Same debate again, motor windings!

Hummina Shadeeba said:
What about it makes it high efficiency?

1. Low phase-to-phase resistance...Each half is 32mOhms, so 16mOhms for the motor as a whole, so it can handle the current with low copper losses. eg 200A makes only 640W of heat in the copper, and 100A, a realistic highway cruising speed requirement, only 160W of heat. That's why phase-to-phase resistance is such a critical statistic for every motor, because it's the only way to predict the controller current limits you need to use.

2. High quality thin (.3mm) steel stator laminations. I know nothing about the grades of motor steel, but the factory literature always bragged about the use of high quality Japanese steel.

3. The low stator slot (24) and pole counts (10 pole pairs) compared to other hubmotors give it a low operating frequency, which combines with the high quality stator steel to result in quite low iron losses. eg at 100kph (62mph) the motor spins at 1089rpm with the tire size I use and iron losses are only 151W. Compare that to my V1 QS273 with it's .5mm lams and higher than a 205 slot and pole count that has no-load losses of over 800W at the same rpm. I'd have to run HubMonster at over 3000rpm to get to the same iron losses as my QS at 1000rpm. #2 and #3 combine in the motor stats as low no-load current for a given rpm. Voltage X no-load current is your iron losses.

4. Small air gap and curved magnets. If I remember correctly the magnetic gap is about .5mm . I think these close build tolerances fall out in the performance results as better torque per amp and less torque ripple (cogging torque), which I see in the form of these large hubmotors (15kg) coasting better than any hubmotor I've used.

As an example, at whatever wind resistance and grade cause the motor to draw 100A at 100kph, Miles' spreadsheet tells me the power output at the wheel is right at 6,000W and the motor is creating only 319W of heat (an amount of heat easily dissipated at that speed) for an overall efficiency of 94.9%. Then with the near doubling of power required for a 25% increase in speed to 125kph (78mph), HubMonster is putting out 11,656W on input of 12,270W , at slightly higher efficiency of 95%, and again it should be able to shed the 614W of heat generated, so continuous cruise at that speed is possible as long as your battery can last. Now at the well over 30kw peak input I run one of my HubMonsters at (260A on a 31s pack), I did ventilate the motor with exterior blades on the air output side, and in regular use it's high efficiency makes it my coolest running HubMonster that typically below 70°C with the highest I've seen being 100°C with hard acceleration out of every switchback really attacking a 20% continuous grade to visit the windmill farm on top of a mountain.
 
Hummina Shadeeba said:
got lots of pics? maybe it could be improved with another winding pattern or more in. being the motor winding topic.

I wouldn't dare trying to mod the stator of the 6 in my possession. The rotor is another story, I've removed 1.5kg of weight from the rotor for a mid-drive installation I'm working on. Their earlier 2 speed model that I call the MiniMonster has a slightly lower peak efficiency will get a rewind though. I have 13 of those, and the mechancal 2 speed switching inside the motor was also their weakness due to added resistance and contact switches not capable of high current. In addition to dumping the series/parallel switching, I want to increase the Kv to between 30 and 40rpm/volt to make it easily capable of handling 25-30kw with efficiency approaching that of HubMonster in a smaller size and weigh about 8kg.
 
Interesting summary, John. The advantages of low phase-resistance and pole count often seem to be countered by a motor that's very hard on controllers, but with yours being split in to 2 x 3 phases, it neatly side-steps that issue.

No doubt the low iron losses also contribute to low drag when coasting too.
 
Punx0r said:
No doubt the low iron losses also contribute to low drag when coasting too.

they're one and the same :p
 
sn0wchyld said:
Punx0r said:
No doubt the low iron losses also contribute to low drag when coasting too.

they're one and the same :p

I guess the torque ripple is part of the iron losses, but PunxOr was right to point out that all of the iron losses contribute, since it's probably more the torque drag of eddy current and hysteresis that we feel while coasting than the torque ripple (cogging torque), since it plateaus at pretty low rpm. I didn't know ripple and cogging were the same thing until making that post and looked up a definition of torque ripple, so I focused on getting that in the post.

