Why Air Can’t Do It Alone: Five Production Cycles Where Pneumo-Hydraulic Feed Control Is the Right Engineering Answer

By Luisa Piovanelli — Technical Lead, Enerfluid SNC, Brescia

Most of the time, when an engineer asks me how to power a station, my honest answer is: use air. It is already in the building. It costs nothing while the station sits idle. It runs no power unit, throws off no heat, and the worst thing it leaks is more air. For a huge amount of assembly work, air is simply the right answer, and the engineer who reaches for it first is usually right.

So this is the question I find myself coming back to. If air is this cheap, this available, this clean — why do whole categories of precision work quietly refuse to run on air alone? Riveting. Clinching. Press-fit insertion. Die clamping. Peck-feed drilling.

For a long time I assumed the answer was force. It isn’t. These five cycles cover an enormous range — from around 400 lb of feed load on a small drill carriage up to something near 30 tonnes on heavy clinching. If raw force were the dividing line, they would have nothing in common. What actually connects them is quieter, and for me it is the more useful thing to understand: each one needs the speed, the compact size, and the economy of compressed air — but air alone cannot give the controlled, repeatable behaviour the operation needs at the one moment that decides whether the part is good or scrap.

In four of them — riveting, clinching and self-pierce riveting, bushing and bearing insertion, and die clamping — what’s missing is consistent end-of-stroke force, on a defined stroke, delivered intermittently. In the fifth, peck-feed drilling, what’s missing is consistent feed velocity and a controlled retract on a moving tool. Different symptom, same disease. And once you see the disease clearly, you can also see exactly where the cure stops working and something else has to take over — which, for me, is the part that matters most, because it is the part most suppliers will never tell you.

This piece walks through all five. By the end I hope you don’t remember five separate applications. I hope you hold one idea in your hand and recognise it anywhere a machine has to do exact, repeatable work in a window too short and too critical to trust to air.

The one idea underneath all five

Air springs. Push a column of compressed air against a rising load and it does not pass your intention along faithfully — it gives way, stores energy like a spring, then releases it unevenly as the load changes. For most movement in a factory this doesn’t matter at all. It is even useful, because that springiness is what makes pneumatics forgiving and safe. But each of these five operations has a short, defined region — usually the last 25 mm or less — where the real work happens, and in that region the behaviour has to be identical on every single cycle. There, springiness is not forgiving. It is the failure.

Oil does not spring. At the pressures these operations run, oil is effectively incompressible. A sealed oil column holds its position even when an outside force pushes against it. The technical literature says it plainly: because oil can be treated as incompressible at these relatively low pressures, the cylinder will not move even if outside forces push against it. For me, that one sentence is the whole engineering case. Everything below is a variation on it.

So the design move is the same across all five, and I like it precisely because it refuses to throw away what already works: keep air as the energy source — already plumbed, intermittent, no power unit heat — and let oil do the precision work. You do this in one of two ways. Either you intensify air pressure into a short, stiff, high-pressure power stroke (a pneumo-hydraulic intensifier cylinder), or you use a sealed oil column to meter the velocity of an air-driven tool (a hydraulic checking cylinder, also called a feed-rate controller or hydraulic speed regulator). In the first family, oil makes force repeatable. In the second, oil makes motion repeatable. In both, the energy stays cheap and intermittent.

Four conditions, together, tell you when this trade is the right one:

1. The duty is intermittent. The work happens for milliseconds to seconds per cycle, not continuously. There is nothing for a full hydraulic power unit to do most of the time except run, heat its own oil, and waste energy.
2. The power region is short. Usually the last 25 mm or less is what matters mechanically. You don’t need precision over the whole stroke — you need it in a small, defined zone.
3. The operation is sensitive to compressibility. Peak force, or feed rate, can’t be left to air, because the part’s quality is decided exactly where air is least trustworthy.
4. The air supply already exists. Adding an electric power unit at every station means spending money, again and again, to make power you already have on tap.

When all four hold, you are in pneumo-hydraulic territory. Now watch the same idea come up from five different directions.

