If you actually know anything about electric motor technology, and you don’t hate me yet? Buckle up…this article is going to be a bumpy ride. I think I have a firm command of the terms that define the more common aspects of electric motors (found in part-1 of this series, click here), and now…I’ll try to explain some of the more obscure aspects of their design and application (as it pertains to electric bikes).
If you like a lot of pictures to make the explanations more clear? this article may not be for you…here’s an index.
Phase/Battery current ratio
Swept Magnetic Area / SMA
Tangential Magnet Speed / TMS
Magnet Speed per Meter [traveled] / MSpM
Phase/Battery current ratio
Most common controllers allow the owner to adjust the motor-phase / battery amps current ratio. If you don’t know what that is, the motor-phase current can be set a little higher than the battery current, and 1.5:1 seems to be pretty common.
From madin88, in Austria:
“…High phase amps = good acceleration from zero to mid speed
High battery amps = good acceleration from mid to top speed
For higher efficiency, lower phase amps is better…”
From NeilP, in the UK:
“I would say…use XPD software rather than Lyen / Keywin…then, you can turn block time down to 0.1 seconds”
From gwhy! in the UK:
“The way I have adjusted it in the past is:
Set the battery current to something safe, like 40-50A (depending on the type of battery), set the over-current detection to 0.0s (this is very important), set motor-phase to approx 1.5 times the battery current (1.5:1).
You need to attach a watt meter. On a slightly uphill road (maybe a 2-4% grade), accelerate hard (from a dead stop) up to top speed, or until it stops accelerating but throttle is still at WOT (this may take a while, so you need a longish stretch of road).
Then check the watt-meters’ max current pull. If the max current pulled is less than your set battery current, then…you need to increase the phase current, maybe by 10A. Or, if the set battery current is reached, then reduce phase current by 10A. You keep increasing or decreasing the phase current until the max current pulled from the battery is what you have set it to. Once you have found a battery/phase ratio where the max battery current is always reached? Then this will be the ratio you would use when increasing the battery and phase current together.
If the phase current is too low, then you will never reach full speed or max battery current. My controllers use a ratio of around 1.7:1 for the motors, gearing , my riding style and total weight of my bikes
Another method I have used, is set at a much lower battery limit (maybe 20A), and a phase limit of 30A with the wheel off the ground go WOT and then start applying the brake to slow the driven wheel down (loading the motor). When the wheel is approximately half the max speed, watch the battery current on the watt meter. This should go to your set max battery current, and should stay there as the driven wheel gets slower and slower until the controller cuts out (locked rotor fault protect).
The same applies when you increase or decrease phase current until you see the set max battery current limit hit, always just before the controller cuts out. Each test needs to take around 5-seconds, so it’s a much quicker and easier method of finding the optimum phase setting for the motor. Once you have found your optimum phase ratio you can always turn the battery current down and keep the phase limit at the optimum”
“if you go too high with the phase current, you could pop your controller, if the motor is really chugging and the throttle speed is being dragged down too much. A standard 12-FET controller should be OK for around a safe max of 150A phase current…if I was you, I would play it really safe and do no more than 100A phase current.
So, if your Battery current is 50A, set phase current to no more than 150A (3X battery current) or be extra safe, and use 50A battery and a motor-phase of 100A (2X battery current ). As NeilP said, set the block time down to zero, this will should limit the current as fast as possible if an over-current situation occurs. The optimum settings will just be the least stressful (but still working 100% correct)…”
From AlanB, in the USA:
“…[more] Phase current makes [more] torque, up to magnetic saturation. Then, it still makes [more] torque, but…it increases at a much lower rate [per added amp]. It is best to set max phase current to no more than the point where magnetic saturation begins (and is compatible with wiring and connector capacity), though…you might want to set it slightly higher, but heat is going to be a problem if you run at that level very long. Don’t worry about the ratio, worry about the max current. Calculate the I-squared-R heating and think about your motor dissipating all that heat…
Phase current “squared” makes heat in the motor. So, you will quickly overheat with high phase currents [that are above saturation] for only a modest improvement in torque, because the square grows so fast. The battery current determines the max power, and should be set for battery, BMS, and wiring capacity (or less).
