Commutator Bar Repair

Most commutator bar problems are not “brush problems” and not “surface problems”. Once a bar has moved, cracked, or lost support, you are choosing between doing a real structural repair, replacing the commutator, or accepting repeat failures. Everything else is just delay.

Why commutator bar repair is rarely just cosmetic

A lot of material online treats commutator work as stoning, skimming, and undercutting, as if the copper ring were one homogenous part. In practice, you are dealing with individual bars, each with its own mechanical support, insulation stack, and riser connection. When one of those bars goes wrong, the surface finish is only the visible symptom.

Specialist guides on surface maintenance quietly admit this. Morgan’s service notes, for example, say that if the bars are moving under normal operating conditions, the unit should go to a repairer for bar replacement or re-seasoning before any surface work happens at all.  EASA’s loose-bar guidance mirrors that: tightening or reworking wedges without addressing the actual bar support is not a long-term fix. 

So when you think “bar repair”, read that as “local rebuild of the mechanical and electrical path for this segment”, not “clean it up in the lathe”. That shift alone pushes your work past most generic how-to pages.

Fast triage: surface job or structural bar job

You already know how to check runout and brush contact. So instead of re-explaining, treat triage as three blunt questions.

First question: does the commutator hold its shape when you lightly skim it? A stable, well-bonded bar pack will machine cleanly and stay round; a commutator with loose bars tends to bounce in subtle ways, leaving chatter, local flats, or a “washboard” feel that your fingers notice before your gauges do. Surface maintenance documents describe this indirectly as the point where truing stops improving the running condition. 

Second question: are the symptoms tied to one bar, a group, or the whole ring? A single dark bar, a bar with repeated brush burning, or one bar that keeps developing high mica usually signals a local mechanical or brazed-joint problem. Broad grooving or threading over many bars is more often brush grade, environment, or general commutation conditions. Industry repair notes classify grooving, streaking, and threading in exactly that way. 

Third question: how close are you to minimum diameter and minimum undercut depth. OEM and maintenance references give hard limits on how far you can skim before the commutator is no longer serviceable.  If you are already near that limit, any bar repair that needs another skim is really a replacement decision in disguise.

If all three answers are “not bad”, you probably have a surface job. If any one of them is a problem, you are doing commutator bar repair in the strict sense.

Reading the bar damage instead of just looking at the colour

The copper tells a story, but only if you read it past the obvious.

A dark bar that is still smooth and flush with neighbours often comes from slightly different current distribution, sometimes due to minor winding imbalances or brush holder position. That is usually not a bar repair case. A dark bar with localized pitting, raised metal at the leading or trailing edge, or recurring carbon build-up after a skim is different; this hints at local heating and arcing, and that usually tracks back to a loose bar, poor riser connection, or broken strands in that coil.

Longitudinal cracks along the bar, especially near the riser, suggest thermal cycling plus mechanical stress from the riser connection. On older machines with silver-brazed risers, that can be the first clue that the joint is fatigued or that past repairs overheated the copper.

If you see a step change in bar height, even a few hundredths of a millimetre, pay attention. Loose bars tend to “stand up” slightly under load, then sit back once cooled. Surface-maintenance PDFs warn that movement or distortion of bars under normal running should trigger commutator replacement or re-seasoning, not just stoning. 

Grooving and threading across many bars, on the other hand, often trace back to brush grade, humidity, or contamination rather than bar structure. Good DC motor repair guides treat these as surface issues that can be corrected by machining and cleaning, as long as the underlying commutator is still solid. 

So: individual, repeat trouble on one bar points you toward structural repair. Uniform patterns across the ring usually mean conditions or brush selection.

Anatomy of a real commutator bar repair

Online instructions tend to jump straight to steps. In the shop, the work feels more like a negotiation with the hardware.

You start by locking in the “as found” condition. Dial-indicator readings on diameter and runout at several axial positions, bar-to-bar flash test results, insulation resistance, and photos of brush tracks under normal load. This is not paperwork for its own sake. It defines how much distortion you are allowed to add with your repair before the customer starts to feel it.

Once the armature is in the lathe and supported the way it runs in the machine, you do a light skim, just enough to reveal what the copper wants to show. If the commutator machines cleanly and stays round, you have a stable base for bar work. If not, you are at the edge of what “repair” can honestly claim.

For a single damaged bar, classical references from GE and others describe a straightforward but delicate process: the old segment is removed, kept as a template, and a new segment is machined from solid copper, because segments are not interchangeable between commutators.  The new bar has to match not only width and height but the exact curvature and taper of the existing ring, or your skim later will sacrifice far more copper than necessary.

Removing the bar usually means cutting through the riser joint, separating the coils cleanly, and extracting the bar without scraping the mica on its neighbours. This is where imperfect logic sneaks in: you sometimes have to damage a little more to avoid damaging a lot. It feels wrong to dig into good insulation beside the bad bar, but if that frees the segment without prying, you may preserve the mechanical support where it matters.

Once the new bar is in, it must be locked mechanically before you trust any brazed or welded joints. That can mean re-wedging, reinstalling filler pieces, or renewing the commutator compound behind the bar, depending on the design. Only when the bar is mechanically solid do you restore the electrical path by brazing or welding the riser and reconnecting the coil leads. High-end repair shops treat that braze as a critical joint, with controlled temperature and quenching, because overheating changes copper hardness and can warp nearby bars.

Finally, you skim again, undercut, chamfer, and polish. But now the skim is small, almost a finishing pass, because you already matched the bar to the existing ring. If you need a heavy skim to make the new bar blend in, the template was wrong, or the commutator was already too far gone.

When to stop saving bars and replace the commutator

There is a point where “bar repair” is just short for “delaying replacement”. It is not the same point for every plant, but there are some practical markers.

