Field Weakening VS Gear Box VS Higher voltage battery

[halcyon_m]

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.

I think you may be overlooking the marginal gains you get from a multi-speed transmission unless you really need the low end torque. Lets consider an example:

if you did install a transmission with a 10:1 reduction and a second gear with a 5:1 reduction.
in the 10:1 (1st gear) you would get 2,000Nm at 100RPM, 790Nm at 600RPM and 295Nm at 1,200RPM
in the 5:1 (2nd gear) you would only get 1,000Nm at 100RPM, 940Nm at 600RPM and 395Nm at 1,200RPM

For anyone used to analyzing transmission torque-speed curves, it makes sense that you would shift to 2nd gear above 500RPM, or about 1/6th the top speed because that's where you have more torque available. Let's assume that the top speed of this machine is 120MPH (like liveforphysics' wicked quick electric bike). This means you should shift gears at just north of 20MPH.

Here's the real question, do you NEED the extra torque provided by the low gear? Personally for sizing the gear ratio of my electric conversion, I realized that I could spin the wheels at low speeds using a single ratio, and still hit a top speed of 127MPH, which is likely quicker than I ever will get it especially considering it's been sitting for a year... but I digress. Maybe I'm biased, but in my case, I had enough torque to overcome the traction of the rear wheels and I didn't need more than that.

If you can find a use for 2000N-m, then I can see the transmission being useful, or if the sweet spot of efficiency was more towards the low speed range, like it is for most engines.

 

[kenkad]

halcyon_m,
Are you also saying that for an AF that rotor skewing has no effect on characteristic current? I cannot imagine why it would not. I can believe that timing advance does not enter the picture because you are only testing the characteristic current when appearing as a generator. Are you at ORNL or UTK?
kenkad

 

[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.

I think you may be overlooking the marginal gains you get from a multi-speed transmission unless you really need the low end torque. Lets consider an example:

if you did install a transmission with a 10:1 reduction and a second gear with a 5:1 reduction.
in the 10:1 (1st gear) you would get 2,000Nm at 100RPM, 790Nm at 600RPM and 295Nm at 1,200RPM
in the 5:1 (2nd gear) you would only get 1,000Nm at 100RPM, 940Nm at 600RPM and 395Nm at 1,200RPM

For anyone used to analyzing transmission torque-speed curves, it makes sense that you would shift to 2nd gear above 500RPM, or about 1/6th the top speed because that's where you have more torque available. Let's assume that the top speed of this machine is 120MPH (like liveforphysics' wicked quick electric bike). This means you should shift gears at just north of 20MPH.

Here's the real question, do you NEED the extra torque provided by the low gear? Personally for sizing the gear ratio of my electric conversion, I realized that I could spin the wheels at low speeds using a single ratio, and still hit a top speed of 127MPH, which is likely quicker than I ever will get it especially considering it's been sitting for a year... but I digress. Maybe I'm biased, but in my case, I had enough torque to overcome the traction of the rear wheels and I didn't need more than that.

If you can find a use for 2000N-m, then I can see the transmission being useful, or if the sweet spot of efficiency was more towards the low speed range, like it is for most engines.

Understood. Now if you did not jump a bunch of gears as in your example by doubling the ratio and went to a more logical step, such as a 3:1 reduction and then to a 2:1 reduction it paints a very different graph.

 

[halcyon_m]

halcyon_m,
Are you also saying that for an AF that rotor skewing has no effect on characteristic current? I cannot imagine why it would not. I can believe that timing advance does not enter the picture because you are only testing the characteristic current when appearing as a generator.

I don't know where I would have made comments about the Axial Flux (AF) machines so far, but regardless, I looked up the paper you suggested earlier and it's quite interesting. I didn't find exactly what you are referring to, so I'm going to take a guess. Correct me if I'm wrong.