Cycle 1 — Riveting: when end-of-stroke force has to be identical every time

What happens, mechanically. A solid, semi-tubular, or self-piercing rivet is upset inside a stack of sheets to form a permanent joint. One cycle runs: approach, as the ram travels free; contact, as it meets the rivet head; upset, as it applies forming force; a brief dwell, held at peak force to let the material flow settle; then retract. Force matters in the upset and the dwell. Impact velocity at contact matters too, because over-driven impact work-hardens softer rivet alloys. And consistency matters most of all, because rivet height, head flushness, and joint strength all depend on hitting the same peak force to the same depth, cycle after cycle.

The force and motion reality. This is not a vague range. Industrial press riveting is catalogued across roughly 4 to 50 tons (about 35 to 445 kN) in air-over-oil rivet presses, with a typical intensified power stroke around 0.5 in (12.7 mm). Self-pierce riveting, which overlaps here, runs 20 to 80 kN — corroborated by an 80 kN electric setting tool on the market — with a pierce-and-flare forming sub-phase of 100 to 300 ms and a total cycle including feed of roughly 1 to 4 seconds. In aircraft and structural solid riveting, gun classes run from 1× to 5× by stroke, with 3× the largest reliably matched to 3/16″ rivets and 4×–5× corresponding to 1/4″ rivets. As for repeatability, published manufacturing-tolerance studies on rivet head position in automotive structural SPR cite ±0.3 mm as a working figure.

Why pure pneumatic fails here. Air is compressible, and on a riveting stroke that shows up three ways. End-of-stroke force drifts with line pressure, ambient temperature, and how much the air in the cap end yields against the rivet before it deforms it — so the same machine gives a different peak today than it did this morning. Impact-driven setting carries the documented work-hardening failure: solid rivets hit too hard work-harden and will not properly set and clinch the material. And an air hammer is essentially a bang-bang on/off valve with little throttling, where a true riveting hammer has a valve that allows some control — a pure air ram simply does not govern the work part of the stroke. There is even an end-of-stroke bounce problem: pneumatic pistons can bounce several times against a hard stop unless cushioned, and on a rivet already at peak yield, that bounce blurs set height. Tellingly, one major pneumatic-cylinder maker sells stacked multi-piston cylinders specifically to reach rivet-upset forces of 4,800 lbf and beyond — a quiet admission that a single-piston pneumatic ram of practical bore cannot reliably hit upper-end rivet forces from 90 psi shop air.

Why full hydraulics is the wrong correction. You could solve all of this with a hydraulic power unit, and you would be over-engineering it. Riveting stations are intermittent — seconds of contact per minute at most. A full power unit running continuously to feed one rivet press wastes energy, drifts oil viscosity, demands a 25-plus-gallon reservoir against the 1 to 2 gallons in an air-over-oil booster, needs high-voltage power, and brings oil-leak risk into an assembly cell. An intensifier, by contrast, is self-limiting: intensifiers do not need relief valves because they simply stall at maximum pressure — something a continuously-pumped power unit will never do for you.

Why pneumo-hydraulic is the right middle. The air-over-oil cylinder runs a fast approach on air, then switches automatically to a high-pressure power stroke on intensified oil the moment the moving ram meets resistance. Catalogued units in this class run roughly 1 to 20 tons, with a total stroke of 2 to 8 in but a power stroke of only 0.25 to 1 in, approaching at 6 to 10 in/sec and driving the power stroke at 1 to 2 in/sec, all off 30 to 100 psi shop air. Force is set by adjusting input air pressure — readable, repeatable, and decoupled from valve timing. The sealed oil holds the stroke rigid through the dwell. No power unit, no continuously running motor, intermittent energy draw.

Where it goes wrong — and how you would know. A spongy stroke means trapped air in the oil; that compressibility makes accurate mid-stroke stopping and smooth speed control hard to achieve, and the fix is purging on full strokes with tanks mounted above the cylinder. Cavitation comes from retracting the power-stroke piston without retracting the ram. Punch-through over-travel is a specific trap: piercing applications must provide an external resistance to the rod and tooling before breaking through, or the rod runs out and pulls a vacuum into the high-pressure chamber. And at intensified pressures with a leaking rod seal, the dangerous mode is dieseling — when a mixture of air and oil is compressed in a hydraulic cylinder it can ignite and burn, or even explode. None of these are reasons to avoid the technology. They are simply the things I would tell you before you find them yourself.

