Field Weakening VS Gear Box VS Higher voltage battery

[Doctorbass]

What about the formula-E ?

I guess they certainly choose the best technology since it's in the race industry! :| I would have some difficulty to believe that a multi billion dollars industry could do the bad choice?? is it ?

I wonder why they choose to use a 5 or 6 gear transmission ? combined with their 200kW motor.

They seem to run pretty high RPM the sound is just like our RC motors and a dentist mill !! 5-10kHz+

It would be great to debate on that...

Doc

 

[gogo]

What about the formula-E ?

I guess they certainly choose the best technology since it's in the race industry! :| I would have some difficulty to believe that a multi billion dollars industry could do the bad choice?? is it ?

I wonder why they choose to use a 5 or 6 gear transmission ? combined with their 200kW motor.

They seem to run pretty high RPM the sound is just like our RC motors and a dentist mill !! 5-10kHz+

It would be great to debate on that...

Doc

Just a guess, but gearboxes, in this context, are expendable equipment. You simply can't make durable gearboxes for a consumer product that cost less overall than just using a big motor instead. Its a trade of cost and reliability for short term performance. An analogy would be dragster motors that are rebuilt after every run. Sure, you can do that, but its not a practice that translates to a production product.

 

[major]

What about the formula-E ?

I guess they certainly choose the best technology since it's in the race industry! :| I would have some difficulty to believe that a multi billion dollars industry could do the bad choice?? is it ?

IMO, Formula-E made a number of bad choices. It appears they did things their way, not the best way. It would not surprise me to learn that it was dictated to use a shifting gearbox and the design team had no option. Perhaps they will open the spec and allow innovation someday then we'll see the technological superior systems actually competing.

For all the problems the TTXGP had, they actually put on some decent electric gran prix shows with the electric racers qualifying within a few percent of top gas racer times. Mission, Lightning, Brammo, Muench, MotoCzysz. How many of those bikes had multi-ratio gearboxes?

 

[zombiess]

This thread is full of "if you aren't doing it my way it's the wrong way". With that said, here comes my 2 cents.

Pick what works best for your application. Field weakening is nothing moren than a tool. Gear boxes are also tools, so are big motors. Depending on your application, you might use any of the 3 mentioned tools or something completely different. It's all pros and cons and the engineer has to weigh what works best for situation.

I personally like big direct drive hubs for mostly level street use. If my application was off road hill climbing, I might go mid drive. If I needed good hill claiming and high top speed I might use a mid drive with a gear box. If I had no constraints I might choose a really big motor in mid drive or direct drive.

Different applications require different tools. One size does not fit all. We get into a lot of measurabating on this site sometimes and forget there is no one best method for every application, but there are better and worse ones.

BTW, I am a proponent of field weakening. It's a nice option to have in a controller. Even if you decide you don't want to use it (you simply turn it off).

Open your minds.

Spec a drive system for this vehicle. I want it haul 1000lbs in addition to its own 1000lb weight, go 150mph top speed and climb a 20% 10 mile grade. Motor(s) can only weigh 75lbs total and can not be custom.

Better break out that box of tools.

 

[John in CR]

Open your minds.... and climb a 20% 10 mile grade.

It's always opened, but since there is no such road I rejected the problem.

FWIW, since you and Luke signed off on field weakening, I'm in, but I do need to be able to turn it on and off without running back to my computer.

 

[Punx0r]

Spec a drive system for this vehicle. I want it haul 1000lbs in addition to its own 1000lb weight, go 150mph top speed and climb a 20% 10 mile grade. Motor(s) can only weigh 75lbs total and can not be custom.

75lb including any gearing?
Expected service life?
Is the 150mph on level ground or up the 20% grade?
What CdA?

Can you provide an example of an equivalent ICE vehicle? Your allowed weight for the motor is only 3.8% of the total vehicle weight and this doesn't seem much.

 

[zombiess]

1. 75lb including any gearing?
2. Expected service life?
3. Is the 150mph on level ground or up the 20% grade?
4. What CdA?

1. yes.
2. 42
3. both
4. 42

BTW, those were just random specs for examples. When you have a problem to solve, it's often best to use the right tool for the job that fits within your allotted budgets.

 

[onloop]

What would be cool is a constantly variable KV motor. Using field weakening theory. It can deliver more torque OR top speed based on load. It is completely automatic. Or maybe a torque button / speed button. Simply set the top and bottom limits that is should vary between.