I'm going to assume that you're not talking about magnet position skewing (to shape or blend the BEMF and smooth out torque ripple), and I am going to assume you are talking about intentionally misaligning the rotors on either side of a single stator. In this case you would be blending the two magnetic fields so that they are partially destructive (think constructive vs. destructive wave theory) and end up making a magnetic profile that is of a smaller magnitude, producing a smaller AC voltage.

In this case, the characteristic current WOULD change. Since the characteristic current is essentially the voltage divided by the inductive impedance, both scaling linearly with speed, then you would decrease the characteristic current by decreasing the voltage.
I_characteristic= voltage/impedance = (kv[V/(rad/sec)]*speed[rad/sec])/(L[Henries]*speed[1/sec])
If you substitute henries for V-seconds/amp, then then you have the following:
I_characteristic=(kv[V/(rad/sec)]*speed[rad/sec])/(L[V-sec/A]*speed[1/sec]), cancel out the speed, the V's, the seconds, and you get a unit of current. A little math-heavy, but i'm just reinforcing the point that if you change the voltage constant of the motor (kv) by skewing the rotors, then you will change the characteristic current.

Here's caveat #1: I'm not sure you need to care so much about characteristic current in a machine like this because field weakening through the inverter, or by other means, is used to ensure that the voltage you have available to control the motor is enough to do the job. By changing the voltage constant of the motor mechanically, no inverter solution is needed, no d-axis current is needed to push back against the magnets. The voltage is already low enough.

Here's caveat#2: with the axial flux machines, the 'air gap' (region of the machine without a direct iron/magnetic pathway) is typically much larger than radial flux machines, for mechanical reasons, and the difficulty of getting a sufficient amount of iron between the windings on the stator. If this is one of the exceptions, then I'm wrong for sure, but with such a large magnetic gap the inductance is typically quite low, which pushes the characteristic current quite high. This means it would be difficult to weaken the field with the controller/inverter and so it's good that there's a mechanical means.

Side-bar: I had a phone call with Alan Cocconi, the mastermind of AC Propulsion years back and one of the things he planted in my mind was the idea of changing the air gap mechanically to reduce the voltage constant. In this case it would mean pulling the rotors away from the stators. This is something that the solar car motors do today even, if I heard correctly. All neat ideas, and for free if you don't count the cost of mechanical complexity and reliability issues. The obvious benefit is the full use of inverter current to producing torque.

 

[halcyon_m]

[/quote]Understood. Now if you did not jump a bunch of gears as in your example by doubling the ratio and went to a more logical step, such as a 3:1 reduction and then to a 2:1 reduction it paints a very different graph.[/quote]

Curiosity got the best of me and I ran the numbers for a 6:1 and a 9:1 (same as a 2:1 and a 3:1 secondary reduction with a primary reduction of 3:1)
and here's what it looks like. Very much closer for sure. All depends on what you need it for, I guess.

 

[speedmd]

Thanks for that. It does significantly extend the torque curve on the bottom end. But agree it does not appear to have as much effect on the top end. 312 vs 265 at around 1350 rpm. Speed is very different than RPM in the two. Scratching my head a bit how to best picture it. Appears that the bigger the gear jump the better. Significant differences in torque vs ground speed. Very diminished returns given the added complexity unless the motor size / current is restricted for racing specs.

 

[halcyon_m]

No problem. The RPM is the wheel speed, or vehicle speed. Same thing (except for burn-out).
I think this is a pretty decent example all around that flux weakening is almost a complete substitute for a multi-speed transmission for most use-cases. In general, it extends the usable range of speed of a motor while making full use of the power capabilities of the inverter.

It certainly is a bit of a shift from DC motors, where you have a more square torque-speed profile... or trapezoid... or triangular if you dare go full voltage a zero speed

One thing we haven't mentioned is the efficiency of a gear box. Most good gear boxes are around 95% efficient at transmitting the multiplied torque to the output shaft. Bad ones are in the neighborhood of 80% and get hot, or make a lot of noise, and fail quickly. If you look at some of the torque figures, the torque lost in a poor gearbox could more than make up for the difference in figures we were seeing for high speed torque.