Cycle 2 — Clinching and self-pierce riveting: when the transition is the whole game

What happens, mechanically. Clinching forms a cold mechanical interlock with no fastener at all (the method is covered by DIN 8593); self-pierce riveting drives a semi-tubular rivet that pierces the top sheet and flares into the bottom one. They are closely related cold-joining processes, and they share a defining feature: a stack-height-sensitive transition. The punch and die close, a blank-holder applies clamping force first, then the forming stroke does its work — deep-draw and squeeze for a clinch, pierce and flare for an SPR. Joint quality is judged by interlock, residual die-side material thickness, and head position. For me, the quality lives entirely in how cleanly that transition from clamping to forming happens — and that is exactly the moment air can’t be trusted.

The force and motion reality. Clinching forces run from single-digit kN on thin sheet up to 200 to 250 kN for stacks approaching 7 mm total thickness, and one manufacturer’s customer-specific C-frames quote a joining force of 35 to 300 kN outright. SPR runs 20 to 80 kN. The forming sub-phase lasts 100 to 300 ms, and the lightest, fastest tools reach up to 3 joints per second. As an industry voice put it: cycle time for a self-pierce riveting system is generally the same as spot welding steel, and it takes about 1.3 seconds for a complete rivet-and-feed cycle at a 30 mm tool opening — with the sourced range running roughly 0.33 to 1.3 s per joint depending on tool opening and system. Peer-reviewed work pins specifics further: in 5052 aluminium SPR, peak riveting force lands between 4.96 and 9.88 kN depending on die geometry; in 980 MPa steel-and-aluminium stacks, a 5 kN blank-holder runs against 250 kN of test-frame headroom.

Why pure pneumatic fails here. The force required — 15 to 300 kN depending on stack — is more than a practical single-bore pneumatic cylinder can deliver from shop air. But the force ceiling is only half of it. The process depends on that stack-height-sensitive transition, and air’s compressibility makes the punch decelerate unpredictably through the deep-draw or pierce phase, so the interlock and the neck thickness drift out of tolerance. Some processes are stricter still: one setting method patents a defined time at a constant clamping force, in the range of 0.1 to 2 seconds at 2 to 10 kN — the kind of rigid hold a sealed oil column delivers and air simply cannot.

Why full hydraulics is impractical here. A body-in-white line carries dozens to hundreds of these joints per body, often on robot-mounted guns. Equipping each gun with its own power unit is not realistic, and a robot cannot carry a power unit at all. The field splits between high-precision, high-capex servo-electric at the top end and lower-capex, clean, self-contained pneumo-hydraulic below it — with air-over-oil typically covering up to around 30 tons before full hydraulic takes over.

Why pneumo-hydraulic is the right middle. The positioning is exact: this class of cylinder is ideal for piercing, riveting, notching, clamping, marking, coining, and assembly where short-stroke, high-speed, high-force operation is wanted. The short power stroke — a fraction of an inch up to a few inches — maps directly onto the few millimetres of plastic flow the joint needs. And where traceability matters, load-cell-plus-position monitoring gives closed-loop force-and-distance feedback on a pneumo-hydraulic gun, without the cost of a full servo system. This is proven ground, not a novelty — the air-over-oil intensifier was the historical default for low-volume clinch-riveting long before anything else arrived. As one shop engineer recalled, they ran modular punches off an air-over-oil intensifier unit, pneumatic-pedal operated, riveting and punching aluminium extrusion. That is the field memory of a technology that simply works.

Where it goes wrong. Cavitation appears when no external resistance is provided after breakthrough — the rod keeps travelling and pulls a vacuum in the high-pressure reservoir. Under-specified force margin produces shallow interlocks and weak pull-out strength; the standing advice is to multiply your application force requirement by 1.25 to 1.50 to make sure adequate force is available. Too-thin die-side material that cracks the button usually traces to force overshoot from too little deceleration in the final millimetre — which is, once again, the compressibility problem wearing a different mask.

