Most guarding on Indian shop floors was added after the machine was installed, usually after an incident or an audit finding. That sequence is what produces guards that operators defeat: doors that block the loading position, fences that force a long walk to clear a jam, interlocks bypassed with a spare key taped to the panel. A guard that gets defeated protects nobody and still costs you the money.
Guarding designed alongside the machine behaves differently. It is built around how the operator actually works, so there is no reason to defeat it. This guide covers what the law and the standards require, how the numbers behind safety distances are derived, how to choose between guard types, and what to budget.
What applies in India
Two layers matter. The statutory layer is the Factories Act, 1948 and the Factories Rules of the state you operate in. Section 21 requires secure fencing of dangerous parts of machinery. Section 22 restricts work on or near machinery in motion. Section 23 restricts who may work on machines notified as dangerous. Section 24 requires striking gear and effective means of cutting off power. Your state's rules add specifics, and your factory inspector enforces them.
The engineering layer is the ISO machine-safety family, adopted in India as IS/ISO standards and universally expected by multinational customers and their audits:
- ISO 12100 — risk assessment and risk reduction. The parent standard. Everything else follows from it.
- ISO 13857 — safety distances to prevent upper and lower limbs reaching hazard zones. This is where fence heights and mesh apertures come from.
- ISO 13855 — positioning of safeguards with respect to the approach speed of the human body. This is where light-curtain mounting distances come from.
- ISO 14119 — interlocking devices associated with guards, including how to design against defeat.
- ISO 14120 — fixed and movable guards: construction, fixings, viewing panels.
- ISO 13849-1 (or IEC 62061) — the performance level or SIL the safety-related control system must achieve.
If your customer is an automotive OEM, expect their supplier manual to reference these by number and to ask for the risk assessment file, not just a photo of the fence.
Start with the risk assessment, not the fence
ISO 12100 sets a hierarchy, and the order is not optional:
- Design it out. Reduce the hazard by design — lower force, lower speed, remove the trap point, move the load position outside the danger zone. Free, permanent and cannot be defeated.
- Safeguard. Guards and protective devices for what you could not design out.
- Inform. Markings, warnings, instructions and training for the residual risk. Weakest, and never a substitute for the first two.
For each hazard you identify severity of injury, frequency and duration of exposure, and the possibility of avoiding it. Those three feed the required performance level (PLr, a to e) under ISO 13849-1. A trapping hazard an operator is exposed to every cycle usually lands at PLd, which drives real architecture decisions — dual-channel circuits, monitored contacts, a safety relay or safety PLC — not a single limit switch wired into the standard PLC.
How safety distances are actually calculated
Two different calculations, often confused.
Fixed guards — ISO 13857
A fixed barrier stops the hazard being reached. The standard tabulates the safe distance from the hazard for a given barrier height and opening size: taller fence, smaller distance; larger mesh aperture, larger distance. A 40 mm mesh needs far more standoff than a 20 mm mesh at the same height. This is why cheap large-aperture mesh often ends up costing more floor space than fine mesh would have.
Light curtains and other detection devices — ISO 13855
A detection device does not stop reach — it detects it and stops the machine in time. The minimum distance is:
S = (K × T) + C
where S is the minimum distance from the detection plane to the hazard, K is the approach speed constant, T is the total stopping time of the whole chain, and C is an intrusion allowance that depends on the device's detection capability.
For perpendicular approach to a vertical light curtain, K is taken as 2000 mm/s where the calculated distance is 500 mm or less, and 1600 mm/s where it exceeds 500 mm. For a detection capability of 40 mm or finer, C = 8 × (d − 14) mm, where d is the detection capability in millimetres — so a 14 mm finger-detection curtain gives C = 0, while a 30 mm curtain gives C = 128 mm. Coarser devices carry a much larger fixed allowance.
The term people get wrong is T. It is not the machine's stopping time alone. It is the device response time plus the safety relay or safety PLC response time plus the valve or contactor response time plus the mechanical run-down of the machine, and it grows as brakes and valves wear. Measure it with a stop-time analyser on the actual machine rather than adding up datasheet figures, and re-measure it periodically.
