Latest / eMotors: Electric Revolution / The Axial Flux Revolution: Redefining Electric Propulsion Engineering
Transcript
- 0:10What if I told you that the the motor in your electric car could
- 0:14basically be shrunk down to the size of a standard dinner plate,
- 0:18but still packing a punch to completely out accelerate a Top
- 0:22Fuel dragster? I mean honestly, if you told me
- 0:24that, I'd say it sounds like some pretty aggressive
- 0:26exaggeration from APR department.
- 0:28Right. It sounds totally made-up, but
- 0:30well, welcome to the deep dive, everyone, because today our
- 0:33mission is to decode this radical, really fundamental
- 0:38shift that's happening right now in electric vehicle technology.
- 0:41Yeah, we are unpacking the rise of the axial flux motor today,
- 0:45which is it's a fascinating topic.
- 0:48It really is. And for you listening, we've
- 0:50brought an incredibly dense stack of sources to pull from so
- 0:54we can actually understand the physics here.
- 0:56We're looking at a highly technical academic paper from
- 0:58the University of Nebraska. We also have the performance
- 1:00specs and some frankly wild hype videos from YASA.
- 1:04Right, the engineering firm that Mercedes-Benz just acquired.
- 1:06Yeah, and we've got a beautifully designed Italian
- 1:09engineering presentation on motor cooling.
- 1:12Plus a really brilliant, highly skeptical tear down blog post by
- 1:16an electric motor design engineer, a guy named Antia
- 1:20Leahy Coynan. Yes.
- 1:22So to even appreciate why this, you know, dinner plate sized
- 1:26motor is disrupting the whole industry, we kind of have to
- 1:29establish the baseline. Right.
- 1:30Exactly. For decades, the entire EV
- 1:32industry, and really the broader industrial world as a whole, has
- 1:35relied almost exclusively on what we call radial flux motors.
- 1:39And visually, if you picture a radial motor, it's cylindrical
- 1:42like it's elongated. It's shaped like a sausage,
- 1:45basically. A sausage, right?
- 1:46So we're essentially looking at why an industry built entirely
- 1:49around these elongated metal sausages is suddenly like
- 1:53pivoting to ultra compact flat pancakes.
- 1:55That's a great way to put it in the Italian presentation
- 1:58visualizes this shift perfectly. These new axial flux designs the
- 2:01pancakes. They cut the total thickness of
- 2:04the motor by up to 80%. 80% that is a massive space saving, but
- 2:09the shape change is really just a byproduct of the magnetic
- 2:11topology, isn't it? Like we need to look at what's
- 2:13actually happening to the electromagnetic fields inside
- 2:16the casing. Right.
- 2:17So in a standard radial motor, the sausage the magnetic flux
- 2:20travels perpendicular to the axle of rotation.
- 2:23OK, perpendicular. Yeah.
- 2:25So the magnetic forces are basically pushing inward and
- 2:27outward radially between the spinning rotor and the
- 2:31stationary stator that's wrapped around.
- 2:32It OK. So if we think about this
- 2:34mechanically, it's a question of leverage.
- 2:37A radial motor is applying its magnetic force pretty close to
- 2:40the center shaft. Exactly.
- 2:42It's a bit like trying to spin a heavy merry go round, but you're
- 2:45grabbing the bars right near the center post.
- 2:48You have to exert a tremendous amount of energy to get the
- 2:50thing moving. You really do, And by contrast,
- 2:53in an axial flux motor, the electromagnetic flux travels
- 2:57parallel to the axis of rotation.
- 2:59So the magnetic forces are moving side to side.
- 3:01Yes, side to side, straight down the length of the shaft.
- 3:04And because of this layout, the rotating discs sit face to face
- 3:08with the stationary stator, so the magnetic interaction happens
- 3:11at the outer diameter of those discs.
- 3:14Oh wow. OK, so taking the merry go round
- 3:16analogy further, the axial motor is like standing on the grass
- 3:19and pushing the merry go round from the very outer.
- 3:21Edge, which is so much easier. Right, you maximize your lever
- 3:24arm, you get substantially more rotational force or torque for
- 3:28the exact same amount of electromagnetic push and the.
- 3:31Underlying geometric advantage of that outer edge leverage is
- 3:34just immense. But here's the thing that often
- 3:37surprises people when they study motor topology.
