How to Spec Press Force (and Get a Ballpark You Can Take to Management)

Start from the joint

Every press job is one of a short list of operations: press-fitting, staking, swaging, crimping, coining, broaching. Each has a real force the part demands. A press-fit's force comes from the interference between shaft and bore, the materials, and the engagement length. Staking and swaging come from the volume of material being moved. That force can be calculated from the part print, and it can be measured directly by running the part on an instrumented press. Bring the part and the print, and the force question gets answered with data. Schmidt sizes presses from measured seating force routinely, which is why our first question is never "how many tons were you thinking?"

What a force-distance curve tells you

Run the joint once on an instrumented press and you get a force-vs-stroke curve: force on one axis, ram travel on the other. It shows the force ramping as the interference engages, the peak at seating, and a signature that repeats when the joint is right. A spike early in the stroke says misalignment or galling. A peak that wanders part to part says the tolerances are wandering too. The curve is the sizing data, and later, on a monitored press, the same curve becomes the quality record, so the trial run pays twice.

Add headroom, then pick the mechanism

Size the press so the worst-case seating force sits comfortably inside its rating instead of at it. A sound habit is to keep the expected peak below roughly two-thirds of the press's rated force, which leaves room for part-to-part variation, tooling wear, and the day the supplier's bore comes in at the tight end of tolerance. Then the mechanism is the real choice:

• Manual, rack-and-pinion: the operator drives the ram through a gear, giving constant force through the whole stroke. The proven workhorse for low-volume benchtop assembly.
• Manual, toggle: the linkage multiplies force sharply at the bottom of the stroke, where a stubborn press-fit needs it. Same bench, different force curve.
• Pneumatic and hydro-pneumatic: compressed air (or air over oil) when the force outgrows an operator's arm, from a couple of kN up to 220 kN on hydro-pneumatic H-frames, in the same benchtop-to-station footprint.
• Electric and servo: PLC-integrable, programmable force and position, no compressed air, and smooth at low speed. Servo presses run from a fraction of a kN to 250 kN with position resolution to 0.1 µm. This is the class the automated line and the audited joint both point at.

For scale across the whole family: hand-lever presses start at a few ounces of force and the servo side reaches 25 tons, all sharing the same monitoring platform. (A kilonewton, if you think in pounds, is about 225 lbf; 10 kN is roughly a ton.)

When monitoring is the requirement

If the part is medical, aerospace, or safety-critical, the machine's real spec is the record it leaves. A monitored press plots the force-distance curve for every cycle and judges it against tolerance windows, good part or bad part, green or red, per part number. On Schmidt's current platform (the PressControl 700 and 7000 as of 2026) that means up to 12 QA observers per part dataset, storage for 24 part datasets so changeover is a recipe load instead of a re-teach, and process documentation that connects to an MES for the audit trail. If your quality system will ever be audited on this joint, buy the record, and let the tonnage follow.

What over-buying costs

Tonnage you don't need is money spent three times: the bigger press, the bigger footprint and guarding around it, and the resolution you gave up (a 100 kN press proving out a 2 kN joint is measuring with a bathroom scale). The force study that stops the guess costs a phone call and a part shipment. This is why "how much force do I actually need" is a development question that comes before the purchase, and it's a question we'd rather answer with your part on a press than with a bigger number for safety's sake.

One caveat before you size anything

If the part can't take straight-line force at all (a fragile housing, an electronic assembly, soft metal that flows), pressing is the wrong method and no tonnage fixes that. Orbital forming rolls the material over under continuous pressure instead of pushing through it. That fork comes before sizing; our pressing and forming page walks it.

A worked sizing, start to finish (illustrative numbers)

A bronze bushing pressed into an aluminum housing, 0.03 mm interference, 12 mm engagement. Calculated seating force lands near 1.8 kN; a trial on an instrumented press measures peaks of 2.0 to 2.2 kN across ten parts as tolerances stack differently. Two-thirds headroom on 2.2 kN says the press should be rated 3.5 kN or better, so this is a 4 kN-class machine: a monitored electric press if the joint needs a record, a bench rack-and-pinion if it doesn't and volume is low. The 20 kN machine a tonnage guess would have bought is the wrong tool for it.

What a ballpark needs from you

Three things: the part (or its print), the operation (press-fit, stake, swage), and whether the joint has to be documented. With those, a budgetary number comes back without a force study; with a sample part, the number comes back with the force measured. Volume and cycle rate refine it from there. If the press will sit at a load station, our guarding guide covers the light curtain question that follows.