An axis gets drawn with a ball screw on it because the machine before it had one, and nobody reopens the question. Sometimes that is right. Just as often the same axis could have been built with a trapezoidal screw, a smaller motor and no brake, and the difference would have paid for something that changed the result.
This decision sits upstream of every other screw question. Sliding contact or rolling contact: settle that first and the questions about lead, diameter, grade, preload and end machining have a frame to sit in. Settle it wrong and no care downstream gets the axis back.
Every number below is a typical value published in the trade, given for orientation. None comes from one customer's project, and the number that governs your order is on the datasheet for the part you buy.

1. The short answer
A lead screw carries its load on sliding flanks: the nut is threaded to match, and the two surfaces slide past each other as it travels. A ball screw puts hardened balls into matched grooves in screw and nut and lets them roll. Everything below follows from that one difference.
- Efficiency. Typical published figures put sliding screws in roughly the 30–50% band and ball screws around 90%. Behind the gap is friction: sliding pairs commonly run at 0.1–0.2, a ball screw at a rolling coefficient one to two orders of magnitude lower.
- Wear. A sliding nut wears with every metre travelled, showing up as rising backlash. A ball screw wears slowly, then reaches a fatigue limit you can calculate before buying.
- Holding. A fine single-start trapezoidal screw commonly will not backdrive. A ball screw almost always will. On a vertical axis, that one line decides whether a brake goes on the motor.
Short travel, low speed, frequent stops, position held unpowered: the trapezoidal screw is usually correct, and the ball screw is money spent for nothing. Continuous duty, higher speed, or accuracy over long travel: the reverse is true.
2. What is different inside the nut
Both are a screw and a nut. What changes is the load path.
Trapezoidal and acme. The nut is bronze, cast iron or polymer, threaded to match. Load crosses the whole flank contact area, and the surfaces slide at the velocity set by the lead angle. Wear is not a failure mode here but the mechanism, and the materials put it on the cheap replaceable part.
Ball screws. Screw and nut each carry a groove; hardened balls run in the two and return through a recirculation path in the nut. Contact is a small rolling patch at high local stress, but the surfaces do not slide at feed rate. Rolling needs clean lubricant — a ball screw usually dies of contamination or starvation, not of wear by design — and its groove needs protecting from side load and misalignment.
Naming, declared once. "Trapezoidal" means the 30° metric form covered by ISO 2904 and DIN 103; "acme" the 29° imperial form. Ball screw terms follow ISO 3408. The same words carry the same meaning from here on.
3. Efficiency, and where the rest of the torque goes
Of the torque put into a screw, what does work becomes thrust; the rest becomes heat.
With a lead angle λ and a friction angle φ, the standard machine-design relation for forward efficiency is η = tan λ / tan(λ + φ), where tan φ = μ / cos α for a thread of half flank angle α. A ball screw needs no correction: its effective friction is a rolling coefficient, commonly published around 0.003–0.01, which is why its efficiency sits near 90% across a wide lead range, and why it reverses nearly as efficiently.
For a sliding screw the input that matters is μ, and that is where published numbers spread — greased bronze, steel on polymer, dry running and temperature all move it. So the relation gives orientation, not prediction:
- At 30–50% efficiency, half to roughly 70% of the input torque becomes heat in the nut for as long as the axis moves — which is why sliding screws are specified for intermittent duty.
- The motor is sized on torque at the screw, not thrust at the table, so a sliding axis commonly needs the next frame size up for the same thrust and speed, and pays for it twice.
Illustrative example. Tr20×4 lifting 2 kN on greased bronze, μ taken as 0.15: λ ≈ 4.0°, φ ≈ 8.8°, so η ≈ 31%. Input torque is about 4.1 N·m, of which roughly 1.3 N·m lifts and 2.8 N·m becomes heat; a ball screw at 90% needs about 1.4 N·m for the same load — about three to one. Change μ and the ratio moves too: about 2.2 to one at μ = 0.10, about 3.6 to one at μ = 0.20.
