Is it anything more than the cars transponder indicating that a car is in/near its starting box?
Could this be the reason for a 'tolerance' that is mentioned by the FIA?
Brian
There are sensors in the track.
Re: Race start monitoring system
Posted: 11 Jul 2017, 10:55
by alexx_88
So a driver is only deemed to have moved after his car passes a point in the grid slot or?
Re: Race start monitoring system
Posted: 11 Jul 2017, 18:01
by hardingfv32
The traditional transponder system puts out a signal that is cone shaped. There is distance over which signal once received will be maintained. I would say this is where the FIA gets the systems tolerance. If this is the case then the car can move a little and not leave the transponder's transmission boundaries.
The reaction time is completely irrelevant for a jump start, a reaction of 0.0s would be perfectly legal - the 0.2s shown was not from official FIA data but just shown by the FOM.
The ONLY thing determining whether there's a false start is the sensors.
There's a loop in every starting box, every car has a sensor, that loop is in front of the senser and is measuring the signal strenght over time, since the sensor is not in the same place on every car it will be different every time, but every car has a "zero" distance strenght.
The data produces a graph, signal strenght on y, time on x - there's a red and a green line, one for the 5th red, one for 'green' - the car is not allowed to move before the vertical green line.
The car starts moving towards the loop, when the sensor is above the loop the graph will have it's peak and then the strenght goes down to 0.
The system has not once failed in over 20 years.
Ideal start:
Bottas did indeed move 0.06s before 'green' BUT there is an error of margin for the system and a tiny allowed movement (both a secret so that teams don't optimize their starts for it) - Bottas was within the tolerance.
So i hope this clears things up and there won't be another few pages with unnecessary discussions ...
here's a bit from the sky post-race analysis with brundle mentioning that graph i tried to draw with the info i had from that article
and since i was bored ... here's basically what's happening ... in this case with some small (exaggerated in the pic) movement before green ... sensor loop in orange, transmitter in blue
Re: Race start monitoring system
Posted: 11 Jul 2017, 22:30
by hardingfv32
I am assuming car movement is the criteria for the jump start judgement.
So the car is the 'transmitter' and the starting box loop the 'receiver'. Signal strength is determined by the distance between the transponder and the loop. So for this system to work the signal strength has to be measured after the car comes to a stop in its starting box. A change in this signal strength means there has been movement.
So my question would be how accurately can signal strength be measured? Is this where the tolerance is created?
Brian
Re: Race start monitoring system
Posted: 12 Jul 2017, 00:59
by RZS10
Such a system can be quite precise and i assume there's ways to filter the signal but in this case it seems to be irrelevant since the tolerance they have for starts isn't within the system but it's merely a certain distance (or rather change in signal strength) they allow.
One of the reasons given in the article is that the cars move slightly when they engage gear (which means the sensors can pick up those 'jerks') another is probably that some of the starting grids are on inclines.
The FIA won't reveal how much movement is allowed though but i can only guess that they have put some thought into that so it won't give false positives for false starts.
Re: Race start monitoring system
Posted: 12 Jul 2017, 11:43
by notsofast
When you see the word "tolerance", think "uncertainty" instead. The point is not that the FIA is certain that a car is moving, and that they are giving the driver a small amount as a freebie. The point is that the FIA cannot be certain that the car is moving. All measurements have uncertainties, at all levels. On top of that, the transponder can move even when the tires are not rotating. An acceptable level of certainty is achieved only when the measurement exceeds the tolerance.
When you see the word "tolerance", think "uncertainty" instead. The point is not that the FIA is certain that a car is moving, and that they are giving the driver a small amount as a freebie. The point is that the FIA cannot be certain that the car is moving. All measurements have uncertainties, at all levels. On top of that, the transponder can move even when the tires are not rotating. An acceptable level of certainty is achieved only when the measurement exceeds the tolerance.
Nope, the fia explanation was that it was a tolerance and not inherent to the uncertainty. They can accurately measure movement but allow some freedom because the car can still have some movement inherent of putting it in gear or something like that, only when a driver moves more than x cm (x being a secret) the fia considers the car to start. Also x is bigger than the precision of the system. Luckily for Bottas he started moving (supposedly) before the red light went off but only moved this x allowed tolerance 0.201s after the start so it was deemed legal.
As in RZS10 brilliant posts you can imagine the graph is already moving before the time and thus is bellow the detection limit of the system but still within the tolerance parameter artificially designed to prevent small movements to trigger false start alarms.
Re: Race start monitoring system
Posted: 12 Jul 2017, 18:18
by hardingfv32
I challenge this 'secret' the the FIA is withholding.
There is no reason that the teams could not figure out the movement tolerance of the 'triggering' aspect of this start system. They have the transponders and can determine the frequency being used.
Does anyone have any knowledge about how accurate one can measure very small distances using radio frequencies?
I challenge this 'secret' the the FIA is withholding.
There is no reason that the teams could not figure out the movement tolerance of the 'triggering' aspect of this start system. They have the transponders and can determine the frequency being used.
Does anyone have any knowledge about how accurate one can measure very small distances using radio frequencies?
Brian
How would the teams be able to figure out the imposed tolerance without resorting to trail and error (which itself is unreliable in this case). Sure they could deduce the measurement precision but the FIA has said the tolerance is some arbitrary amount larger than the measurement precision.
Re: Race start monitoring system
Posted: 12 Jul 2017, 19:23
by hollus
I feel a bit weird about this arbitrary precision. It makes sense, but "tolerances" are well defined in floor deflection, tire pressure, wing flexing, etc.
This sounds a bit like giving a goal in soccer when the ball "crosses the line", which would be defined, maybe, as x cm overlap between ball and line (x kept secret), or as the center of the ball crossing the line by x cm, or as only x cm of the ball remaining over the vertical of the line... sounds weird.
I can see why they do it, though.
Re: Race start monitoring system
Posted: 12 Jul 2017, 20:10
by hardingfv32
It would be my pinion that a 'secret' arbitrary precision could lead to questions of fairness. Does this type of officiating then start to flow into other rule sets… kind of a slippery slope situation.
Since the FIA has not had many secret rules before… I wonder if this is just not a coverup of a inadequate start monitoring system.
When you see the word "tolerance", think "uncertainty" instead. The point is not that the FIA is certain that a car is moving, and that they are giving the driver a small amount as a freebie. The point is that the FIA cannot be certain that the car is moving. All measurements have uncertainties, at all levels. On top of that, the transponder can move even when the tires are not rotating. An acceptable level of certainty is achieved only when the measurement exceeds the tolerance.
Nope, the fia explanation was that it was a tolerance and not inherent to the uncertainty. They can accurately measure movement but allow some freedom because the car can still have some movement inherent of putting it in gear or something like that, only when a driver moves more than x cm (x being a secret) the fia considers the car to start. Also x is bigger than the precision of the system. Luckily for Bottas he started moving (supposedly) before the red light went off but only moved this x allowed tolerance 0.201s after the start so it was deemed legal.
As in RZS10 brilliant posts you can imagine the graph is already moving before the time and thus is bellow the detection limit of the system but still within the tolerance parameter artificially designed to prevent small movements to trigger false start alarms.
I agree on the tolerance, but I think it's still uncertainty depending on the variable you try to assess. If that variable is movement, then the tolerance is much bigger than the measurement uncertainty. However if the variable is "the car has indeed started" I think some noise raises and there's where the tolerance probably come from.