 

[halcyon_m]

I agree with anyone here who says "it depends", because it does.

Field weakening, flux weakening, field oriented control, all the same thing. It's accomplished in permanent magnet motors by making a magnetic field on the windings that opposes the magnetic field from the permanent magnets, weakening their effect. The cost of doing so is current in the windings, or extra copper heating. The effectiveness of this approach depends very much on how the electric motor is designed, and having a good amount of magnetic saliency helps.

Salient motors are rotors with the magnets buried within the rotor iron, which allows a tight enough air gap to allow effective field weakening. The saliency comes primarily from the iron in the rotor that is between the magnetic poles of the rotor, where the magnetic path is easier (less magnetic reluctance than the path containing the magnets), but the important part is that the air gap is small. Something called 'characteristic current' is the most relevant parameter, which is the current the motor will put into the windings if you short them at any speed above some minimum value. A motor with a characteristic current equal to the current rating of the motor has a constant-power-to-speed ratio that is effectively infinite, meaning you can get the rated power at any speed because it is possible to push back on the magnets completely with the current available from the controller.

Speed is obviously restricted by mechanical constraints and iron losses above a certain speed that start significantly reducing efficiency, but one popular example of this type of control is in pretty much every hybrid car in use today, where the top speed is about 3x the 'base speed' where flux weakening starts.

Most hobby motors have the magnets on the surface and don't usually have field weakening ability because their characteristic current is much higher than the rated current of the motor. (If anyone would like to try the shorted winding current as a function of speed test with any of your hobby motors, I'd be very interested to see the results)

So, assuming you have a motor that is capable of field weakening, do you use a gear box with it?
To differentiate, I don't see any way around a gear reduction, typically. Motors you can buy and motors that fit an application rarely have the speed profile of the wheel you are trying to spin. Therefore, an appropriate gear ratio is a must.

However, the question becomes, does the added weight of multiple gear ratios, and the time spent shifting gears, and the mechanical complexity of shifting gears weigh negatively in your application versus added heat in the windings of the motor? It depends a lot on how you use it. I'd love to come up with a solid answer, but I think we need a case to study before any definitive answers are able to be presented.

 

[John in CR]

Halycon,

Would outrunners with curved magnets for a constant and small magnetic gap be good candidates for field weakening?

 

[Arlo1]

Interesting thread.

 

[Miles]

Would outrunners with curved magnets for a constant and small magnetic gap be good candidates for field weakening?

Steel spacers between the magnets would probably help but would also increase flux leakage...

 

[Miles]

Something called 'characteristic current' is the most relevant parameter, which is the current the motor will put into the windings if you short them at any speed above some minimum value. A motor with a characteristic current equal to the current rating of the motor has a constant-power-to-speed ratio that is effectively infinite, meaning you can get the rated power at any speed because it is possible to push back on the magnets completely with the current available from the controller.

Thanks halcyon.

 

[kenkad]

halcyon_m,
Can you post a link to a paper or article that describes how the experiment is done to measure this 'characteristic current'? This paper 'Analysis of a Novel Coil Design for Axial Flux Machines' references differing aspects of salient design of the rotor (around page 4). There is a lot of 'it depends', especially when it involves design complexity/cost. For AF types, rotor skewing (dual rotor types), accurate timing advance control, multiple three-phase groups, multiple coil rings, etc., all play significant roles. Some affect the motor design and some affect the controller design.
kenkad

 

[major]

Something called 'characteristic current' is the most relevant parameter, which is the current the motor will put into the windings if you short them at any speed above some minimum value. A motor with a characteristic current equal to the current rating of the motor has a constant-power-to-speed ratio that is effectively infinite, meaning you can get the rated power at any speed because it is possible to push back on the magnets completely with the current available from the controller.

Thanks halcyon.

I also found that to be an interesting paragraph. Most of those who read it may miss the point that power-to-speed ratio is torque. So is halcyon saying motors have infinite torque?

 

[halcyon_m]

Halycon,

Would outrunners with curved magnets for a constant and small magnetic gap be good candidates for field weakening?

In general, I would think they would. However, with the nickel-coated magnets that are common, there is a potential for magnet heating from the rapid changes in flux through the various parts of the magnet that would cause eddy currents. I think this is one reason why the early honda IMA motors (insight and civic hybrid) used a 'captured magnet' variety, which essentially used part of the steel that made up the rotor to loop over the magnet. This did allow some of the magnet flux to loop back through the steel keeping the magnet, but the vast majority was able to cross the gap and be useful. The motor I'm speaking of is one of the rare examples where a motor with very little magnetic saliency was able to accomplish flux weakening.