I set the phase current to control the front wheel lift (if nothing else limits it first). No point in throwing the front wheel skyward too quickly, that’s just wasted torque….”
Swept Magnetic Area / SMA
I don’t know the proper term for this, so…I just invented this term, and I’ll use it until I stumble across the correct term (if you are a professional motor-design engineer, email me. Please understand that, I will still ignore you, but…at least then I can say that real engineers emailed me).
I needed a way to compare two similar hubmotors, which both use the same common off-the shelf lamination, which results in a 205mm diameter stator. I already covered my “rogues gallery” of hot rod hubmotors in this article on large Direct Drive (DD) hubbies.
So here’s a direct drive hubmotor list (off the top of my head) to help explain what I mean. All of these motors have a 205mm diameter stator, and the numbers listed below are the width of their stators.
28mm, MXUS 1000W
35mm, Edge 1500W (click here to read about this motor)
45mm, MXUS 3000W (click here to read about this motor)
50mm, QS 205/50H V3
The wider the stator, the more copper mass (in wire) is wrapped around the stator-teeth, which results in a greater ability to use higher amps without overheating. However, even if they are all being fed the same exact amount of volts and amps, the larger magnetic area that results from a wider rotor magnet (which interacts with the stator electro-magnets) will result in more wheel-torque.
(edit: a larger diameter motor, like the QS 273, adds a lot more weight, and also shortens the available spoke length)
I’ve already listed the width of the magnets, so…next comes the circumference. Of course, circumference would be “the diameter times Pi”, which for our purposes can be rounded-off to 3.14 (for the OCD among us, 3.14159265 is more exact). This means that 205mm X 3.14 = 643.7mm, and…this is the circumference of the magnetic “air gap” where the electrical power is converted to rotary movement, and produces torque.
This common motor-stator circumference multiplied by the width of the various stators, will give us the active magnetic area in squared millimeters (I know there are small gaps between the electromagnets, but they are fairly equal between all of these hubs, so their area is negligible in these calculations, when comparing one motor to the other).
There are 100 square millimeters in one square centimeter, so moving the decimal point over two spaces will result in the motors’ useable magnetically-active area being listed in square centimeters.
The rule of thumb has been that;…if your motor is getting hot under your loads, you need a bigger motor. If it’s running very cool, you could probably get by with a lighter and less expensive motor. If it only gets warm, then…it’s “just right”
But now? this calculation can show you the ratio of how much more torque (or less torque) a given motor has, compared to another…IF…the permanent-magnet strength and input watts to the stator are the same.
180.2, MXUS 1000W V2
225.2, Edge 1500W
289.6, MXUS 3000W V2
321.8, QS 205/50H V3
Tangential Magnet Speed / TMS
This is a real engineering term, not to be confused with the rest of the bullshit terms I used in this this article.
In casual conversation, “motor speed” is often referring to the RPM’s (Revolutions Per Minute). As a result, motor design engineers had to invent a phrase that specifies how fast the permanent magnets in the rotor are passing by the electro-magnets in the stator (which can be turned on and off as needed, in order to make the damn thing spin).
I try to find the proper term for anything that I feel is important enough to write about, and TMS is one of those instances where I accidentally stumbled across something when reading a lot of obscure technical papers on motor design. (*most of which I still don’t understand. Edit: change “most” to all).
[This section of this article has exceeded the maximum allowable “word per picture” ratio, so…here is a 4-minute video of kittens and dogs]
To try and make this as simple as possible: A large-diameter Direct Drive (DD) motor that is spinning at “X” RPMs (like a Magic Pie?), has a much faster TMS than a small diameter DD hub (like the rare Crystalyte G-series, or a Conhis?). Imagine both are spinning at the same RPM’s, but…the larger-diameter motor has a faster / higher TMS. So…should the magnet speed be measured in feet per second, or meters per second? (here’s a hint: Benjamin Franklin liked bangin’ French chicks)
A common 26-inch diameter tire at 28-MPH (45 km/h) is spinning at 362-RPMs, and…if using a DD hubmotor with a common 205mm diameter stator (643.7mm circumference, listed below), the TMS is 0.6437 meters per revolution X 362 RPM’s, or…233.0 meters per minute (equal to 764.5 feet per minute).