So... it all depends. I like to think about it that there are many bad ways to design something. Efficient systems, where the prius is a good example, has a motor, controller, and gear train that has been designed to a high degree to ensure it is having the best trade-offs as possible. That's toyota though, and them making good cars does nothing to help you have fun on your bike, or skateboard, or personal car conversion.

 

[kenkad]

halcyon_m,
Yes, rotor skewing, not rotor magnet skewing. Our design air gap is 0.6-0.7 mm. Minimized iron in the stator to minimize mutual inductance.

Hopefully we can do some testing of the motor using a fixture in a lathe as the drive. I want to see the Bemf at various RPMs so we can properly model the drive waveform. I will bring up the characteristic current reference idea at that time. Having the Bemf wave distortions, then we can determine how to optimize (distort the drive waveform) the drive lookup table as a function of RPM. Check out timing advance as a function of RPM to maximize torque. Finally, test the efficiency of dropping three-phase groups to see if torque is linear with the number of three-phase groups driving. This is so much fun!
kenkad

 

[speedmd]

No problem. The RPM is the wheel speed, or vehicle speed. Same thing (except for burn-out).
I think this is a pretty decent example all around that flux weakening is almost a complete substitute for a multi-speed transmission for most use-cases. In general, it extends the usable range of speed of a motor while making full use of the power capabilities of the inverter.

Thanks for entertaining the comparison. The earlier comparison of the 10:1 and 5:1 shows the lines crossing harder at about 500 rpm. The taller gear taking over with more torque after roughly 40 MPH (motorcycle size tire estimate) and holding roughly a 100 Nm advantage much past the rev limit of the higher reduction alone. The one data point that looks to be a comparison 295 vs 395, is a significant jump in torque and this looks to be well over 80 MPH. Assuming simple efficient fast acting two speed setup, there may be some significant advantage on both ends of the speed range. Something like 6:1 shifted to a 2:1 may show even more of an effect. May not be worth the additional complexity but very interesting comparison.

 

[halcyon_m]

Fyi,
The torque values listed are all in pairs for the same rpm in each gear.

 

[halcyon_m]

Kenkad,
I'd be interested to take a look at your motor. What sort of application are you going for and what sort of volumes? If it gets too off topic, pm me

 

[liveforphysics]

For anyone used to analyzing transmission torque-speed curves, it makes sense that you would shift to 2nd gear above 500RPM, or about 1/6th the top speed because that's where you have more torque available. Let's assume that the top speed of this machine is 120MPH (like liveforphysics' wicked quick electric bike). This means you should shift gears at just north of 20MPH.

You may not have awareness, but I was not WOT until about 100' out. I kept letting air out of the tire and dropped a 1/10th of a second each time I did, yet still never hit a point where I could use WOT until higher speeds. More torque off the line would have just made me need to launch even more gently, and it's already tricky to launch this without spinning up. My next version of the bike will feature a drag slick and a second size6 controller.

Here's the real question, do you NEED the extra torque provided by the low gear? Personally for sizing the gear ratio of my electric conversion, I realized that I could spin the wheels at low speeds using a single ratio, and still hit a top speed of 127MPH, which is likely quicker than I ever will get it especially considering it's been sitting for a year... but I digress. Maybe I'm biased, but in my case, I had enough torque to overcome the traction of the rear wheels and I didn't need more than that. Yes. This is what the P85 Model S offers as a driving experience. Enough torque to exceed the traction of the tires effortlessly anywhere from 0-50mph if you select the option to disable traction control, and yet it still would do >130mph effortlessly.

If you can find a use for 2000N-m, then I can see the transmission being useful, or if the sweet spot of efficiency was more towards the low speed range, like it is for most engines.

