Spindle Utilisation: What 25%, 45% and 70% Mean on an Indian Floor

Spindle utilisation is the share of available time your spindle is actually cutting metal — not switched on, not loaded, cutting. Most small-to-mid Indian machine shops sit far lower than their owners think: industry estimates put the typical band at roughly 25–50%, while well-run shops push past 70%. The number matters because moving from 25% to 45% on one machine can be worth over ₹11 lakh a year without buying a single new machine — it is capacity you already own and aren't selling.

By the Mama Editorial Team. "Spindle utilisation" (also called spindle-cutting time or in-cycle time) is the honest cousin of machine utilisation: it ignores the hours the machine merely stood powered and counts only the hours it earned. Here is what the common benchmarks mean on an Indian floor, and what each band is worth in rupees.

Key takeaways

• Spindle utilisation = cutting time ÷ available time. It is stricter than "machine on" — idle-but-powered doesn't count.
• Owners guess ~70–80%; measured studies find far less. MachineMetrics and others place many discrete shops in the 25–30% range; typical job shops run ~25–50% (industry estimates — see below).
• The three bands: ~25% = large hidden capacity, usually setup/wait/micro-stops; ~45% = a competently run mixed shop; ~70%+ = a well-managed, scheduled operation nearing practical ceiling.
• The lever is the gap, not a new machine. In the worked example, 25%→45% is worth ~₹11.5 lakh/year/machine in extra contribution.
• The spindle stops for a short list of reasons — setups, material/crane waits, tool changes, inspection, micro-stops — and most are attackable without capex.

What spindle utilisation measures (and what it doesn't)

Available time is the time the machine was staffed and scheduled to work — one shift, two shifts, however you run. Spindle utilisation is the fraction of that time the spindle was in cut. It deliberately excludes:

• Setup and first-off approval
• Load/unload between parts
• Waiting for material, tooling, programs, cranes or sign-offs
• Tool changes and gauging
• Micro-stops and idling-but-powered time

That strictness is the point. A machine can be "utilised" 90% of the shift — switched on and attended — and have a spindle that cut only 35% of it. Utilisation flatters; spindle-cutting time tells the truth. It maps directly onto the Availability and Performance factors of OEE: the same lost hours, expressed as the share the spindle was earning.

The benchmarks — treat as directional, not gospel

There is no reliable India-specific published survey of spindle utilisation, so treat all numbers as industry estimates, not a benchmark to quote at an auditor:

• MachineMetrics, from instrumenting large fleets, has repeatedly reported average discrete-manufacturing machine utilisation in the high-20s to low-30s percent, against owners' self-estimates of 70%+ (machinemetrics.com).
• JITbase and Modern Machine Shop benchmarking put typical job-shop spindle-cutting time well under half of attended time, with top-quartile shops materially higher.
• The gap between gut-feel and measured reality is the consistent finding across sources — the direction is far more trustworthy than any single percentage.

For an Indian floor, read the bands like this:

Note the ceiling: a job shop with small batches and constant changeovers cannot reach the 85–90% of a dedicated high-volume line, and shouldn't aim there. The realistic target is to climb one band.

Why the spindle isn't cutting

Separate the causes you can attack from the ones that are legitimate. The spindle is idle because of:

• Setup and changeover — the single biggest bucket in small-batch work. Attackable: SMED, pre-staged tooling, standard work.
• Waiting — material not at the machine, crane busy, program not ready, quality sign-off pending. Attackable: scheduling and material discipline.
• Tool changes and gauging — partly legitimate, partly avoidable with tool-life management.
• Micro-stops and minor stoppages — the least-logged, often the largest hidden loss.
• No work / no operator — a scheduling and manning problem, not a machine problem.

Only a slice of this is unavoidable. The rest is capacity you paid for and left on the floor — the same hours that quietly inflate your machine hour rate and keep the idle-but-energized machine drawing its standing power cost.

What each band is worth (worked ₹ example)

One CNC, two shifts. Figures are indicative October 2026 estimates — plug in your own.

Available time, double shift ≈ 16 h × 26 days × 12 months ≈ 4,800 hours/year. Say each true cutting hour earns ₹1,200 of contribution (sale value of the machined time minus variable cost — use your own rate).

• 25% → 45% adds ~960 cutting hours = ~₹11.5 lakh/year on one machine.
• 45% → 70% adds ~1,200 cutting hours = ~₹14.4 lakh/year.

No new machine, no new building, no extra sanctioned load. Just selling the hours you already own. Across a ten-machine shop, one band of improvement is a crore-scale number — which is exactly why it belongs in front of the owner, not buried in a maintenance log. For the stop-by-stop ₹ logic behind it, see the real cost of downtime and OEE benchmarks for Indian factories.

The levers that move the band

1. Attack changeover first. In small-batch shops it is the biggest bucket. Pre-stage tooling and fixtures, standardise the setup, measure changeover minutes.
2. Fix material flow to the machine. Half of "waiting" is a blank or a crane that should have been there. Kit and sequence ahead of the spindle.
3. Shorten the quality loop. First-off sign-offs that take 40 minutes stall the spindle; move inspection closer.
4. Count micro-stops. You cannot cut what you cannot see — the hidden bucket is only addressable once it's measured.
5. Measure the bottleneck first. A high spindle utilisation on a non-constraint machine just builds WIP; raise cutting time on the machine that gates the shop's output.

You can't improve what you don't measure

Every lever above depends on one thing: an honest, continuous record of when the spindle was cutting and why it stopped. On an old fleet with no PLC that is the hard part — there's no signal to read. The three practical ways to get one (operator logs, spindle current sensors, cameras) are compared in rupees in Tracking Old CNCs Without a PLC. A camera already pointed at the aisle has an advantage the others don't: it sees not just that the spindle stopped but why — the missing blank, the busy crane, the operator at the tool crib — which is exactly the reason code you need to attack the right lever.

That is what Mama does. Record a short phone walk of your shop; Mama reads the feeds — your existing cameras over RTSP/ONVIF with no replacement, or cameras we add — and every morning the owner gets a plain-language WhatsApp note: which machines sat idle yesterday, and what the lost cutting hours were worth in rupees. It watches machines, not people. Send a floor video and you get back where the hours go, a camera plan for the blind spots, and a proposal — the first step to moving your floor up a band.

FAQ

What is a good spindle utilisation rate? It depends on batch size. High-volume dedicated lines can exceed 80%, but a small-batch job or auto-component shop with frequent changeovers realistically targets the next band up — from ~25% toward ~45%, or ~45% toward ~70%. Treat these as industry estimates, not fixed benchmarks.

What's the difference between machine utilisation and spindle utilisation? Machine utilisation asks whether the machine was switched on and attended; spindle utilisation asks whether it was actually cutting. A machine can be 90% utilised and still have a spindle cutting only 35% of the time — the gap is setup, waiting and micro-stops.

Why is my spindle utilisation so low? Usually setups and waiting, not the machine itself. In small-batch work, changeovers plus material/crane/sign-off waits and un-logged micro-stops routinely consume more than half of attended time.

How much is low spindle utilisation costing me? Multiply the extra cutting hours a higher band would give by your contribution per cutting hour. In the indicative example, moving one machine from 25% to 45% is worth about ₹11.5 lakh a year — capacity you already own but aren't selling.

How do I measure spindle utilisation on an old machine with no controller? Measure a proxy for cutting: operator logs (quick but coarse), a current sensor on the spindle drive (objective, needs a panel cut-in), or a camera watching the spindle and chip flow (no contact with the machine, and it also shows why it stopped).