Field note

Keypad Backups, Memory Modules, and Software Clones

A drive fails during a production week, and the replacement waiting on the shelf is the part of the event that everybody planned for. What few plants plan for is that the parameters in the failed drive were the only copy that existed, so a swap that should take hours becomes a recommission that takes days while someone attempts to reconstruct an acceleration ramp from memory. The backup, in that light, is the only durable record of what the drive actually was during production, which differs from what the drive was on the morning it was commissioned, and the two diverge within days of startup as the process teaches the crew what the drive needs.

Three Artifacts, Three Modes of Loss

Keypad backups, memory modules, and software clones tend to be discussed as interchangeable, although they hold roughly the same information and fail in markedly different directions. A keypad or panel backup lives inside the removable operator interface; it is the fastest of the three to take, requires neither a laptop nor a software license, and travels with the panel, which is precisely its weakness, since panels are borrowed for the next commissioning job and do not always return. ABB’s assistant control panels illustrate the constraint well, offering two user backup slots and, once both are occupied, asking the user to choose one and replace it, so that each new backup competes for space with the one taken before it (ABB, n.d.-a). A memory module, by contrast, is a physical object that outlasts the drive it came from, and its entire advantage depends on someone having labeled it, because an unlabeled module in a drawer with several others cannot be matched to a drive by anyone except, perhaps, the person who removed it.

The software clone, pulled with the manufacturer’s tool, is the only one of the three that can be named, dated, stored away from the panel it came from, copied, and compared against what is running now. It is also, in practice, the one most likely to reside on a single laptop, and laptops are replaced on schedules that have nothing to do with the drives whose configurations they hold. Most plants appear to maintain one of these three forms, and few seem to have chosen which one deliberately, which suggests that the form of the backup is usually a frozen accident, a decision made once by whoever commissioned the line and never revisited.

The Habit at the Panel

The parameter change most likely to cause harm is rarely the one made at commissioning. It is more often the change made at 2 a.m. to stop a nuisance trip, made by a technician who was correct, under pressure, with the line down and a supervisor waiting, after which the ramp has been lengthened, the trip has stopped, the line is running, and nothing has been written down, because writing it down was never the emergency. Six months later that drive fails, the commissioning backup is restored because it is the backup that exists, the ramp returns to its original setting, and the line begins tripping again for reasons nobody connects to the restore. Neither the drive nor the restore is at fault in that sequence; the backup simply described a drive that ceased to exist on the night of the change.

The corrective habit is small. When a parameter is changed with the intention of leaving it changed, a backup should be taken before the technician leaves the panel, which on ABB’s assistant control panels is a matter of selecting Menu, then Backups, then Create backup, a task that runs while the technician waits (ABB, n.d.-a). Two details determine whether the habit produces anything useful later. The panel clock should be set at commissioning under Menu, Settings, Date and time, since a backup without a reliable timestamp cannot be sorted against the others (ABB, n.d.-a). Where the control unit is powered from an external 24 V supply, saving parameters to permanent memory becomes a separate step, which on the ACS880 is parameter 96.07, Parameter save manually (ABB, n.d.-b); a change that was never written to permanent memory exists neither in the drive after a power cycle nor in any backup taken of it.

The Annual Pull to a Shared Location

The habit at the panel protects the drive, although it offers no protection against losing the panel, the drawer, or the person who knew which module belonged where. Once a year, every drive’s configuration should be pulled with the manufacturer’s software onto a laptop and then moved immediately to a server or shared location that maintenance and engineering can both reach. Plants without a software license are not excluded from this practice, since ABB’s panels allow backup files to be copied to a PC over the panel’s USB connection using an ordinary file manager (ABB, n.d.-a).

The location matters more than it first appears. A folder on one engineer’s laptop functions as that engineer’s private archive rather than as a plant program, and it lasts only as long as the engineer’s tenure. Maintenance needs the file during the night shift, engineering needs it when specifying a replacement, and where the two groups keep separate copies, one of them will eventually restore the wrong file. A single location with read access for everyone who might need it at night resolves that tension. File names should allow a stranger to use them, carrying the asset tag, the drive model, the firmware revision, the date, and the name of the person who took the backup; the firmware revision is the field most often omitted and the one most likely to decide whether a restore is sensible two years later. An annual interval is a minimum rather than a target, since any plant engaged in active process development is changing drive parameters more often than that, and the appropriate interval is however often the configuration actually changes.

What a Restore Actually Returns

A restore is a set of choices rather than a single action. ABB’s assistant control panels present restore options that include all parameters, a parameter restore group, application items, user sets, and production data items (ABB, n.d.-a), and a menu constructed that way implies that the parts of a drive’s configuration do not necessarily travel together. Whether the settings of a PROFINET or EtherNet/IP adapter return within the selected scope is a platform question, as is whether the configuration of a safety option returns at all, or returns in a state that requires revalidation before the machine runs. The manual for the platform resolves both questions in principle, and the restore screen on the installed drive tends to resolve them faster.