With the awesome EV's on the market now improving so rapidly, mainstream folks will soon be exposed to the bliss of silent silky smooth 1spd driving. Makes every other system feel like a massive compromise in driving/riding experience.

 

[Biff]

one thing that hasn't been mentioned is that when designing a system, it is important to balance acceleration and cruising speed and efficiency. Most electric vehicles are not expected to operate at their top speed, high speed usually means higher than sustainable power levels, or extremely reduced range, it is not a big deal to sacrifice a lot of efficiency at top speed, to provide better acceleration. Different motors have different abilities for field weakening. One good report I read was the difference between the Chevy Spark EV, and the Fiat 500e . Both cars have approximately the same specifications, vehicle weight and power, but the Fiat is much more punch. This is because the way that the designers took advantage of field weakening.

https://autos.yahoo.com/news/fiat-500e-vs-chevy-spark-ev-electric-car-110034071.html

That article describes the difference in acceleration provided both motors have about the same power as "gearing" but when you get down to it, the important thing is actually the torque curve and the ratio between constant torque rpm range, and constant power (field weakening) rpm range. They are correct in stating that it is the gearing choice that provides the better off the line acceleration even though the motor has a lower maximum torque specification, but gearing has little to do with it other than selecting the correct gearing to map the motor torque curve to the wheel speeds that you want. I think I read another article that had this same comparison, but was a little smarter about realizing what was the actual difference, but it might have just been me reading between the lines.

The Renault Twizzy has something like a 1:4 constant torque to constant power ratio, i.e it goes into field weakening at 1/4 of its top speed. This gives it loads of off the line torque for the power rating of the motor, and the car is designed to operate at around 1/4 of its top speed so it ends up being super fun to drive and does what is expected of it really well.

-ryan

 

[speedmd]

With the awesome EV's on the market now improving so rapidly, mainstream folks will soon be exposed to the bliss of silent silky smooth 1spd driving. Makes every other system feel like a massive compromise in driving/riding experience.

Hi Luke

I can see more clearly what your saying with the recent plots above. Gear it and settle on the KV for what your doing mostly and let the controller do the rest. Interesting, how much torque can be added to the torque curve at the low speed end if you are often in need of "blasting off" with a smaller battery pack or pulling out tree stumps. Possibly something like a trials bike where you want explosive acceleration from a stop while your trying to clear extremely tight sections and need as much torque as you can muster, and then switch to normal mode to cruse down the road at normal traffic speeds. Certainly not every day situations.

 

[Punx0r]

Mr Halcyon,

Thank you for posting the graphical example of the torque curves for two ratios. I have recently been converted from a believer in multiple ratios to singles-speed.

This has been in large part to arguments made by Mr liveforphysics, who recently described the benefits of very low inductance (single turn) motors as a way to ensure adequate torque is available at almost any speed you wish.

One thing that would help my understanding is knowing why flux weakening is necessary? Why not just have a very low inductance motor and use current limiting, such that even at high speeds the motor is able to draw enough current (from your limited pack voltage) to produce lots of torque?

Specifically, is field weakening a good solution to better utilising current (non-ideal) available motor/controller technology? Based on the idea that it is wasted current in that it produces no work, or have I missed the point and flux weakening will always be a good idea?

 

[johnrobholmes]

From my limited experience, when a motor is built for such low inductance and high KV the controller would have much harder problems controlling the motor. This is a very good case for field weakening to increase top speed instead of higher KV or taller gearing. Taller gearing would reduce zero rpm torque at the wheel, and faster KV may cause controller failure. But in cases where the controller can take faster motors and the motor is not near heat soaking then add voltage or increase KV for better performance.

 

[halcyon_m]

Mr Halcyon,

Thank you for posting the graphical example of the torque curves for two ratios. I have recently been converted from a believer in multiple ratios to singles-speed.