“Feet per minute” would provide more discrete steps in the systems we often find ourselves using for electric bikes…however…with the global trend towards the metric system? I think meters per minute is the metric that will prove to be the most useful in calculating TMS between one motor and another.
Saturation (max amps?)
There’s probably a way to form an equation so you can calculate ahead of time exactly how many amps you can feed a given DD hubmotor without overheating it, however…The “fast and dirty” way to find out how many amps you can shove down it’s throat (before spending a lot of riding-time with the stators’ copper mass being over-saturated) is to log the amps you used, and the temperatures that result.
Brushless motors can operate in a wide voltage range, which is part of their appeal. If you upgrade the battery and controller, you can take your system to a higher voltage, and the motor doesn’t mind. Anything above 20S is pricey, and anything above 24S is going to get VERY expensive for the controller and charger. Using 24 cells in “Series”, at a nominal 3.7V per cell, has its “average” at 88.8V nominal. A 24S battery fully-charged to 4.2V per cell means a high voltage of 100.8V, hot off the charger.
If you are worried about safety, and you don’t like the fact that higher-voltage electricity can penetrate dry skin, stick with 14S or lower (51.8V nominal, and 58.8V fully charged), and…the lower, the safer. If your skin is sweaty and salty? At that point the volts don’t matter, because even low volts will fry your finger if the system you’re touching [without gloves] has high amps.
But…let’s get back to amps and saturation. It’s actually not a bad design goal to go into saturation a “little bit” once in a while. However…if your user-profile spends a lot of time in saturation? Your motor will be soaking-in more heat than it can shed.
This is less of an issue for mid drives, because the operator can control what gear the bike is in, and that controls how many RPM’s the motor is spinning at [hint: motors love high RPMs].
However, a large DD hub is by definition always at low RPMs when starting out at a red light, when it turns green. I mean that, if you live in a relatively flat area, it’s not a bad thing to saturate the copper mass with high amps for five seconds or so (when accelerating), as long as you allow a full minute of just cruising, in order for everything to cool off a little.
Where is the point of saturation? There are many factors, but…if you really want to know, get a temp probe. Start at reasonably low amps. Raise the amps 5A at a time, run your usual commute, and watch the max temps. Every time you raise the amps, you will “feel” more acceleration, but…there will come a point where…even though you still feel more power? You also notice a significant bump in heat. That point is the “quick and dirty” point of saturation.
The reason this is worth taking the time to sort out is because…my hills might be longer and steeper than your hills. Why not get the maximum performance that your components are capable of, if it only takes a minor adjustment? You can likely go into saturation just a little bit once in a while, but…if you go too far, and for too long? Any one of several things can happen from the heat-soaking that results.
- Rubber seals on axle-bearings might deform/melt, and the bearing will die soon after that, from grit and water-intrusion
- Weak Chinese wire-joints will re-melt at low temp due to using cheap “cheese grade” solder instead of the good stuff.
- Generic Hall-sensor dies, [buy three authentic Honeywell Halls to replace all of them]
- The clear enamel on the copper wires in the stator will bake off, and wires short out. The wires will look dark and they smell bad. It’s called turning your motor into “wall art”
- Cheap Chinese neodymium magnets will begin to de-magnetize. The top speed will go up a little (altered Kv), but…the power will go down at all RPMs
Due to these reasons, keep the motor below 200F / 93C. If your DD hubmotor is getting hot too often? Add more copper mass by going to a larger motor. You could also add cooling mods, like oil-cooling, and ventilating the side-plates. Plus maybe upgrade to a model of hubmotor with an aluminum stator support to act as a heat-sponge…and maybe add Ferro-fluid with motor-rim fins? but…that info will be in an article for 2017.
