Key takeaways
- A handover must transfer four states in writing, machine, material, quality and open actions; a verbal briefing transfers the fourth and loses the first three by the second break.
- Changeover settings that are not recorded are the top handover risk on a press brake, because a 0.3 mm backgauge offset chased on one shift produces parts that are 0.4 mm under on the next.
- If one handover in a hundred loses a machine setting and a lot is 380 parts, a 250-day year puts about 5,700 parts a year into a state nobody can trace.
- Record the last-off sample of the outgoing shift and the first-off approval of the incoming shift against the same report number; without that pair, a defect appearing at 07:00 cannot be attributed to a shift.
Related reading: the 5S on a fabrication floor: what it changes in practice and the Calibration register: how instruments stay traceable take the same problem further. See the process and inspection scope on the capability pages, or send the drawing through the RFQ route.
At 22:10 the brake operator added 0.3 mm to the backgauge to bring a flange back inside tolerance. He wrote it on the whiteboard beside the machine, and the whiteboard was wiped at the start of the next shift because nothing on it was a controlled record. The night shift ran 380 parts at the original offset. At 07:00 the morning inspector found the flange 0.4 mm under nominal on a sample of five, put the pallet on hold, and the whole lot went to the sorting bench.
The shift handover is the control that was missing. It is a documented, signed transfer of four states between the outgoing and incoming shift: machine state, material state, quality state and open actions. Everything else here is detail about those four states, the record that carries them, and the failure each one prevents.
The plant, quality and sustainability hub covers the other systems that hold a shop together, from calibration to incoming inspection.
What a shift handover has to carry on a sheet metal line
It has to carry the four states in writing, and the test of a handover record is simple: can a competent operator who has never seen the job before set the machine to the same condition, from the paper alone, without asking anyone.
Machine state is the hardest of the four, because a formed or welded part depends on settings that are invisible once the part is in a pallet. On a press brake that means the program number and revision, the backgauge offset, the ram depth, the tooling station in use and the shim stack. On a press it means the ram or stroke setting, the die height, the sensor positions and the counter reading. On a robotic weld cell it means the program revision loaded, the fixture in the station, the torch and tip condition, the wire batch and the gas flow rate.
Material state is the one that looks trivial and is not. A coil change at 02:00 in the middle of a lot, a part counted against the wrong lot number, or a bundle of CRCA sitting beside a bundle of DD grade with one tag between them, all produce parts that are dimensionally correct and metallurgically wrong.
Quality state carries the last-off result, any pending first-off approval, open hold tags, nonconformance numbers, deviation approvals and the status of gauges that have been withdrawn from use. Action state is everything that is due but not done: a maintenance request on a worn clamp, a die that needs a punch grind, a consumable that will run out inside the next shift.
The fields that make the record usable
A handover sheet with eleven fields and four blanks is a form. The table below is the working set for a press, brake or weld cell, with the field that an auditor or a customer quality engineer will actually ask to see.
| Field | What the outgoing shift writes | Ganesh limit or note | Where it is held |
|---|---|---|---|
| Program and revision | Brake program, press tool number, robot program revision | Revision, not just program number, because edits are the silent change | Handover sheet, tied to the lot number |
| Setting deviations | Backgauge offset, ram depth, shim stack, sensor position | Any deviation from the setup sheet must state the reason and the measured result that justified it | Setup sheet annotation plus handover line |
| Material identity | Grade, heat or coil number, sheets remaining, lots completed | Coil change mid-lot is recorded as a separate lot, never merged | Lot card and handover line |
| Last-off result | Two or three characteristics measured on the final part | Sample kept with the lot until the first-off of the next shift is signed | Inspection record |
| Hold and nonconformance | Tag numbers open, quantity, location, disposition pending | Hold area counted against the register at every handover | Hold register |
| Gauge status | Gauges in use, gauges withdrawn, calibration due this week | A gauge withdrawn for calibration is never left on the bench | Gauge register |
| Consumables | Wire batch, gas pressure, powder batch, pretreatment stage | Batch numbers, not "topped up" | Process record |
| Open actions | Maintenance, tooling, deviation approvals, safety items | Each action has an owner and a due time, not a shift | Action log |
The revision column does more work than any other on a press or a brake. A program that was edited at 21:00 to clear an interference on one station changes the ram depth for every part made after it, and a handover that records "program 4471" without the revision cannot distinguish the two states.
