Chuck Yung
EASA Senior Technical Support Specialist

With a steady increase in random wound AC motor sizes and the obvious superiority of the form coil winding, one area where we can help improve customers' motor reliability is by redesigning those large random wound motors to accept form coils. Most repairers would agree that machines rated larger than 600 hp (450 kW) should be designed as form coil machines. Likewise, those rated over 2 kV will be much more reliable as form coil machines.

No one wants to rewind a motor using 60 #14 AWG (62- 1.6 mm) wires in hand. With an abundance of niche suppliers of stator laminations, the cost and practicality of converting a random wound motor to form coil are available to nearly all service centers. Replacement laminations can be punched, laser-cut or water-cut, and supplied with very reasonable delivery times.

 See Figure 1. 

It’s worthwhile to consider this conversion from an overall view first. The slot geometry must change, of course. But there are several opportunities possible with this change. Use EASA’s AC Motor Verification and Redesign Program – Version 4 to calculate the magnetic flux densities and current density. Focus initially on the backiron density, as it’s often practical to increase the slot depth to accommodate more copper. Consider the tooth width and slot width together, as any increase in copper width and/or groundwall insulation comes at the expense of tooth density. 

Copper area can nearly always be increased, with the gain in reliability greater than the ratio of the change in conductor cross-sectional area. The volts/coil stresses are much better managed in a form winding, as the volts/turn are limited to the volts/coil divided by the turns/coil. [The volts/turn of a random winding can be as high as the volts/coil.]

Redesigning a motor/generator from random winding to form winding is an iterative process. The calculations are simple, and you should take a couple of passes through EASA’s AC Motor Verification and Redesign Program to optimize the changes. With that in mind, use the “Editor” “Clone” buttons; that way your different iterations are each saved for review. It’s helpful to save each iteration as the job number -A, then -B, etc.

The first step is to use the AC Motor Verification and Redesign Program to determine the magnetic flux densities (airgap, tooth and backiron), as well as the volts/coil, for the original random winding. Keep the EASA recommended limits in mind, and try to stay comfortably below those limits (airgap density upper limit of 65k [kilo-lines per square inch], tooth density of 130k, and backiron 132k). Reviewing those numbers, assess how much room there is to work with – especially backiron density. The airgap density, of course, should not change unless there is also a horsepower change involved.

The slot width can be determined by:

[(Bore i.d. + slot depth) π)/number of slots] – tooth width

The lower the tooth density, the more potential to increase the slot width.

Depending on the backiron density, the slot depth can usually be increased by 20% or more, allowing approximately 0.060” (1.5 mm) tooth above the wedge groove and a wedge groove of 0.090” - .125” (2.2 mm-3 mm). That leaves the usable slot depth/depth below the wedge. Note: For sufficient core stiffness, the backiron dimension should be at least 80% of the slot depth.

To size the conductor, keep in mind the circular mils/amp of the random winding. The goal should always be to increase that, both for better efficiency and reliability. Determine the wire area in hand and size the rectangular wire to be used based on the slot width, with a 10% increase in wire area over the original random winding. To avoid a decrease in power factor, use full-length, tightly fitted magnetic wedges. (For ease of insertion use multiple shorter length wedges; they just need to cover the full length of the slots.) Power Factor (PF) is reduced at less than rated load, and PF correction capacitors are available if the customer does want to improve PF.

Groundwall insulation should follow the conservative suggestion in Table 1 below.

Work from the preliminary slot dimensions, with the groundwall requirement from Table 1, and determine how much conductor can fit. It’s prudent to give yourself a 0.010” (0.25mm) allowance for the coil to be smaller than the slot, so the coils can be easily inserted. Remember, no matter how exacting the supplier of the new core is, the slot width will not be perfect; it’s impossible to stack a thousand or more laminations perfectly.

Now run the preliminary form coil design data through the AC Motor Verification and Redesign Program and check the results. The tooth and backiron densities should remain below 130k, preferably by a comfortable margin. The circular mils/amp should have increased, in most cases. You can increase the conductor width – at the expense of tooth density. Never decrease the groundwall insulation below the recommendation in Table 1. If backiron density is comfortably low, then increasing the slot depth under wedge will permit an increase in wire thickness.

• Always work from standard rectangular wire sizes. If the coil vendor has to order special wire sizes, that will delay delivery time for coils.
• Nothing says you must use the same connection. Form windings often lend themselves to a change in circuits without risk of increased volts/coil stress.

Let’s work through a real-life example taken from the EASA database. The original motor is an 800 kW, 4-pole with 72 stator slots, and 25 #14 AWG wires in hand, rated 480 volts at 1073 FLA.

Core length 21.25” (540 mm)
Turns/coil:  2
Bore diameter: 18.50” (470 mm)
Wire: 25 #14
Backiron: 2.75” (70 mm)
1-12 pitch
Tooth: 0.375” (9.5 mm)
4Y connection
Slot depth:  1.94” (49.3 mm)

There are several things apparent when glancing at this data. First, 2 turns with a 4Y lends itself to a connection change to 4-delta; this will increase the turns and thereby reduce the wire size. Second, the 1-12 pitch is very narrow for a 72-slot, 4 pole. Entering the data into the AC Motor Verification and Redesign Program, the densities are as shown.

