Rewinding a Wound Field Motor for More Torque

[PatB]

I have an old Marx 3-rail loco with a nicely cooked wound-field universal motor. The armature windings are definitely shot, and the field windings may be. 

 

As I need to rewind it anyway, I would like, if possible, to improve its torque output, in particular its starting torque. I don't need its factory supplied ability to hit 15 squillion rpm at 9V. 

 

I will be running this on DC, with the field coil connected via a bridge rectifier to allow simple reversing. 

 

If I remember my rather rudimentary motor theory correctly, a greater number of turns of thinner wire on the armature poles of a permanent magnet motor gives greater torque but lower speed. Is this correct? 

 

In this case, being a universal motor, should I also adjust the field winding to suit? If so, would it require more or fewer turns? More would seem to make sense. 

 

I realise that this is going to be a bit of an experiment, and I've never rewound a motor before. However, the Marx mech is so simple and chunky and of such low value it seems an ideal candidate to learn on. 

Edited April 30, 2021 by PatB

[Nearholmer]

Magneto motive force is a function of the current and number of turns, so if you want more torque, you need more At.

 

By reducing wire diameter, you will be able to fit-in more turns, but you may find that the smaller cross-section reduces the current that can be carried without excessive heating, so it might become self-defeating. Also, the iron core of the windings won’t have a linear characteristic, so you may find that increasing At yields less benefit than you expect - if it’s operating near saturation now, it will yield barely any.

 

You could tackle either armature or “field” winding, or both. Probably easiest to start with the “field”, because sometimes the armature winding is in quite a complex configuration (maybe not in a cheap motor, though).

 

It might have to be a case of trial and a few errors to get the characteristic you want.

 

Another thought is that old, cheap motors often had very poor magnetic circuits, with huge air gaps, so a large proportion of whatever flux you make might go straying-off all over the place anyway!

 

Another, further, thought: doesn’t using a rectifier to “dc” the motor reduce the torque? If I’m right about that, why not just fit a manual reversing switch?

[PatB]

2 hours ago, Nearholmer said:

Magneto motive force is a function of the current and number of turns, so if you want more torque, you need more At.

 

By reducing wire diameter, you will be able to fit-in more turns, but you may find that the smaller cross-section reduces the current that can be carried without excessive heating, so it might become self-defeating. Also, the iron core of the windings won’t have a linear characteristic, so you may find that increasing At yields less benefit than you expect - if it’s operating near saturation now, it will yield barely any.

 

You could tackle either armature or “field” winding, or both. Probably easiest to start with the “field”, because sometimes the armature winding is in quite a complex configuration (maybe not in a cheap motor, though).

 

It might have to be a case of trial and a few errors to get the characteristic you want.

 

Another thought is that old, cheap motors often had very poor magnetic circuits, with huge air gaps, so a large proportion of whatever flux you make might go straying-off all over the place anyway!

 

Another, further, thought: doesn’t using a rectifier to “dc” the motor reduce the torque? If I’m right about that, why not just fit a manual reversing switch?

Thanks for the reply Kevin. As with all things, I suspect there is no perfect solution. Fortunately the Marx motor is both very simple and large enough for me to actually see what I'm doing, so having to have a couple of goes isn't a disaster. 

 

I'll probably do the armature first as it can be removed and replaced without doing more than taking the brush plate off. Proper access to the field may require the side plates separating (it's basically a pancake motor), which will require me to sort out a wheel puller of some kind. Again, not particularly hard, but an additional step. Of course, if the field's done medium rare as well, I'll need to do that anyway. 

 

I'm not sure how good the fundamental motor design and construction is. Although old King Louis was quite fanatical about cutting costs, this seems to have been achieved by leaving stuff off (reversers, headlamps, smoke units, front couplers (or rears on cabeese), con rods, detail etc). What remained seems to have been quite well engineered. That said, it's still a toy, designed in the mid-1930s, when the design of small motors was still not a fully mature technology, even if the designer cared about efficiency. 

 

You may be right about the effect of a rectifier. If nothing else, it's going to result in some diode drop (2x I think), which won't help, and which I hadn't thought of. A switch would help with that, but my ultimate goal is reliable reversing without Hand of God intervention, as far as possible. This may, of course, be an impossibility with 80 yo toys, but I find the challenge interesting. 

