How Should You Back Up and Restore PLC Programs?
To back up and restore a PLC program, save the engineering project together with its hardware configuration, required software versions, firmware details and authorised access arrangements. Document the restore sequence, then prove the package on suitable spare hardware or in an approved maintenance window. Treat the backup as ready for recovery only after a controlled test confirms the agreed behaviour.
What should a usable PLC backup contain?
Define the backup around a practical question: could an authorised engineer restore this controller using the saved package and its instructions?
Use the following checklist as the starting point. Mark each item as included, separately controlled or not applicable, with a reason.
| Item | What to capture | What to verify |
|---|---|---|
| Engineering project | Complete editable project and its approved revision | It opens without missing components |
| Controller identity | Model, hardware revision and installed firmware | The intended replacement is suitable |
| Hardware configuration | Rack layout, modules, addresses and communication settings | It matches the documented installation |
| Software dependencies | Engineering software version, device packages and libraries | They are available and usable |
| Access arrangements | Required permissions, credential owner and recovery route | An authorised person can obtain access |
| Operating data | Agreed recipes, settings and retained values | Each value has an approved recovery source |
| Connected equipment | References to relevant HMI, drive and other device backups | Ownership and restore order are clear |
| Recovery instructions | Connection method, sequence, checks and rollback plan | Another authorised engineer can follow them |
Keep an inventory alongside the package. Record when it was captured, who checked it, which machine it belongs to and whether its restore test passed.
Separate an archive of an old project from the package approved for current recovery. Both may be useful, but their status should be unmistakable.
Is an upload from the PLC enough?
Do not assume that a successful upload gives you a complete editable engineering project. What can be recovered from a controller, including comments, symbols and source information, depends on the platform and how the project was originally stored.
Where available, retain the approved engineering project as well as a fresh controller capture. Compare them using the platform's supported tools and investigate differences before declaring either one the recovery copy.
Record what the comparison actually checked. Avoid a broad statement such as “everything matches” if the check covered only part of the project.
If the original source is missing, document the limitation and agree what recovery work remains. An upload should not quietly become an approved master merely because it is the only file available.
How should you capture the backup?
Arrange the capture as a controlled maintenance activity:
- Identify the equipment. Confirm the machine, controller and connection point before connecting.
- Agree the permitted actions. Establish whether online access is allowed during production and who authorises it.
- Record the installed state. Capture hardware identity, firmware, software version and relevant configuration details.
- Collect the project and dependencies. Include the files and packages needed to open and work with it.
- Compare and review. Investigate differences between the saved project, controller capture and approved change records.
- Protect the result. Store the approved package with access controls and a separate recoverable copy.
- Record its status. Distinguish “captured”, “checked” and “restore-tested”.
Keep this session focused on capture. Handle any proposed controller change through a separately approved procedure.
How do firmware and software versions affect restoration?
Compatibility must be checked for the exact controller, replacement hardware, project version and engineering software involved. Do not assume that the newest available firmware or software will accept an older project unchanged.
Record exact versions rather than labels such as “current software”. Include required device descriptions, libraries, installation media and licence arrangements where applicable.
Before approving a replacement controller, check its compatibility against the saved project and the manufacturer's guidance. Resolve any required conversion or firmware change before the emergency replacement.
If a conversion is necessary, preserve the original package and test the converted version separately. Give each a clear status so maintenance staff know which package has been proved for which hardware.
How should passwords and access be handled?
The recovery instructions should explain how an authorised engineer obtains access without placing readable passwords in a general project folder.
Identify the credential custodian, approved storage location and out-of-hours access process. Include access to the controller, engineering workstation and any other system required for recovery.
Check access before the restore exercise. Confirm that the person performing the test has the necessary permissions and usable software licence.
Also check for protected project components and any recovery restrictions. Protection and password-recovery options vary by platform; establish the supported route rather than assuming access can be recovered later.
For operator-level access on screens, see which HMI permissions operators should have.
What about recipes, settings and connected devices?
Create a separate recovery decision for operating data. For each relevant value, specify whether to restore a saved value, enter an approved starting value or obtain the current value from another authorised system.
Depending on the platform and transfer method, downloading a project may initialise or overwrite some controller data. Verify the exact behaviour before a restore.
Review recipes, calibration settings, counters and other production-dependent values with the process owner. A value that was correct when captured still needs approval for the intended restart.
Define the boundary around connected equipment too. List which HMI, drive, robot or other device configurations the recovery depends on, who owns their backups and how they will be checked.
Put these dependencies and acceptance conditions in the recovery specification. Our guide to what a functional design specification should contain explains how to make responsibilities and expected behaviour explicit.
For approving and releasing recipe versions, see how to control machine recipe changes.
How do you run a controlled restore test?
Prefer an isolated test arrangement with suitable hardware. If the test requires the production machine, agree a maintenance window, safe working conditions and restart authority first.
Use an approved procedure covering isolation, output control and the people responsible for each step. For safety-related controls, define the test and any required validation with a qualified assessor.
A useful restore test follows this sequence:
- Agree acceptance criteria. List the communications, operating sequences, alarms, interlocks and data checks to demonstrate.
- Confirm the starting condition. Record the test hardware, firmware and isolation arrangements.
- Prepare rollback. Identify the recovery route if the attempted restore fails.
- Restore from the saved package. Follow the written instructions and record any missing file, permission or undocumented step.
- Check before operation. Verify the intended configuration and approved starting values.
- Test behaviour in stages. Perform the agreed checks under controlled conditions before authorising normal operation.
- Record the outcome. Keep observed results, unresolved issues and approval with the package.
Opening the project is one check; transferring it is another. Define success around the agreed recovery behaviour, with evidence for each acceptance criterion.
For production testing, include recovery and restart checks when planning a PLC retrofit around a shutdown window. Give rollback a clear decision point within that window.
What common mistakes should you avoid?
- Keeping only a file called “latest”. Identify the machine, revision, capture record and approval status.
- Leaving software preparation until a breakdown. Check installation files, dependencies and licence access during the test.
- Restoring old operating values without review. Assign an approved recovery source for each relevant setting.
- Treating a transfer as a complete test. Verify the agreed machine behaviour and connected equipment.
- Keeping the only copy on one laptop. Maintain a separate recoverable copy under the site's access rules.
- Keeping an untested change as the recovery master. Preserve the previously approved package until the replacement passes its checks.
Repeat the relevant recovery checks after changes to hardware, firmware, software, protection or operating configuration. Set the review interval according to the site's maintenance and change-control needs.
Related reading: managing PLC and HMI program versions.
Frequently asked questions
Can we restore to a different PLC model?
Treat this as an engineering change until compatibility has been established. Check the manufacturer's migration guidance and test the resulting configuration and behaviour before approving it for service.
Can we test without stopping production?
Use suitable isolated hardware where available. Agree which checks that arrangement proves and which still require a controlled test on the installed machine.
Who should own the backup?
Name a plant-side owner responsible for availability and approval. Define the integrator's responsibilities, handover contents and access arrangements in writing.
When is the backup ready?
When the package is complete, authorised access works and a recorded restore test meets the agreed acceptance criteria. Keep any remaining limitation visible alongside the recovery instructions.