- 3:39It's. Not a new idea, is it?
- 3:41No, not at all. This isn't some new concept born
- 3:44in a modern Tesla lab. Michael Faraday's very first
- 3:48electrical machine in 1831 was essentially an axial flux
- 3:52design. 1831 That is wild. And Nicola Tesla even patented a
- 3:56highly detailed disk motor back in 1889.
- 3:59OK, wait, I have to stop right there and just look at the
- 4:01timeline. If Faraday and Tesla understood
- 4:03the leverage advantages of this parallel flux over a century
- 4:06ago, why on earth did the industrial world spend the next
- 4:09100 years standardizing the less efficient radial sausage?
- 4:12That is the big question, and the answer is that there was a
- 4:15massive showstopper. It came down to mechanical
- 4:18deflection. Deflection meaning things
- 4:21bending out of shape. Right.
- 4:23For any electric motor to operate without just destroying
- 4:25itself, you absolutely need an air gap between the spinning
- 4:29rotor and the stationary stator. They can't touch OK, but in an
- 4:33axial motor you have these large flat metal discs facing each
- 4:37other and they are generating A terrifying amount of magnetic
- 4:41attraction. Oh I see, because they're flat
- 4:43wide plates the magnetic field isn't just trying to spin them
- 4:46right, it's desperately trying to snap them together like 2
- 4:49magnets on a fridge. Exactly, and maintaining A
- 4:52microscopic, perfectly uniform air gap across a large spinning
- 4:56disk without the metal bending or warping under that immense
- 5:00attraction that. Sounds like a metallurgical
- 5:02nightmare. It was for decades,
- 5:05manufacturing a housing that was rigid enough to keep those
- 5:08plates from crashing into each other, but still light enough to
- 5:11put in a car. It was just economically and
- 5:14physically prohibitive. So it wasn't a failure of the
- 5:16physics, it was a failure of material science.
- 5:19The manufacturing tolerance is back then just couldn't handle
- 5:21the deflection. Precisely.
- 5:23And the breakthrough that finally solved this wasn't
- 5:25necessarily a structural one, it was actually magnetic.
- 5:28Oh, really? How so?
- 5:29The commercialization of super strong rare earth permanent
- 5:33magnets, specifically neodymium magnets.
- 5:37They completely change the math. Because they're so much
- 5:39stronger. Yeah, because these magnets
- 5:41generate such a ferocious field, engineers can design machines
- 5:45with slightly wider, much more forgiving air gaps.
- 5:48So you can leave enough physical space to accommodate a tiny bit
- 5:52of deflection without the plates ever making contact.
- 5:55Right, which suddenly makes mass production viable.
- 5:58The materials finally caught up with Tesla's blueprints.
- 6:01That is incredible. Which I guess brings us to the
- 6:05modern specs from YASA. Yes, the YASA numbers.
- 6:07Because their prototype numbers claim a peak power density of 59
- 6:11kilowatts per kilogram, which to me seems almost entirely
- 6:16detached from reality. It is a truly staggering metric.
- 6:19I mean, for context for you listening, a standard EV motor
- 6:22right now might hit what, 15 to 20 kilowatts per kilogram?
- 6:25Yeah, around there. And a Top Fuel Nitro methane
- 6:28dragster is putting out roughly 30 kilowatts per kilogram, So
- 6:32Yasa claiming 59 requires A fundamental reimagining of how
- 6:35much power you can extract from a specific volume of mass.
- 6:39And that 59 number is exactly what triggered entirely a
- 6:42coinin's tear down. As an electric motor design
- 6:45engineer, he looked at 59 kilowatts per kilogram and felt
- 6:49it was just perplexingly high. Like he thought they were lying.
- 6:53Pretty much. So rather than just taking the
- 6:55marketing materials at face value, he decided to reverse
- 6:58engineer Yass's entire design through rigorous software
- 7:01simulation. And they call it a yokeless and
- 7:03segmented armature, right? That's what YASSA stands for.
- 7:06A lot of the online discussions focus really heavily on that
- 7:08yokeless aspect. I do, yeah.
- 7:10The assumption is that simply deleting the heavy iron backing
- 7:13yoke that you find in radial Staters removes so much dead
- 7:16weight that the power to weight ratio just naturally skyrockets.
