Winning Gran Turismo in Real Life and the Laguna Seca Test Rig

  |   Agile, Blog, Test Driven Development, WIKISPEED Shop & Manufacturing, Xtreme Manufacturing   |   No comment

Winning in Gran Turismo video games is accomplished by crossing the finish line first in the Gran Turismo World Cup. The Cup is a series of 5 lap races, open to any car in the game.

For most of my life, I had no idea these games could be “won”. Like many people, I enjoyed hours of tuning and modifying a vehicle like one I drove in real life until it was one of the fastest cars on the planet, but never was able to win so many races as to unlock the Gran Turismo World Cup.

Well, it’s time to win it. In real life. And here is the plan. Step 1: Define the test (Test Driven Development):

Gran Turismo 2 was my favorite. The list of cars in the game wasn’t just collector cars and top tier racecars, but included minivans and economy cars and therefore included the cars in which I learned to drive and several of the cars I drove during formative college years while playing Gran Turismo. Plus, it was simpler and easier to jump in and play. So let’s start by winning Gran Turismo 2 (GT2) in real life.

The Gran Turismo 2 World Cup consisted of several races on made-up race tracks, and one race on a real-world race track: the legendary Laguna Seca in California. To be competitive, you needed to build-up a car until it could complete 5 laps of Laguna Seca at 1 minute and 18 seconds or faster. Alright amazing video game, challenge accepted. The fastest car we could just go out and buy will lap the track in 1:28, or 10 seconds off our GT World Cup pace. ( So, buying a car is out. We’ll build a car up to do it, just like in Gran Turismo.

By the time a car is ready for the GT World Cup, it is capable of 2 to 3 lateral gravities of acceleration. This is accomplished by having air push or suck the car to the track at 2 to 3 times the car’s weight, so only the actual car weight is pulling sideways via centripetal force while the tires feel 2 to 3 more times that weight pushing straight down, making them stick. The car can slow down at a rate of 2 to 3 g’s as well, and accelerate at 1g. The acceleration number is achieved by the car producing right about 3 horsepower for every 100 lbs. it weighs, with the driver on board.

Step 2: Run the test in 5 seconds or less (TDD on a CI Server in Hardware):

Now we have the recipe of the car we need to build. If it weighed 1000lbs, it would need to generate 1000lbs of forward force which typically indicates 300 hp. It would need to be pushed from the side with 2,000 lbs. to 3,000 lbs. of force and not move, even while pivoting the front wheels slowly which brakes some of the traction. And it will need to be pushed from the front with 3,000 lbs. of force and not budge. You might note already a specific pattern in these tests; I’m writing them as static tests that can be performed while the car is parked. Why?

Zipping a car around a race track to test it requires a race track, which is a tremendous investment in time and money to book track time and get to the track. Near our build server factory in Lynnwood, WA, USA, booking track time is $400 to $4,500 USD for one day. To test at any time, we would need to own a track, which is a tremendous impact on habitat for the natural world unless we innovate a completely new forest racetrack concept (which would be cool and maybe we will in the future). It also requires a racing driver that is consistent, so the changes the team makes to achieve higher speed can be isolated from the driver tiring or otherwise varying. And it is dependent on the temperature and moisture of the track, and wind conditions.

That’s where the Laguna Seca Test Rig comes in.

With simple hydraulic presses, made from bottle-jacks for less than $100 each, we can push on the wheels of the car to test lateral stick. We can push on the front and rear of the car to test braking. By bracing these jacks against a pair of 2,500+ lbs. weight scales like the type we use to weigh our cars, we can measure how much pressure the car can take from the side until the tires start to slide. We can measure how much pressure the car can push on the front, and resists when braking. We can measure lateral, acceleration, and braking g’s on a stationary car in the footprint of a parking space. Aerodynamic downforce is simulated by hydraulically pressing down on the aerodynamic load points then pressing the car sideways to measure lateral acceleration. The aerodynamic loading is then confirmed by CFD (computational fluid dynamics) of the airflow around the vehicle shape, run on a tablet laptop right there on the test rig. By then insulating this space, we can control temperature and moisture variables and test them systematically. We can put blocks of different track surfaces under the tires, mist then with water or even oil-slick and test the vehicle dozens of times a day. This is our Continuous Integration Server, and our goal is to automate all of these tests.

Our intent is to make these rigs available in our store if other groups find them or the concept useful. We would happily just buy one, but seeing nothing like it on the market it looks like we have to make our own. Luckily this will be super fun, and full of very fun tests.

incidentally, the downside of testing many configurations in short order is that we accumulate many test vehicles in short order. To that end, we are currently exploring a 64 car garage as our build facilities are already choked with cars. The Rotary Parking Tower seems to be the best answer, housing 8 cars in 2 parking spots worth of footprint for $6,000 including the cost of installation. So, 8 of those might be in our near future ( And, hopefully before too long a forest test track.

Joe Justice is the President of Scrum@Hardware at Scrum Inc., and the CEO of WIKISPEED Inc.

WIKISPEED is a registered road-legal automotive manufacturer with operations in 23 countries.

No Comments

Post A Comment