BTW sn0wchyld, I must apologize for still not getting one of the Mini's rewound, since I've been promising that for years now. Every time I think about it I get sidetracked about how I'm going to mod the cast steel shell to mount a sprocket or pulley. You got any ideas? Since my Nuc 24's are in route and my Nuc 12's aren't ready yet, maybe I should go with a MiniMonster instead, since geared down to 80-90kph top end with this road and trail bike using 21" moto wheels the 30-40kw potential of dual Nuc 24's would be going to waste. Of course I could put the 24's on my 105kg scooter that now has a pair of 12's and turn it into a real motorcycle slayer. I pull away from the vast majority already, but the much harder hitting 24's would really show them what's possible with electrics :twisted:
 
John in CR said:
sn0wchyld said:
Punx0r said:
No doubt the low iron losses also contribute to low drag when coasting too.

they're one and the same :p

I guess the torque ripple is part of the iron losses, but PunxOr was right to point out that all of the iron losses contribute, since it's probably more the torque drag of eddy current and hysteresis that we feel while coasting than the torque ripple (cogging torque), since it plateaus at pretty low rpm. I didn't know ripple and cogging were the same thing until making that post and looked up a definition of torque ripple, so I focused on getting that in the post.

BTW sn0wchyld, I must apologize for still not getting one of the Mini's rewound, since I've been promising that for years now. Every time I think about it I get sidetracked about how I'm going to mod the cast steel shell to mount a sprocket or pulley. You got any ideas? Since my Nuc 24's are in route and my Nuc 12's aren't ready yet, maybe I should go with a MiniMonster instead, since geared down to 80-90kph top end with this road and trail bike using 21" moto wheels the 30-40kw potential of dual Nuc 24's would be going to waste. Of course I could put the 24's on my 105kg scooter that now has a pair of 12's and turn it into a real motorcycle slayer. I pull away from the vast majority already, but the much harder hitting 24's would really show them what's possible with electrics :twisted:

All good mate - could have done it myself but like you, keep getting sidetracked. I'm moving to a different country for a while (work) - so I mgiht be able to find some locals that can re wind for a non absurd price, will let you know how it goes, but yea ill be pairing it with a nuk24 too once done, and possibly trying to squeeze it into a surron. Will likely look at liquid cooling too if I go that far, you know... in for a penny in for a pound.

I'm not so sure 30ish kw will be 'wasted' on a mx though, many of the top end mx bikes have 40+. Granted neither of us are primed for racing hahah but should still make for a very fun (if short) ride.

NOW STOP GOING OFF TOPIC!!?!?!??11!

:D
 
Hummina Shadeeba said:
In my looking i didn’t see anything about cogging torque and torque ripple plateauing at higher speed. Why would that happen?

I picked that up from Justin in a discussion in the ebike technical section years ago when he was talking about the breakdown of iron losses. Damned if I can find the thread though. I don't have the foggiest idea why cogging torque becomes a fixed amount of drag, but it makes sense though if you've ever tried to pedal through the drag of an unpowered DD hubmotor.
 
sn0wchyld said:
I'm not so sure 30ish kw will be 'wasted' on a mx though, many of the top end mx bikes have 40+. Granted neither of us are primed for racing hahah but should still make for a very fun (if short) ride.

It's going to be an on/off roader, and since I have near zero dirt riding experience I'm talking about light trail use. I'm running dual Nuc 12f's on a heavy scooter geared (tire size) for a top speed of 130kph with 0 field weakening. It does pretty well, so geared for a top end in the 80's on a lighter bike with a shorter wheelbase and much higher CG running the same voltage should enable me to turn the 12f's all the way up and activate the boost feature too for wheel lifting launches. When I want truly violent launches I just hop on my little low slung super commuter that uses the lower Kv MidMonster with a 31s pack that's never lost from a red light. The only thing to come close was a 1 liter gasser that was reving up before the green. Of course he came flying by at about 150m.


sn0wchyld said:
NOW STOP GOING OFF TOPIC!!?!?!??11!
:D

LOL, who me?
 
Here's a pic showing the long transitions on high efficiency HubMonster.

Long transitions on HubMonster.JPG


Here's a close up of the windings on the wire side. Why I wouldn't rewind, since I don't think any way I could fit more copper without going to square or rectangular and I'd need experience first.
Wire side up close HubMonster.JPG
 
sn0wchyld said:
Punx0r said:
No doubt the low iron losses also contribute to low drag when coasting too.

they're one and the same :p

STOP ATTACKING ME!!! I DEMAND RESPECT AND ADORATION!!!

Ahem. Sorry. I meant: you are of course correct :)

There was some good discussion on cogging recently: the net force averages out to zero (attraction-repulsion-attraction etc). At speed the rotor inertia would give a flywheel effect, smoothing it out. IIRC there are some losses (hysteresis & eddies in the magnets and iron) associated with this, but they're already accounted for in the core losses.
 
Punx0r said:
sn0wchyld said:
Punx0r said:
No doubt the low iron losses also contribute to low drag when coasting too.

they're one and the same :p

STOP ATTACKING ME!!! I DEMAND RESPECT AND ADORATION!!!

