Lionel 2333 New York Central Diesel F3 Performance Testing

Thanks so much for doing this Mike.  Your 2333's performance absolutely ROCKS!   The 9:1 gear ratio over small 0.88" drivers means there's plenty of RPM and torque, even at slow speeds.  It's also impressive that those old Pullmor motors lug down below 600 RPM without stalling.  I've tested a lot of Pittmans in O-gauge locos that won't do that!

Unlike many modern dual-motored locomotives, the gears in your 2333 are back-driveable.  This means, if one motor begins to stall, the second motor can "push" the whole loco down the track a little, which serves to jog the first motor until its next armature pole is attracted by magnetic flux from its field.  It's almost like having a six-pole motor, so you are benefitting from about 20 usable power pulses per inch of travel.

If you used the full voltage of your ZW, I'm pretty sure you could have achieved 80 mph even with a trailing consist.

If you happen to test an MTH F3 made in the late 1990s with two vertical can motors and four rubber tires, I'm willing to bet that it won't run this well.  I was never impressed with the slow-speed performance of mine on sharp tubular curves.  Even if it didn't stall, there was a noticeable slow-down in the curves.  Meanwhile, your 2333 exhibits minimal speed variation.  So much for progress!  Another print magazine that's focused on 3-rail O used to publish speed tables like yours in their product reviews.  I can tell you, in the era before speed control, very few of the locos they tested were able to sustain 5 mph.

To me, your post is proof that Lionel's F3 "growlers" were king of the 3-rail mountain for over 50 years, and in many ways, still are!

So far, my only criticism of the Lionel 2333 is this: the dummy A-unit is very heavy, at 2 lbs 11.7 oz, it has a significant amount of rolling resistance, and although it looks good, it degrades the performance of the powered unit.

I suspect it would do even better if you had wider curves. I'm thinking that 11 cars around tight O-31 Curves is a lot of resistance. It's just about chasing it's own tail car

Amazing an almost 80 year old “toy” performs so well.

@Sullyman626 posted:

Amazing an almost 80 year old “toy” performs so well.

I am fairly new to this hobby. I had a Lionel train set in the early 1960s, which I kept all these years in very good condition, and in the original box. But back in October I was cleaning out my Father-in-laws basement and found several large boxes covered with probably sixty years worth of sawdust - and in them I found a treasure trove of old Lionel trains, tracks, transformers and all sorts of accessories. So I started to get interested in this - to see if I could get them running. I found so many helpful videos on Youtube, and found so much helpful maintenance and repair info on this forum. To my amazement I also found that parts are still readily available for these old trains. So, I started doing some more searching amongst my storage boxes for a locomotive that was given to me by my Grandfather back in the early 1960s. I found it - a diecast locomotive with the number 999 on it. I thought it was a Lionel, but a web search soon revealed that is was Marx locomotive. It never ran - even back in 1964 when I got it - but I never threw it away because it belonged to my Father when he was a boy. I found some on-line repair videos, and ordered some new parts - a replacement gear, and some motor brushes, and soon got it working again. It is one of the very early Marx 999s with the open slots on the cowcatcher. I believe it is a late 1930s vintage locomotive. It certainly is amazing that these old trains are so durable - especially the Marx 999 - which seems like it is virtually indestructible. The 999 is easy to fix, and they are so ubiquitous. Marx must have sold many of them.

Magic of the horizontal drive.

@Ted S posted:

Thanks so much for doing this Mike.  Your 2333's performance absolutely ROCKS!   The 9:1 gear ratio over small 0.88" drivers means there's plenty of RPM and torque, even at slow speeds.  It's also impressive that those old Pullmor motors lug down below 600 RPM without stalling.  I've tested a lot of Pittmans in O-gauge locos that won't do that!

Unlike many modern dual-motored locomotives, the gears in your 2333 are back-driveable.  This means, if one motor begins to stall, the second motor can "push" the whole loco down the track a little, which serves to jog the first motor until its next armature pole is attracted by magnetic flux from its field.  It's almost like having a six-pole motor, so you are benefitting from about 20 usable power pulses per inch of travel.

If you used the full voltage of your ZW, I'm pretty sure you could have achieved 80 mph even with a trailing consist.