This has been in large part to arguments made by Mr liveforphysics, who recently described the benefits of very low inductance (single turn) motors as a way to ensure adequate torque is available at almost any speed you wish.

I'm not familiar with the background for that sort of argument. Can Mr Liveforphysics chime in or someone provide a link?
In general, multiple turns allows the use of higher voltage and lower current to achieve the desired power. More turns means the more voltage you need to push the current required, but also means less current is required to get the torque you want. Number of turns is very much like a gear ratio for current. A motor cares about the magnetic field generated from amp-turns (number of amps in the wire, multiplied by the number of turns), and motors will provide a certain amount of torque based on this amp-turn value. If you have more turns and less current or less turns and more current, makes no difference to Mr. Motor. Changing the number of turns doesn't buy you anything out of the ether, but it does allow you to shift the operational range to what is most beneficial for your application. Power is power, no matter how you cut it, whether is 4S4P pack or a 16S1P pack you choose to drive the solution, there will be a number of turns that is best. From our perspective, we typically make due with what's available, sizing the controller and battery voltage to suit.

One thing that would help my understanding is knowing why flux weakening is necessary? Why not just have a very low inductance motor and use current limiting, such that even at high speeds the motor is able to draw enough current (from your limited pack voltage) to produce lots of torque?

Let's say you had a 1-turn motor versus a 5-turn motor. Let's also say you want to produce torque that requires 10 Amp-turns from the motor. The 1-turn motor would require 10A and the 5-turn motor would require 2A. The 1-turn motor would produce 1/5th the back-EMF of the 5-turn motor, so the 1-turn motor would be better suited for a low-voltage application versus the 5-turn motor that would be better suited to a higher voltage. Your application dictates the practicality of it all. For example, controller current ratings and wire sizes become an issue on the lower end, where controller voltage ratings and wire insulation become an issue on the top end. (and brush arcing for brushed motors). I hope this helps answer why there is no silver bullet solution to turn number. If 1 turn for a particular motor seems to work out given available controllers and wire size and battery voltage, then that's the magic number in that case, but certainly not in all cases. The short answer is that you may not have enough current available to get the torque you want if you lower the number of turns. On the flip-side, you may limit your top speed with too many turns. If you have the ability to weaken the flux, then you can extend the speed to higher ranges without sacrificing too much power. Also, low inductance motors can have high current pulsations from the controller, so high in fact that the pulses may exceed the current rating of the controller far before the average value would, requiring higher switching frequencies which lowers efficiency.

Specifically, is field weakening a good solution to better utilising current (non-ideal) available motor/controller technology? Based on the idea that it is wasted current in that it produces no work, or have I missed the point and flux weakening will always be a good idea?

Field weakening was a tool that was discovered and exploited to increase the usability of electric motors. After the discovery, some motors were designed to take advantage of this, and from the previous discussion, I hope you can see how it can be used beneficially. For example, if the number of turns of the prius motor were changed to allow the same top speed without the use of field weakening, the peak torque would be reduced by the same amount as the reduction in turns. In this case it would be 4x, so you would have 25% of the peak torque shown, so 50N-m, but it would be available at top speed. Conversely, if you left the number of turns the same and didn't use field weakening, you would only be able to run the motor to 3000RPM, 1/4 the top speed. Both of these are trade-offs I would never choose. So, flux weakening is a process that has been shown here to be very useful in providing almost the same power over a 4:1 ratio without shifting any gears. It's like electrmagnetic gear shifting, and it's exactly what has enabled the hybrid cars we use today to be what they are. If we didn't have flux weakening, the motors would have to be 4x the size they are to compensate, and I just don't think you could fit that under the hood of a compact car as part of the transmission.

