One of the most common questions that comes up when discussing Pure Watercraft is: why did we build so many of our components and capabilities in-house, when there are engineering consulting firms, electric motor manufacturers, and huge battery companies already serving the electric car business?
The short answer is that through the process of testing and iterating, we learned that owning the capabilities ourselves was the only way we could make something truly great – the only thing worth doing.
As we took a closer look, we saw that people weren’t buying higher-powered electric outboards because they were too heavy, expensive, and unreliable. They cost so much that even when you took the much lower operating costs into account, they’d never come close to paying for themselves. Repair shops said they were even less reliable than gas outboards. Boaters said they were loud; the motor eliminated emissions but noise pollution remained. We needed to achieve simultaneous competing objectives: high performance, low cost, near silence, and rock-solid reliability. Tesla was an inspiration, having broken the previous inverse relationship between cost and performance, so we learned from them and set out to create a better way to experience the water.
In the beginning, we partnered with an outside development firm to engineer our first electric runabout.
It emulated the fast-prototyping method used successfully to launch Internet companies, and gave us speed, which we believed was critical to validating our product idea. In the end, what we got out of it was a pricey prototype. On the positive side, through this partnership we secured a founding engineer with the insight he had gleaned from being embedded with the outside development firm. But we also realized the limitations of working with outside engineering. Much of the learning comes from the details of development and was captured externally. There also was a misalignment of incentives; because we had skin in the game, we were focused on finding paths that would lead to viable products, while engineering firms are in the business of serving the stated needs of the client, regardless of the viability of the product. After the first prototype, we concluded that we needed to build the core capabilities in-house, even if it took more time. The electric vehicle field was so new that available people with outstanding skills and experience were rare. We’d have to start with what we learned from the runabout project, and hire smart people who could learn the field. It was starting to look like a long road, but the only way to develop groundbreaking products.
The Battery Pack
We started with the battery pack because it is 50% of the weight, cost, and complexity of an electric vehicle, and a critical factor in performance. At first, we reached out to battery companies that manufacture and sell cells, which we would use to build into battery packs – similar to the way Tesla does.
But the big manufacturers didn’t return emails from a marine propulsion start-up. We next looked at companies that built and sold battery packs made using cells from the big manufacturers, but they cost 2-4X as much as the cells themselves as a result of the battery pack IP. If we took their prices as an input cost, then our product was going to cost the customer far more than if we developed the battery pack IP ourselves. We quickly learned that we couldn’t build a competitively priced product if we outsourced the component with the most significant IP.
Besides the economics, we were concerned with performance. The Tesla battery pack was the gold standard, and none of the packs we saw held a candle to theirs in terms of energy density and capacity retention. A boat is more weight-sensitive than a car, so we had to be even more focused than Tesla on getting the most energy per pound.
But how would we build a safe pack, with high energy density, at a cost that was commercially viable, if the cell manufacturers wouldn’t return our emails? How could we build a battery pack when we had never done so? How could it outperform the packs built by big companies? We had to secure access to cells, and learning. Leading-edge battery pack technology wasn’t published. There were tear-downs of electric car battery packs and how-to’s written by hobbyists, but our goals were different than theirs. They were trying to convert an old car into an electric proof-of-concept, and we were building a commercially viable, high density battery pack that would last years.
So we found a partner to co-develop a battery pack with us. We could learn from them, and get a working battery pack quickly. Together, we chose a standard cell type, the 18650 used in power tools and Tesla cars. We thought we could outsource the battery management system (BMS), because every battery pack has similar needs. So we used the same BMS we used in the runabout. However, we later found that the BMS didn’t work on cold days – we had to use a hair dryer to warm it up before we could use it. It weighed too much, it cost too much ($1500, when the chips on it cost < $100) , and we couldn’t improve it because it was someone else’s technology.
We then took a new approach. We tried to use sub-modules from our partner, and take on a bigger share of the development ourselves, but it wasn’t enough. The result was still high cost and heavy. We had to scrap that and start over.
Finally, we got the top battery cell provider to return our emails after 5 years of pursuing them, and they agreed to supply us cells. We would be spending less than half as much on cells, for the best cells available. Our internal engineering build-up was paying off, as we now had the expertise to develop a radical new design. By developing our own BMS, we could use a far smaller set of circuit boards, built to fit our pack, instead of the mess of 97 circuit boards that the off-the-shelf BMS required. This was the fourth battery pack we developed, and finally, we had one that really excelled. It beat our gold standard – the Tesla Model 3 pack – in energy density. The flywheel of component and team development was finally whirring, after 7 years.
Electric motors have been around for about 150 years: you would think there is a wide variety of motors available at reasonable prices. However most on the market are used in fixed applications, where neither weight nor size is very important, and they can use A/C power from the grid. Most of those used in mobile applications have large diameters and are heavy. Only the motors that have very high volumes of existing products using them are produced at low cost – for example, motors for quadcopter drones. If it’s more powerful than a drone, but less powerful than an electric car, then there are only low-volume, high-cost motors out there.
We looked into building our own motor, as Tesla had done. But where to look? We reached out to one of the gurus of motor design to point us in the right direction. When we explained our objective, he knew what we needed to do, and shockingly, he said he’d do it himself. We could now build a motor that, compared to those we had found, weighed about half as much, had about 1/3 smaller diameter, and cost 2/3 less per unit. We did it. Our process was unconventional but we created a motor with weight and performance only found in scientific motors, for a fraction of the cost.
In our effort to build the other components, we were often met with similar issues that we faced with the motor. The propeller became a component of surprising importance. Most electric outboards used generic off-the-shelf propellers, but we worked with a skilled propeller designer who ran computer simulations of the performance of design and materials choices. The electric motor’s ability to deliver full torque across a wide range of RPMs enabled a propeller with 30% more efficiency than the off-the-shelf ones.
In the end, our product achieved 3X the propulsion per pound of best electric outboards, and more than 2X the propulsion per dollar.
This could not have been achieved by buying our components off the shelf, as most others did. In designing our own components, we integrated components that we had previously imagined as separate, thus reducing the number we needed. Our relationships with manufacturing partners grew stronger as we sourced more and more components from them.
It would be neat to say that that’s why we built our own components, but we really stumbled into it, through the fog of uncertainty, feeling our way, learning at every stage.