If you happen to test an MTH F3 made in the late 1990s with two vertical can motors and four rubber tires, I'm willing to bet that it won't run this well.  I was never impressed with the slow-speed performance of mine on sharp tubular curves.  Even if it didn't stall, there was a noticeable slow-down in the curves.  Meanwhile, your 2333 exhibits minimal speed variation.  So much for progress!  Another print magazine that's focused on 3-rail O used to publish speed tables like yours in their product reviews.  I can tell you, in the era before speed control, very few of the locos they tested were able to sustain 5 mph.

To me, your post is proof that Lionel's F3 "growlers" were king of the 3-rail mountain for over 50 years, and in many ways, still are!

Ted

I noticed that the 2333's worm gear drive is reversible, i.e. not self-locking. So that means rotating the worm gear can cause the worm to rotate. You called this back-driveable. That is an interesting characteristic of the 2333 - and I wonder if backdriveability would improve the low speed performance of the MTH Santa Fe F3 with the DC permanent magnet can motors. I mentioned in my review of the MTH Santa Fe F3, that the worm gear drive was irreversible and self locking. In fact, most worm gear drives are irreversible and self-locking. The worm gear drive on the Lionel 2333 has a very high lead angle of 45 degrees. This is what makes it reversible.

Mike

Hot rollers.  Got me thinking about connecting the rollers on my dummy unit to the powered unit with an umbilical cord.  Cuts the current load in half.

Perhaps not for the faint of heart since gears and bushings have to be knocked out, I have added ball bearings to the horizontal motors. Ball bearings greatly reduce internal friction of the motor and allows the armature to spin much more freely over that of the oilite bushings. Furthermore, the motors are much more quiet.

For the 2343 motors 3/16 id x 5/16 od x 1/8 is at the idler gear end of cast metal housing. The ball bearing is a perfect fit.

At the armature end a 7 mm id x 11 mm od x 3 mm ball bearing is used.

An 11 mm od bearing is a kissing fit for the motor housing.

A 3/16 inch nylon bushing is inserted into the ball bearing.

Two bearings are used due to the mass of the armature.

The ball bearings are secured with some cyanoacrylate in the motor housing. Here is the motor assembled and in operation.

And here is the 2343 assembled and in operation with its ball bearing drive.

The 2343 locomotive is much more quiet than it was before and has smoother and slower starts. There is still some noise from the idler gears as they make a clicking noise when they engage each other.

@WBC posted:

The 2343 locomotive is much more quiet than it was before and has smoother and slower starts. There is still some noise from the idler gears as they make a clicking noise when they engage each other.

Other than noise reduction and smoother starts, has there been a noticeable improvement in performance?

@Mossback Mike posted:

Other than noise reduction and smoother starts, has there been a noticeable improvement in performance?

The motor is still essentially a 1930's design and as such is still limited. The ball bearings bring the motor a bit more to 21st century standards but the motor still has the small motor brushes relative to the armature size, flat faced commutator, three rotor poles and two stator poles. These mechanical specifications do not change. My units drew 4.25 amps before the conversion so I was quite proud when the amps reduced to 2.0 to 2.5 (it bounces around as the unit rolls down the track) but that seems to be in the same neighborhood that you posted. The TMCC electronics do not get nearly as hot as before the ball bearing installation. Traction doesn't change and would not be expected to.  As can be seen, the 2343 locomotive crawls at quite low speeds; close to that of can motor modern production. So I should add that it sustains lower speed operation than before.

Another difference is that I put pick up rollers on the rear truck so there are a total of 4 pick up rollers which greatly helps low speed operation.

I'm sure that bigger motor brushes with greater surface contact area would greatly reduce amperage draw and lead to smoother operation. A drum commutator would also greatly improve performance by reducing friction and improving brush contact. More stator and rotor poles would also be great, like the universal motors that can be seen in a high quality electric drill or hair dryer but now we are talking a completely different, while still universal, motor.

Here are a couple more videos of the 2343 pulling some passenger cars.

And now with a railsounds B unit.

@WBC posted:

The motor is still essentially a 1930's design and as such is still limited. The ball bearings bring the motor a bit more to 21st century standards but the motor still has the small motor brushes relative to the armature size, flat faced commutator, three rotor poles and two stator poles. These mechanical specifications do not change. My units drew 4.25 amps before the conversion so I was quite proud when the amps reduced to 2.0 to 2.5 (it bounces around as the unit rolls down the track) but that seems to be in the same neighborhood that you posted. The TMCC electronics do not get nearly as hot as before the ball bearing installation. Traction doesn't change and would not be expected to.  As can be seen, the 2343 locomotive crawls at quite low speeds; close to that of can motor modern production. So I should add that it sustains lower speed operation than before.