Cycle 3 — Bushing and bearing insertion: the one-shot you cannot take back

What happens, mechanically. A bearing, bushing, or pin is press-fitted into a housing or onto a shaft using controlled axial force, using elastic interference. The cycle is align, approach, a slow sustained press through the interference depth, then a seat-and-dwell to a force or position stop, then retract. Force matters from contact all the way through insertion. And here consistency matters more than almost anywhere, because this is a one-shot operation: you cannot back the bearing out without damaging it. There is no second attempt. I take that seriously — when there is no second attempt, the margin for “good enough” disappears.

The force and motion reality. Low-force work — soft housings, polymer bushings — runs 1,200 to 5,000 lbf (about 5 to 22 kN). Typical bearing insertion runs 2.5 to 15 tons (about 22 to 134 kN); a press in that range handles most bearing insertions. The seating force itself follows F = p × A × µ, where p is interference pressure, A the contact area, and µ the friction coefficient — so the same nominal fit can demand very different force as area and friction vary. The field rule on velocity is blunt, and I agree with it: apply press force slowly and steadily, and if the bearing does not move within the first 2 to 3 mm of travel, stop and check alignment.

Why pure pneumatic fails here. The peak force is inconsistent as the press meets the seated bottom — a bearing needing 8 kN might see 6 kN one cycle and 11 kN the next, enough to crack a die-cast aluminium housing or disfigure the mating part. A pneumatic ram offers no mid-stroke holding rigidity, so it drifts back under load if paused. It gives you no force-distance signature without bolt-on instrumentation, so a bearing going in cocked, riding a burr, or hung up on a chip is caught at the next station, not in the cycle that made it. And the cocking risk is the dominant failure: alignment of the bearing to the bore is critical to prevent a cocking motion during insertion, which can damage or ruin both bearing and housing — and a pure air ram’s high initial velocity makes starting crooked more likely, not less.

Why full hydraulics is excessive here. A pneumo-hydraulic press at 2.5 to 15 tons covers most bearing work; beyond roughly 30 tons full hydraulic begins to dominate, but most stations never get there. A power unit at every press-fit station means oil mist, heat, energy draw, and a 25-gallon reservoir, for an operation that runs for a second or two at a time.

Why pneumo-hydraulic is the right middle. The two-stage motion maps directly onto press-fit mechanics — fast approach, then an automatic power stroke at ram contact. The oil column’s rigidity holds position through the dwell, which lets a pressure switch or transducer confirm “press complete” rather than guessing. Force-distance monitoring gives an in-process press signature, so a cocked or hung-up bearing produces a signature anomaly you can alarm in the cycle rather than discover downstream. The principle is established enough to appear in patents: one engine-block bearing-cap press specifies a first air cylinder and a second hydraulic cylinder — an explicit two-actuator design that mirrors air-over-oil function exactly.

Where it goes wrong. Cocking is the headline failure, and the first 2 to 3 mm is your diagnostic window. True brinelling — forcing load through the rolling elements instead of the loaded ring — dents the raceway permanently. Burr push-through and over-pressing into soft housings are both, at root, the pure-pneumatic over-pressure problem; regulated peak force from air-over-oil is the corrective. And a spongy stroke after maintenance means trapped air — purge on full strokes, tanks above the cylinder.

Cycle 4 — Die clamping: where the oil moves to the power source

What happens, mechanically. Dies are secured to press platens in stamping, forming, and die-casting, or to mould platens in injection moulding. There are two distinct contexts here, and they call for opposite answers — which is why I always ask which one we are actually talking about before recommending anything. In production-cycle clamping, the die stays clamped through every shot or stroke for hours or days, set once at die change. In quick-die-change clamping — the SMED context — the act of clamping and unclamping is itself a production cycle, done many times a day, automatically. That cycle runs approach, set hydraulic pressure to the bolt or stud, lock for the duration of stamping, unlock, withdraw. Force is the whole game, and consistency matters because an under-clamped die walks under impact and eventually slips.