Concerning temps, brushless motors can take a lot of abuse, but…the battery and controller work better when they are warm, however…it is bad for them to be too hot, or too cold. There are many factors that affect the actual “best” temps, but…100F to 120F is a really great zone to aim for (38C-50C). Warm is good, hot and cold are bad.
Magnet Speed per Meter traveled (MSpM?)
I know this will sound arrogant, but…This is another term I just made up out of thin air. I created it to help me mathematically compare a given Direct Drive hubmotor (DD), to a geared hub, and then to a mid drive.
If the input watts are the same, the worst-case scenario would be a small-diameter DD hub in a large diameter wheel (Crystalyte 408 in a 29’r?). It will struggle to get up to its designed top-speed. And…the method that the controller will TRY to use, is to keep applying high amps until it gets there. Since this combo will accelerate slowly, the continuous high amps will result in frequent overheating.
The next step in increasing the magnet-speed, is to switch to a smaller diameter wheel, or…upgrade to a larger diameter DD hub, OR…maybe both. In part-1, we saw that many motors are now available in a variety of Kv’s, which means it’s designed top speed can differ, even when the same volts are applied.
If we switch to a larger diameter of DD hub, away from the smaller 408 (like any one of the 205mm diameter stator’ed motors listed above), and then swap it to a smaller diameter wheel (like a 20-inch bicycle rim, or the very similar 16-inch moped rim), we can also then specify a much faster Kv version of that motor. The result is that…the final form of this theoretical wheel can have the same top speed as the “bad” example, however…it will have a much higher magnet speed, per meter that the bike travels.
Geared hubs have a distinct advantage over a similarly-sized DD hub at any power level below approximately 1500W. Inside the motor housing there is a mechanical gear-set, which allows the motor to spin about 5 times for each wheel revolution. This allows a small and light motor to make a LOT of torque from an average amount of watts.
Geared hubs all have a one-way clutch inside too, and that is the component that sometimes fails when any power above 2000W is applied firmly. If you get one of the larger geared hubs (like the BMC), a common power level is 48V X 25A = 1200W, and the most I would risk using on those is roughly 60V X 30A = 1800W (with an added temp sensor, and…maybe oil-cooling).
[You may notice that in those two examples, the amps are about half the volts. This is because geared hubs have a poor heat-shedding path, so amp-heat will be the limiting factor. If you want more power than 1800W (using a geared hub), I’d recommend going to a 2WD]. If you want more power from a motor that has poor heat-shedding, raise the volts more than the amps…
Getting back to MSpM, if…you are limited to 1800W, a geared hub is lighter (compared to a similarly-sized DD hub), and the 5:1 gearing means the very fast-spinning motor will run more efficiently. How much more? The popular BMC has a stator that is 135mm in diameter, so 135 X 3.14 = a circumference of 424mm (times five geared revolutions) = 2120mm of magnet travel per wheel revolution.
Let’s compare that to the common DD hub circumference of 205 X 3.14 = 644mm of magnet travel per wheel revolution…which equals 644 (DD) vs 2120 (geared). If the common geared hubs were matched up against a DD hub that had the same diameter and width of stator (SMA), it would be a huge mismatch, with the large geared hub winning. It is only when you begin using more than 2000W that a DD hub begins to provide the type of performance that a geared hub cannot compete with (due to limited heat-shedding).
Maybe someday a vendor will make a geared hub with more rollers in the clutch (to lower the “load per roller”, which would raise the clutches’ power rating) and also widen the stator to the point where it can only hold a single-speed freewheel. Then…it could provide more torque per the same input watts. However…right now? the existing examples are capped somewhere near 1800W per motor. Even at the same input watts (heat-limited), a wider stator would provide more SMA torque. If such a geared hub was designed from the factory to have oil-cooling, it would not leak (I worked for a dozen years in the hydraulics industry, trust me on this)
[I realize that common DD hubs have more leverage due to the length of the magnetic gap radius, but I’m trying to keep this as simple as possible for now. A 205mm diameter has a roughly 30% more radius “leverage” compared to a 135mm diameter stator from a geared hub]
Next…The most common way to increase the motors “magnet speed per meter” of wheel travel is to configure the system with a non-hub motor. It would typically be mounted in the frame, and connected to the rear wheel by a chain. If you want an even higher magnet speed (per wheel travel), you could add a jackshaft to form a dual reduction. This results in the same wheel-speed you designed the original system for, but…the motor can spin even faster.