Why a verbal handover loses the information that matters
It loses it because the person who changed the setting knew why, and the person receiving the handover has no reason to ask about a number that is not on the sheet. The information exists in one head, for one shift, and it is gone by the second break.
The arithmetic makes the exposure visible. A three-shift operation has six handover boundaries a day. At 250 working days that is 1,500 handovers a year per line. If one in a hundred loses a machine setting, that is fifteen events a year. At 380 parts per lot, it is 5,700 parts a year produced in a condition nobody can reconstruct, and each of those events costs a sort, a rework decision and a customer conversation.
The 1-in-100 figure is an assumption and it is worth arguing about. What is not an assumption is that the cost scales with lot size and with how long the wrong setting runs before anyone measures. Both of those are controlled by the handover, not by the skill of the operator.
The first fifteen minutes of the incoming shift
A handover that is read is worth more than a handover that is written. Twelve to fifteen minutes at the machine, in this order.
- Read the record before touching a control. The outgoing shift's numbers are the starting point, not the machine's current state.
- Verify the machine against the record: backgauge zero, program revision on screen, ram depth, tooling station, fixture pin positions.
- Physically check the last-off part against the recorded result. A part that does not match its record is a stop.
- Run the first-off, measure the two or three characteristics that decide fit, and sign it before production resumes.
- Count the hold tags against the hold register and confirm the location of each pallet.
- Confirm consumables and process state: wire batch, gas pressure, powder batch, pretreatment tank concentration and temperature.
- Sign the handover, both operators, with the time.
Step three is the one that gets skipped, because the last-off part is usually at the bottom of a pallet. Keep it on top of the lot, in a labelled bag, and the check takes forty seconds.
The four failures this prevents, with the physical cause
Each of these is a recorded failure mode in a sheet metal plant, and each is caught by one field on the handover sheet.
Setting drift on a press brake. A flange measured 0.4 mm under nominal in the morning. The cause was a backgauge offset added on the previous shift to compensate for springback on a different batch of material, and left in the program. The physical mechanism is ordinary: the offset changes the bend line, the bend deduction changes, and every part after it inherits the error. The handover field is the setting deviation with its reason.
A coil change inside a lot. AA grade CRCA replaced DD grade at 02:00 when the coil ran out. Both are cold reduced steel, both measure 1.2 mm, and the draw quality is different: a part that formed cleanly on DD splits at the radius on AA. The handover field is material identity, and the control is to close the lot at the coil change rather than merge it.
A hold that ships. Covered in the callout above. The mechanism is physical, a label on the blind side of a pallet, and the control is counting tags against a register at every boundary.
Pretreatment chemistry drifting across a shift. A degreasing stage that lost temperature overnight still cleans a part well enough to look clean. The failure appears after cure, as a coating that cross-hatches off in flakes, and by then the parts are finished and packed. The handover field is the process state, with the tank temperature and concentration recorded at the start of each shift rather than at the start of the week. The seven-tank pretreatment process sets out which stages are sensitive to that drift.
A fifth belongs on the list for welded work: a robot program edited at the end of a shift to clear a fixture interference, with no revision recorded. The next shift runs the same program number and a different weld path.
What Ganesh runs
Ganesh has manufactured since 1992 and operates five units across Mumbai (Vasai) and Nagpur with more than 200 people. The production set that a handover has to describe is broad: mechanical presses from 20 T to 300 T, a 250 T punching line, a 12 kW fibre laser with a 40 ft x 16 ft bed, a 200 T five-axis brake with a 4,450 mm bed forming to 10 mm, three CLOOS robotic MIG cells and more than 30 manual MIG/MAG, spot and SPM stations.
The setting-dependent equipment is the reason this discipline matters here. Formed parts come off the bending capability with stored programs on the Delem control, which is only useful if the revision that ran is the revision on the record. Welded assemblies are built in fixtures designed and built in the in-house tool room, so a fixture pin that has been replaced between shifts is a geometry change that has to be written down.
