A mid-size discrete line (one to three cells, one HMI per cell, a single conveyor or process train) should plan for a 6 to 12 week calendar when staged by zone, or 8 to 16 weeks when run as a parallel/shadow cutover, per a December 2025 engineering roadmap [S3].
Method choice, not controller brand, is the dominant variable. A "like-for-like" swap with vendor conversion utilities lands near the short end of that range; a phased refactor with FAT plus staged cutover sits in the middle; a full parallel rack with hard cut pushes toward the long end. The 90-day phased model published in May 2026 (assessment, design, cutover, LOTO) is consistent with the 12-week figure for a single line [S7].
Scope tiers and the weeks each one eats
Three scope tiers map cleanly to elapsed time, not engineering hours, and they are the first thing to fix before quoting a schedule [S1][S2]. A single-cell PLC upgrade (one controller, one panel, unchanged I/O) usually clears in 2 to 4 weeks of calendar once parts are in hand. A mid-size line, defined here as one to three cells with shared conveyor or molding line glue, runs 6 to 12 weeks staged, or 8 to 16 weeks on the parallel path. A plant-wide modernization spanning multiple platforms and a network overhaul is measured in quarters, often 4 to 9 months, because the I/O survey and network redesign dominate once logic conversion is no longer the bottleneck [S1][S3].
The dominant schedule-killers, in rank order, are field I/O documentation gaps, vintage network gear, and HMI tag rebuilds. Legacy code conversion is rarely the slowest step if the source program is documented; the slowest steps are re-mapping racks, re-terminating field wiring, and rebuilding HMI screens to the new tag database, all of which must be verified on a factory acceptance test (FAT) before any field cut [S2][S3].
Method A vs B vs C, on a criteria matrix
Three execution patterns show up in field guides, and they trade cost, downtime, and risk against each other in a way that maps to a mid-size line's appetite [S3]. Method A, like-for-like swap with vendor conversion tools (e.g., RSLogix Project Migrator translating PLC-5 or SLC-500 ladder to ControlLogix), is the fastest path because I/O and field wiring stay put. Method B, staged migration by zone, is the typical "brownfield refactor": 6 to 12 weeks calendar, moderate cost, and disruption limited to a per-zone maintenance window. Method C, parallel rack plus shadow run with emulation, runs 8 to 16 weeks, costs the most in duplicated hardware, and is the right answer only when the line cannot accept a single extended outage.
The decision rule is simple: if a 24 to 72 hour per-zone window is acceptable, pick Method B. If the line is the bottleneck and any extended stop bleeds revenue, pay for Method C. If the existing I/O and wiring are already in good shape and the code is well-commented, Method A still has a place on single-cell upgrades where downtime must be minimized and the modernization is mostly a parts-availability fix [S3][S5].
What the calendar actually contains, phase by phase
Five phases recur across the 90-day and 12-week roadmaps, and their proportions are stable enough to plan against [S3][S7]. Discovery and risk register typically take 1 to 2 weeks and should produce a documented I/O list, a tag inventory, and a network topology drawing before any code is touched. Controls design and emulation usually take 2 to 3 weeks, including the target-platform mapping, the function-block library, and the offline simulation against the legacy program. Panel or rack build plus FAT take 2 to 4 weeks and are the highest-risk phase for rework because real field devices get connected for the first time. Field install plus site acceptance test (SAT) is staged or parallel, running 1 to 3 weeks of cutover windows. Ramp, training, and stabilization close out the schedule in 1 to 2 weeks with the legacy PLC kept available as a fallback [S2][S3].
A PLC migration roadmap that collapses phases 1 and 2, or skips FAT, will run shorter on paper and longer in practice. The reason is mechanical: a stage-3 FAT catches the I/O mapping and HMI tag errors that a stage-4 field cutover would otherwise catch at 2 a.m. on a live line, and the cost ratio between fixing them in the panel shop versus on the plant floor is roughly 10:1 on most projects I have watched [S2][S7].
What compresses the schedule, and what blows it up
Three factors reliably pull the mid-size line toward the 6-week end of the range, and each is a decision made before kickoff rather than a heroic effort during execution [S6]. First, a complete and current I/O list with verified wire labels removes the discovery-phase guesswork that typically inflates phase 1 by a week. Second, standardizing on one vendor across the plant collapses training time and spares inventory, and is the single biggest predictor of a smooth migration because the same engineer can move between racks without context-switching. Third, keeping the legacy PLC in service during FAT of the new system allows parallel testing instead of serial testing, which is the difference between 8 and 12 weeks on most staged projects [S4][S6].
The schedule-killers are equally consistent. Missing schematics, a programmer who left years ago, and unlabeled field wiring push the I/O documentation step from 1 to 2 weeks into 3 to 4 weeks, and they cascade. A master-follower drive or synchronized motion section, common on conveyors and stacker cranes, adds servo tuning time that pure discrete migrations do not face, and any upgrade that includes a new safety PLC under a safety PLC regime will need a documented risk assessment, validation per ISO 13849-1 or IEC 62061, and a separate SAT, which alone can add 1 to 2 weeks [S1][S3].
Downtime, rollback, and the line-acceptance window
For a mid-size line, the cutover window itself is short, usually 24 to 72 hours per zone, because the heavy work is done in FAT before the line ever stops. The "no-downtime" promise some integrators advertise is not literally true; what they mean is that the line's overall OEE impact is minimized by staging the work zone-by-zone and running the rest of the line on the legacy automatic molding line controls until each zone is proven [S3].
Rollback discipline is the second-order requirement. The legacy PLC should be kept powered, with its program intact and its I/O still terminated, until the new system has run a full production shift with the planned throughput and quality numbers. Skipping that holding period, or treating it as a formality, is how 12-week projects become 16-week projects. The published 6-week Emerson upgrade for a North Carolina conveyor OEM, which retired GE 90-30 and 90-70 controllers, is a real example of what compressed execution looks like when the I/O list is current and the legacy program is well-commented; the published 90-day phased model is a real example of what realistic execution looks like when the line is mid-complexity and documentation is partial [S7][S8].
Decision checklist before locking the schedule
Before committing to a 6, 12, or 16-week calendar, four questions resolve the rest of the plan. Is a 24 to 72 hour per-zone stop acceptable, or does the line need a parallel/shadow run? Are the field wiring, I/O list, and HMI tag database current, or does phase 1 need to absorb documentation work? Is motion control in scope, or is the migration purely discrete? And is a conveyor sorting line safety PLC part of the scope, because safety validation adds a fixed 1 to 2 weeks that no amount of parallelism can compress [S3][S6].
Answer those four, and the method, the calendar, and the budget fall out. If the answers are "yes, current, no, no," a mid-size line lands at 6 to 8 weeks on Method A or B. If they trend the other way, plan 12 to 16 weeks on Method C with a rollback path tested in FAT, and a legacy PLC kept warm on the plant floor for at least one full production shift after each zone cutover. Two trackable signals to watch in the next 6 months: integrator-published post-mortems on 90-day phased rollouts, and any vendor move that bundles conversion utilities with safety-controller validation, since both directly tighten the lower bound of the 6-week figure.