I think one of the main reasons that hobby motors are not typically designed to take advantage of field weakening (aside from the increased tolerances and change in manufacturing process) is that most of these motors are intended for use with propellers, where the load profile is more like the right side of a parabola on the torque-speed graph, trending into the top right where field weakening doesn't help. It allows more to the right (higher speed) but at the expense of height (amount of torque). Passenger prop airplanes don't have multiple-gear transmissions, and some don't even have gear reductions, so it seems obvious that they are not the best application to take advantage of an electromechanical transmission effect that is afforded by field weakening. That being said, I do wish a hobby company would provide a motor that was capable of field weakening so it could be applied more easily to small electric vehicles.

Note: this entire discussion has been aimed towards AC permanent magnet motors. Induction motors, like the ones found in the Tesla, AC20 and AC50, etc., all rely on the inverter to make the field (induce a field) in the first place, so they already make use of field weakening (or making less field), some would say better than permanent magnet machines, but that's a discussion for another thread.

 

[major]

... For example, if the number of turns of the prius motor were changed to allow the same top speed without the use of field weakening, the peak torque would be reduced by the same amount as the reduction in turns. In this case it would be 4x, so you would have 25% of the peak torque shown, so 50N-m, but it would be available at top speed. ....

Isn't this an unfair example? If you were to reduce the turns in the motor to 25%; 1) you could reduce the size of the motor, or 2) you could increase the turn size (conductor cross sectional area) by a factor of 4. This second option gives the capability of running higher motor current and thereby producing higher torque, by a factor of 4.

In fact I think this is the argument from Mr. LiveForPhysics. Reducing the turns, increasing the conductor size and increasing the motor current for the same motor volume then gets you that same (original 200Nm) torque over the entire speed range (12,000RPM) without the field weakening. Yes, I know this requires a bitch of a controller.

Don't get me wrong, I am not against field weakening. But, as implied in the term weaken, you no longer have maximum utilization of that field (which amounts to motor space). This is basically why true maximum power is not obtainable in the weakened field state, IMO.

 

[halcyon_m]

... For example, if the number of turns of the prius motor were changed to allow the same top speed without the use of field weakening, the peak torque would be reduced by the same amount as the reduction in turns. In this case it would be 4x, so you would have 25% of the peak torque shown, so 50N-m, but it would be available at top speed. ....

Isn't this an unfair example? If you were to reduce the turns in the motor to 25%; 1) you could reduce the size of the motor, or 2) you could increase the turn size (conductor cross sectional area) by a factor of 4. This second option gives the capability of running higher motor current and thereby producing higher torque, by a factor of 4.

In fact I think this is the argument from Mr. LiveForPhysics. Reducing the turns, increasing the conductor size and increasing the motor current for the same motor volume then gets you that same (original 200Nm) torque over the entire speed range (12,000RPM) without the field weakening. Yes, I know this requires a bitch of a controller.

I believe it's a fair comparison. I was assuming that the wire size would be increased, though I didn't state it explicitly. That's typically what is done though to make use of the area available for copper no matter how many turns are used. In the example I presented, if you did use 1/4 the number of turns, I don't see how you could reduce the size of the motor. You still need the magnetic area (same stack length, same machine dimensions) to get the same power out of the motor. If you can spot something I've got wrong, let me know, but if you changed any dimension of the magnetic elements of the motor, then that would also change the torque and/or power rating.

Also, let's dig into the example you mentioned. For example, let's say that to get 200Nm, you need 1,000Amp-turns, and that the machine in the prius normally had 4 turns. If we reduce the number of turns to 1-turn, and increased the copper cross-sectional area of the wire 4x to fill up the usable area, then you would need 1000A to produce rated torque versus 250A in its stock form. It would be the same amount of torque, with the same amount of electrons passing through the slots in the stator, but it would be at a 4x lower voltage. To get any more torque by using more current, the windings would heat up more than in the sock case making it not a fair comparison.

If you wanted to preserve the top speed of the motor to 12000RPM, the controller would have to provide the same 650V, which would increase the power rating of the controller by a factor of 4 (650V, 1000A) as well, which is likely the largest drawback to going this route. That is to say, the physical size and cost of the converter to support this configuration.