 

I also suspect that I may not be helping my locos' starting by feeding them nice, clean DC from a laptop brick. Maybe I need to slap a rectifier onto a big AC transformer, and use the output without further smoothing, to nudge the armature round. Shame though, as they're so lovely and smooth on DC. As an experiment, I may (very carefully) try my working engines on an original Marx variable transformer controller, fed via a suitable step down adapter (which I already have for another purpose) to drop Oz mains voltage down to the appropriate 110V, to see if they have more grunt when fed as originally intended. 

[Nearholmer]

I’ve heard from many people that these old universal motors perform better on ac than dc, although I have no personal experience of comparing, and am at something of a loss to understand why that might be (not that I’ve actually sat down to think about it properly).

[PatB]

17 hours ago, Nearholmer said:

I’ve heard from many people that these old universal motors perform better on ac than dc, although I have no personal experience of comparing, and am at something of a loss to understand why that might be (not that I’ve actually sat down to think about it properly).

I suspect it's less about the juice being AC as such, and more about its unsmoothed nature acting as a low(ish) frequency pulsed supply. Hence my thoughts on trying a "rougher" power source. I also seem to remember that a rising or falling current in a coil creates a stronger magnetic field than a steady one, so that may also be a factor.

 

As I said, my locos seem much smoother and quieter on DC than what I've seen on YouTube running on AC, but I've not had means of directly comparing pulling power and starting reliability between the two. Now that I've got a couple of beefy transformers (as well as the original Marx Death-o-Matic) I can have a bit of a play with alternative means of feed. I'm a bit set on DC though, for the simple reversing. If I can make it work it's simple anyway. 

[Nearholmer]

4 hours ago, PatB said:

I also seem to remember that a rising or falling current in a coil creates a stronger magnetic field than a steady one, so that may also be a factor.

Are you sure you’re not getting mixed-up with “rate of change of flux” and transformer-effect, which is how an induction motor works?

[PatB]

That would probably be it, but I've tended to assume it to also be part of the reason why solenoids seem to snap over more vigorously on AC. However, beyond a fairly basic level, AC electrickery always made my eyes cross when I encountered it during my engineering studies and so, along with the more esoteric thermo stuff, the details always went out of my head pretty much instantly on completion of any necessary exam. 

[AndyID]

As Kevin says it's the ampere-turns that produce the torque for a given magnetic circuit and that's a function of the cross sectional area of the winding regardless of the wire gauge. Changing the wire gauge lets you alter the motor's operating voltage but not the torque produced.

[PatB]

Thanks Andy. That was what I was starting to suspect.

 

However, I've just spent a few minutes digging out my biggest AC power supply (5A at 17V max output), and connecting it up via a somewhat beefed up (and modified to use locally available components) Roger Amos controller, with full wave rectification but no smoothing capacitor. The result is that an uncooked standard Marx loco appears to be starting 100% consistently. It's running a bit more noisily than when fed smooth DC, but the wiggly waveform seems to be doing the trick as far as overcoming "sticktion" is concerned. Well, well; who'd have thunk it?

 

Bottom line is that it looks as if I'll be rewinding the dead motor as standard, or at least with standard wire thicknesses (~0.3 mm on the armature, ~0.6 mm on the field). There might be space to get a few more turns in, but I'm now confident that a standard motor, suitably fed, will do the job, so it's not a disaster if things don't fit.

 

Now I just need to package up the controller circuit a bit more neatly than its current breadboard arrangement.

[Nearholmer]

Does this mean that you've tested AC vs Smooth DC vs Lumpy DC, and arrived at an "order of preference" for good operation?

[PatB]

1 hour ago, Nearholmer said:

Does this mean that you've tested AC vs Smooth DC vs Lumpy DC, and arrived at an "order of preference" for good operation?

Just smooth and lumpy DC. Although I've got a couple of AC power supplies (the one I've just tried and another that might be a tad marginal, current wise), I haven't, until getting the Marx deathtrap a couple of weeks ago as part of a set, had a means of speed control. Consequently, I've been running things on DC, either simple testing with unmodified locos using their electromechanical reversers, or actual operation with a bridge rectifier feeding the field to allow simple DC reversing. 