4. Accuracy, and how the two wear
Ball screws are bought to a lead accuracy grade — a permitted travel variation over a reference length, on the drawing. Trapezoidal screws usually are not: many catalogues quote no grade, and what governs the axis is the clearance in the nut, which the user takes up again.
- Sliding. Backlash starts at whatever the nut was machined to and grows with distance travelled. Anti-backlash nuts take the clearance out at the cost of higher friction torque and faster wear. Accuracy here is not fixed; it drifts towards more lost motion.
- Rolling. Backlash is set by preload rather than by wear, and stays where it was set until the balls and raceways pit. What ends a ball screw is fatigue, calculated from a dynamic load rating and an equivalent load before anything is bought — the method is set out in how a ball screw's load rating and life are worked out. Preload and clearance sit on a different line from the grade, as the preload and backlash discussion explains.
Honest counterweight: on a short axis with a coarse requirement, a trapezoidal screw's own error may sit below the rest of the machine's error budget — then the screw is not the constraint, and accuracy bought in it buys nothing, which is the point the accuracy grade article makes inside the ball screw family.
5. Self-locking and backdrive: the vertical axis question
A screw holds a load without a brake when the friction angle is at least as large as the lead angle — the same φ and λ, with the condition φ ≥ λ, which is the self-locking condition used throughout.
- Fine single-start trapezoidal screws commonly satisfy it. Small leads give a lead angle of a few degrees against a friction angle of roughly 6° to 12°; coarse and multi-start leads can fall outside it, so "trapezoidal" is not by itself an answer.
- Ball screws essentially never satisfy it. That is not a defect; it is what 90% efficiency means. A ball screw axis backdrives, which is also why it can be pushed by hand.
Two consequences that belong on the design rather than on the purchase order:
- A vertical ball screw axis needs a holding device — a motor brake, a counterbalance, or a self-locking gearbox. Specify it at design time; adding one later is a different job.
- Self-locking is a characteristic of the friction pair, not a safety device. Lubrication, wear, temperature and vibration all move it, and where a load can fall, safety rules in most markets expect a separate holding device. That applies to a trapezoidal vertical axis as well as to a ball screw one: the holding device is specified against the applicable machinery safety rules and your own risk assessment, not against the friction in the nut. The friction that holds the load is the friction counted in section 3.
6. Speed, duty and noise
Ball screws. Speed is limited by the nut's DN rating and, on long shafts, by the shaft's own bending resonance; both limits, and how to tell which applies, are set out in why ball screws vibrate at high speed.
Sliding screws. Speed is limited by the nut material: bronze and polymer nuts carry a pressure–velocity limit, and past it wear climbs steeply. Published PV figures differ by material and maker, so take the number from the nut supplier.
Noise is diagnostic: a ball screw at speed produces ball recirculation noise, and a long fast shaft whips if the speed, diameter and length combination is wrong. A sliding screw is quiet at low speed and squeals or stick-slips when the lubricant is wrong.
7. Maintenance and service life
Ball screws. Life is a calculation — dynamic load rating, equivalent load, and a rated life in revolutions checkable against the duty cycle before the order. What ends them early is contamination, lubrication starvation and misalignment: seals, wipers and a lubrication interval do more than any datasheet figure.
Trapezoidal screws. Life is a wear budget, and the nut is the consumable — bronze needs lubrication, polymer runs dry and tolerates dust better, and the screw rarely fails. Wear arrives as backlash before it becomes a fault: you get warning.
8. Cost: a multiple, and then the whole axis
For the same nominal size and length, trade commentary commonly puts a ball screw assembly at roughly two to five times a trapezoidal one. The multiple widens with grade, length, end machining and small quantities, and narrows with volume. That is an order of magnitude, not a price list; your quotation governs.
