Most print shops do not buy a machine for one product. They buy for an order book that mixes flat boards, small parts, bottles and promotional items, and then discover that the purchase decision was made on print speed rather than on changeover.

This guide compares a flatbed and cylinder printer for mixed production: what each machine actually covers, where rotation is the only answer, how changeover and fixtures decide the economics, how to model capacity per product family, and the three mistakes that most often turn a mixed order book into a disappointing investment.

What a mixed order book actually contains

Mixed work is best described by geometry, not by product name. Panels and boards, small rigid parts, curved or cylindrical objects, and thin flexible pieces each impose a different handling problem.

Start by counting your own jobs in those four categories, weighted by revenue rather than by order count. Most shops find that flat rigid work dominates by value even when cylindrical items dominate by interest, because bottles and drinkware attract attention while signs and panels pay the rent. That weighting is the first input to the machine decision, and it is usually different from what the shop assumed before counting.

Then note which categories share a preparation route. If acrylic, PVC and metal all pass through the same cleaning and priming steps, one machine can cover them with a single qualification library. If cylindrical work needs a different nozzle strategy, different fixtures and its own quality checks, it is a separate production line in practice, even if it stands next to the flatbed.

Where a flatbed wins

A flatbed wins on substrate variety, part size and handling simplicity. Rigid parts sit on the table, are held by vacuum or a fixture, and print in one pass without a rotation step.

That single-pass geometry removes the registration problem that rotation introduces. A board cannot drift around an axis, so artwork placement is a function of the file and the loading position rather than of a rotating fixture, which is why flatbed work is easier to standardise across operators and shifts. Platform size then decides what can be nested: a wide-format machine in the 2500 by 1300 mm class accepts full sheets and multiple small parts in one cycle, which cuts cycle time per part on short-run work.

Thickness range and weight limits are the boundaries worth checking. Machines are quoted with a media thickness range and a maximum weight per square metre, and those two numbers decide whether a heavy stone or glass panel can be run at all. Where the order book includes such material, confirm both numbers against the actual parts before comparing prices.

Where a cylinder machine is the only answer

A cylinder machine is the only practical route for tapered and cylindrical parts that must be decorated around the circumference with consistent registration and a controlled seam.

Rotation is not a convenience in that case; it is the mechanism that presents a continuous surface to the printhead. Clamping, diameter range, taper compensation and length limits become the specification that matters, and they vary enough between machines that a sample test on your own part is the only reliable comparison. A machine quoted for bottles from Ø60 mm to Ø170 mm with an optional range to Ø240 mm, and with seven standard fixtures plus custom tooling, describes a capability envelope rather than a universal answer - your part has to sit inside it.

Cylinder work also changes the workflow around the machine. Loading takes longer per part, first-article checks involve the seam, and packing has to protect a curved printed surface rather than a flat one. Those differences belong in the capacity model, because they decide the cost per accepted piece just as much as print speed does.

Can one machine cover both?

Only if the fixture holds quality and cycle time.

Some flatbeds accept rotary fixtures and hybrid machines exist. The practical test is whether the combined configuration holds quality and cycle time on your round parts, not whether a fixture can physically hold them.

Ask three questions of any dual-purpose claim. What is the taper range the rotation system can follow? How is the part registered after loading, and how much of the circumference can be printed without a seam or a repositioning step? What cycle time does the fixture deliver on your part, including loading and unloading? A supplier who answers those three with measurements is describing a capability; one who answers with a photograph is describing a hope.

Then decide honestly whether a compromise machine is better than two machines. For a shop with a small share of cylindrical work, a flexible fixture may be the right answer, because the alternative is capital parked most of the week. For a shop where drinkware is the margin, a dedicated cylinder machine usually pays back faster than any attachment.

Changeover is the real constraint

On mixed work, changeover usually costs more than printing. Fixture setup, file and profile checks, first-article approval and cleaning are repeated every time the product changes.

Those minutes are invisible in a specification sheet and very visible in a quotation. A job that takes four minutes of printing and twelve minutes of setup has a cost structure driven by setup, and if the order book changes products six times a shift, the machine's rated speed is almost irrelevant. This is why shops that succeed with mixed work invest in fixtures and standard work before they invest in a faster machine.

The practical countermeasures are unglamorous: reusable fixtures with recorded registration positions, a photo of the correct loading arrangement at the machine, standard files and profiles per product family, and a first-article check that takes a known number of minutes. Each one removes setup time from every repeat order, which is where the margin in mixed work actually lives. Where the machine guards and maintenance isolation matter for the operators, the expectations published by HSE in the United Kingdom and by OSHA in the United States describe how a production machine of this class is assessed after installation.

Modelling capacity for mixed orders

Model capacity per product family and then add the families up in the sequence you plan to run. Four numbers do most of the work: parts per cycle, cycles per hour including loading and curing, changeover minutes, and reject rate.

Build the model in a spreadsheet that a production manager can edit, not in the purchase business case alone. For each family, calculate accepted parts per hour after rejects, then apply the changeover cost for each switch between families. The output tells you whether the constraint is the machine, the fixture, the operator or the drying and packing step - and it usually identifies a bottleneck that no machine purchase can fix.

Where a batch decision has to be defensible rather than approximate, base the inspection plan on a documented scheme: the acceptance-sampling method published in the NIST engineering handbook is a practical reference for how many pieces to check and how to read the result. Where the machine will be placed on the market in the European Union, the compliance scope also belongs in the plan: machinery is covered by the Machinery Regulation (EU) 2023/1230, which the European Commission describes on its machinery legislation page.

What buyers get wrong

Three errors recur in mixed-order purchases, and each is preventable with a pilot rather than an argument.

• Buying on headline speed. Rated output is measured on one product, one layer structure and one pass count. The number that matters is accepted parts per shift across your actual mix.
• Ignoring fixture and changeover cost. A cheap machine with slow changeover can cost more per part than a higher-priced machine with better tooling, particularly on short runs.
• Assuming an attachment equals a second machine. Rotary fixtures extend a flatbed's coverage; they rarely match a dedicated cylinder machine on taper, cycle time or seam control.

The other quiet error is buying for the order book you hope to win rather than the one you have. A pilot on the three products that represent most of your current value produces better information in a week than a year of specification comparison, and it also produces the fixtures and settings that make the first orders profitable.

Flatbed vs cylinder printer: a decision table for mixed order books

Use the table below to test your own mix, then verify the conclusion with a pilot on the products that carry most of your value.

Once the machine decision is made, the rest of the work is standardisation: recorded fixtures, approved settings per family, and a first-article routine that a new operator can follow. That combination, not the speed figure in the brochure, is what makes a mixed order book profitable. Where the parts will be sold into markets with substance restrictions, keep the ink and primer documentation with the qualification record - the requirements summarised under the European Union's REACH regulation are part of that file. The printhead technology is worth understanding for the same reason: standard industrial piezo printheads are documented by their makers, and Ricoh's industrial inkjet head documentation explains how head geometry and channel configuration affect the detail you can hold on small parts.