Ahem. Sorry. I meant: you are of course correct :)

There was some good discussion on cogging recently: the net force averages out to zero (attraction-repulsion-attraction etc). At speed the rotor inertia would give a flywheel effect, smoothing it out. IIRC there are some losses (hysteresis & eddies in the magnets and iron) associated with this, but they're already accounted for in the core losses.

Thanks for the laugh with that first line. Starting a day with a laugh is always good.

It sounds like I must have something wrong and now have to go find what Justin was talking about. For some reason I had it stuck in my head that there were 3 components of the iron losses... cogging torque, eddy currents, and hysteresis. Not that I know much now, but back then most of the technical stuff was flying over my head, so I must have misabsorbed something.
 
I think I remember the post in question and Justin was making the point that a lot of the "cogging torque" people feel is just trying to overcome magnet repulsion and attraction. As you push through the magnet resistance and get to the 'other side' then the magnets push away with similar force. So that it ends up like climbing a series of tiny hills, dump energy to go up and get energy back out when you go down. So much of the energy that goes into pushing past a 'cog' goes into propelling the bike further.

There are still losses, of course. You are never going to get 100% of the energy back out of the motor due to heat losses. The big take away from his post was that one of the advantages of a more efficient motor is less cogging action to go along with less losses. Because otherwise it's usually more cost-effective to just get a cheap motor and slightly bigger battery pack then get a high efficiency motor. It's relevant because there are more advantages to being efficient then just power savings.



This sort of stuff is why 'seat of the pants' measurement and personal experience doing X, Y, or Z is usually a huge waste of time. Humans just are very very shitty at estimating effects like acceleration and speed. It has to do with how our depth perception works, estimating distance, and physiological things like that. And then because of that relying on sensation of movement and trying to correlate that to power output is just a exercise in compounding errors.

So to compensate our brain uses various 'proxies' to estimate speed unconsciously. Things like noise amount, pitch of noise and vibrations. We interpret high vibrations as 'fast'. This is why people think they ride faster with high-pressure skinnier tires on bicycles, despite the fact that 100+ PSI tires often have quite a bit higher rolling resistance on anything that isn't a glassy smooth concrete. It just 'feels' fast, but the vibrations they feel and noise they hear is actually watts being wasted. It's also why you feel like you are flying on a bicycle going 40, but in a big highly dampened car going 40 feels like it's slow as shit.

Basically... If you can't or won't measure something then it doesn't happen and isn't really relevant. Experience without measuring and record keeping isn't really useful.
 
Justin knows a lot about motors and puts a lot of time and effort into researching and writing about them, so if there's anything he says that contradicts what I've said then definitely go with what he says :wink:

I always think what the average lay ebike user calls cogging is really the drag of the iron losses.
 
sn0wchyld said:
Punx0r said:
No doubt the low iron losses also contribute to low drag when coasting too.

they're one and the same :p


Miles said:
Keeping it simple...

Copper losses (resistance losses) are the penalty for torque.

Parasitic losses (iron losses etc.) are the penalty for speed (angular velocity).

The point of maximum efficiency is where the copper losses are at parity with the parasitic losses.
 
Punx0r said:
Justin knows a lot about motors and puts a lot of time and effort into researching and writing about them, so if there's anything he says that contradicts what I've said then definitely go with what he says :wink:

I always think what the average lay ebike user calls cogging is really the drag of the iron losses.

I'm talking about something I read around 10 years ago, and interpreting it with a real lack of knowledge before sticking it in my permanent file. I would read through technical discusses understanding little and trying to pick up what I could. If there's no 3rd component to iron losses that tapers off in addition to eddy currents and hysteresis losses that clearly don't taper off, then I got something wrong.

sleepy_tired's "If you can't or won't measure something then it doesn't happen and isn't really relevant. Experience without measuring and record keeping isn't really useful." is about the equivalent to me of "If you don't have a video, it didn't happen." Both attitudes are laughable AFAIC. If I was writing a scientific paper, sure detailed measurements are required. I'd rather spend my time building another bike than collecting and processing data, and leave that to those who enjoy it.
 
sleepy_tired said:
"cogging torque" people feel is just trying to overcome magnet repulsion and attraction. As you push through the magnet resistance and get to the 'other side' then the magnets push away with similar force. So that it ends up like climbing a series of tiny hills, dump energy to go up and get energy back out when you go down. So much of the energy that goes into pushing past a 'cog' goes into propelling the bike further.

There are still losses, of course. You are never going to get 100% of the energy back out of the motor

This is my understanding too.

The inefficiency due to cogging comes from energy converted into vibrations and noise (fairly minute in the scheme of things) , which is lost and can't be converted back into the "downhill" force.