- 7:20Which makes sense on the surface, but it naturally
- 7:22addresses that assumption directly.
- 7:23He breaks down why weight reduction is really only a
- 7:25fraction of the story. So taking the iron out helps,
- 7:28but it's not the magic bullet. Right.
- 7:31It helps, certainly, but the true defining benefit of Yass's
- 7:35topology is how it circumvents the geometric scaling penalty.
- 7:40The geometric scaling penalty. Let's unpack the physics of
- 7:42that. In a traditional radial
- 7:44cylinder, if an engineer wants to increase the power output,
- 7:47they need to increase the electric loading.
- 7:49Exactly. And to do that you have to pack
- 7:52more copper wire into the stator.
- 7:55But since you're constrained by the outer casing of the motor,
- 7:59the only way to fit more copper is to dig deeper slots into the
- 8:02iron core. Moving toward the center, Yes.
- 8:04Which creates A vicious cycle. By digging deeper slots toward
- 8:08the center shaft, you're actively shrinking the inner
- 8:11diameter of the air gap. You're literally reducing your
- 8:14rotor size to make room for the copper.
- 8:16Wow. So going back to our merry
- 8:18ground analogy, you're basically forcing yourself to push closer
- 8:21and closer to the center post. You are cannibalizing your own
- 8:23mechanical leverage just to fit the wiring.
- 8:26Exactly. You're trapped by the geometry
- 8:28of the circle. You trade torque for copper, but
- 8:32in an axial motor the geometry works with you.
- 8:34If you want to increase the electric loading in an axial
- 8:37motor, you simply deepen the slots along the axis.
- 8:41The magnetic air gap diameter doesn't change at all, the motor
- 8:44just gets like a few millimeters longer.
- 8:46Oh I see. So there is 0 geometric penalty
- 8:49for adding copper. None.
- 8:51You maintain your maximum outer edge leverage while drastically
- 8:55increasing the magnetic field strength.
- 8:56OK, that lack of a scaling penalty really explains the
- 9:00foundation of the power output. But Yasa implements another
- 9:04specific architectural choice that Anedi highlighted in his
- 9:08blog, right? Yes, the dual rotor
- 9:10configuration. They sandwich a single
- 9:12stationary stator between two spinning rotors.
- 9:15Yeah. Ice Cream Sandwich model.
- 9:16Exactly, and that configuration is often referred to in the
- 9:19industry as the YASA trick because the magnetic flex flows
- 9:22actually straight through the stator core.
- 9:25Utilizing both sides of the stator essentially splits the
- 9:27machine in 2. Wait splits.
- 9:29It in two, Yeah, you have the required electric loading per
- 9:31air gap that keeps it's the inductance low and the
- 9:34efficiency super high while still compounding the torque
- 9:37from both rotors at the same time.
- 9:39But wait. Looking closely at Anti's
- 9:41reverse engineering data, there's a discrepancy that
- 9:43really stands out to me. The 250 kilowatt limit, yes.
- 9:48When he simulated that specific configuration, one stator
- 9:51sandwiched between two rotors, he pushed the simulated current
- 9:55absolutely as high as the software model could tolerate,
- 9:58and the resulting power output capped AT250 kilowatts.
- 10:02But Yossi's documentation claims 750 kilowatts total to hit their
- 10:0759 kilowatts per kilogram power to weight ratio.
- 10:11A single stage falls short by a factor of 3.
- 10:14Did Yossa lie? They didn't lie, but anti Iran
- 10:17head first into the ceiling of what we call magnetic
- 10:19saturation. Magnetic saturation?
- 10:21What is that? Basically, the steel laminations
- 10:23inside the core can only carry a finite amount of magnetic flux
- 10:26before they become fully saturated.
- 10:28Like a sponge that can't hold any more water.
- 10:30Exactly like that. Once you hit that threshold,
- 10:33pumping more current into the copper coils doesn't generate
- 10:36any more magnetic force, it just creates massive amounts of waste
- 10:39heat. The metal physically cannot hold
- 10:41any more magnetism. So if the physics of magnetic
- 10:44saturation prevent a single stage from exceeding 250
- 10:47kilowatts, the only way URSA achieves 750 kilowatts is, well,
- 10:52how do they do it? Modularity.