The screw is not the axis, and the honest comparison is at axis level: the sliding screw usually brings a larger motor and drive; the vertical ball screw brings a brake; both need guides, a coupling, support units, end machining and assembly time. Counted together the gap often narrows, and sometimes neither screw bought as parts is the answer: a linear module integrates screw, guide and support in one envelope, and on a short axis can cost less than the parts plus the machining to fit them.
9. The decision list
Choose a trapezoidal screw when: travel is short, speed low and duty intermittent; the axis must hold position unpowered; the environment is dusty, dry or washed down and a nut running dry is an advantage; the axis is adjusted by hand; budget binds and the duty cycle is light.
Choose a ball screw when: duty is continuous or the travel per cycle is long; speed brings a sliding nut's PV limit into reach; efficiency matters — motor size, heat, energy over the machine's life; life must be predictable in writing before the purchase; accuracy and repeatability under preload are requirements.
Two ways it goes wrong. Choosing the trapezoidal screw for a high duty cycle because the load looked light — wear tracks distance travelled, not how heavy the load felt. And choosing the ball screw for a 150 mm manual slide.
10. When neither is the answer
- Long travel at high speed. Any screw gets long, heavy and resonance-limited; past a certain stroke-and-speed combination, a belt-driven axis or a rack and pinion is the answer.
- Long travel, guided load, moderate speed. A belt-driven linear module.
- Short envelope, high thrust, dirty environment. An integrated linear module puts screw and guide in one body.
- Two positions and nothing between. A simpler actuator does it for less money than either screw.
Screws turn rotation into thrust efficiently and precisely, but they are not the only way to move a load.
11. Frequently asked questions
Can I replace a lead screw with a ball screw without changing anything else? Rarely. Mounting dimensions differ, the nut is not interchangeable, the motor may be larger than the torque requires, and a vertical axis that relied on friction needs a brake. It is a redesign.
Is a lead screw ever more accurate than a ball screw? No, not on lead accuracy: a ball screw is bought to a grade, a trapezoidal screw usually is not. Over a short travel a trapezoidal axis with a good anti-backlash nut can still repeat well.
Are trapezoidal screws really self-locking? Commonly yes for fine single-start leads, no for coarse or multi-start ones. The condition is the friction angle against the lead angle, and friction moves with lubrication, wear and temperature. Treat it as a property of the pair, not as a holding device.
Which one lasts longer? It depends on duty rather than type. A ball screw has a calculated fatigue life and tends to be ended by contamination or overload. A trapezoidal screw wears with distance travelled: on light intermittent duty it can outlast the machine, on continuous duty it will not.
Do ball screws need lubrication? Yes. Seals, wipers and a lubrication interval do more for service life than any datasheet figure, and most early failures trace to contamination or starvation.
12. Data sources
- ISO 2904 and DIN 103 — metric trapezoidal screw threads, the 30° form referred to here as "trapezoidal".
- ISO 3408 — ball screws; grade and acceptance terminology follows ISO 3408-3, as elsewhere on this site.
- Screw efficiency relation, and the efficiency bands and friction coefficients in section 3 — standard treatment in general machine design references plus typical published figures; the friction coefficient is not standardised, hence bands. Published bands differ by source — some quote 20–40% for sliding screws, others 30–70%; the band used here sits inside that spread, and no conclusion above depends on where its edges fall.
- Cost multiple in section 8 — an order-of-magnitude range widely quoted in trade commentary; not Pinsi prices and not a price list.
- PV limit in section 6 — a property of the nut material; no figure is quoted.
Every figure above is a typical published value, not the specification of any part; the number that governs is the one on your datasheet. This article is general technical guidance, and is not a substitute for a calculation done on your own axis.
Last updated: 2026-10-22
13. Choosing between the two for a specific axis?
Send the stroke, the speed, the thrust, the orientation, and whether the axis must hold position unpowered. Those five answers settle the type before anyone talks about price.

+8615622924499
+8615622924499