Skip navigation
Wind 'em Up and Let 'em Go

Wind 'em Up and Let 'em Go

When it comes to motor efficiencey, 'rewinding' has long been considered a naughty word. Not so, says a recent study Claims that rewinding inevitably decreases the efficiency of induction motors have reached mythic status. Thanks in part to a handful of studies of mostly smaller motors (no larger than 30 hp or 22.5kW), many in the electrical industry have been led to believe that efficiency can drop

When it comes to motor efficiencey, 'rewinding' has long been considered a naughty word. Not so, says a recent study

Claims that rewinding inevitably decreases the efficiency of induction motors have reached mythic status. Thanks in part to a handful of studies of mostly smaller motors (no larger than 30 hp or 22.5kW), many in the electrical industry have been led to believe that efficiency can drop by as much as 5% when a motor is rewound. Efficiency supposedly suffers even more with repeated rewinds. In an effort to dispel these myths, the electric motor repair industry, led by the Electrical Apparatus Service Association (EASA) and the Association of Electrical and Mechanical Trades (AEMT) undertook a study to find answers to efficiency questions for customers and others. The research focused on motors with higher power ratings than those in previous studies — those most likely to be rewound. The results are based upon independent efficiency testing before and after rewinding.

A key finding of the study was that the best rewind/repair procedures maintain motor efficiency within ±.2%. The multiple rewind stage of the study determined that efficiency actually improved .2% on average when best practices were followed. Publication of a comprehensive report and a new set of best practices in the coming months should help ensure comparable results for the entire industry.

Objectives of the study. The study involved 23 low- and medium-voltage, 2- and 4-pole motors ranging from 50 hp to 300 hp (37.5kW to 225kW) ( Fig. 1 above). Both 50-Hz and 60-Hz IEC and NEMA designs were tested. Open drip-proof (IP 23) and totally enclosed fan-cooled enclosures (IP 54) were included. EASA and AEMT had several objectives, including the following:
  • Determine the effect of rewinding/repair on motor efficiency.

  • Investigate the effect of repeated rewinds on several motors.

  • Identify and compile a set of rewind/repair procedures that will maintain or improve motor efficiency.

  • Identify rewind/repair procedures that could degrade motor efficiency.

  • Identify and compile a set of best practices for motor rewinds.

Round robin testing and testing protocol. To ensure accurate test results, a 30kW IEC motor was efficiency tested first by the University of Nottingham. A second round of tests were conducted by U.S. Electrical Motors, St. Louis; Baldor Electric Co., Fort Smith, Ark.; and Oregon State University, Corvallis, Ore.

The IEEE 112 Method B (IEEE 112B) method of efficiency testing was used. All testing used the loss-segregation method — at no load and full load — which allowed for detailed analysis. As a benchmark, the results were compared with those of round robin test programs previously conducted by members of the National Electrical Manufacturers Association (NEMA). Interestingly, the first NEMA tests were conducted without standardized test procedures and yielded results that varied by 1.7 points of efficiency. This variation was subsequently reduced to .5 points of efficiency when NEMA members used standardized test procedures.

The efficiency variation, per IEEE 112B, for the round robin tests of the 30kW motor didn't exceed 0.9 points. This was achieved without standardization and compares favorably with the 1.7% variation of the NEMA tests.

Loss analysis. This study was conducted in three stages: six motors were stripped and rewound in Stage 1; 10 low- and medium-voltage motors and one high-voltage motor were stripped and rewound in Stage 2; and in Stage 3, four motors were stripped and rewound multiple times, and two motors were burned out five times and then rewound.

A low burnout temperature of 660°F (350°C) was used for Stage 1, which made it difficult to remove the coils and clean the slots. This resulted in mechanical damage to the stator core like flared teeth at the ends of the stator laminations, which increased the pulsating, or zig-zag, losses. Stray load losses increased significantly for these machines. No attempt to control winding resistance was made for the six machines rewound in Stage 1.

The burnout temperature range was increased to 680°F to 700°F (360°C to 370°C) for Stage 2, which made the coils easier to remove and required less cleaning of the stator slots. The winding resistance of the new coils also was controlled to avoid appreciable increases in stator I2R losses. The burnout temperature and the winding resistance controls for Stage 3 were the same as in Stage 2. Results for all three stages are shown in the Table).

