Rewinding armatures - questions

[mmagliaro]

Related to my ongoing project of upgrading old Rivarossi can motors with a new armature and magnets (yes, I am still developing this), I have some good-running examples, but I'm scientifically intrigued by rewinding one of my armatures (the replacement upgrade one, not the original Rivarossi) to see if I can make it better.

There have been discussions on armature windings before in Railwire and years ago in the Atlas Forum.

What started me down this path is noticing that the 3-pole motor used in the Life-Like/Walthers 0-8-0 has always had exceptional slow-speed performance, especially for a "simple" 3-pole open-frame motor.  It will start and turn at well under 100 rpm, even under some load, which is positively stupendous for such a motor. 
I note that the resistance of each coil is about 52 ohms, a good double the 20-24 ohms in my replacement 3-pole armatures for my Rivarossi project (or for typical Atlas/Kato motors).
It has thinner wire.  This leads me to conclude that they put a lot more windings of thinner wire on there, which gets you a much stronger magnetic field, more torque, and lower RPMs.

So I'm going to try rewinding one of my test 3-polers with #42 or #44, with double the turns that were originally on there.

Question:
#36 wire (typical motor gauge) is rated at about 400 ohms/ft.  Of course, in an N Scale motor, there are many feet of wire and the resistance is still only about 20 ohms.  So what the heck is going on?  How is the resistance only 20 or 50 ohms with all those feet of fine gauge magnet wire?

[C855B]

Question:
... #36 wire (typical motor gauge) is rated at about 400 ohms/ft. ...

Uh... 400 ohms per 1000 ft.   

[MK]

This leads me to conclude that they put a lot more windings of thinner wire on there, which gets you a much stronger magnetic field, more torque, and lower RPMs.

This is correct.  I have rewound armatures of 540 sized motors used in RC applications back in the 90s.  Back then I needed high torque, low RPM 540 sized motors for a twin drive RC tugboat.  All that were out there were RC car motors with 10-20 turns for each armature.  They wanted speed and short run time to win races.  So I got a bunch of them on closeout since I didn't care what turns they had.  I rewound them with 32 gauge wire.  I think I was able to fit 100 turns per armature (compared to the usual 10-20).  Now I can run my RC boat forever since the current draw had dropped tremendously and torque increased to the point where gear reduction was not needed.  Much simpler with a straight drive setup.

[peteski]

As I remember from school, strength of the magnetic field around a wire conducting current is directly related to the current flowing through it (voltage does not come into play).  But voltage and resistance comes do into the equation (Ohm's Law).   The number of turns also determines the strength of the magnetic field.

I'm just not sure how it all fits together.  I would have thought that for example a coil with 100 turns of thicker wire conducting 100mA woudl produce the same strength magnetic field as a coil wound with 200 turns of a thinner wire conducting 50mA.  But  this is just my gut feelign, not based on any scientific calculations.  Maybe someone with more knowledge about the subject can enlighten us.

[mmagliaro]

Uh... 400 ohms per 1000 ft.   

Ha!  THANK YOU!  What a marooon I am.   Okay, so that clears that up!

[mmagliaro]

As I remember from school, strength of the magnetic field around a wire conducting current is directly related to the current flowing through it (voltage does not come into play).  But voltage and resistance comes do into the equation (Ohm's Law).   The number of turns also determines the strength of the magnetic field.

I'm just not sure how it all fits together.  I would have thought that for example a coil with 100 turns of thicker wire conducting 100mA woudl produce the same strength magnetic field as a coil wound with 200 turns of a thinner wire conducting 50mA.  But  this is just my gut feelign, not based on any scientific calculations.  Maybe someone with more knowledge about the subject can enlighten us.

I think you are correct.  The hope would be that if a thinner wire were to double the resistance (and halve the current), it wouldn't do that until you get something like 4 times as many turns on there.  And half the current flowing through 4x as many windings would still give you a much stronger field.

We're going to find out.  I have spools of magnet wire coming my way.

[peteski]

I think you are correct.  The hope would be that if a thinner wire were to double the resistance (and halve the current), it wouldn't do that until you get something like 4 times as many turns on there.  And half the current flowing through 4x as many windings would still give you a much stronger field.

We're going to find out.  I have spools of magnet wire coming my way.

One thing I might have failed to consider is that with smaller diameter (round) wire there is less empty space (aiir) in between turns of the winding, so the coil is more dense (more copper per square millimeter). That will result in stronger magnetic field.

[Maletrain]

I am not understanding how this effect is happening.  Looking at the resistance of wire as a function of diameter (I compared #36 to #40 on a chart which doesn't go smaller than #40), it looks like the resistance is exactly the reciprocal of the cross sectional area of the wire strand.  So, if you decrease the wire diameter such that you reduce its cross section area by 50%, you would get 2X the resistance.  And, if you filled the area of the coil to the same cross sectional density of copper with the smaller wire, then you would be twice as many turns.  But, for a fixed voltage, you would get half as much current through the smaller wire.  So, the number of amps X turns would be exactly the same.  And the strength of an electro magnet is determined by the number of amps x turns, so they should be the same.