One of the tricks motor designers use is segmenting the magnets into a grid so that the eddy currents don't circulate even within the magnets themselves. Also, it is popular to use non-conductive magnet coatings (they need a coating to avoid oxidizing the NdFeB rare earth material), which lowers losses as well.

Also, with any time you lessen the gap, you're encountering a higher chance of magnetic asymmetries from gap inconsistencies, but also more of a chance to have the rotor interfere with the stator. In larger machines, the windage drag of the gap is also a consideration that all point to using a gap around 0.75mm or so from what I've seen.

 

[halcyon_m]

Would outrunners with curved magnets for a constant and small magnetic gap be good candidates for field weakening?

Steel spacers between the magnets would probably help but would also increase flux leakage...

Agreed, but that doesn't mean that something couldn't be done. With any buried magnet machine (interior permanent magnet, or captured magnet) is a trade-off of having steel bridges as part of the structure to be the 'tension cables' to bear the weight of the spinning magnets, particularly for the inrunner type where the magnets want to fly out to the stator. But, those bridges are designed to be as thin as possible to decrease the amount of flux leakage to a minimum practical amount.

Obviously having a solid piece (not laminated) would be the worst for losses as it would be an eddy current nightmare, but it's along the right track. In fact, one of my friends actually used JB Weld (which until that time the concept of 'steel' in JB Weld had just been an analog for strength in my mind) to fill the gap of an outrunner, then sanded/finished to create a minimum amount of clearance. He reported that the voltage constant increased significantly (KV decreased), which I thought was quite cool. There was likely some filings that further contributed to leakage between magnets, but an interesting idea for sure.

Another colleague of mine used iron powder in a 500w wind turbine design (the one using a bunch of magnets and an automotive wheel hub for the mechanics) to increase the voltage of the machine.

All interesting ideas for sure. They would take a bit of refinement to become practical for manufcacturing, but we are mostly just spit-balling here, right?

 

[Miles]

One of the tricks motor designers use is segmenting the magnets into a grid so that the eddy currents don't circulate even within the magnets themselves. Also, it is popular to use non-conductive magnet coatings (they need a coating to avoid oxidizing the NdFeB rare earth material), which lowers losses as well.

Also, laminating the rotor yoke. I did see one paper reporting that a laminated rotor with unsegmented magnets actually increased rotor losses, though. I'm not sure why...

 

[halcyon_m]

Something called 'characteristic current' is the most relevant parameter, which is the current the motor will put into the windings if you short them at any speed above some minimum value. A motor with a characteristic current equal to the current rating of the motor has a constant-power-to-speed ratio that is effectively infinite, meaning you can get the rated power at any speed because it is possible to push back on the magnets completely with the current available from the controller.

Thanks halcyon.

I also found that to be an interesting paragraph. Most of those who read it may miss the point that power-to-speed ratio is torque. So is halcyon saying motors have infinite torque?

I guess I should have been more explicit. I did -not- mean that motors have infinite torque. I meant that the ability of a motor to deliver constant power (where torque decreases while speed increases) is available over a range of speeds. For example, look at the 2010 Prius motor shown below. (it's from the oakridge national lab report on the 2010 prius)

The key elements to not here are that:
1) the base speed is around 3200RPM. This is the maximum speed at which full torque can be applied. In this case torque is limited by the current rating of the inverter.
2) the torque decreases above 3200RPM, as the inverter is only capable of delivering so much current, and some current has to be dedicated to 'pushing back against the magnets', something referred to as negative d-axis current, where the d (direct) axis is in line with the magnets. Only the remainder of the current can go in the q-axis (quadrature axis, 90 electrical degrees from the d-axis, aligned between magnetic north and magnetic south), where the d and q axis currents add as orthogonal vectors to be less than the current limit of the inverter. Example: if current limit is 100A, and 60A is needed for flux weakening, then only 80A can be used to create torque in the q-axis, and this comes from the equation where: square root of (60^2 and 80^2) is 100. Remember the 3,4,5 right triangle?
3) the characteristic current for this motor is slightly more than 1, which is evidenced by the torque capability decreasing faster than the speed is increasing, giving less and less power as speed increases beyond 3200RPM. However, not all is lost. There is still 50kW available at 6000RPM, where there was 64kW available at 3200RPM.
4) imagine if you only had the use of this motor to 3200RPM without flux weakening and you had to shift gears to get to higher speeds. Not only would you increase the mechanical complexity, but you would also miss out on being able to operate the machine in the area of peak efficiency. Ok, so this is cheating, but only a little, because the machine was designed to take advantage of flux weakening, so the peak efficiency could have been designed at a lower speed, but this was a very well designed machine that is able to provide nearly constant power over a 4:1 speed range.