On the positive side, if you had the battery power to support higher power levels than the stock values, then torque values in the upper right hand area would be available as well, but if not, then the converter would have to be controlled to limit power at to higher speeds so as to not over-draw power from the battery pack and possibly blow fuses.

I'm not trying to say any one perspective is wrong, but it's a matter of trade-offs in a lot of different areas and ultimately comes down to the specific goals and perceived risks are.

Don't get me wrong, I am not against field weakening. But, as implied in the term weaken, you no longer have maximum utilization of that field (which amounts to motor space). This is basically why true maximum power is not obtainable in the weakened field state, IMO.

You point out something that is very true, that the maximum field is no longer available, which, you could say is wasting iron as it doesn't utilize the flux carrying capabilities nearly as much. The trade-off for this is allowing higher speed operation, where the drop-off in torque is met with an nearly increase in speed, which keeps the power nearly constant over a large speed range. Whether the power over a given speed range is exactly the same or falls off is in large part due to the pairing of the current rating of the inverter and the characteristic current of the motor. If they are mismatched at all, then the power is not constant above the base speed (corner speed, 3000rpm in this case).

So this brings up a good point here: Motors are designed magnetically to provide a given amount of torque, but they are powerful when they can spin quickly. Speed constraints come in various forms, but mostly mechanical concerns of it flying apart, so there's a practical limit to power of the motor. The other two limitations to power mentioned here are the source (battery, usually) and the controller. All of these things cost money, and weigh more, as you get to higher power levels and I see that as a major reason why flux weakening is popular versus simply changing the turns to a low enough value so you never run out of voltage... but that doesn't mean it's not a viable option.

 

[liveforphysics]

In fact I think this is the argument from Mr. LiveForPhysics. Reducing the turns, increasing the conductor size and increasing the motor current for the same motor volume then gets you that same (original 200Nm) torque over the entire speed range (12,000RPM) without the field weakening. Yes, I know this requires a bitch of a controller.

You got it Major.

Shift the burden to the motor controller. The motor controller is the piece of this puzzle making drastic performance improvements in the ability to provide continuous current, and will only get better with each new astounding generation of ever lower RdsOn mosfets with gate designs that enable faster switching with reduced losses.

Think about it. Would you rather spend $200 in parts and add 5lbs of weight to make a controller leveraging modern MOSFETS that does triple the phase current (and wind the motor with a few less turns), and get to be a 1spd with incredible acceleration performance and power potential at all points in the RPM curve seamlessly, OR, spend that $200 in parts and 5lbs to add stages of mechanical loss and complexity and add a bunch of new failure modes and NHV etc.

If you are making an EV, and you get to design your motor for your application, if you ever end up with multiple gears in your system, you designed the wrong motor for the application. It's really that simple.

If you are making a motor designed for an RC helicopter work both for pulling your loaded cargo-bike up steep hills as well as assist at 40mph on the highway, yes, you will likely find productive use of the crutch of multiple gears. It doesn't mean multiple gears are a good idea as design-intent for something, this would be like designing something that functions only because of some very non-ideal band-aid solutions being added. It simply means, if you select a motor type far enough away from your applications ideal motor, a transmission may enable making the wrong motor choice into something that still provides a level of function that meets your needs.

 

[onloop]

I started this thread to find out more info on squeezing additional performance out of my electric skateboards builds.. as an electric skateboard designer one of the major constraints with eboard builds is size & weight, of... um, well... everything! as we are strictly limited by the VERY SMALL gap beneath the deck & the road... so squeeze we must

Can I firstly say the technical information that is flowing in is fantastic so thanks ES community...YOU ROCK!

So far what I gather from this thread is that it would seem as though the little "hobby type" Brushless outrunners us skateboard builders are using just aren't designed to take advantage of the

"electromechanical transmission effect that is afforded by field weakening"

... I loved your wording so much I just had to include your quote!