The force and motion reality. Hydraulic T-slot clamps run 15, 30, 45, and 70 kN at 350 bar (5,000 psi), with an 8 to 12 mm clamping stroke, Viton seals, and a 200 °C rating. Hydraulic nut and ledge clamps span single-clamp forces of 8,000 to 10,000 lbf up to double-clamp forces of 22,000 to 26,000 lbf. T-slot bolts are grade 8.8, 10.9, or 12.9; cast-steel clamps are heat-treated to 85,000 to 135,000 psi tensile. The dies themselves range from hundreds of kilograms to tens of tonnes for stamping, and from 50 kg to several tonnes for injection moulds. The numbers matter because they explain straight away why air alone is hopeless here.

Why pure pneumatic fails here. To reach just 70 kN from 90 psi shop air you would need roughly 175 in² of piston — about a 15 in bore. That is impractical on its face. Worse, a stamping die transmits cyclical impact straight up through its clamps, and an air cushion lets the joint micro-walk on every stroke, loosening it until the die slips. And pneumatic pressure bleeds, so it cannot hold reliably without continuous power — which is exactly why hydro-mechanical clamps with a locking screw release hydraulic pressure entirely during stamping and hold mechanically.

Where full hydraulics is genuinely correct. This is the one cycle of the five where, for large production presses, a 350-bar electric power unit is the right and standard answer — and has been the workhorse for more than fifty years. I think being honest about that boundary is part of what makes the rest of this worth reading.

Where pneumo-hydraulic wins instead. On small and medium presses, on retrofits, and wherever clamping is intermittent — one or two clamp events a shift — running a 350-bar power unit is overkill. The dominant pattern is an air-driven hydraulic intensifier: shop air feeds a built-in pressure regulator where the operator sets the air pressure, which in turn sets the maximum hydraulic oil pressure, delivering up to roughly 4,785 psi from about 24 SCFM of shop air at under 75 dBA. The structural insight I’d want you to take away is this: in die clamping the pneumo-hydraulic component sits on the power source, not on the clamp. The clamp is hydraulic; the intensifier feeds it during clamp and unclamp, then sits idle through production. The oil has moved upstream — but it is the same idea doing the same job.

Where it goes wrong. Die walking and slip come from inadequate total clamping force against the dynamic load — the clamp model and quantity must be chosen to exceed the total static-plus-dynamic force. Bolt fatigue comes from incorrect preload, and the torque figures are specific (on the order of 125 ft-lb for a 10,000 lbf clamp, 350 ft-lb for 20,000 lbf). An under-specified intensifier silently produces insufficient clamp pressure — the most dangerous failure, because nothing looks wrong until the die moves. Sight glasses on retrofit intensifier reservoirs are not optional.

Cycle 5 — Peck-feed drilling: where the logic inverts

This cycle earns its place by turning the previous four inside out, and for me the inversion is the proof that we are dealing with a real principle and not just a story about presses. In the first four cycles, oil intensifies force. Here, the hero component is not a force intensifier at all — it is a hydraulic feed-rate controller, a hydraulic checking cylinder. The drive is pneumatic. The control of motion is hydraulic. Force matters only as a downstream consequence of feed rate.

What happens, mechanically. A drill bit feeds into the workpiece in a series of partial advances — pecks — separated by partial retractions that lift the bit clear of the cutting zone to clear chips and let coolant in. One peck runs: rapid advance (fast and cheap), a transition to controlled cut feed, the cut feed itself (controlled penetration to the next peck depth), then a retract — full, for chip clearance, or fractional, to break the chip — with an optional dwell. On the final peck comes breakthrough, and at that instant feed rate must be tightly controlled or the drill grabs and breaks. This is a motion-control problem first. The force is derived: feed rate, drill geometry, and material set the thrust, and if any of them drift, the drill breaks or the hole is wrong.

The feed-rate reality. Peck depth is typically 1 to 3 times the drill diameter per peck. Hydraulic checking cylinders regulate velocity smoothly from roughly 0.01 in/sec up to about 1.4 in/sec depending on model and load. Load capacity in this class tops out around 1,200 lb including impact for a standard regulator, down to 400 lb for the smallest, while comparable air-return regulators span roughly 200 N to 9,800 N (about 45 to 2,200 lbf). The rule of thumb is clean: drilling deeper than three times the drill diameter often requires peck drilling. One control mechanism is described precisely — a one-way ball clutch prevents automatic extension of the plunger rod, with a deliberate slight extension of about 0.005 in to keep the drill bit from impacting the work, an especially important feature with smaller tools. And the internal fluid is silicone, filtered every stroke, sealed by a rolling-diaphragm seal, with published cycle life of 10⁷ cycles without leaking — a number that does an enormous amount of believability work in a single figure.