So, I am coming to the same conclusion that Peteski posted: there must be substantially more than 2x the number of turns with the smaller wire to make the magnet stronger.  That seems plausible, mainly because of the edge effects on packing circles into small rectangular spaces.  So, Max should probably see how many turns are on an original armature and then see how many he can fit on the same armature with smaller wire.

But, more turns will increase resistance, which will decease current, and thus amps x turns will not increase linearly with the number of turns.  On the other hand, the turns on the inside of the coil are shorter than the ones on the outside, so their resistance is not the same per turn as you add more turns.  The additional turns on the outside should be more than average resistance when measured per turn.

Which brings me to another question:  Could it be that those Atlas motors have smaller armature cores, so that the innermost turns are smaller in diameter and thus shorter than the other motors?  That could provide more turns per ohm of wire resistance, so more amp-turns.

 

[peteski]

We power our motors with variable voltage (0-12V) so the increased resistance is actually beneficial, because it results larger usable throttle range.  That is also why the slow-speed motor will run slower than the fast motor at the same voltage.

For example, I suspect that Atlas slow-speed motors use the same armature as the older "fast" motors, but are wound with thinner wire.  That results in denser winding with higher resistance, producing  slower RPMs than the "fast" motors within the 0-12V range.  This results in apparent better slow speed performance, and slower maximum speed (at 12V).

One of these days one of us should take a fast and slow speed motor and compare the number of turns per pole, and the wire diameter.

[mmagliaro]

We power our motors with variable voltage (0-12V) so the increased resistance is actually beneficial, because it results larger usable throttle range.  That is also why the slow-speed motor will run slower than the fast motor at the same voltage.

For example, I suspect that Atlas slow-speed motors use the same armature as the older "fast" motors, but are wound with thinner wire.  That results in denser winding with higher resistance, producing  slower RPMs than the "fast" motors within the 0-12V range.  This results in apparent better slow speed performance, and slower maximum speed (at 12V).

One of these days one of us should take a fast and slow speed motor and compare the number of turns per pole, and the wire diameter.

Actually, way back when, there was a long thread in here about the scale speed vs older Atlas motors, and the wire is indeed thinner on the scale speed motors.

Here is the old thread from 2015, we puzzled over a lot of these same questions:
https://www.therailwire.net/forum/index.php?topic=35052.0

[conrad]

From a wiki site, not my very old memory:

For solenoids the field strength is B=unI  where u is permeability, I is current (related to voltage and resistance, and in a motor other factors) and n is the turn density (turns per length of solenoid).

Conrad

PS trivia - 1 Tesla equals 10,000 Gauss

[peteski]

Actually, way back when, there was a long thread in here about the scale speed vs older Atlas motors, and the wire is indeed thinner on the scale speed motors.

Here is the old thread from 2015, we puzzled over a lot of these same questions:
https://www.therailwire.net/forum/index.php?topic=35052.0

Yes, this was discussed at length, and the question of unwinding the armature and counting the turns was brought up.  And the other thing that was brought up is that with BEMF present in a rotating armature, things get complicated.  But the base logic that you brought up here again (wire thickness and number of turns) still play an important role.    And no, I didn't read all 7 pages of that old thread. 

[mmagliaro]

Well, let me tell you, unwinding an armature with those blobs of epoxy in the way is no easy feat.  (The blobs in the skewed gaps between the metal armature sections).
I was able to cut most of that junk away, but stilll had to be content unwinding only a little at a time, break, unwind some more,
and just "estimate" a 1/2 turn when the wire breaks and I have to pick up again with a new piece.
340 turns.  Yikes.   It's going to be fun trying to wind, say, 500 turns of thinner wire on there.  My goal is to wind as much as I possibly can so that it completely fills the form for maximum magnetic field.

[peteski]

Good luck Max!  Since I assume it is a 5-pole armature, each winding spans 2 poles, so you need to make an allowance for 2 winding to coexist on the same pole.  I would not want to attempt to do that. 

[mmagliaro]

Good luck Max!  Since I assume it is a 5-pole armature, each winding spans 2 poles, so you need to make an allowance for 2 winding to coexist on the same pole.  I would not want to attempt to do that. 

No, this is only a 3-pole.  My Rivarossi retro-fit project uses 3-pole armatures that I could buy cheaply in quantity off eBay.
I have not been able to find quantity 5-polers that would fit the same can.
But the replacement 3-pole is a skewed armature with more windings and a larger coil section, so it does give the motor
more power than the original (plus I have the neodymium magnet ring). 

This rewinding thing is a distraction from my focus, but I'm super curious as to why the LL 3-pole is so great.