Preliminary recommendations. As stated earlier, a list of best repair/rewind practices based on the results of the study will be published early next year. In the meantime, the following preliminary recommendations can help in maintaining the efficiency of rewound motors.

Burnout process. It's important that this process be tightly controlled. Burning out at a temperature below 680°F (360°C) will not fully release the varnish, making coil removal difficult. The additional force required to remove the coils may cause mechanical damage, such as splayed teeth, which will increase the stray losses. Burning out at more than 680°F (360°C) for organic core plate or 750°F (400°C) for inorganic core plate may increase the risk of damaging the interlaminar insulation and increase the core losses.

Coil removal. Cutting the end winding with an air chisel resting against the core end laminations or causing the end laminations to flare when pulling the coils out of the slot will increase the stray losses (Photo).

Conductor cross-sectional area. When changing wire gauge due to stock availability, never reduce the cross-sectional area because doing so will increase stator I2R losses. In fact, you should require that the cross-sectional area be increased whenever possible. Mean length of turn. Failure to control mean length of turn (MLT) may increase stator I 2R losses. Whenever possible, you should try to decrease the MLT. You can do this by reducing the straight section of the coil (where it exits the slot) to the minimum required for the slot liner cuff or tracking length. Remember to keep coil extensions as short as possible. Changing winding configuration. Don't change a two-layer winding to a single-layer winding. At first this may appear attractive since there are half as many coils to wind and connect. However, this could increase stator I 2R losses and will increase stray losses. Suitably trained personnel should control all winding conversions. Bearing lubrication. Over lubrication of bearings will increase friction losses and shorten bearing life. Mixing incompatible greases will increase friction losses and impair lubrication. Use sealed-for-life bearings or open bearings with manufacturer's recommended grease fill. Don't replace open or shielded bearings with sealed bearings unless they're the non-contact type. Ensure that bearings have the same internal clearance as original bearings. Decreasing the clearances could increase the friction losses. Fans. Replacing a fan with a different design could increase windage losses, raise temperatures, or increase I 2R losses. New fans should meet the designer's original intent. Catastrophic damage. It may be unwise to repair motors with catastrophic damage since there could be a reduction in efficiency. The customer's operational needs, however, may necessitate the repair despite the possibility of a reduction in efficiency. The cost of lost production may greatly exceed the cost of lower efficiency. In such cases, encourage the customer to purchase a new motor as soon as practical. Core losses. Due to the wide variety of electrical magnetic steels used by manufacturers, it's impossible to set rigid rules for core test acceptance. However, measuring core loss before burnout and after core stripping and cleaning will identify significant increases in core losses. Levels of repair. There are five levels of repairs dictated by the amount of work necessary to correct the damage sustained. Motors that require repair as noted in Levels 1 through 3 can usually be rebuilt and maintain their original efficiency levels. In the case of irreparable damage, it's not always possible to restore motors to their original efficiency levels. Motors that require a Level 4 or Level 5 repair should be evaluated based on the amount of damage that can't be completely repaired. Level 1. Basic reconditioning includes the replacement of bearings, cleaning of all parts, and replacement of lubricant. This level of repair also includes the addition of seals and other accessories as agreed upon with the customer. Level 2. This includes all of the items mentioned in Level 1 with the addition of varnish treatment of stator windings, repair of worn bearing fits, and straightening of bent shafts. Level 3. This includes all of the items mentioned in Level 1 as well as rewinding the stator, which involves replacing windings and insulation. Level 4. This includes rewinding of the stator plus major lamination repair or rotor rebarring. It may be necessary to replace the stator laminations or restacking of laminations. Shaft replacement would normally fall into this category. In short, Level 4 involves major repairs that are costly enough to justify examining the option of replacement. Level 5. This level includes motors that would normally be replaced except for special circumstances faced by the customer, such as no spare or unacceptable lead time for a replacement. Level 5 candidates include misapplied motors, those with inadequate enclosures, and pre-U-frame motors.

As these five levels imply, the damage resulting from a motor failure varies widely, as do the associated repair costs. An evaluation process that fails to consider the various levels of repair is too simplistic to yield an accurate assessment.

Bonnett is an education and technology consultant with EASA, St. Louis. Gibbon was with Dowding & Mills PLC, Birmingham, U.K., when the study was done.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.