Another difference is that I put pick up rollers on the rear truck so there are a total of 4 pick up rollers which greatly helps low speed operation.

I'm sure that bigger motor brushes with greater surface contact area would greatly reduce amperage draw and lead to smoother operation. A drum commutator would also greatly improve performance by reducing friction and improving brush contact. More stator and rotor poles would also be great, like the universal motors that can be seen in a high quality electric drill or hair dryer but now we are talking a completely different, while still universal, motor.

WBC

Your work on this has got me thinking - maybe modern model railroad technology relies to heavily on electronics to solve basic mechanical power transmission problems. Now if model locomotives were built to the same standards and precision of an inertial guidance system, I suppose the cost would be prohibitive. But, there has to be some reasonable break point where better use of miniature ball bearings and better gearing would provide affordable improvements.

Mike

@Mossback Mike posted:

I wonder if backdriveability would improve the low speed performance of the MTH Santa Fe F3 with the DC permanent magnet can motors. I mentioned in my review of the MTH Santa Fe F3, that the worm gear drive was irreversible and self locking.

Mike I have no doubt that back-drivability would improve the low-speed performance, as then the motors could actually help each other.  Lionel used this to good effect with their Legacy Liondrive diesels introduced circa 2007.  In my experience, these models have smoother starts and better performance in the 1 to 5 smph range, compared to dual-motored diesels from MTH, Atlas, K-Line etc.  Of course most newer locos have some type of "speed control."  Which makes the performance of the 2333--stock or modified--that much more impressive!

@WBC the performance of your upgraded diesel is AMAZING!  I don't have the guts to try the ball-bearing modification myself.  If I get krazy glue anywhere near my motors, that's not going to end well!  However I wish that you or someone would sell these upgraded motors at York, or offer a mail-order service, perhaps with a core exchange so that you would have a sustainable supply of old motors to upgrade.

Mike, I agree wholeheartedly with the comment in your post immediately above: "modern model railroad technology relies to heavily on electronics to solve basic mechanical power transmission problems."  Note that in the 2nd post of this thread, I predicted that a '90s MTH F3 woudn't measure up to the performance of the 1948 model.  I posted that back in MARCH--nine months before you actually bought one and tested it.  How did I know?  Because I spent (wasted?) a lot my own of money on "new and improved" trains during the '90s, and frankly, came away a little disappointed.  IMO real improvements in performance didn't come until 2000 with speed control.  Which, IMO, is a band-aid solution broadly applied to what your testing clearly shows to be an inferior design.  Hubble bubble!  Furthermore, time has shown that first generation of speed control introduced failure modes (i.e., swollen capacitors on the 5-volt Proto 2 boards, the Odyssey "lurch," exploding magnets, etc.), leaving us to spend even more money on repairs and upgrades.  There was an old lady who swallowed a fly... we all know how that story ends!

If the "scale revolution" of the 1990s hadn't happened, more folks like WBC might have devised and marketed unequivocal improvements to the enduring, mass-produced trains of our childhood.  Ball-bearing upgrades.  Redesigned brushes and brushplates.  Perhaps even a 5-pole armature.  It's all still possible. Presently, postwar trains are being overlooked to the point of irrelevance because they're not "scale," and the perception (in some cases deserved) is that they run like toys.  But they're far more affordable and serviceable than what's being made today.  AFAIC, this thread proves beyond any doubt that the best postwar locos--potentially with modest upgrades--still have a place on operating layouts instead of the display shelf!

@Ted S posted:

@WBC the performance of your upgraded diesel is AMAZING!  I don't have the guts to try the ball-bearing modification myself.  If I get krazy glue anywhere near my motors, that's not going to end well!  However I wish that you or someone would sell these upgraded motors at York, or offer a mail-order service, perhaps with a core exchange so that you would have a sustainable supply of old motors to upgrade.

