Chassis Development

For two years I led the chassis subteam of Northwestern Formula Racing, which designs and builds the frame, steering, drivetrain and cockpit of the racecar. During those years I ended up designing all four systems at least once, along with tons of smaller projects that cropped up along the way. Given my background in education, I set a goal at the beginning of my time as lead that I wanted my new members to become the best engineers at Northwestern. Though only time will tell if I succeeded there, the early results look pretty promising: almost half of this year's formula exec board came off my chassis team from last year. 

In this page I'll walk you through the projects I worked on and managed as chassis lead. Let's start with the big one: the frame. I designed and manufactured the frame for the 2015 car, going all the way from initial geometry selection to final welding. Designing a racecar frame is all about systems integration. You have to package a driver and powertrain in a way that's comfortable for both, and do it with a geometry that creates the motion your suspension people want. I've described it to people before as a giant game of connect the dots, where you make your own dots. I started making and connecting the dots in Solidworks in the summer before the school year started, and had the placement of all the tubes completed before school started. Then it became a period of pushing around frame nodes to slightly different positions until everything fit and the suspension geometry worked correctly. With points mostly confirmed, I began finite elements analysis (FEA) on the frame to set factors of safety under various loading conditions, and determine the all-important torsional stiffness value, which affects the way the front and rear suspension interact.

With the design complete, I built a frame jig out of 8020 aluminum to hold the tubes in place for TIG welding, and got to work welding the frame together. Three months, and many late nights later, we had a frame.

For more on the NFR15 frame, read my official design documentation here

With one frame under my belt, I entered my second year as chassis lead knowing a bit more about what I was doing. The first lesson: don't have the chassis lead also do the entire frame. The combination of doing the frame and leading a team through other projects is pretty over the top. Fortunately, I had built up my subteam well enough by 2016 that I could devote two other people on the team just to the frame, one taking the lead on design and the other on manufacturing. This let me step back into more of an advisory role while I focused on aero, steering, and other projects that needed attention.  The second lesson I learned was to focus more heavily on systems integration, and less on mass reduction, than I had in the past. The 2016 frame was designed almost entirely around packaging and serviceability, and I can say with confidence that it is by far the best chassis we have ever made. 

I kicked off the focus on packaging with a full cockpit mockup, collecting data on where drivers preferred everything from the steering wheel to the headrest to be. This mockup drove a number of design decisions for 2016, and let us design the frame with confidence that drivers would enjoy sitting in it.

The packaging requirements got more interesting around the suspension and drivetrain, leading to a number of brainstorming sessions in early fall until we settled on a design that worked. It required a bit of structural gymnastics around the rear of the frame, but the rewards of well packaged systems were well worth it. This particular frame-suspension-drivetrain layout worked out so well that we're keeping it almost unchanged in 2017.

The last major concern when integrating other systems in to the frame is the attachment points themselves. We focused pretty heavily on these in 2016, making sure that each mounting point could be easily serviced, had good tolerances, and could survive the loads being put through it. I came up with the idea of having the company that coped our frame tubes cut square tubes into clevises to make suspension mounting points, which turned out to be extremely successful. The square profile was naturally strong in bending and buckling under non-axial loading conditions, and enforced the separation between the two clevis faces to within a few thousandths of an inch. These tube segments were welded onto the frame using a single jig, called the universal jig, that could enforce global tolerances all across the car. With well thought out brackets and frame jigging, we cut our suspension hardpoint tolerances to less than 25% of anything we had hit before.

With all the frame tubes tacked into place, the process of fully welding all the joints was upon us. It had taken me over a month of continuous welding to complete this part of the process the year before, making it an obvious target for improvement to speed up the build. Welders generally spend much more time setting up welds than actually welding, especially on structures like spaceframes where you end up welding at strange angles in tight quarters. To get around this, I built a frame rotisserie, which allowed people welding the frame to rotate it to whatever orientation they needed. It took Russell and Charlie a week to finish welding the frame on the rotisserie, less than a quarter of the time it took me without it.

With all brackets attached and welding complete, we had our new frame ready to race. You can see from all the tabs welded onto it just how many things have to attach to this frame.

Building the frame was a huge project both years, but it wasn't the only one we tackled. My second year as lead I took on the steering and differential mount systems as well, both of which came with their share of interesting engineering challenges. Steering on our car had been a weak point in the past. It was hard to move, there was compliance in the system, and it even failed at speed once on the previous year's car. I knew going into it that steering needed a ground up redesign.

The redesign started with suspension geometry, where our suspension lead was able to reconfigure his geometry to eliminate major bending loads on the steering rack. These loads are what had caused the previous car's steering rack to fail. We selected a new rack with wider mounting points, which further reduced bending loads, and increased the stiffness of the system. 

Next up after geometry was the bearing setup. We had previously used bronze bushings, which were prone to friction and compliance. I replaced the bushings with a combination of axial and thrust needle bearings, which could be preloaded to adjust the tightness of the steering feel. These bearings sat around a double universal joint, which changed the direction of the steering column from its nearly vertical position at the rack to its nearly horizontal position for the driver. This combination of preloaded bearings and double U joint worked so well that our cockpit design judge at FSAE Michigan called it the smoothest steering he'd felt that year. 

For more on steering, read my official design documentation here

Though drivetrain components are definitely part of the powertrain, chassis team designed and manufactured the differential mounts for the two years I led the team. We're hoping to finally pass that project over to powertrain this year, but in the past only chassis and suspension teams had experience doing the finite elements analysis (FEA) required to design complex structural components. I designed the system my second year as lead, and led a new member through her first major engineering project designing a component in it as well. 

The drivetrain holds the title of having the largest load on the car, a huge tension load on the chain between the differential and engine sprockets. Anything it goes through is going to be fairly highly stressed, which leads to an obvious goal: reduce the number of parts in the load path. To do this I redesigned the whole system to mount to the engine block, eliminating the frame from the picture. Without frame members in the way, we could package a larger sprocket and have more freedom with packaging in the rear area. I used a few iterations of finite elements analysis to reduce the weight of the system, eventually settling on a design that looked like this.

I designed all the major differential mount components to be machined from a single piece of 7075-T651 aluminum, an alloy with an extremely high yield strength. 7075-T7351 was chosen for the rotating sprocket adapter, as it is more resistant to fracture propagation due to fatigue. Using one of our school's CNC mills, I machined all the parts out from their aluminum stock. Last up was final assembly and attachment to the frame. On the road this system worked as planned, but there are ways to simplify the system by using a slightly wider model of our current differential, which we will be incorporating next year.

For more about the differential mounts, check out my design documentation here

None of this would have been possible without my amazing chassis team members. Thanks for making these last two years something to be proud of.

NFR16 Chassis Team

Charlie Steingard, Russell "Rascal" Ohnemus, Ellen Zhao, Ana Acevedo, Derya Arin, Aaron Janick, Andy McIntosh

NFR15 Chassis Team

John Harris, Charlie Steingard, Russell Ohnemus, Ana Acevedo, Charlie Tokowitz