Introducing India’s Most Powerful Axial Flux Motor: Flux350

A motor is at the heart of the fundamentals of an electric vehicle and how could Tresa call itself Tresa Motors if it doesn’t make motors?

Axial Flux motors are key to the long-term success of the EV industry. So much so that Mercedes Benz recently acquired British axial flux motor company Yasa Motors, undoubtedly the leader in this space. Axial Flux motors are at the core of what we’ll do at Tresa as well. They are, at minimum, 1/4th the volume, and half the weight of radial flux motors (which are already 1/4th the weight of ICE engines), saving precious weight and volume, which means greater range.

Let’s delve deeper into what they are.

There are 2 major motor topologies: radial and axial flux motors. The difference in their names represents their magnetic flux direction, which is radial (perpendicularly) in radial flux and axial (parallel) in axial flux motors, to the axis of rotation. In the case of radial topology, the flux path is much longer compared to axial motors, which forces designers to make bigger size motors. The magnetic field remains strong along the small path in axial motors, increasing active winding copper, making the motor efficient and power-dense. The fact that the flux path can be unidirectional in axial flux motors, we can use CRGO steel plates (that have high magnetic permeability: allowing the magnetic flux to go pass through easily), to enforce even more concentrated flow, leading to low hysteresis and eddy current losses, further reducing cooling requirements.

Now, back to the actual design and starting with Delta Engineering: you need to have “something” to begin with to optimize it. There were far too many problems staring at us before we started targeting motor optimization:

1. 1. 1. How do we actually design it? What are the parameters? Where do we start?
      2. Where do we get the components machined? There are a lot of components that need to be machined with high precision.
      3. Tresa’s motors are required to run north of 4500RPM, delivering 600Nm of torque. This means we need extremely high magnetic fields. 
      4. How do we actually install the magnets on the rotor? Without a solid mechanism, magnets just wouldn’t stay at their desired locations.
      5. How do we cool this beast? 
      6. How do we get the copper coils on the stator? Since this is a new topology, there are no manufacturers that would have ready-made setups!
      7. Where do we get the CRGO steel plates (we needed ~0.2mm thick plates) and how do we make a motor stamping out of it? How will we stack some 100 of them on top of each other and hold them in place?

Breaking down the problems helps solve them in a structured manner. Here’s how we solved these:

1. 1. 1. Simulate, simulate, and simulate! Since we had to commit to a design, and making a motor is a major investment, we spent a good 11-12 odd months on simulations and research (more so because ANSYS, or any other simulation software cannot be used for simulating the topology specific to our motor, finally we had to write our own macros to link 100+ dependent and independent parameters between Maxwell and MATLAB to get the simulations up and running). Our team built the whole motor in Ansys, and all the simulations put together would have taken some 60 days of computing! Several design parameters like power target (defining inner and outer radius), number of turns, copper coil diameter,  width between any two magnets, stamping size, the total number of CRGO steel plates in the stator lamination stack, basic cooling system, etc were tested before arriving at the final configuration. Parallely, we kept looking for manufacturing partners and got a few amazing ones! We also spent a lot of time making a few parts in-house on our own CNC machine.
      2. Fortunately, Bangalore has some of the best precision manufacturers in the country, thanks to ISRO, other automobile companies and Bosch/Crompton/ABB, etc! A lot of work from them gets outsourced to private vendors. Peenya and Rajajinagar became the hot spots for us to go and get our job done. ZetWerks was also used for a few of the components, but direct vendor relations were important, so they were prioritized. A lot of big manufacturers turned us down, especially for the stator lamination production, it was too unique, and not something that gets used in radial flux motors. We are really thankful to our manufacturing partners who believed in Tresa.
      3. Finally, we picked Neodymium magnets due to their higher magnetic flux. 
      4. We considered the industry-standard thermal gluing process (Loctite AA 3342). And immediately placed the order, but never got it on time. We ended up using Loctite AA 326. 
      5. Simulation again! A lot of work went into designing the cooling mechanism for the motor, and again, simulation helped. We know how we could improve it further in the next version(we’ll share more details about it in a future blog post about RJ3 Motor), but went ahead with this design, sticking to Delta principles.
      6. Special jigs and supports were designed to wind the copper onto the stator. This was tricky but with some effort was solved in time.
      7. CRGO stator was a big challenge. Machining loose 100s of thin CRGO plates stacked on top of each other was a pure engineering challenge. We achieved it by creating a special contraption to keep it in place, and finally laser welding them together.