 

Now that I've got a Marx transformer, and a suitable 220-110V step down unit, I'll see about getting around to trying a couple of unmodified locos as Louis intended and seeing how they go. One of the locos is a (relatively, in a Marx context) uncommon tin diesel which is a bit too nice for even a philistine like me to attack with wire cutters and soldering iron. 

[PatB]

I've noticed one odd thing. One of my DC'd locos has what Marx aficionados call the "small" or "one-way" motor. This is an el-cheapo variant that was used, primarily, in later bottom of the range units. It's got a shorter pickup shoe, paxolin rather than steel side plates, a consequently slightly different electrical return path, and no separate brush plate, making it much harder to dismantle. The armature and field look to be the same as the full size motor, though as the armature won't come out for inspection without separating the side plates it's hard to be sure. Anyhow, this motor variant, on smoothed DC, was staggeringly fast and smooth. Running light, it would become airborne on 027 curves on 9V, whereas the traditional large motor locos needed another 3-6V to achieve the same feat. However, on unsmoothed DC it is a decidedly reluctant runner. Although it seems to start reliably it's slow, noisy and generally unhappy. It intrigues me that there can be so much variation in motors which, on the surface at least, are electrically similar. 

[Nearholmer]

Laminations? Commutator angle of offset from windings? Air gaps?

[PatB]

On 06/05/2021 at 15:06, Nearholmer said:

Laminations? Commutator angle of offset from windings? Air gaps?

Conceivably any of those. Or maybe just the vagaries of ancient, cheaply constructed toys.

 

Bit of an update. I did a fairly quick and dirty armature rewind this afternoon. Standard is 60 turns of ~0.3 mm wire per pole. Without trying very hard I got 70 turns on and, with more attention to layering them neatly, I reckon I could get 80 turns on fairly comfortably. As a test of my workmanship I stuck it in an uncooked motor and it's now trundling up and down my auto-shuttle track quite happily. It seems slower at a given controller setting than an unmodified mech. I'm not sure about torque output. I haven't the room at the moment to really load it up and test rewound against standard. TBH, I'm just ridiculously chuffed that it works. I can now play around knowing that, if I'm daft enough to cook another one, a couple of bucks worth of wire and an hour or two will see it as good as new again.

[AndyID]

Good for you. Adding more turns ought to improve the torque a bit but you might need to increase the field magnetic flux too.

 

I'm guessing it's connected as a shunt motor where the coil is across the armature. You might try messing about with it in series like a full size traction motor. One potential snag with that is the speed might go out of control of there's no load on the motor 

 

[Nearholmer]

It’s probably series-connected already.

 

By Googling, I’ve yet again accidentally got deep into HRCA’s ‘members only’ website (they really need to sort security!).

 

If this link works for you, you will find it mega-interesting.

 

https://www.hrca.net/wiki/0-Gauge_Maintenance_Manual#ELECTRIC_MOTORS

 

[PatB]

10 hours ago, AndyID said:

Good for you. Adding more turns ought to improve the torque a bit but you might need to increase the field magnetic flux too.

 

I'm guessing it's connected as a shunt motor where the coil is across the armature. You might try messing about with it in series like a full size traction motor. One potential snag with that is the speed might go out of control of there's no load on the motor 

 

I'll be rewinding the field on the cooked motor as soon as the wheel puller I've ordered turns up so I can get the mech apart safely. I could probably get the wheels off by levering, but I don't need to so I've decided not to risk it. I could probably rewind it in situ, but it would be very awkward, involving feeding metres of wire through a narrow, sharp edged slot, guaranteed to scrape off the enamel in places. Looking at it, there's bags of space to increase the windings proportionate to those on the armature. 

 

Of course, the question arises that, if it's likely to be beneficial, why didn't Marx do it? I'm going to choose to believe that it was down to cost control. When you're building many thousands of motors, a couple of feet extra wire in each will add up to something substantial enough to be noticed by the bean counters. And if your locos have enough oomph to pull the 3 or 4 cars in the average set at 160 mph, why make them capable of more? 