Choosing the guard type
| Fixed guard | Access needed only for maintenance. Removable only with a tool. Cheapest, most reliable, first choice wherever it does not obstruct production. |
|---|---|
| Interlocked movable guard | Routine access for loading, changeover or clearing. Door opening commands a stop; on machines with run-down, add guard locking so the door cannot open until motion has ceased. |
| Light curtain | Frequent manual loading where a door would slow the cycle. Needs clear stopping-time data and enough standoff — if the distance does not fit, the curtain is the wrong answer. |
| Safety laser scanner | Mobile equipment, AGVs and area protection where the footprint changes. Configurable zones, but slower response than a curtain, so a longer distance. |
| Two-hand control | Presses and similar single-operator cycles. Protects the operator's hands only — it does nothing for anyone else nearby, so it is rarely sufficient alone. |
| Pressure-sensitive mat or edge | Detecting presence inside a guarded area, or trapping hazards on moving guards and doors. |
Designing guards operators will not defeat
ISO 14119 has a whole section on motivation to defeat, because it is the commonest failure mode in the field. What works:
- Guard the hazard, not the machine. Fencing the whole cell when only one axis is dangerous creates long walks and slow access, and that is what drives bypassing.
- Put material access where material moves. Chutes, transfer openings and tunnel guards sized to ISO 13857 let parts in and out without opening anything.
- Make the interlocked door the fastest route for the access that actually happens several times a shift, and reserve bolted fixed panels for the ones that happen twice a year.
- Use coded or high-coded interlock devices where defeat is plausible, and mount them so a spare actuator cannot simply be held against the sensor.
- Give maintenance a legitimate path — trial or inching mode with an enabling device and reduced speed — instead of leaving bypassing as the only way to do their job.
- Keep sightlines. Polycarbonate panels let supervisors see the process and remove one more reason to open a door.
Budgeting for guarding
Indicative planning bands, not quotations. Guarding is usually 3–8% of a machine's cost when designed in, and considerably more when retrofitted around a machine that never allowed for it.
| Mesh fencing with posts | Roughly ₹2,500–5,000 per running metre of standard-height panel, depending on mesh, finish and post spacing. |
|---|---|
| Interlocked access door | Roughly ₹25,000–80,000 per door including frame, interlock device and wiring; guard-locking devices sit at the upper end. |
| Safety light curtain pair | Roughly ₹60,000–2.5 lakh depending on protected height, resolution and functions such as muting or blanking. |
| Safety relay or small safety PLC | Roughly ₹15,000–1.5 lakh depending on channel count and whether the architecture needs programmable safety. |
| Risk assessment and safety file | Engineering time, typically a few days per machine; often the single highest-return line in the list because it prevents over-guarding. |
The costly mistakes are predictable: guarding an entire cell because nobody analysed which part was dangerous; selecting a coarse-resolution light curtain and then discovering the required standoff does not fit the floor space; and specifying PLe architecture where the risk assessment only called for PLc.
Retrofitting guarding to existing machines
Older machines usually need more than a fence. Common findings on an audit of a legacy line:
- No monitored stop category — power is cut by a single contactor with no feedback, so a welded contact goes undetected.
- Emergency stops wired through the standard PLC rather than a safety circuit.
- No means of isolating and locking off energy for maintenance, including stored pneumatic and gravity energy.
- Stopping time never measured, so any light curtain distance on site is a guess.
Sequence a retrofit as: survey and risk assessment → measure actual stopping times → design guarding and the safety control architecture → install during a planned shutdown → validate each function and document it. Validation matters: a guard that has not been function-tested against its risk assessment is not evidence of anything in an audit.
Where we come in
We design and build guarding as part of the machines and lines we supply, and we retrofit it to existing equipment — risk assessment, guarding design, safety circuit architecture, fabrication, installation and validation, with the documentation your customer audits will ask for. See our guarding and safety products, or send us photos and a layout and we will tell you what the gaps are before quoting.