If the "scale revolution" of the 1990s hadn't happened, more folks like WBC might have devised and marketed unequivocal improvements to the enduring, mass-produced trains of our childhood.  Ball-bearing upgrades.  Redesigned brushes and brushplates.  Perhaps even a 5-pole armature.  It's all still possible. Presently, postwar trains are being overlooked to the point of irrelevance because they're not "scale," and the perception (in some cases deserved) is that they run like toys.  But they're far more affordable and serviceable than what's being made today.  AFAIC, this thread proves beyond any doubt that the best postwar locos--potentially with modest upgrades--still have a place on operating layouts instead of the display shelf!

@Mossback Mike posted:

WBC

Your work on this has got me thinking - maybe modern model railroad technology relies to heavily on electronics to solve basic mechanical power transmission problems. Now if model locomotives were built to the same standards and precision of an inertial guidance system, I suppose the cost would be prohibitive. But, there has to be some reasonable break point where better use of miniature ball bearings and better gearing would provide affordable improvements.

Mike

Gearing plays a big part in motor performance. The performance of the horizontal motors is in large part due to the 9.1:1 gear ratio (means that it takes 9.1 revolutions of the armature to turn the wheels once). The 2028 and 2321 vertical systems are roughly 8:1. The other is the better designed stator. Less cogging takes place when the armature rotates more to turn the wheels once.  Guessing here, at a gear ratio of 12:1 cogging would be largely eliminated even with a three pole armature/2-pole stator.

A lot of technologies are available today (such as ball bearings) that were simply not available in the 1930's, 40's, 50's and even up to the 70's.  I am not sure if a 5-pole armature would make much of a difference with the 2-pole stator. Stator pole design and winding has advanced considerably over the last 30-40 years with designs such as this https://www.mytec.de/en/increasing-the-efficiency/. Something like a 4 or 6-pole field winding with a 5-pole armature would make a huge difference and is technically feasible today; even for motors the size of our "Pullmor". Keep in mind that modern universal motor design is more efficient than DC can motors.  DC can motors are typically around 65% efficient while universal motors are typically 75% efficient.

Thanks for the complement on the 2343 operation. I am not sure if something like this is economically feasible to sell at York without going into actual motor production. 2343, 2028, 622, 681 and such motors run about $50. The ball bearings are about $5 each at the small volumes I buy. In bulk, they would be 25-50 cents. There is the labor and and other little components that need to be added in.  Each motor will probably be in the neighborhood of $100 just to break even once travel costs and booth are factored in. I am not sure if many would be willing to pay that price.

Also, I believe that the motors that Lionel Corp produced were actually manufactured in Italy. So there is a curious mix of metric and imperial. For instance, a 3/16 inch bushing in an 11-mm cavity. That makes fitting and adapting the ball bearings a bit difficult as the ball bearings are either imperial or metric but not both. Thus, there is still an element of kit bashing involved as I search for the perfect combination. However, the ball bearing system that was used for the 2343 type motors did work well. The motors will the brushplate bearing are a bit more of a challenge.

Yes, gearing plays a HUGE role, and as Mike's tests show us, the wheel diameter is an implicit part of the gearing.  The 773 Hudsons with a ratio of 18:1 develop slightly more RPM per inch than horizontal-motored F3s with smaller wheels.  Of course the Hudsons have only one motor.  The F3s have back-drivable gears, so the two motors can "help" each other.  Thus, if you count total power pulses per inch, you could logically add both motors together, and the growlers come out as tops among all Postwar power.  Later diesels with two vertical motors (FMs and later F3s) are almost as good.

There are a few posts on this Forum and in print discussing postwar Lionels that have been geared down.  As you surmised, their performance was spectacular!  Certainly better than most mass-produced trains made in the '90s, and nearly as good as today's stuff with speed control.

To me personally, $100 seems like a worthwhile investment to wring that last drop of performance out of a postwar loco.  Your mention of metric, imperial, etc., shows that you have unique knowledge of the measurements and the process.  You know what parts to buy, which adhesives to use and where to apply them, etc.  You actually wouldn't have to buy the motors, because in my proposed business model folks would send you their motors (or at least send you back the cores for a substantial rebate.)  If special tools aren't required, maybe a bunch of us could pay you in advance for the parts.  Then we could all get together around a table at York and you could lead a clinic on this mod!