Minimum diameter is the obvious one. DC motor maintenance notes usually state that the commutator can only be skimmed down to a diameter specified by the manufacturer, and that normal refurbishing already consumes part of that allowance.  If your repair plan requires another heavy skim, you are trading copper life for short-term uptime, and you should say so directly.

Severe wear and repeated resurfacing that still give poor commutation is another marker. Regeneration guides from motor repair firms are clear: when wear is too severe, the correct remedy is to replace the commutator, which normally means removing the rotor and sending it for full service.  Trying to rebuild multiple bars on a badly worn shell tends to produce compromised geometry, high vibration, and more brush problems later.

Persistent bar movement, even after re-wedging or re-cementing, also points to replacement. If the bar pack has lost its internal compression or the shell rings are loose, each thermal cycle will undo your careful work. You might make it through a test run. You rarely make it through another year of duty.

Multiple burnt bars, especially adjacent ones on the same coil group, can indicate a deeper winding problem rather than just local bar damage. At that point, you are halfway into a rewind problem set anyway. Building new bars onto a winding that is heading for failure wastes effort.

You do not need perfect logic here. You just need a threshold you will actually respect.

Undercutting, edges, and cleaning: the subtle part of bar repair

After any bar repair and skim, you still have to prepare the working surface, or your nice structural work will be hidden under more brush dust within weeks.

Undercutting depth between bars is usually in the range of about 0.5–1 mm for many industrial DC machines, enough to keep mica below the sliding surface but not so deep that the brushes drop into trenches. Repair articles that describe commutator regeneration often point to this depth range after machining.  The exact figure is OEM-specific, and you already know to check that, so we can jump straight to what gets missed.

First, the edges. After undercutting, the top edges of the bars need a small chamfer so the brushes see a rounded transition rather than a sharp corner that strips film. Practical guides on DC motor issues mention chamfering the bar edges together with undercutting as a standard step to reduce arcing and brush wear. 

Second, burrs and copper dust. Even one carelessly handled tool can smear copper across a slot and partially bridge adjacent bars. That bridge may not show up on a cold ohmmeter test but will matter under load. So the undercut is not finished until you have cleaned, inspected, and tested bar-to-bar insulation, not just cut the slot.

Third, deep cleaning between bars. Over time, carbon dust accumulates in the undercut and can create a conductive path. Some maintenance notes suggest cleaning the undercut area using pressurized air while the commutator turns, sometimes after mechanically dislodging stubborn deposits.  This is the kind of small step that dramatically changes the life of your repair even though it never shows on an invoice.

Practical repair planning: defects vs typical repair scope

To keep this grounded, here is a compact reference table you can adapt into your job instructions. It is not a standard. It is simply a distillation of what many shops already practice, stated bluntly.

This type of table is not meant to replace experience. It simply forces a link between what you see and how big a repair you plan, so you are not pretending a bar-level problem is a surface-level task.

Process control and documentation that actually helps the next repair

Documentation is often treated as a formality. For commutator bar repair, it is closer to a diagnostic tool that arrives late.

If you record the “as found” diameter, minimum bar width, undercut depth, mica condition, and bar-to-bar readings before you touch anything, you build a narrative that lives beyond this job. When the same motor comes back two years later with fresh bar damage, those numbers tell you whether the geometry drifted, whether the previous bar replacements concentrated stress, or whether operating conditions changed.

It also helps to track environmental notes: humidity in the motor room, load profile, starting frequency. Maintenance papers keep reminding readers that the film on the commutator surface depends heavily on humidity and operating conditions.  If you continuously see threading in a dry environment, that is not a coincidence, and it may push you toward a different brush grade or enclosure modification, not just more machining.

For bar repairs, include a sketch or photo sequence of the replaced segments, joints, and insulation stack-up. Generic as-built drawings rarely show field modifications, and those modifications matter when someone later tries to understand why only certain bars keep failing.

Small-shop shortcuts worth questioning

There are habits that persist mainly because they work just long enough to appear acceptable.

Filing bars in place, by hand, with no real control over profile, often seems attractive for small tools and low-cost jobs. It can remove local high spots. It also creates flat sections, uneven brush loading, and debris in the undercut that nobody has time to clean correctly. The cost arrives later, as noise, vibration, and brush wear.

Packing damaged undercuts with whatever insulating compound is on the bench, instead of matching the original mica system, can keep two bars from shorting today. It may also change the thermal behaviour and mechanical stiffness of that region. Over time, the odd segment can work differently under temperature, lifting or cracking while its neighbours stay stable.

Using “similar” copper from scrap stock to fabricate new bars, without matching conductivity and hardness to the existing segments, introduces mixed behaviour under load and different wear rates under the brushes. On a light-duty machine you might get away with it. On a high-current industrial drive, the differences can show up as unequal heating or surface finish after a few months.

Each of these shortcuts comes from rational pressure: time, budget, availability. The point is not to condemn them. The point is to see them clearly and to say out loud when a “quick repair” is actually a bet against the machine and your future schedule.

Closing thoughts: treating commutator bars as structural parts

It is easy to think of commutator bar repair as a kind of advanced polishing. But once bars have moved, cracked, or burnt, you are doing structural work on a critical current-carrying component that also defines the mechanical behaviour of the rotor. Standards and guides for rotating electrical apparatus keep repeating that proper repair practices and limits are essential for reliable operation. 

If you treat every suspect bar as a small structural project, with mechanical support, insulation, and electrical connection all considered explicitly, your repairs stop being mysterious. Some units will justify a meticulous bar replacement. Others will clearly fall on the “replace the whole commutator” side of the line. A few will turn out to be simple surface conditioning after all.

The value is in choosing honestly. The copper will keep the score for you.