- 10:54Modularity. Because the magnetic flux moves
- 10:56linearly parallel to the shaft, you aren't forced to redesign a
- 11:00completely new, larger motor, you just stack multiple
- 11:03independent stages on the exact same shaft.
- 11:06Like adding more patties to her. Sandwich.
- 11:08Exactly. And as you realized, you can
- 11:10stack multiple yokeless stators and rotors back-to-back.
- 11:14His simulation proved that by chaining three of these modules
- 11:16together on a single shaft, you easily hit that 750 kilowatt
- 11:21target. Creating a multi layer burger of
- 11:23immense torque. It really is, and because each
- 11:25stage is so incredibly thin, a three stage axial motor is still
- 11:29dramatically more compact than a single stage radial motor of
- 11:32equivalent power. It's a remarkably elegant
- 11:35solution, but it immediately creates A secondary problem,
- 11:38doesn't it? Because running that kind of
- 11:40current through tightly packed copper coils to generate 750
- 11:44kilowatts in a space the size of a few ginner plates.
- 11:46It creates a thermal monster, right?
- 11:49Heat is the absolute bottleneck of all electrical engineering.
- 11:52And to reach that record-breaking power density,
- 11:55Antis calculations showed the current density had to hit 90
- 11:58amps per square millimeter. 90 amps, and Kathy Dryly referred
- 12:03to this in his blog as trying to cram 90 elephants into a space
- 12:08meant for one. It's absurd.
- 12:09The specific heat capacity of copper makes that number
- 12:12completely terrifying. Yeah, if you push 90 amps per
- 12:15square millimeter through those coils without aggressive thermal
- 12:19management, what happens? Well, the energy dumped into the
- 12:22winding resistance would cause the temperature of the copper to
- 12:24rise by 30°C every single second.
- 12:27Every single second. Yep, the insulation on the
- 12:29windings would vaporize and the motor would literally melt into
- 12:33a block of slag in under a minute.
- 12:35Wow. OK, so how do they fix it?
- 12:36Yeah, because the Italian engineering presentation we
- 12:38review details exactly how this thermal bottleneck is solved,
- 12:42and it requires abandoning traditional EV cooling entirely.
- 12:45Right. In a standard radial motor,
- 12:46cooling is usually indirect. You basically wrap a water
- 12:50glycol cooling jacket around the outer iron casing of the stator.
- 12:54So indirect cooling relies on thermal conduction.
- 12:57The heat has to slowly migrate from the copper wire, fight its
- 13:01way through the electrical insulation, pass through the
- 13:03iron core, and then finally reach the coolant jacket.
- 13:06It's a very slow, highly resistant thermal path.
- 13:10Way too slow for 90 elephants per square millimeter.
- 13:13Way too slow. So the Italian presentation
- 13:16shows that YASA utilizes direct oil cooling, specifically a
- 13:20technique called jet impingement.
- 13:22Jet impingement? That sounds aggressive.
- 13:24It is. Instead of waiting for the heat
- 13:26to conduct sootward, they inject dielectric cooling oil directly
- 13:30into the core of the stator. They are aggressively spraying
- 13:33it right onto the bare copper and windings.
- 13:36So the oil physically bathes the copper.
- 13:38Exactly. By directly targeting the
- 13:40primary heat source, you drastically lower the thermal
- 13:43resistance of the entire system. The heat is wicked away the
- 13:47instant it's generated. And the real world application
- 13:50of jet impingement, it kind of changes the entire paradigm of
- 13:53how we drive electric vehicles, doesn't it?
- 13:55Oh, absolutely. If you take a standard
- 13:57performance EV to attract a, or if you're simply doing repeated
- 14:01heavy accelerations on the highway, the copper windings
- 14:04heat up rapidly. And as the temperature rises,
- 14:07the electrical resistance of the copper increases, right?
- 14:10And to prevent catastrophic failure, the vehicles computer
- 14:14intervenes through a process called deerating.
- 14:17Deerating, where the computer artificially throttles your
- 14:20current to save the motor. Exactly.
- 14:22So a motor rated for a peak of 200 kilowatts might suddenly
- 14:25restrict you to 80 or 100 kilowatts until the system cools
- 14:28down. Your performance degrades
- 14:30exactly when you were asking the most of it.
