IIRC The main CFD processing used to be done at night when all the designers had gone home. They used to distribute the processing over ~30 unix workstations. We also used to have HP kit like Williams and Jordan.
Vehicle dynamics modelling was performed on PC's.
>>A microcontroller with a 128MB compact flash card could probably log all the data they need from a race
Maybe back in the 1980's. Today's loggers are doing a large amount of processing. They have around 200 different sensors on the cars, some logged at 10Hz (pressures, temperatures) some at 200, 300, 500 or 1000Hz. At these rates 128Mb is not enough for a race, so the number of channels and their frequency are cut back. This is not so important now that most teams have microwave telemetry.
IIRC the memory was not compact flash, it wasn't fast enough.
>>The real trick is making sense of the data.
No shit:-) In Addition, another trick was making sense of the data in a timely manner, which was my main objective. Just think about qualifying:
1. The car comes in. 2. You download the data from the car (twas by fibre) 3. The race engineer talk to the driver: "The car is rubbish" 4. The race engineer examine the data 5. The race engineer issues instructions to the mechanics 6. Mechanics make changes to the car 7. The driver goes out 8. Repeat 4 times during the (old) qualifying hour
The computing components are all well inside the body work, usually up in the nose or above the driver's legs.
Odd. There may be small junction boxes up front, but most of the electronics should be as close to the centre of the car, and as low as possible. Also there are sometimes cooling problems to take into account whih may affect the siting of the boxes
I know that at Benetton, the steering wheel also needed programming and communicated with the rest of the car via a CAN(N?) network.
Michael Schumacher is part CEO, part engineer, part driver and basically gets what he gets because he is a large reason that Ferrari executed the plan it had. He brought Ross Brawn with him from Bennetton Ford.
I dunno about CEO and engineer, but he did take a large slice of the Benetton team with him...but he forgot about me:-) When Schumacher arrived at Ferrari, they recruited Ross Brawn, Rory Byrne (Chief Designer), a few top of designers, + 1 Aero + 1 electronics guy. It was a blow to Benetton, but you have to remember that the team comprised 300+ people.
When he left it was Benetton Renault. We dumped Ford after winning the championship with them in 1995.
>>Unfortunately the difference between the top team's budget and the bottom team's budget is desparingly huge!
And it is compounded by the fact that succesful teams get huge prize money, free tyres, free travel, free engines (excluding Ferrari, Renault, Toyota etc)
>>What's going to wake up F1 is when you get CART cars which have rules governing the make up of the cars going just as fast as F1 cars
Maybe in a straight line, but an F1 car would leave a CART car for dust if you introduce some (fairly) wiggly bits.
It was a bit before my time, but I can remember reading that Gerhard Berger had, in his 1.5L car, over 1200bhp available to him during qualifying.
Now they are limited to 3L normally aspirated engines, down from 3.5L in 1995.
I used to work for the Benetton F1 team when Schumacher was around.
I joined in 94 as their sole software engineer tasked with writing data analysis, strategy and telemetry sw. 94 & 95 were great years - we won around 60% of that year's GP's and the bonus was $$$$:-)
In those days the on-board data loggers contained 4mb(now 128Mb) and the real-time telemetry was a pitiful 9600 baud, which didn't work 100% on every track (e.g. Hockenheim and Monaco).
I found Schumacher to be a good egg, fairly quiet, polite and interested in your work.
Happy days.
wrt the super computer etc, I'd take that with a slight pinch of salt. F1 teams are prone to exagerate slightly. PC's are adequate for most of their tasks excepting cfd and design work which is usually done on unix boxes.
Slashdot Mirror
A new study shows that companies hire foreign workers for cheap labor, not skill.
Do Bears shit in the woods?
McLaren just opened a new shop. $900 million worth
:-)
They should have spent it on the car
IIRC The main CFD processing used to be done at night when all the designers had gone home. They used to distribute the processing over ~30 unix workstations. We also used to have HP kit like Williams and Jordan.
:-) In Addition, another trick was making sense of the data in a timely manner, which was my main objective. Just think about qualifying:
Vehicle dynamics modelling was performed on PC's.
>>A microcontroller with a 128MB compact flash card could probably log all the data they need from a race
Maybe back in the 1980's. Today's loggers are doing a large amount of processing. They have around 200 different sensors on the cars, some logged at 10Hz (pressures, temperatures) some at 200, 300, 500 or 1000Hz. At these rates 128Mb is not enough for a race, so the number of channels and their frequency are cut back. This is not so important now that most teams have microwave telemetry.
IIRC the memory was not compact flash, it wasn't fast enough.
>>The real trick is making sense of the data.
No shit
1. The car comes in.
2. You download the data from the car (twas by fibre)
3. The race engineer talk to the driver: "The car is rubbish"
4. The race engineer examine the data
5. The race engineer issues instructions to the mechanics
6. Mechanics make changes to the car
7. The driver goes out
8. Repeat 4 times during the (old) qualifying hour
The computing components are all well inside the body work, usually up in the nose or above the driver's legs.
Odd. There may be small junction boxes up front, but most of the electronics should be as close to the centre of the car, and as low as possible. Also there are sometimes cooling problems to take into account whih may affect the siting of the boxes
I know that at Benetton, the steering wheel also needed programming and communicated with the rest of the car via a CAN(N?) network.
Michael Schumacher is part CEO, part engineer, part driver and basically gets what he gets because he is a large reason that Ferrari executed the plan it had. He brought Ross Brawn with him from Bennetton Ford.
:-) When Schumacher arrived at Ferrari, they recruited Ross Brawn, Rory Byrne (Chief Designer), a few top of designers, + 1 Aero + 1 electronics guy. It was a blow to Benetton, but you have to remember that the team comprised 300+ people.
I dunno about CEO and engineer, but he did take a large slice of the Benetton team with him...but he forgot about me
When he left it was Benetton Renault. We dumped Ford after winning the championship with them in 1995.
>>Unfortunately the difference between the top team's budget and the bottom team's budget is desparingly huge!
And it is compounded by the fact that succesful teams get huge prize money, free tyres, free travel, free engines (excluding Ferrari, Renault, Toyota etc)
>>What's going to wake up F1 is when you get CART cars which have rules governing the make up of the cars going just as fast as F1 cars
Maybe in a straight line, but an F1 car would leave a CART car for dust if you introduce some (fairly) wiggly bits.
It was a bit before my time, but I can remember reading that Gerhard Berger had, in his 1.5L car, over 1200bhp available to him during qualifying. Now they are limited to 3L normally aspirated engines, down from 3.5L in 1995.
I used to work for the Benetton F1 team when Schumacher was around. I joined in 94 as their sole software engineer tasked with writing data analysis, strategy and telemetry sw. 94 & 95 were great years - we won around 60% of that year's GP's and the bonus was $$$$ :-)
In those days the on-board data loggers contained 4mb(now 128Mb) and the real-time telemetry was a pitiful 9600 baud, which didn't work 100% on every track (e.g. Hockenheim and Monaco).
I found Schumacher to be a good egg, fairly quiet, polite and interested in your work.
Happy days.
wrt the super computer etc, I'd take that with a slight pinch of salt. F1 teams are prone to exagerate slightly. PC's are adequate for most of their tasks excepting cfd and design work which is usually done on unix boxes.