That is why motors with angled stator teeth and curved magnets have much lower cogging forces, and are much easier to start turning. This is something that is considered important for wind turbines, so they start with minimal air flow.
 
I think of cogging as only a pulling of a magnet away from a tooth not pushing. It is a loss while free coasting and also while powered to a lesser degree with the unpowered teeth i think.

I think it linearly parallels the rpm. Like hysteresis increasing linearly with rpm. Maybe it’s more a loss than the hysteresis? For sure seems so. Can’t isolate them from each other. Depending on ur motor design it can be virtually removed. Mellow hub motor for skateboards does it by assembling the teeth after winding and not needing a slot. no slot and no cogging.
 
Hummina Shadeeba said:
I think of cogging as only a pulling of a magnet away from a tooth not pushing. It is a loss while free coasting and also while powered to a lesser degree with the unpowered teeth i think.

If the magnets were always only pulling then the motor would spin indefinitely on its own ;) If you turn the motor shaft very slowly and you should feel the repulsion transition to attraction via a dwell point.
 
I'm not sure about the low rpm cogging not having losses. Think about spinning a wheel with bearings and how long the last few turns can take when it's spinning down freely with the linear bearing losses. Now take a DD hubbie and give it a spin and it's readily apparent that the losses causing it to slow are different, especially the last revolution or two where it slows even quicker. It sure seems that some loss is non-linear to me. What causes that if cogging nets to zero even at low rpm?
 
Punx0r said:
Hummina Shadeeba said:
I think of cogging as only a pulling of a magnet away from a tooth not pushing. It is a loss while free coasting and also while powered to a lesser degree with the unpowered teeth i think.

If the magnets were always only pulling then the motor would spin indefinitely on its own ;) If you turn the motor shaft very slowly and you should feel the repulsion transition to attraction via a dwell point.
Unpowered the magnets are only pulling the stator steel. I don’t see how any push would be possible without power. What u feel in ur hand is pulling of magnet to the next tooth or pulling away from the last tooth.

There’s “notches” which the rotor has to get through but I don’t believe those notches decrease or increase their resistance at any speed. when it goes very slowly the notches appear to change the resistance but it’s just the lack of momentum needed to over come the notches. Why would cogging be any different at different rpm? The magnetic field isn’t going to change w speed
It depends on the motor of course but with a common rc 12n14p i bet the hysteresis is less a loss than cogging torque when coasting when the teeth aren’t being formed into a polarity really
 
Hummina Shadeeba said:
Unpowered the magnets are only pulling the stator steel. I don’t see how any push would be possible without power. What u feel in ur hand is pulling of magnet to the next tooth or pulling away from the last tooth.

There’s “notches” which the rotor has to get through but I don’t believe those notches decrease or increase their resistance at any speed. when it goes very slowly the notches appear to change the resistance but it’s just the lack of momentum needed to over come the notches. Why would cogging be any different at different rpm? The magnetic field isn’t going to change w speed
It depends on the motor of course but with a common rc 12n14p i bet the hysteresis is less a loss than cogging torque when coasting when the teeth aren’t being formed into a polarity really

Perfect analogy is gravity. Think about riding up and down a hill. Use energy to get up the hill. Return energy going back down. Losses are friction. In the case of cogging, friction manifests itself in vibration and noise, and there will certainly be hysteresis and eddy currents in these small magnetic fluctuations (in addition to the hysteresis and eddy currents produced by the changing electromagnetic fields from powering the coils).
 
Wiki says

“Cogging torque results in torque as well as speed ripple; however, at high speed the motor moment of inertia filters out the effect of cogging torque.”

The language of filtering out the effects of cogging torque is vague

If hooked up a flywheel and measured different speeds coasting times I guess likely could see it coast longer from higher speed.
 
serious_sam said:
Perfect analogy is gravity. Think about riding up and down a hill. Use energy to get up the hill. Return energy going back down. Losses are friction. In the case of cogging, friction manifests itself in vibration and noise, and there will certainly be hysteresis and eddy currents in these small magnetic fluctuations (in addition to the hysteresis and eddy currents produced by the changing electromagnetic fields from powering the coils).

Here is a reference along the same lines as what I was saying:

"5. Consideration of cogging torque
In addition to the winding factor, cogging torque is also an important consideration. Cogging torque creates vibration and noise during operation and acts to disturb the motor away from its desired position when used in servo applications. The cogging torque frequency is also closely correlated with the production of rotor losses both with and without current supplied to the armature."

https://things-in-motion.blogspot.com/2019/01/selecting-best-pole-and-slot.html?m=1
 
Back
Top