 

I'm not too worried about the theoretical possibility of the motor overspeeding. I suspect there's enough friction in the mech that there will always be at least some load on the motor. Besides, even if it did run away to the point of bursting the armature windings or making the bearings glow, I can't see the consequences as being too dire. It's too small to explode dangerously, and it won't be unattended so it's unlikely to burn the house down.

 

9 hours ago, Nearholmer said:

It’s probably series-connected already.

 

By Googling, I’ve yet again accidentally got deep into HRCA’s ‘members only’ website (they really need to sort security!).

 

If this link works for you, you will find it mega-interesting.

 

https://www.hrca.net/wiki/0-Gauge_Maintenance_Manual#ELECTRIC_MOTORS

 

 

That's absolutely fascinating. Many thanks. I've only skimmed the material so far, but will give it my full attention once I've done all the stuff that was less attractive than hand rewinding an armature yesterday . It will be very interesting to compare the differing approaches used on either side of the Atlantic. 

Edited May 9, 2021 by PatB

[PatB]

Another little bit of progress in my quest for a SuperMarx motor. 

 

I split the side plates of the cooked motor this afternoon and had a go at rewinding the field. Standard is 125 turns of 0.6 mm wire. I got 150 on before boredom and finger fatigue made me stop. I reassembled it with the previously rewound armature and tested it on the track. It sort of ran, but made loud clacking noises and repeatedly locked up. Hmmm. 

 

It took me a while to work out what was happening, but eventually I noticed the visible end float on the armature shaft. It appears that the increased magnetic field is dragging the armature sideways hard enough to overcome the, frankly rather hopeful, thrust control provided by the pressure of the brush springs, pulling the spur gear out of mesh and jamming the mech. 

 

OK. Better thrust control needed. A couple of brass washers, and an insulating shim cut from a plastic bottle, saw the end float reduced to something sensible. 

 

Back on the track and, once I'd remembered to put the brushes back in, it chuntered up and down quite happily and fairly reliably, with only the occasional refusal to proceed. Things are improving. 

[kevinlms]

On 06/05/2021 at 17:06, Nearholmer said:

Laminations? Commutator angle of offset from windings? Air gaps?

Perhaps Pat will swap out the laminations for lamingtons? They taste better.

[Nearholmer]

Well, I had to look that up, but yes, they do look tastier than bits of thin iron plate.

[AndyID]

8 hours ago, Nearholmer said:

Well, I had to look that up, but yes, they do look tastier than bits of thin iron plate.

 

Similar but different -

https://shop.christiecakes.co.uk/products/raspberry-jam-eiffel-towers

[Nearholmer]

I’m not wildly keen on coconut in cakes, so here’s a picture of some flapjack, which I am very partial to.

 

 

Does the title of the thread need to be expanded?

[AndyID]

45 minutes ago, Nearholmer said:

 

Does the title of the thread need to be expanded?

 

Probably not. I think rewinding covers it

[PatB]

I'm afraid that, although Oz citizenship rendered me partial to a cheese sausage, it didn't take as far as lamingtons are concerned, and I'd rather have a nice battenburg. 

 

Going back to the subject of ancient motors, after cleaning everything again, and reducing the duty cycle on the auto shuttle, I got 5 hours out of my modified mech before it finally "failed to proceed" without a poke, which was quite pleasing. It seems happy to work for a lengthy period, and then becomes more reluctant. I'd put it down to overheating but it doesn't feel excessively warm to the touch. I've just received some new brushes and springs, so I'll try swapping those. I can't really see anything wrong with the old ones, but it's always possible the springs have lost their temper at some point, and aren't providing enough pressure. These motors are so basic, there's really not a lot else to fiddle with. 

 

Onwards in small steps. 

 

Edit: I've just put in the new brushes and springs. The springs are somewhat different from the existing ones. The originals are (I think) phosphor bronze, based on the colour. The new ones are something silvery (presumably either steel or a coating of something), and are appreciably stronger than the ones I took out, clicking into place quite emphatically. Testing on the track shows noticeably more reliable starting. Previously, when changing direction, the first few inches of travel were a bit tentative. I now assume this was due to the slightly slack brushes taking a few motor rotations to settle into place. This seems to be no longer happening. Now it's properly warmed up it seems to be faster for a given controller setting too. Let's see how sustained the improvement is.

Edited May 29, 2021 by PatB