So it seems like in my specific application & considering the design of the motors - focussing on; higher volts, higher KV (highest feasible KV), greater "single gear" reduction, biggest motors.... would be the most efficient enhancements!.... not so much the "electrickery" side of things..

..............However in saying all of this, my controller definitely has the ability to adjust the "motor timing degrees"
see bottom / middle:

& The manufacturer assures me that these settings will increase the speed of a "Hobby style" BLDC motor...... So I suppose i just need to rig up a electric-skateboard-centric-style-dyno and TEST TEST TEST.... until i find the absolute best outcome......

 

[Punx0r]

I'd like to echo onloop's comments and thank everyone for contributing to the discussion in this thread. I have learnt a lot!

 

[parabellum]

Found a lot of missing needed puzzle peaces in this tread and kind of starting seeing light at the end of the tunnel. This is serious stuff for sticky.

 

[zombiess]

Onloop, do not confuse field weakening with timing advance.

That screen shot shows typical rc programming and those controllers only allow you yo set a fixed advance. While this will increase the speed of your motor you will also lose torque and efficiency at lower Rpm. With field weakening the advance is variable and only happens when the controller. Sees bemf from the motor.

The rc style controllers use something called unipolar pwm. They are trapezoidal control and "dumb" because they do not know anything about phase current.

I think I got the above correct, its early

 

[halcyon_m]

Onloop, do not confuse field weakening with timing advance.

That screen shot shows typical rc programming and those controllers only allow you yo set a fixed advance. While this will increase the speed of your motor you will also lose torque and efficiency at lower Rpm. With field weakening the advance is variable and only happens when the controller. Sees bemf from the motor.

The rc style controllers use something called unipolar pwm. They are trapezoidal control and "dumb" because they do not know anything about phase current.

I think I got the above correct, its early

I think Zombiess is dead-on here. It's tough to know how the software is implementing this timing advance, but suffice to say, it could be spanning the range of possibilities from very clever (which makes me wonder why they would even ask for the angle in the first place) to very dumb, which is what Zombiess is implying.

If it's simply that it looks at the halls, or BEMF zero crossing to reference the commutation angle offset, then you are essentially tuning proper operation (where the voltage waveform lines up perfectly with the current waveform) to a given frequency. Any faster or slower than that frequency (or speed, rather), then the voltage and current become less aligned.

Like above, when I mentioned that motors care about amp-turns, and not much else, from a magnetic perspective, getting the current to line up with the voltage waveform is key to making torque. For example, get it to line up opposite to the voltage waveform and you will have turned your motor into a generator. Get it half way between and you have a lot of rumbling and shaking in the motor as torque goes a bit of forward and a bit of back, but a whole lot of zero average torque.

To get the current waveform to line up, to make your motor 'motor' (forward torque), the voltage from the controller has to lead the voltage from the motor, so that by the time the voltage from the motor has reached a peak value, then the current will have had time to build up to its peak value as well. The advance in degrees is based on speed though, because it's really just a constant time value needed for this head-start of voltage to get the desired current. However, the ideal head-start angle becomes larger and larger as the speed increases. A good controller would be able to sense this and vary the advance angle based on speed, much like an engine ignition timing advance, or a more continuous version of VTEC, for those that are familiar. Unfortunately, I don't think this is a good controller, so you do what engine builders do with their intakes without VTEC, and they just select the intake runner length to improve the performance of the engine in an RPM range where they need the power the most.

Sorry for the ramblings, again, but this is a forum, right? Anyway, I wouldn't suggest trying to change the timing degrees purely for top speed without also testing the detriment to low-speed behavior. You may end up with too much of the rumbling and grumbling motor behavior I talked about before when the controller voltage is really out of alignment with the motor voltage. Other folks here with more experience with RC motors and getting the halls hooked up incorrectly may have more anecdotal advice on what to watch out for.