UNDER THE HOOD
Pat Symonds on F1’s speedy pitstops
Incidents, like buses, often come along together and over the first few races of this season pitstops and wheel retention have become a focus of both the teams and the FIA. It started on the very first morning of pre-season testing when Mclaren lost a wheel on Alonso’s car. The incidents continued through the Haas debacle in Australia, the unfortunate accident that befell Ferrari in Bahrain, and a number of less visible episodes during practice at later grands prix
(in Spain, for instance, Sergio Perez left his box without all his wheels attached properly during Free Practice 2).
Many articles have been written about the choreography involved in pitstops but little about the automation and safety of them, although the systems employed have achieved a level of sophistication that may surprise many viewers.
The sub-two-second pitstop is an amazing achievement and one doesn’t have to look too far back to see that progress has been rapid and spectacular since refuelling was banned for 2010. Before this, with refuelling taking a leisurely six or seven seconds even for a three-stop race, the return on investment for fast wheel changes was negligible. It was more important to ensure that the task was carried out safely and correctly.
The ban on refuelling put the emphasis on wheel-changing acumen, and bit by bit times came down from a typical four seconds in 2011. By 2012 this was below three seconds and in the 2013 US Grand Prix the Red Bull team set a new benchmark of 1.92 seconds. Since then teams have achieved 1.6 seconds in practices, while Mercedes claimed the race record at 1.83 seconds when servicing Bottas in this year’s Chinese Grand Prix.
So how has it been done? Certainly a scientific focus on human performance as opposed to a blind ‘practice makes perfect’ attitude has helped immeasurably, but technology and design effort have played their parts too.
In terms of design it was common some years ago for the axles to exchange torque with the wheels via a series of bullet-shaped drive pegs. The design of these was not particularly sophisticated beyond the thought applied to the differential expansion between the wheels and hub, and ease of location of the wheel onto the axle. While ease of fitting required a loose fit, a tight fit was required to avoid fretting and possible acceleration of wheel loosening if the nut was slightly below target torque.
Today a much more sophisticated, spline-like design is generally used. The nuts themselves have always been of a relatively large diameter but the thread form is unlike anything that would be found in a set of conventional design standards. As one might imagine, the pitch is very coarse, meaning the nuts need very few turns from initial engagement to fully tight – but what may be more surprising is that the thread is also very sloppy. This serves two purposes: one is to reduce friction on the thread, and the other is to allow the inevitable swarf that may be cleaved off the nut to escape freely without jamming the nut.
The real development, though, has taken place off the car in the form of the Automatic Release Systems (ARS) that are now commonplace. The first step of this came in 2008 when Ferrari replaced the “Brakes On/go” lollipop system with manually controlled traffic lights.
Today a fully automatic ARS is instrumented in every respect, such that not only does a central computer check each function but the data is recorded, and video from three separate cameras is used to analyse each move of the pitstop to expose weaknesses and allow development. When I was at Williams we produced a report which could typically run to 16 pages to cover just two pit stops in a race.
The essence of these systems is a series of sensors which monitor every action of the jacks and wheel guns and feed that data to a simple controller that can activate the jacks and the pitstop lights. The technical regulations only allow sensors on the wheel guns to act in a passive manner. In other words, the completion of the wheel tightening must be signalled manually. The rules also require that the mechanic responsible for operating the wheel gun
has a visible means of identifying an incorrectly fitted nut. In the less sophisticated systems this is merely a painted line on the axle which is only visible when the nut is fully home. In the latest systems, such as that used by Ferrari, a pair of sensors monitor the strain in the axle (which is the only accurate way of determining if a nut is tight) and also the position of the nut to ensure it isn’t cross threaded.
By regulation, even with this electronic assistance, it’s still the responsibility of the gun man to press a button to indicate he is satisfied everything is in order. In China this procedure obviously fell down as the go signal was given without the wheel even being changed.
THE TECHNICAL REGULATIONS ONLY ALLOW SENSORS ON THE WHEEL GUNS TO ACT IN A PASSIVE MANNER. IN OTHER WORDS, THE COMPLETION OF THE WHEEL TIGHTENING MUST BE SIGNALLED MANUALLY
Once a positive signal has been received from both wheel guns on an axle, the relevant jack is dropped automatically by a further signal from the control computer to a solenoid on the jack. As soon as both jacks are down the light is armed to turn green. But it will only turn green if there has been no over-ride by an observer or observers who should hold the car if they see something has gone wrong, or if a car is entering the pit directly in front of the exiting car.
With such high-performance systems safety is paramount, and one of my responsibilities at Williams was to sign off a comprehensive Failure Mode Effect Analysis (FMEA) document that investigated the outcome of every possible system failure.
All of this comes at a price, of course. The wheel guns themselves may seem expensive at £3,500 but this is before the teams have changed various parts and added the sensors, all of which pushes the cost up to around £10,000 each.
A front swivel jack costs in the region of £70,000, and once the cost of the controller and the bespoke software is included one can easily see that the improvement from four seconds to 1.8 seconds comes at a cost approaching half a million pounds – and that’s before you account for any servicing or personnel training costs.
Fractions of a second don’t come cheap.