I hope that clears up what I mean by constant-power speed-range (aka CPSR). Modern gear boxes have about a 5:1 range if I'm not mistaken, so if you have a machine with a 5:1 CPSR and a decent efficiency throughout (check out the 90+% efficiency at higher speeds!), then why introduce a multi-gear transmission with added weight?

Of course there are other practical considerations to using a gear box, but I'll let someone else argue that side for now.

 

[major]

....but this was a very well designed machine that is able to provide nearly constant speed over a 4:1 speed range. ....

Thanks for expanding on that for us. Did you mean power instead of speed? As in CPSR?

 

[Punx0r]

Is flux weakening a Band-Aid for a suboptimal motor?

BTW, the multi-ratio Vs. single ratio debate has been done to death. The answer seems to be: With an ideal motor multiple ratios are an abomination that can only hurt power delivery, but with compromised, commercially available motors, they can be beneficial.

Same answer as single ratio reduction Vs. direct-drive.

 

[halcyon_m]

....but this was a very well designed machine that is able to provide nearly constant speed over a 4:1 speed range. ....

Thanks for expanding on that for us. Did you mean power instead of speed? As in CPSR?

Yes, thanks for the catch, major, i'll edit above to correct.

 

[halcyon_m]

halcyon_m,
Can you post a link to a paper or article that describes how the experiment is done to measure this 'characteristic current'? This paper 'Analysis of a Novel Coil Design for Axial Flux Machines' references differing aspects of salient design of the rotor (around page 4). There is a lot of 'it depends', especially when it involves design complexity/cost. For AF types, rotor skewing (dual rotor types), accurate timing advance control, multiple three-phase groups, multiple coil rings, etc., all play significant roles. Some affect the motor design and some affect the controller design.
kenkad

While I'm sure such a paper exists, I can't seem to find one.

In stead, I conducted a quick test on a motor that I'm using for a project I'm working on, and it illustrates the effect I am speaking of. If you short a motor, and spin it with another motor, then the current should reach some maximum value at some speed where the resistance of the windings is small compared to the BEMF in competition with the inductive impedance of the windings. At low speeds, the resistance limits the current, but at high speeds the impedance of the inductance dominates the total series impedance and you get a flat current characteristic. The result is a maximum value of current regardless of speed. For this machine, the characteristic current is around 3.5A peak, (~2.5A RMS). The inductance of this type of machine is characteristically high, so that means that the characteristic current is also quite low.

What does it all mean? You are never going to get more than 2.69A RMS out of this machine, no matter how fast you spin it as a generator. This is one reason why designing a machine with a characteristic current less than the rated current is a bad idea.

So it's a balance. Design the characteristic current above the desired current rating and it's difficult to flux-weaken. Design it below your desired current rating and it will not perform to your design, particularly as a generator. Design it dead on (1 per-unit, i.e. where the characteristic current equals the rated current), then, in theory, the constant-power speed-range (CPSR) is infinite. From that perspective, you can see why toyota designed their motor to have a characteristic current slightly more than 1 per-unit, so they could have some head-room on performance, without sacrificing the high-speed power performance too much.

 

[Miles]

An example comparison study from Infolytica: http://www.infolytica.com/en/applications/ex0181/

 

[speedmd]

Looking at the prius chart, makes me wonder why would you not want to try just a low/high gear setup in addition to the FW controller. At the higher rpms you still have the same power, understood, but torque drops off drastically. You have 190 Nm at 3200 rpms and it drops to what looks like 40 Nm at 10000 rpms. This kills your ability to accelerate or climb as well at the higher rpms. Is it better to shift into a taller gear and have more overall torque available than to just wind out the revs for simplicity sake? If your electrically able to get a 4:1 speed range and add a logical mechanical step in a mechanical reduction (thinking ebike / sprockets/ using a go cart type two speed) of something like a 1:3 to a 1:2 ratio it would even improve the already tremendous speed range and broaden the torque band also.