Why pure pneumatic fails here. Air’s compressibility produces stick-slip. The canonical statement is worth keeping intact: in repetitive drilling applications compressed air often is not the best choice for feed drives, because the slow feed rates usually required can cause stick-slip operation and irregular feed speeds. An air piston runs fast-then-slow as resistance changes, when drilling needs slow-and-constant against rising thrust. And then the worst failure, the one every drilling engineer has felt in their hands: at the point the helix breaks through the thin edge at the bottom of the hole, the drill tends to self-feed — and as one engineer described it, you develop a sense of touch that anticipates imminent breakthrough so you can quickly reduce feed pressure. A pneumatic-only feeder has no sense of touch. It accelerates at exactly the wrong instant. The crude pneumatic-only workaround — PLC ladder logic with limit switches and timer-driven oscillators — is documented, and a proper pneumatic-hydraulic feed unit is priced at around $3,000 to $3,500, which tells you the market has already decided what the problem is worth solving properly.

Why full hydraulics is impractical here. A drilling station does not need 200-bar oil — it needs metered motion. Full hydraulic feed (proportional valve, servo, power unit) is a six-figure solution to a problem a sealed checking cylinder solves for hundreds to low thousands of dollars, with no electricity at the actuator. The self-feeder mounts on a fixture or a robot, where a power unit per drill head is infeasible; the checking cylinder is self-contained, silicone sealed for life, no reservoir, no plumbing.

Why hydraulic feed control paired with pneumatic drive is the right middle. The pneumatic drive provides the muscle — compact, intrinsically safe, continuous-duty. The hydraulic checking cylinder turns that thrust into smooth, regulated velocity by metering silicone fluid through an adjustable orifice as the carriage pushes the plunger. It does not drive; it resists at a controlled rate. The smoothness comes from having no sliding seals and no heavy springs to make the plunger bind or jerk. A built-in ball clutch lets the retract run fast for chip clearance and engages metering only on the cut stroke — exactly the asymmetric profile peck drilling needs. Variants add a fast-feed interval for through-tube drilling (clear the first wall fast, slow for the second) and a dual-speed mode that runs fast through bulk material then controls the final quarter-inch to eliminate composite delamination, which makes it viable on CFRP and titanium aerospace stacks.

Where it goes wrong. Drill grab on breakthrough is the canonical failure. Chip jam at depth means the peck depth is too aggressive for coolant to reach the edge. Composite delamination on breakthrough means feed too high in the last millimetre or two — mitigated by the dual-speed or fast-feed variants. Stick-slip at very low feed is the pneumatic-plus-needle-valve signature. And misapplication has a specific cost: a regulator without the ball-clutch function meters the retract stroke too, producing slow, wasteful cycles and heating the oil. The right tool for a peck cycle is not optional — it is the difference between a clean cycle and a hot, slow one.

A sourcing note worth flagging. The long-standing Kinechek regulators that defined this category reached end of life when their maker ceased operations at the end of 2025, with another manufacturer’s Hydro-Speed regulators positioned as the named successor. One distinction matters more than any spec sheet: a peck-feed cycle needs the air-return type that holds rod position at each peck, not the spring-return type, which loses the at-depth hold. No published numeric cross-reference table exists yet — replacements are matched case by case through engineering — so the practical guidance is a checklist, not a part number: plunger thrust at least 1.25× peak drill thrust; stroke at least the full drill travel; air-return hold for peck work; a rolling-diaphragm seal rather than a sliding one; sealed silicone fluid with no external reservoir; a cycle-life rating in the tens of millions; and a mounting envelope you have confirmed will adapt, because these are similar but not drop-in. That kind of candour — telling you exactly where the easy answer runs out — is, for me, the most useful thing a supplier can offer in a market mid-transition.