While the tooth density limits us, the low backiron density shows promise. We should be able to deepen the slots to accommodate an increase in wire size. Further, increasing the coil pitch looks like a good way to decrease the actual turns which also allows an increase in conductor size.

Slot depth
As a quick start, ratio the backiron density to a conservative 125k, and see how much we can decrease the backiron. That correlates to our possible increase in slot depth:

99.6k / 125k = 0.797

Decreasing the backiron to 2.19” (55.6 mm) permits a corresponding increase in slot depth to 2.500” (63.5 mm); from that we subtract the 0.060”(1.5 mm) and 0.090”(2.3 mm) for the tooth above the wedge groove, and the wedge groove:

Depth under wedge  =  2.500” (63.5 mm) – 0.15” (3.8 mm) =  2.350” (59.7 mm)

Slot width
To determine slot width, the best thing to do is actually measure the average slot width. An alternative is to use the formula below, which is based on the reported tooth width. (Winders are notorious for reporting all teeth as being ¼” (6.35 mm) wide, so check this yourself.)

{[( stator bore + slot depth)π/ slots} 

– tooth width = slot width

{[(18.5" + 2.5")π / 72} - .465″ = 0.451”

{[(470 mm + 63.5 mm)π/ 72} - 11.8 mm = 11.46 mm

The wedge groove should be as simple as possible. Avoid angled wedge sides; the magnetic wedges must fit tightly.

Using the redesign program, an equivalent winding using 3 turns, a 1-15 coil pitch with a 4-delta connection addresses the concerns mentioned earlier.

Raw conductor estimate
To determine the usable slot under wedge, subtract an allowance for bottom stick and separator. I use 0.100” (2.5mm). Divide the result by two coils; from that subtract the groundwall allowance from Table 1. Divide that result by the number of turns to get a maximum wire thickness. Slot width minus the groundwall allowance gives a working wire width. 

To size our conductor, a conservative conductor width is 0.381” (96.8 mm) (0.451” [11.46 mm] – 0.070” (1.8 mm)  =  0.381” (96.8 mm)

Wire thickness

2.500” (63.5 mm) – 0.100”  (2.54 mm) = 2.400” (61 mm)
2.400” (61 mm)/ 2 = 1.200” (30.5 mm) coil height
1.200” (30.5 mm) – 0.070” (1.8 mm) groundwall = 1.130” (28.7 mm)
1.130” (28.7 mm) / 3 turns = 0.376” (95.5 mm) Conductor: 0.376” (95.5 mm) x 0.381” (96.8 mm)

Now we plug in the new core and winding information as a double-check:

Densities
Airgap	56.6k
Tooth	105k
Backiron	126k
Cicular Mils/amp	1026

 

Reviewing this new data using the revised slot measurements, the backiron density is near our limit, and the circular mils/amp are three times the original. It makes sense to decrease the slot depth (and wire thickness) to bring the backiron density down a bit.

Adjusting for a backiron density closer to 110k:

110/126 x slot depth of 2.500” (63.5 mm) gives us a new slot depth of 2.182” (55.4 mm)

The backiron dimension increases to 2.510” (63.8 mm)

The resulting backiron density of 110k is acceptable, so we need to revise the wire thickness:

2.182” (55.4 mm) – 0.150” (3.8 mm) gives a new depth under wedge of 2.032” (51.6 mm)
2.032” (51.6 mm) - 0.100” (2.54 mm) = 1.932” (49.1 mm)
1.932” (49.1 mm) / 2 coils = 0.966” (24.5 mm) coil height, minus 0.070” (1.78 mm) groundwall leaves 0.896” (22.8 mm).
0.896” (22.8 mm) / 3 turns = 0.298” (7.57 mm) conductor 

The densities are now much better:

Densities
Airgap	56.6k
Tooth	106.6k
Backiron	110k
Cicular Mils/amp	797

 

The raw cm/amp of 797 is over twice the original; consulting the rectangular wire size table, one stock wire size is four in hand of 0.144” (3.66 mm) x 0.182” (4.6 mm). Work with your coil vendor to determine the final wire sizes. They may elect to reduce the wire thickness to reduce eddy-current strand losses, and they might decide to use two different wire widths, with two of each size diagonally opposite to provide a more secure coil construction.

Final Core		Winding Data
Slot depth	2.182” (55.4 mm)	Turns:3	Connection:4-delta
Depth under wedge	2.032 (51.6 mm)	Pitch: 1-15
Slot width	0.451” (11.46 mm)	Conductor: (4) 0.144” x 0.182”	0.451” (11.46 mm)
Backiron	2.510” (63.8 mm)

 

To minimize the impact on power factor, I recommend always using magnetic wedges when redesigning from a random winding to form winding. Magnetic wedges need to fit the wedge grooves securely and should extend the entire length of the core. Form coil windings are easier to block securely, and a VPI process keeps moisture and chemicals out of the winding while also improving heat transfer. The end result is a more reliable motor for the customer. In this case, being able to nearly double the circular mils/amp, the efficiency improvement is noteworthy. 

AVAILABLE IN SPANISH

Categories: Technical topics, AC motors, Stators/windings, Form wound, Random wound, Winding connections, Design/theory/application

Tags: Winding connections

Rate this article:

1.0

Print