Once the motor was ready, these were our observations:

• • The motor worked flawlessly! We weren’t expecting it to just work after we connected it to the controller! It was beautiful!
  • There was slight axial vibration in the motor, which must be solved in the future using weight balancing (till we enter precision production for the final design).
  • Due to the non-availability of the AA3342 glue, we went ahead with AA326 and knew if we pushed the motor, the magnets would come off at higher temperatures.
  • Our first motor design was an “out-runner”, so we decided to make our next motor an “in-runner” topology (this was important for us to decide as finally it would also go inside the eAxles of our trucks).
  • The copper windings did not flush well with the inside of the stator and there were air gaps left (mechanical design improvement for the next version).
  • The copper windings did not have tighter tolerance on the external windings, so we immediately 3D scanned a single stator and designed a jig to hold it in shape after winding for the next version.
  • We used Neodymium magnets, which will demagnetize above 80degC, a huge risk! We planned to evaluate Samarium Cobalt magnets, that can go up to 250degC (but with 80% of Neodynum Flux density) in the next version (RJ2).

Next, we wanted to break the motor. It is important to break what we make, to see what actually fails first! We decided to make a rudimentary eAxle with actuator-driven gear shifts to push the motor to extremes. The design inherently had slight axial misalignment to stress the motor design. After running it at the highest possible RPM for quite some time, it finally broke! To check out what happened, we opened it, and we realized 2 magnets came off, and the glue (as expected, AA326 won’t work for us) was to be blamed! We had smoked the motor and it was real fun! We’ve built a culture of not being afraid of breaking things internally! Things should rather fail internally than outside! Fixing this motor, later on, was an easy fix, and from RJ3 onwards, we hold magnets mechanically as well as with high-temperature glue (the industrial practice). It was surreal, and we all had a lot of teary-eyed moments during making it! Minus the simulation time, we finished within our target of 3 months! The video below is dedicated to our motor team and a special advisor, who, even though can acclaim celebrity status due to his achievements and knowledge, decided to commit his life to electric motors, in the humblest way possible! 

India has lagged in the development of things that are fundamental. The language I’m using to communicate with you is not Indian, neither is the platform I’m using to talk to you. None of the programming languages or semiconductors inside the device where you’re reading this are Indian or neither a lot of components going inside the EVs on Indian roads today.

I am not sure if you know this, but most of the electric motors in Indian EVs today have foreign origins: [Prestolite (owned by Broad Ocean Motor Group, China, for Tata/TataComp collaboration), Valeo (a French company, for Mahindra), Mahle (a German company, for Ather), SAIC (a Chinese company, for MG), NIDEC (a Japanese firm, for Hero Electric), Hella (a German Company, for Kinetic Electric), etc.] Some OEMs do have in-house low-power motor manufacturing, but I couldn’t find many details online. Maybe we’ll all hear soon about them. RJ1 (our FLUX150 motor’s version name) was the beginning of something very fundamental for us!

Nothing against any other country, but with more than a sixth of the world’s population, this is unacceptable. We need to build fundamental technologies, not to boast, but to take care of the 90% of the Indians that make less than INR 25000 per month!

360 of the top 1000 and 62 of the top 100 rank holders in 2010 IIT JEE, went abroad. I’m sure the numbers are similar for my 2008 batch. I decided to stay back, a decision that many around me still call foolish.

I don’t think it was foolish for me to stay back. I really want our generation to be the last one to live with this tag called “developing nation”. I want to see India becoming the nation of every country’s dreams. I want to see India becoming a power that strives for innovation, human growth, and global prosperity. That’s the kind of world-domination dream I have for my country, and I think I was born to do something to make it happen. Just like you!