The implication is that the restore test belongs on a schedule rather than in an assumption. A backup should be restored onto a spare drive, or onto a live drive during a planned outage, with someone observing and the running configuration printed beside it. What returns is what the recovery plan can count on, and what does not return becomes a list of settings someone will otherwise reconstruct by hand during an outage, a list far more useful in a drawer than discovered under pressure.

The same test is a reasonable occasion to check one further assumption. ABB’s panel documentation describes an automatic backup that the panel creates a short time after the most recent parameter change (ABB, n.d.-a), which places the function at the panel level. The ACS880 primary control program firmware manual contains no backup and restore section (ABB, n.d.-b), and the firmware-level descriptions of the automatic backup belong to ABB’s ACS580 and ACS380 documentation. The function is therefore documented for the panel and not contradicted for the ACS880, although no ACS880 document confirms it, and the backup list on the panels actually installed remains the most dependable evidence of what is being kept.

In Finality

A spare drive on the shelf is only half an asset, because it holds no parameters, and what converts it into a usable spare is a current configuration file that someone can reach at night. The argument parallels the one for keeping a spare drive’s capacitors healthy, set out in VFD Capacitor Reforming, and it fails in the same manner, quietly, until the day the spare is needed. Configuration backup belongs with the rest of the work described in the VFD Maintenance and Reliability Guide: inexpensive, unremarkable, invisible while it works, and costly only as downtime attributed to some other cause. The practice that follows from the argument is to take the backup at commissioning alongside the items in the VFD Pre-Startup Checklist, to take it again whenever a parameter changes, to pull it to a shared location once a year, and to restore one deliberately so that the plant knows what a restore does.

Most plants appear able to produce a backup, and comparatively few have ever produced a verified restore, which is the gap most worth closing. The broader case for treating configuration as part of the installation rather than as an afterthought is developed in the VFD Maintenance and Reliability Guide and, at greater length, in Before the First Fault. Crews that have never performed a restore under observation can practice one in the on-site training.

References

ABB. (n.d.-a). ACx-AP-x assistant control panels user’s manual (Document 3AUA0000085685, Rev. E).

ABB. (n.d.-b). ACS880 primary control program firmware manual (Document 3AUA0000085967, Rev. Y).

Frequently Asked Questions

How often should VFD parameters be backed up?

A backup belongs at every parameter change that is meant to remain in place, taken at the drive by the person who made the change. Beyond that, an annual pull of every drive's configuration with the manufacturer's software, stored where both maintenance and engineering can reach it, catches what the individual habit misses. Plants that adjust drive settings frequently will need a shorter interval than once a year.

Does a keypad backup save the fieldbus and safety configuration?

That depends on the platform, the firmware revision, and the option modules installed, and it is a question better tested than assumed. Restore menus that offer several separate scopes rather than a single restore of everything suggest that not every part of the configuration travels together. A restore performed on a spare drive is the most reliable way to learn what a given backup returns.

What is the difference between a keypad backup, a memory module, and a software clone?

The three hold broadly similar configuration data and fail in different ways. A keypad backup is the fastest to take but lives in a removable panel with limited storage, a memory module outlasts the drive but is of little use without a label, and a software clone is the only one of the three that can be named, dated, stored away from the drive, and compared against the running configuration.

Where should drive backup files be stored?

Backup files belong on a server or shared location that maintenance and engineering can both reach, rather than on one person's laptop. Each file should carry the asset tag, the drive model, the firmware revision, the date, and the name of the person who took it, since the firmware revision largely determines whether a restore is sensible to attempt later.

Can a backup be restored to a different drive?

Sometimes, within limits set mainly by firmware revision and drive rating. A configuration taken from one firmware revision may not restore cleanly to another, and a configuration built for one frame size needs checking before it is loaded into a different one. Both limits should be confirmed against the manual for the platform before a restore is counted on during an outage.

Does a drive make automatic backups?

Some panels do. ABB documents an automatic backup at the level of its assistant control panel, which means the function depends on the panel and the drive behind it rather than on the drive alone. The backup menu on the panels actually installed is the most reliable evidence of what is being kept.

The annual report: The State of the VFD 2026

Nineteen drive brands assessed, thirteen scored on what they will actually put on the public record. Free, in a Plant Edition for specification work and a Channel Edition for quoting.

Get the report

Wayward Leaders® is a veteran-owned VFD training practice. We teach maintenance teams and plant electricians to install, commission, and troubleshoot variable frequency drives correctly.

Every class ends with a scored competency assessment. The plant gets a way to prove what each technician can do, not a roster of who sat in the room.

The instruction is field-trained. It is drawn from nearly 8,000 VFD, power, and motor documents and from more than 750 commissioned drives captured through personal field experience and data collection. Carl Lee Tolbert, PhD, CMRP, and ATD Master Trainer™, leads it. He has delivered more than 5,000 hours of VFD instruction across 30 years and trained 8,000 industrial professionals, from International Paper to the U.S. Navy.

The premise is simple: VFDs do not fail. Installations fail them. The curriculum is built backward from that.

Installation, commissioning, troubleshooting, and fault diagnosis guides are published openly at waywardleaders.com.