Kees van der Grint on race starting software

[Per]

This story is already 10 days old but I don't think it has been picked up by international motor sport media. So I thought I'd share it with you.

Former Bridgestone F1 guru Kees van der Grint was interviewed on RTL Autovisie, a car talkshow on Dutch commercial TV station RTL 7. He had something interesting to say regarding certain teams' race starts.

A video (in Dutch) is available here. Below is my translation of (part of) the conversation. They were discussing the allegedly illegal Volkswagen software and changed the topic to cheating with software in F1.

- Rob Kamphues (presenter): You worked with Michael Schumacher. I don't think he was the most honest either. In any case I remember that when he was driving for Benetton, they had software that was only activated when he pressed certain pedals at the start. Or am I lying?
- Kees van der Grint: I don't know [whether you are lying].
- RK: Yes you do! You know that very well!
- KvdG: I believe that his team mate at the time, Jos Verstappen, has had a few stories to tell about that. But he's not the only one. In my days [1998-2010] it would also happen that we were wondering why a certain car was able to start so astonishingly fast. It put a huge pressure on my design as a Bridgestone man. There were allegations that it was in the tyres, which were loaded with a sort of elastic band. But it turned out to be a software manipulation, whereby the driver didn't have to do anything any more and would shoot off like a rocket as soon as the race director pushed a button. In the track there is a [electrical] loop to detect a jump start. The moment [the race director] pushes the button, the loop is switched off, but that was also the signal to go.
- RK: They [the cars] detected that?
- KvdG: Yes, they detected that. So that was...
- RK: This story has never been written anywhere!
- KvdG: No, well, then you have a scoop.


Is this something any of you have heard of before? Unfortunately he doesn't point at any particular team, although we know he was by far the closest to Ferrari... Which years might he be talking about?

[NL_Fer]

That story has been discussed in that period. He is talking about Renault and Michelin, where the launch control was triggered bij de magnetic loop in the ground, that disabled, when the red lights went off.

When Charly Whiting pressed the button, Alonso and Fisichella were launched.

[bill shoe]

Fascinating. I don't remember hearing that claim before.

[PlatinumZealot]

I remember those Renault rocket launches! I remember too, people used to think it was the V-angle of the Renault engine that gave that extra grip off the track.. no one quiet demystified it until now... Very interesting that we now hear that there was some sort of feedback mechanism in the car's software.

[Phil]

if I was to develop such a 'cheating launch software', I'd program it using an algorithm that would include some offset using a random value, perhaps between 0.1 and 0.3 seconds to make it seem arbitrary. Or up to 0.4.

In anycase, would be interesting to see video footage where one can measure the time it took for the lights to go off and certain cars to launch. Then do that over all cars over every race to measure the reaction timing of all drivers/cars. Perhaps there's a trend to see which cars are generally quicker "off-the-line" which would give some weight to launch software.

[bill shoe]

if I was to develop such a 'cheating launch software', I'd program it using an algorithm that would include some offset using a random value, perhaps between 0.1 and 0.3 seconds to make it seem arbitrary. Or up to 0.4.

Yea, in professional sprint running and drag racing, a minimum human reaction time of 0.2 seconds is assumed, therefore if you react within 0.2 seconds after the gun or starting lights, then you are assumed to have gotten a lucky guess and a false start. The ideal reaction is 0.20001 seconds after the start. However, the further reaction time of a drag car is also 0.2 seconds, so drag racing actually has a a total 0.4 second "no-go" time after the nominal start. So if the FIA was analyzing carefully you would want 0.4 to 0.5 seconds delay with random variation. If they were not analyzing carefully you could get down towards 0.2 seconds, but I'm guessing you will never get better that because it would require better car reaction time than a drag race car.

If the allegations are all true then I would guess most of the advantage was from the software-optimized launch rather than having slightly quicker reaction times than competitors.

And I guess this trick (if true) would qualify as a defeat-device. An odd name for something that helps you avoid defeat.

[Cold Fussion]

I don't see how this assumed false start makes any sense, if you are only going on luck rather than reaction then you will more often then not jump start.

[BanMeToo]

Right, so it's to discourage you from only going on luck...

[Cold Fussion]

Right, so it's to discourage you from only going on luck...

That's the point, if you rely on guess work you will more often then not jump the start, which you are already penalised for.

[Brian Coat]

Interesting:

How does this sound for a swag of what the induction loop could give:

Elimination of pure driver reaction time: Up to 0.2 (?) seconds
Insurance against a driver having a 'late light': up to another 0.2?
No reduction of car reaction time

Up to 0.4s

Isn't that 'up to' 3+ car lengths by the time you hit 100 mph?

Despite multiple uses of 'up to' I have no affiliation to the broadband industry.

[PlatinumZealot]

if I was to develop such a 'cheating launch software', I'd program it using an algorithm that would include some offset using a random value, perhaps between 0.1 and 0.3 seconds to make it seem arbitrary. Or up to 0.4.

Yea, in professional sprint running and drag racing, a minimum human reaction time of 0.2 seconds is assumed, therefore if you react within 0.2 seconds after the gun or starting lights, then you are assumed to have gotten a lucky guess and a false start. The ideal reaction is 0.20001 seconds after the start. However, the further reaction time of a drag car is also 0.2 seconds, so drag racing actually has a a total 0.4 second "no-go" time after the nominal start. So if the FIA was analyzing carefully you would want 0.4 to 0.5 seconds delay with random variation. If they were not analyzing carefully you could get down towards 0.2 seconds, but I'm guessing you will never get better that because it would require better car reaction time than a drag race car.

If the allegations are all true then I would guess most of the advantage was from the software-optimized launch rather than having slightly quicker reaction times than competitors.

And I guess this trick (if true) would qualify as a defeat-device. An odd name for something that helps you avoid defeat.

Driver's are really trained like distance runners and not short sprinters then! Short sprinter's reaction time cut off is 0.1 seconds. Any faster than that and it is false start. Still the average for short sprinter is about 0.16 seconds... I bet those Renaults were launching off the line less than 0.05 seconds. A system like that usually can pre-load the clutch and effectively line-lock the front brakes.

[bill shoe]

Driver's are really trained like distance runners and not short sprinters then! Short sprinter's reaction time cut off is 0.1 seconds. Any faster than that and it is false start. Still the average for short sprinter is about 0.16 seconds... I bet those Renaults were launching off the line less than 0.05 seconds. A system like that usually can pre-load the clutch and effectively line-lock the front brakes.

I stand corrected. 0.1 second is really fast!

[Brian Coat]

This supports there being about 0.2 available for RT elimination, with snoozer insurance as a bonus.

When U Bolt ran 9.58, another sprinter "left on him" by over three hundredths but he still had enough power to make it an easy win at the stripe.

I was interested in the 0.05s "line-loc" launch idea.
That's a pretty small clutch to be feeding oomph for as long as all the reds are on?

[PlatinumZealot]

I was only guessing they "line locked" the front brakes to keep the car from rolling forward against a pre-loaded clutch. When I say pre-loaded clutch, I mean the first clutch is engaged just enough to transfer a slight torque but not enough to over power the brakes and move the car. During this time torque would be dissipated in the form of heat. As soon as the computer detects the start signal the clutch will be already engaged for the first stage of the launch, almost instant torque transfer to the ground.

[Brian Coat]

Ah, thanks.

If I understand correctly this limited pre-load/pre-kiss strategy might work (e.g. thermally) but the instantaneous vehicle reaction effect at brake release would be limited accordingly.