Capturing Anomalies During Competition

3 weeks ago by Vanessa Knivett

Why log data just some of the time when you can do it all of the time? Single-seater, fully-electric, Formula Student cars are particularly space constrained but even though every gram counts in a race situation, the Kvaser datalogger attached to Carnegie Mellon’s contender stays in place for the ride.

Carnegie Mellon Racing (CMR) has used a Kvaser Memorator to monitor the reliability of their self-built circuit boards for two years now, since team manager Samuel Westenberg joined the group, to be precise. “We had some strange bugs that we were only seeing in a race situation, so we attached the Kvaser Memorator to the car permanently so that it records whenever it is in a drive state,“ recounts Westenberg.

At CMR, which comprises some 50 students, most components are manufactured by the team in the University’s own machine shops, including the chassis and structural components. Notes Westenberg: “Only the electric motor, tyres and wheels are bought in. Just occasionally we get a bottleneck run out of milling machines, so will outsource the manufacture of certain parts, but our ambition is to design and build everything ourselves.”

“We chose the Kvaser unit because I had worked with it at Tesla. It was the tool I was given on my first day and proved very easy and reliable...”

The approach extends to the electronics with most PCBs on the car made in-house. With 10 to 20 different sensors (many duplicates) and 100 sensors in total, CMR’s 2017 competitor had one CAN network, though the team is likely to move to two in the [2019?] iteration, as the bus loads increase. Reveals Westenberg: “During the Spring 2017 race season, we encountered various bugs in the motor control loops, limiting their stability. If they shut down, a reliable CAN analysis solution was needed to help us find the source of the problem and intervene.”

“We chose the Kvaser unit because I had worked with it at Tesla. It was the tool I was given on my first day and proved very easy and reliable, so we got hold of one for our team. It has built-in triggers and filters, plus the start and stop logging based on specific CAN messages has proved invaluable, meaning that we haven’t had to hardwire an external switch. We haven’t used the t programming features yet but [that’s on the cards for 2018].”

Kvaser Memorator 2xHS v2

In October 2017, CMR froze the design for the 18e iteration. The next step is ‘rolling chassis’ in March 2018, where all the mechanical parts on the car are integrated and tested before the electrical systems are introduced. The next race event will be in June 2018 in Toronto, Canada, a combination of static events such as presenting the team’s business plan to a group of ‘investors’, and dynamic events that evaluate endurance i.e. by finishing a 22km race, acceleration over a fixed distance, maneuverability and of-course, energy efficiency. We wish the team well and will be watching the results with interest!

Kvaser donated a Kvaser Memorator datalogger to CMR and has provided tool support to the team. If you have a University project that would benefit from similar sponsorship, please visit our University page [linked] to find out how we can help.