Up until somewhere around 2000W, there are significant benefits to having the motor drive the multiple sprocket-sets of the rear wheel (with the rear wheel having at least 3 gears minimum). Then, when using any power level above 2600W, the rider will experience significant driveline wear, if…you are using bicycle components. Using #219 Kart chain, and #420 motorcycle chain (and sprockets) are the hot rod options that are the most popular for extra-high power, and you can’t run those through a derailleur.
Non-Hub motor systems
Up until somewhere around 2000W, it is a huge benefit to efficiency, to give the motor the use of the Bicycles’ gears. A common method is to use a Bottom-Bracket (BB) crankset with two chainrings. The motor drives one chainring, and then…the other chainring drives the gears on the rear wheel. However, if you don’t want the pedals to spin all the time? you have to incorporate a freewheeling crankset, like the kind used in a trials bike.
You ‘could’ do this at power-levels above 2000W, but…the chainrings and chains will wear out faster, and sometimes even break. For power levels above 2000W (and also up to “the sky is the limit”?), you might want to run a one-speed chain to the rear wheel. I have seen set-ups like this driving the LEFT side of the rear wheel (recommended), which leaves the stock pedaling system intact…or…with a chain driving the right side of the rear wheel.
If someone mounted a DD hubmotor on the frame, and then attached it to the rear wheel with a 1:1 ratio…would there still be any benefits? Yes!
With the unsprung weight of the motor being taken out of the wheel, it would make the rear suspension more nimble and responsive. Also, it would be easier to incorporate some cooling mods with no spinning spokes in the way.
However, I feel the biggest benefit to this configuration is in the fact that you can take a cheap and robust DD hub, and run a completely different ratio to the rear wheel. Remember earlier how I touted the benefits of the geared hub over a DD at power levels up to 2000W? Imagine a DD hubmotor running a 55T/11T (5:1 ratio) #219 chain to the rear wheel. It would have all the benefits of a common geared hubmotor (with its internal 5:1 gear ratio), but…it would NOT have a 2000W power limit.
Heat would not be much of an issue, so the Leafbike 1500W motor (with 35mm wide stator, adding Ferro-fluid) could provide an insane amount of wheel-torque, if there was even just a 3:1 ratio (or more) between the motor RPM’s, and the rear wheel RPM’s.
“…Every direct drive ebike hubmotor is also acting as a generator as it turns. The separation between motor/generator is merely if you happen to be feeding it power…or… drawing power from it’s rotors’ induced changing magnetic fields creating back-electromotive-force in the windings…”
For some reason, BEMF is how the most common controllers sense if the motor is at its max system RPM, or if the controller needs to add more amps to get it up there.
Inductance, and voodoo magic
I’m probably going to get some hate mail on this because I really don’t understand inductance, but…I’m going to add this here anyways. Inductance is measured in a unit called “Henries” (I’m not making this up…unlike the rest of this article). And…if the amount of inductance is very small, it is measured in “micro-Henries” (μH, or uH).
According to this video here, when any “iron-core” electromagnet coil is energized, it creates a strong and focused magnetic field (which we want). However…when we suddenly de-energize it?…the collapsing magnetic field creates a very brief voltage spike (in that phase-group) which can possibly damage sensitive electronics.
This is similar to how a car from the early 1960’s uses two concentric coils to convert 12V into a high-enough voltage, to jump across the air-gap of the spark-plug tips. Old car spark plugs need lots of coil-inductance feeding them, but…electric bike controllers don’t.