The works hold ISO 9001:2015 and ISO 14001:2015 certification, and both bear on the handover. Clause 8.5.1 of ISO 9001 requires production under controlled conditions, which is what a signed setting record evidences; clause 8.7 requires nonconforming outputs to be controlled, which is what a counted hold register demonstrates. On the environmental side, the pretreatment line and the effluent treatment plant run continuously, so the shift record is also the operational control that ISO 14001:2026 clause 8.1 expects for a process whose failure is not visible in the product.
First article inspection and PPAP are available, and both depend on the handover record being intact: a first article taken from the shift after an undocumented change is a first article of a different process.
What to put in the handover standard
A one-page standard per production area, written with the operators who will fill it in.
- The exact fields, in the order they are filled, with an example line for each.
- The rule for a setting deviation: reason, measured result, and who authorised it.
- The rule for a coil change: close the lot, open a new one, and record both.
- The last-off sample location and the requirement to keep it with the lot.
- The hold register format and the counting step at every boundary.
- The consumable and process-state fields for that area, with the numbers recorded rather than a tick.
- The signature and time rule, and what happens if the outgoing operator has already left.
- The retention period and where the record is filed, by lot number.
The last item is where most shops are weakest. A handover folder by date is easy to keep and useless to retrieve. A handover tied to the lot number is the only version that answers the question a customer asks, which is not "what happened on the night shift in March" but "what was the machine set to when my parts were made".
When this is overkill, and when it is not enough
For a single-shift shop with one machine and one operator, a formal handover record is paperwork. The control that matters there is the setup sheet and the last-off, and the operator carries the rest in their head because there is nobody to hand over to.
For a plant running three shifts on a part with a 0.4 mm tolerance on a formed flange, it is not enough on its own. The handover transfers the state; it does not verify that the state is capable. That is the first-off and the sampling plan, and the inspection routine for repeat press work sets out the sampling that goes with it.
Ganesh does not run a proprietary handover software product and does not sell one. What we can show a buyer is the record for a lot: the settings, the samples, the hold status and the signatures, retrieved by part number rather than by date.
Send the drawing with the characteristics that decide fit, and the engineering team will return the route, the in-process checks and the record set that ships with each lot. Start with a quotation request that names the part family and the required volume.
Frequently asked questions
How long should a shift handover take?
Twelve to fifteen minutes at the machine, of which at least half is writing. A handover that takes three minutes of talk and no writing is a conversation, not a record, and it loses the machine settings first because those are the hardest to describe out loud.
What is a last-off sample and why does it matter at handover?
The last-off is the final part checked by the outgoing shift against the drawing. Paired with the incoming shift's first-off, it brackets the handover window, so a dimension that moves between 22:00 and 06:00 can be attributed to a setting change rather than blamed on the material.
How long should shift handover records be kept?
Agree the retention in the contract. Twelve months is a practical minimum for dimensional and setting records; automotive and railway programmes often ask for part life plus one year. The record has to be retrievable by part number and date, not only by shift.
Should the handover be paper or digital?
Either, provided it is signed, time-stamped and attached to the lot. Paper survives a network outage and is faster to write at the press; a terminal is better for retrieving six months later. The failure mode is the same in both cases: a field left blank because it did not seem to apply.
Does a robot cell need a handover if the program cannot change?
Yes, because the program is not the only variable. Fixture condition, wire batch, gas flow, tip condition and the position of a locating pin all change between shifts, and the robotic weld record has to name which of them were verified at the start of each shift.
Sources and further reading
- ISO 9001:2015 — Quality management systems: requirements, including clause 8.5.1 on control of production and clause 8.7 on nonconforming outputs · reviewed 3 Nov 2026
- ISO 45001:2018 — Occupational health and safety management systems, including clause 7.4 on communication and clause 8.2 on emergency preparedness · reviewed 3 Nov 2026
- ISO 14001:2026 — Environmental management systems, including operational control of the pretreatment line and effluent treatment plant across shifts · reviewed 3 Nov 2026
Need a production control plan for a repeat part?
Send the drawing, the annual quantity and the characteristics that decide fit. We return the route, the in-process checks and the records that travel with each lot.
Request a quote