- 14:32It's super frustrating, but according to the cooling data,
- 14:35jet impingement eliminates this fade almost entirely.
- 14:39A200 kilowatt YASA Motor can run continuously at 150 kilowatts.
- 14:44That's maintaining 75% of its absolute peak power.
- 14:47Yes, all day long without feeding.
- 14:49And the Italian presentation provides the ultimate proof of
- 14:52concept for this, don't they? Yeah, Mercedes-Benz integrated
- 14:55this specific motor technology into their AM GGTXX prototype.
- 14:59And then they deployed it to the Nardo high speed test track in
- 15:03Italy, which is notoriously brutal on powertrains.
- 15:07Right. For those who don't know, the
- 15:08Nardo ring is a 12.5 kilometer continuous bank circle designed
- 15:14for sustained maximum velocity. They ran the prototype for over
- 15:1840,000 kilometers over the course of just 8 days.
- 15:21That is wild. It gets better.
- 15:23The vehicle maintained cruising speeds of 300 kilometers per
- 15:27hour in 35°C ambient heat. So sustaining 300 kilometers per
- 15:32hour in almost 100°F heat with minimal downtime because they
- 15:36used ultra fast charging to keep it on the track nearly 24 hours
- 15:39a day. It's an absolute torture test
- 15:41for stator insulation, and the telemetry data showed zero
- 15:44thermal degradation 00. The stator and the rotors
- 15:48remained entirely locked within their optimal thermal operating
- 15:50windows, entirely due to the efficiency of the direct oil.
- 15:53Cooling that is just incredible. Now the ability to sustain those
- 15:57loads in such a compact package naturally means the first
- 16:00commercial applications are dominating the ultra premium
- 16:02sector. Oh yeah, right now we are very
- 16:04much witnessing A supercar pioneer phase, right?
- 16:08Ferrari is relying on these axial designs in both the SF 90
- 16:12and the 296 GTB. McLaren utilizes them in the
- 16:16Artura. And from a vehicle architecture
- 16:18standpoint, the packaging flexibility these motors offer
- 16:20is just unprecedented. Because the axial motor is 50%
- 16:24lighter and requires only 20% of the longitudinal space of a
- 16:27radial equivalent. Exactly.
- 16:29Ferrari engineers literally slide this tiny pancake motor
- 16:33straight into the existing gap between the V8 internal
- 16:37combustion engine and the transmission bell housing.
- 16:40They seamlessly integrate hundreds of horsepower of
- 16:43electric hybrid boost without having to lengthen the wheelbase
- 16:46at all or alter the complex suspension geometry.
- 16:49It acts almost like an oversized, highly active clutch
- 16:52plate. It solves the packaging dilemma
- 16:54that has plagued hybrid sports cars for two decades.
- 16:57But, you know, viewing this solely through the lens of hyper
- 17:00cars is kind of a mistake, I think, because the academic
- 17:03paper from the University of Nebraska highlights a commercial
- 17:06pivot that demonstrates the true scalability of this tech.
- 17:08Yes, that section of the paper completely reframes the whole
- 17:12narrative. It does.
- 17:14Fasco Australia is currently mass producing a 900 Watt single
- 17:17sided axial flux motor and it's not for a supercar hybrid
- 17:21system. No, it's for swimming pool
- 17:23filtration pumps. Pool pumps transitioning from
- 17:26Nardo track dominance to suburban backyard pool pumps.
- 17:30That requires examining a massive disruption in
- 17:33manufacturing economics. How are they doing that?
- 17:36Well, high performance radial motors are typically
- 17:38manufactured using progressive die stamping.
- 17:41You take massive sheets of highly specialized electrical
- 17:44steel, and you use incredibly expensive machinery to punch out
- 17:48the complex circular laminations that make up the stator core.
- 17:51But there's an inherent flaw in stamping circular components out
- 17:54of square steel sheets, right? The scrap rate you end up
- 17:57wasting a staggering percentage of premium expensive steel, and
- 18:00the progressive die machines themselves represent a massive
- 18:02capital expenditure. So what if Fasco do differently?
- 18:05Pasco analyzed the axial flex topology and realized it allows
- 18:09for a completely different manufacturing process.
- 18:12Because the axial core can essentially be a continuous
- 18:14ring, they employ A numerically controlled punch and wind
- 18:17machine. A punch and wind machine, yeah.