“Inductance is measured in Henries and it evens out the current pulses to normal levels. Desirable motor inductance is around 50-150μH. Figures lower than 20μH, apart from causing excess heat, start to make serious problems [voltage spikes/overshoot].
Some large RC motors and, especially, various “cheap” coreless motors have inductances of less than 5μH. Also, you don’t want inductance too high, because it will give problems at high RPM’s. You can add air core inductors to a low-inductance motor to increase inductance. [building the magnetic field takes some time, even if it’s measured in fractions of a millisecond. But, at high RPM’s, the coil is energized for such a short time, that there is less current creating the actual work of moving the rotor, because some of the time is wasted during the building up and drawing down of the magnetic field]
When inductance is too low, you can run into a situation where current limiting does not kick in fast enough and overshoots. The lower the inductance, the faster the overshoot rises” [end quote]
Do I understand inductance? Hell no! I just read that if you don’t get the inductance right, you can fry your expensive hot rod controller. Always match (and adjust) your controller to the motor you want to use. I am always a “late adopter” of new tech, and I only use components that are proven to work well together (using accepted settings).
Experimenting can be fun, but…but trust me when I say that I have found out…experimenting is always expensive!
V = I × R I = V / R R = V / I
R = Resistance [in Ohms], discovered by a German, Georg Ohm
V = Voltage, discovered by an Italian named Alessandro Volta
I= Current [in Amps] from the French “Intensité de Courant”, from a French scientist named André-Marie Ampère
OK, so…a German, an Italian, and a Frenchman walk into a bar, and…the bartender says “Is this some kind of joke?” (in this joke, the bartender is from the USA, the only country mentioned that put a man on the moon just to prove that they could, and then…refused to build the moon base we were promised)
In plain english, raising the volts increases power without a huge increase of heat (however it increases the motor RPMs if the Kv isn’t changed, and the battery pack might also be larger and more expensive), but…raising the current/amps will dramatically increase the motors’ torque, and also the heat.
Delta / Wye configuration
Each coil (or group of coils) that forms an electromagnet, has two wire-leads coming from it. Since current will only flow through a complete circuit, you can join one leg from each of the three coil groups. Doing this allows you to configure only three wires (out of a possible six) to exit from the hollow axle. And…doing that allows each wire to be as fat as possible within the allowed space. As a result, the common controllers have only three phase leads (typically colored BGY, for Blue Green Yellow ).
Now? most hot rod hubmotors have the motor-phase leads exiting the axle from a slot near the center, instead of at the axle-tip. This makes the axle much stronger (which is important in a hot rod)
Wye is sometimes called “Star”, because if you are using a motor with five coils per phase-group (which is common in industry), the electrical engineering drawing will look like a 5-pointed star.
Delta will produce a motor that has 1.7:1 more RPMs per volt, compared to that same motor configured in Wye/Star. Delta is faster. Wye runs a little cooler.
Delta will have a 1.7:1 faster kV than the same motor terminated in Wye, but…Wye pulls less current on start-up. A long time ago, some ebikers pondered Delta/Wye switching “on the fly” to help their low-current battery packs to provide better performance (Wye at low speed, Delta at high speed).
But now? high current batteries make that a thing of the past. Which is better? Most ebike DD hubs (that you can actually buy) are terminated in Delta, but…it all depends on what you are going to do with it.
I thought about trying to learn more about hysteresis, iron losses, core losses, phase resistance, reluctance, permeability, copper losses, etc…but…this article is way too long already, and those motor characteristics will probably never affect any motor system you are likely slap together in your garage, so…
Please address all “hate mail” to inmate #41, Kansas state correctional facility for the mentally unstable. You will get extra points if you add a pack of Marlboro’s. If I get enough of those, I can trade them for a huge DIY tattoo of “Ebike Lyfe” across my back.
Or…you could go to our Facebook page, and start a conversation with me. Don’t waste your time by simply posting that I am an idiot…I already know that. Tell me exactly HOW I am an idiot, so I can learn.
Written by Ron/spinningmagnets, November 2016