[Edit: The only Pullmor-motored locos geared at 12:1 were the die-cast FM switchers of the late '90s.  And I believe the extra gear reduction was accomplished by using a two-lead worm on the motor shaft, instead of the high-angle three-lead worm that dates all the way back to 1947.  So ironically (in light of Mike's questions above), despite having a lower gear ratio, a '90s FM switcher might not fare as well in testing as the F3 because its gears aren't back-drivable!  Excellent topic!! ]

Nice analysis.  You have a very good 2333.  Mine can get only to about 35smph and it draws a lot more current.   Now I have a benchmark to work toward when I refurbish it.

@Ted S posted:

[Edit: The only Pullmor-motored locos geared at 12:1 were the die-cast FM switchers of the late '90s.  And I believe the extra gear reduction was accomplished by using a two-lead worm on the motor shaft, instead of the high-angle three-lead worm that dates all the way back to 1947.  So ironically (in light of Mike's questions above), despite having a lower gear ratio, a '90s FM switcher might not fare as well in testing as the F3 because its gears aren't back-drivable!  Excellent topic!! ]

H12-44 is a great little locomotive. Cast metal and very heavy. The locomotive would crawl along at quite low speeds. The combination of low gearing, magnetraction and tires meant that it would pull 35 scale cars easily. I did that which was not smart.  The derlin worm wheel stripped and I could not replace it.  Since I had the parts I rebuilt the locomotive with a 623 motor and appropriate worm wheel so it is essentially like a 623 switcher.

@WBC posted:

H12-44 is a great little locomotive. Cast metal and very heavy. The locomotive would crawl along at quite low speeds. The combination of low gearing, magnetraction and tires meant that it would pull 35 scale cars easily. I did that which was not smart.  The derlin worm wheel stripped and I could not replace it.  Since I had the parts I rebuilt the locomotive with a 623 motor and appropriate worm wheel so it is essentially like a 623 switcher.

Years ago, I used to design machinery - and it was axiomatic that you generally wanted to design a gear train with a sacrificial gear in a location that made it easy to replace. So while most of the gears might be hardened steel, the sacrificial gear might be brass or a milder steel. The sacrificial gear would fail before any other gears. I am surprised that a premium locomotive like your H12-44 would use delrin gears. You were unable to replace the failed worm wheel?

@TrainCzar posted:

Nice analysis.  You have a very good 2333.  Mine can get only to about 35smph and it draws a lot more current.   Now I have a benchmark to work toward when I refurbish it.

When I found my 2333, it was in a box, buried deep under a bench in the basement woodworking shop. It was covered with at least 6 inches of sawdust, and hadn't been run in over 50 years. I was amazed at how easy it was to bring it back to life.

@Mossback Mike posted:

When I found my 2333, it was in a box, buried deep under a bench in the basement woodworking shop. It was covered with at least 6 inches of sawdust, and hadn't been run in over 50 years. I was amazed at how easy it was to bring it back to life.

Mossback,
this is terrific. did the bearings fit without modification? I have been tinkering with my pullmors, specifically with the hope of improving the way they run/roll.  I have a pair of NYC and SantaFe 2343's. They were stored for about 35 years and I have fully disassembled and rebuilt the SantaFe and the NYC is next. Mine do NOT roll/run as smooth or slow as yours do in the video. I'd like to find out more about the bearing conversion you have there

@Mossback Mike posted:

You were unable to replace the failed worm wheel?

Lionel built the H12-44 only once in Santa Fe, Chicago and Northwestern, and Pennsylvania. Parts are no longer available for it.

@WBC posted:

Gearing plays a big part in motor performance. The performance of the horizontal motors is in large part due to the 9.1:1 gear ratio (means that it takes 9.1 revolutions of the armature to turn the wheels once). The 2028 and 2321 vertical systems are roughly 8:1. The other is the better designed stator. Less cogging takes place when the armature rotates more to turn the wheels once.  Guessing here, at a gear ratio of 12:1 cogging would be largely eliminated even with a three pole armature/2-pole stator.

A lot of technologies are available today (such as ball bearings) that were simply not available in the 1930's, 40's, 50's and even up to the 70's.  I am not sure if a 5-pole armature would make much of a difference with the 2-pole stator. Stator pole design and winding has advanced considerably over the last 30-40 years with designs such as this https://www.mytec.de/en/increasing-the-efficiency/. Something like a 4 or 6-pole field winding with a 5-pole armature would make a huge difference and is technically feasible today; even for motors the size of our "Pullmor". Keep in mind that modern universal motor design is more efficient than DC can motors.  DC can motors are typically around 65% efficient while universal motors are typically 75% efficient.

