel on those surfaces, and no ledge or pocket where dust collects unseen.
Power and compressed air differ several-fold between the machine types in section 1, so get the utility sheet before the quote is accepted, not after delivery. A machine specified for 200 V three-phase — the FP-100S is — needs a transformer on a plant with a different supply, and a machine that wants 6 bar at its inlet needs a compressor that holds 6 bar at the machine after line losses, not only at the compressor.
Confirm supply voltage and frequency, air pressure and dryness, floor loading and the access route into the building on the quote, so the commissioning engineer is not the one who discovers them.
Ask where the nearest technician who has commissioned this model is based, and price the spares with the machine. The parts that wear are the sealing jaws or ultrasonic horn, the cutting knives, the doser cups or auger, and the thread and tag feed, and a stopped line waits on whichever of these is not in stock. Get a recommended spares list priced into the quote with the lead time on each item, remote diagnostic access, and commissioning and training days in writing.
Compare machines on what one good finished bag costs over the machine’s working life, because the purchase price is paid once and the reject rate, the giveaway and the operator count are paid on every bag. Build this figure for each quote:
Cost per good bag = (machine price ÷ years in service + annual labour + annual filter, tag, thread and envelope material + annual energy and air + annual maintenance and spares) ÷ good bags per year
Good bags per year, the denominator, does most of the work. Rejects leave the denominator and also consume the filter, tag, thread and tea that went into them, so a machine that rejects more at a higher rated speed can deliver fewer good bags per hour of material than a slower one. Giveaway sits in the material line: overfill to cover a wide dosing tolerance is tea paid for and not sold, and it costs most on the teas that cost most per gram.
Write the acceptance test into the purchase contract, run on your tea, your filter material and your bag size, and tie the final payment to passing it, because a factory test on the supplier’s standard tea proves that the machine works, not that it works for you. The table lists what to agree in advance and what to measure.
| Test area | What to agree and measure |
|---|---|
| Sustained output | Good finished bags per elapsed hour, with the treatment of routine stops and replenishment defined before the run. |
| Fill accuracy | Net tea weight excluding packaging: average, variation, and the proportion outside agreed limits — not the average alone. |
| Bag and envelope quality | Closure integrity, particle leakage, tag and string attachment, appearance, and brewing performance. |
| Product handling | Leaf damage, dust generation, bridging or blockage, and blend consistency through the run. |
| Changeover and cleaning | Time from the last acceptable bag of one product to the first acceptable bag of the next, including cleaning and startup waste. |
| Inspection and rejection | Demonstrated detection and removal of the defects the supplied inspection system is intended to identify. |
Run the test twice: at the supplier’s factory before shipment, and again at your plant after installation on your power, your air and your operators. The numbers move between the two runs, and the second set is the one that goes into the cost per good bag.
Send every supplier the same enquiry with the finished bag specified first, so the quotes come back comparable and each supplier is quoting your machine rather than its nearest standard model. The enquiry carries:
Send a sample of each tea with the enquiry, so the supplier’s answer on dosing is based on your product and not on its description.