The part most vendor content won’t tell you: where each cycle stops being the answer

Here is the real test of whether anyone understands this technology — not where it wins, but where it loses. Pneumo-hydraulic has a comfort zone in every one of these cycles, and knowing exactly where you cross out of it is not an admission against interest. For me it is the most trust-building thing an engineer can show, because it proves the recommendation is driven by physics and not by what the supplier happens to sell. I have lost work before by saying a product wasn’t the right fit — and I would do it again, because I won’t put my name behind a component I’m not sure of.

Riveting stays in the pneumo-hydraulic zone up to roughly 50 tons of force, on power strokes of an inch or less, at moderate cycle rates. You cross into full hydraulic or servo-electric past about 50 tons, on continuous-cycle multi-rivet aerospace lines, or wherever a servo force-distance closed loop is mandated.

Clinching and SPR suit air-over-oil up to around 30 tons sustained, on short power strokes, robot-mountable. The boundary is crossed beyond about 30 tons sustained, or where high-end body-in-white work demands servo-electric traceability.

Bushing and bearing insertion fits the pneumo-hydraulic range from 1 to 30 tons, on strokes of 25 to 150 mm. Beyond a single 30-ton press, or where medical and aerospace traceability requires full servo force-distance recording, full hydraulic takes over.

Die clamping is the special case, because the clamp itself is hydraulic and the pneumo-hydraulic role lives on the power source — an air-driven intensifier feeding the clamp circuit during clamp and unclamp, idle through production. That is right for small and medium presses with few intermittent clamp events. Large continuous high-tonnage stamping with many clamps is where a dedicated 350-bar electric power unit is standard.

Peck-feed drilling is bounded not by raw force but by plunger thrust capacity — roughly 1,200 lbf in the standard checking-cylinder class, up to about 2,200 lbf for the larger air-return units. Beyond that thrust, or on CNC machining centres with servo feed, or in gun drilling, you have left the category.

Notice what this list refuses to do: it refuses to give you one clean number. The existing rule of thumb — a single-actuator boundary in the low-hundreds-of-kN range into full hydraulics — holds honestly for the three force cycles (riveting, clinching, press-fit). But for die clamping the real boundary is press tonnage and clamp count, and for peck drilling it is plunger thrust, not force at all. Pretending a single figure covers all five is exactly the kind of tidy oversimplification that signals a vendor talking, not an engineer thinking. The honest answer is cycle-specific, because the five cycles are solving the same physics problem from genuinely different directions.

One last caveat, because for me leaving it out would be dishonest: continuous high-speed production can be done with pneumo-hydraulic — units in this class quote multi-million-cycle service intervals and lifetime warranties. The practical limit is thermal, not mechanical. When cycle-rate multiplied by power-stroke energy exceeds what one or two gallons of sealed oil can dissipate without viscosity drift, you have found the real edge — and that edge is set by heat, not by the headline force number.

The thread, pulled tight

Five operations. One idea. Whether the job is upsetting a rivet, forming a clinch, seating a bearing, holding a die, or feeding a drill, the same tension runs underneath every one of them: air is cheap, fast, clean, and already in the building — but air springs, and each of these operations has a short, decisive window where springing is simply not allowed. Oil does not spring. So the answer, every time, is to keep air for the energy and hand the precision to a sealed oil column — either to intensify force into a stiff, short, repeatable power stroke, or to meter the velocity of a moving tool so it never grabs at the worst possible moment.

Hold that idea, and the five applications collapse into one. You stop reaching for a catalogue and start reasoning from the physics. You can recognise the pattern anywhere a machine has to do exact, repeatable work in a moment too brief and too critical to trust to compressible air — and, just as quickly, you can tell when that moment has grown large enough, continuous enough, or traceable enough that air should step aside and let full hydraulics take the load.

That, for me, is the whole point. Knowing where the answer is right is useful. Knowing where it ends is what I would actually trust an engineer for. If you are weighing one of these cycles and you are not sure which side of the line you sit on, that is exactly the conversation we like to have — bring us the application, and we will tell you honestly where it lands.