The decision in one line
Retrofit when the mechanics are sound. The control is replaceable; worn ways, screws and spindles are not — and no control upgrade restores accuracy the machine has physically lost.
Costing a retrofit properly starts with an honest mechanical assessment, not a control specification. This guide covers what drives the number and where estimates go wrong.
Assess first — this determines everything
| Check | Good sign | Changes the economics |
|---|---|---|
| Ballscrew backlash | Minimal and consistent | Significant or varying |
| Way condition | Even wear, good geometry | Scoring, stepped wear |
| Spindle | Quiet, low runout | Bearing noise, runout |
| Squareness | Within tolerance | Out and not adjustable |
| Crash history | None significant | Repaired structural damage |
Measure backlash and check geometry before quoting anything. The most common way retrofits disappoint is discovering during commissioning that the machine cannot hold tolerance. Software compensates for repeatable error. It cannot compensate for wear that varies along the axis, which is exactly what worn screws and ways produce.
What the money goes on
- Control hardware — industrial PC or dedicated controller, operator panel, enclosure.
- Drives — usually replaced for modern fieldbus interfacing and available spares.
- Motors — sometimes retained if condition and encoders permit; otherwise a significant line item.
- Feedback — encoders, or glass scales if you want closed-loop position on the axis rather than the motor.
- IO and safety — modern IO modules, safety relay or safety PLC, E-stop chain, guard interlocks.
- Wiring and enclosure work — routinely underestimated; old wiring is usually replaced wholesale.
- Software engineering — kinematics configuration, PLC logic for auxiliaries, HMI, tool changer sequencing.
- Commissioning — on-site tuning, accuracy verification, test cutting, operator training.
The three factors that move the number most
1. Tool changer complexity
This is the single largest swing factor. A machine with no changer is straightforward. A carousel adds meaningful work. An arm-type changer with orientation, multiple sensors and mid-cycle fault recovery can approach the cost of the entire axis control work.
2. Whether motors and drives are reused
Retaining serviceable motors saves substantially. But reusing motors with obsolete encoder interfaces frequently costs more in interfacing effort than new motors would — so this needs evaluating rather than assuming.
3. How much undocumented behaviour exists
Machines accumulate quirks over decades. Undocumented PLC logic, special cycles, and workarounds operators use daily all have to be rediscovered and reimplemented. This is where schedules slip.
Budget by tool changer and by how much nobody wrote down. Axis count matters far less than either.
Where estimates commonly go wrong
- Auxiliary functions underestimated. Coolant, lubrication, chip conveyor, guarding — individually trivial, collectively substantial.
- Existing programs assumed to run unchanged. If decades of proven G-code must work as-is, that constrains dialect choices and adds compatibility work.
- Safety treated as an add-on. A new control system needs a properly engineered safety chain with certified components, not the reused original.
- No allowance for mechanical work. Assessment frequently reveals a screw or bearing that should be replaced while access is easy.
- Training omitted. Operators know the old control. Productivity drops until they know the new one.
- Documentation skipped. Cheapest to produce during the work; expensive to reconstruct later.
Retrofitting a fleet changes the arithmetic completely. The engineering — kinematics, PLC logic, HMI, post-processor — is largely done once and reused. Machine two onward costs a fraction of machine one. If you have several similar machines, retrofit them as a programme rather than individually.
What you get beyond a working control
- Available spares — the original reason for most retrofits.
- Modern program transfer — network or USB instead of tape or serial.
- Look-ahead and smooth motion, which old controls with shallow buffers simply cannot do.
- Compensation for repeatable geometric and backlash error.
- Diagnostics and logging that make future faults findable.
- Connectivity for production monitoring if you want it.
That look-ahead point is worth emphasising: many retrofitted machines run 3D surface programs dramatically faster afterwards, not because the mechanics changed, but because the old control could not buffer enough blocks to maintain feed.
When to replace instead
- Mechanical wear requiring a rebuild on top of the retrofit.
- The machine no longer suits the work you do — wrong size, wrong spindle, wrong capability.
- Retrofit cost approaching a good used machine in sound condition.
- Structural damage or repeated crash history.
Have machines with good iron and dead controls? Tell us the machines and what they must do — we will say plainly if replacement is the better call. See our CNC controller service and the practical retrofit guide.