- 18:20So instead of stamping circles out of sheets, they take a
- 18:22single continuous ribbon of steel.
- 18:25The machine punches the necessary slots for the copper
- 18:27wire into the ribbon, and then it just tightly coils that
- 18:31ribbon upon itself like a roll of tape to form the stator core.
- 18:34Like a roll of tape. So the only steel wasted in that
- 18:37entire process is just the tiny rectangular chips punched out
- 18:41for the slots. Exactly.
- 18:42The material efficiency approaches 100%, it utilizes
- 18:46cheap machinery, generates minimal waste, and the resulting
- 18:49pool pump motor is so compact that the integrated electronic
- 18:52controller physically takes up more space than the motor
- 18:55itself. That is amazing.
- 18:57It replaces these bulky, inefficient induction motors
- 19:00with something vastly superior at a lower price point.
- 19:03We have really tracked A phenomenal engineering
- 19:05trajectory today. We really have.
- 19:07I mean, we started with Michael Faraday's early designs, which
- 19:10were limited by the deflection nightmare of 19th century
- 19:12metallurgy. We explored how neodymium
- 19:15magnets unlock the geometry, allowing YASA to bypass the
- 19:19scaling penalties. We looked at the jet
- 19:21impingement, cooling, taming, the heat of all those copper
- 19:24elephants. Right, enabling Nardo track
- 19:26dominance. And finally, how punch and wind
- 19:29manufacturing actually democratizes the entire concept
- 19:33for mass adoption. The technology adoption curve is
- 19:36moving exponentially, but integrating these motors into
- 19:40transmissions or pool pumps is really just an intermediate
- 19:43step. Oh, really?
- 19:44There's another step. Oh yeah, the ultimate highly
- 19:47provocative end game for the automotive industry regarding
- 19:50axial flux technology is full in wheel integration.
- 19:54In wheel integration. Wait, Because the motor profile
- 19:57has been reduced to the size and tape of a dinner plate, there's
- 20:00no longer a physical requirement to keep the powertrain
- 20:03centralized under the hood. Exactly.
- 20:05The entire Dr. unit can be mounted directly inside the hub
- 20:08of the wheel itself. Wow.
- 20:10From an engineering perspective, that eliminates immense amounts
- 20:13of parasitic mechanical drag. You completely delete the
- 20:16transmission, the drive shaft, the differential, the hash
- 20:19shaft. You deliver the torque directly
- 20:21at the point of contact with the road.
- 20:23And if you push the implications of in wheel motors to their
- 20:26logical conclusion, it fundamentally alters what a car
- 20:29actually is. If the whole mechanical
- 20:31drivetrain is housed within the four wheels, the vehicle chassis
- 20:35becomes completely empty. Think about the crash safety
- 20:38structures. Without a massive, rigid engine
- 20:41block sitting up front, engineers can completely
- 20:44reimagine crumple zones. You no longer have to design
- 20:47around preventing a solid block of metal from being driven into
- 20:51the passenger cabin. It.
- 20:52Frees up the entire chassis for batteries or radically spacious
- 20:55cabin. The volumetric efficiency just
- 20:58transforms. You could shrink the exterior
- 21:01footprint for dense urban environments while expanding the
- 21:04interior space. It's a total paradigm shift in
- 21:07urban mobility and industrial design.
- 21:09And it is entirely dictated by a decision to pivot the direction
- 21:12of the electromagnetic flux 90°. We are truly transitioning from
- 21:16the era of the cylinder to the era of the disk, and the ripple
- 21:19effects will reshape the infrastructure of how we move.
- 21:22It's going to be wild to watch it unfold.
- 21:24It really is. We want to sincerely thank you,
- 21:26the listener, for joining us on this incredibly technical
- 21:29journey today. Exploring the mechanical
- 21:31realities behind the electric revolution is only possible
- 21:35because of your dedication to digging deeper than the surface
- 21:38level marketing hype. The boundaries of engineering
- 21:41are rarely absolute, you know. They are usually just waiting
- 21:44for a shift in perspective. Beautifully said.
- 21:47Keep scrutinizing the machines around you, keep demanding to
- 21:50know how the physics actually work, and keep questioning the
- 21:53limits of what is possible. Until next time, stay deeply
- 21:56curious.