Thanks for the complement on the 2343 operation. I am not sure if something like this is economically feasible to sell at York without going into actual motor production. 2343, 2028, 622, 681 and such motors run about $50. The ball bearings are about $5 each at the small volumes I buy. In bulk, they would be 25-50 cents. There is the labor and and other little components that need to be added in.  Each motor will probably be in the neighborhood of $100 just to break even once travel costs and booth are factored in. I am not sure if many would be willing to pay that price.

Also, I believe that the motors that Lionel Corp produced were actually manufactured in Italy. So there is a curious mix of metric and imperial. For instance, a 3/16 inch bushing in an 11-mm cavity. That makes fitting and adapting the ball bearings a bit difficult as the ball bearings are either imperial or metric but not both. Thus, there is still an element of kit bashing involved as I search for the perfect combination. However, the ball bearing system that was used for the 2343 type motors did work well. The motors will the brushplate bearing are a bit more of a challenge.passenger train of aluminum

AMT/Auburn and then KMT/Kusan produced F units in the 50s with a 7 pole Pittman motor - in combination with a traction tire they can really pull (the Pittman motors are mounted vertically like the later Lionel F unit drives). They have hefty square brushes, much larger than the brushes in a Lionel F. Dual motor a pair and you can lug a good, long string of AMT aluminum passenger cars. Pulling power is great, coasting ability is nil - a heavy hand on the throttle when attempting to stop a train will result in a derailment - some ball bearing action would probably make a big difference. Here's a link to a thread where I resuscitated an AMT F unit that needed some help - you can see the Pittman motors in some of the shots:

https://ogrforum.com/...nit-flea-market-find

The AMT / KMT F-units sound interesting.  The fact that they don't coast implies that they probably use self-locking gears.  Perhaps a 2-start worm instead of the 3- or 4-start back-drivable worms used by Lionel.  A worm with fewer starts translates to a lower (numerically higher) gear ratio and slower speeds.

@MTN would you please be kind enough to twirl one of the motors in your AMT with your finger, and calculate the gear ratio for us?  The easiest way to do this is to put a mark on one pole of the armature, and another on the wheel tread.  Ink marker is easy to see and can be wiped off with rubbing alcohol.  Look forward to reading about what you find, and how it compares to Mike's specs on the 2333!

@Ted S posted:

The AMT / KMT F-units sound interesting.  The fact that they don't coast implies that they probably use self-locking gears.  Perhaps a 2-start worm instead of the 3- or 4-start back-drivable worms used by Lionel.  A worm with fewer starts translates to a lower (numerically higher) gear ratio.

@MTN would you please be kind enough to twirl one of the motors in your AMT with your finger, and calculate the gear ratio for us?  The easiest way to do this is to put a pencil mark on one pole of the armature, and another on the wheel tread.  Ink marker is easier to see and can be easily wiped off with rubbing alcohol.  Look forward to reading about what you find, and how it compares to Mike's specs on the 2333!

I'll have to dig one of my AMT F units out - no layout up and all trains are buried in tubs. The AMT/Auburn/KMT/Kusan F unit is not a speed demon - it was geared to lug. The Pittman motor has a worm gear cut into the armature shaft - I found a loose motor and took some snaps which show the armature and another showing the field and one of its brushes:

I used WBC tip and just finished adding another set of roller pickups to my 2333.  No more sparking and seems to run smoother.  Thanks WBC.

@CSXJOE posted:

I used WBC tip and just finished adding another set of roller pickups to my 2333.  No more sparking and seems to run smoother.  Thanks WBC.

I am happy that the extra pair of rollers worked out. That collector assembly is getting difficult to find.

@MTN posted:

I'll have to dig one of my AMT F units out - no layout up and all trains are buried in tubs. The AMT/Auburn/KMT/Kusan F unit is not a speed demon - it was geared to lug. The Pittman motor has a worm gear cut into the armature shaft - I found a loose motor and took some snaps which show the armature and another showing the field and one of its brushes:

It is difficult to say by just looking at the photos, but to me it appears the stator does not engage the rotor well.  I think the stator found on the Lionel horizontal or 622 series of motors is better designed.  The design of the stator is just as important as the design of the rotor.

WBC, I had one in my junk box.