Hacker News new | past | comments | ask | show | jobs | submit login
Formula 1: The secret aerodynamicist reveals design concepts (bbc.com)
175 points by clouddrover 39 days ago | hide | past | web | favorite | 66 comments

Interesting to watch qualifying. The Mercedes was definitely more of a handful so it looks like it has more corner entry over steer which will make the car pivot around the Apex and allow for sooner throttle application.

The Ferrari was more composed but on turn 14 Vettel struggled to get the car to the Apex perhaps showing signs of under steer. Less rotation, harder to get the power down, and more likely to destroy the front tires if you keep pushing too hard.

The Mercedes also looked to not enjoy bumps or curbs as much but as bouncy as it was it was still planted. And though Hamilton had to do more counter steering inputs than we are used to in formula 1 he was still quickest in qualifying and the oversteer looked like it was completely manageable and just the right amount. Probably around 15%-20% which is easy to control and quite fun to drive.

Reading the article it seems like the analysis is spot on. Not enough front grip because of the front wrong on the Ferrari which led to a tendency to perhaps under steer. What looked promising in testing was obviously Mercedes just not showing their full hand.

Still different tracks will suit different cars so will see how the first 3-4 races play out before giving the championship to Mercedes. Last year Ferrari was leading just about till the summer break and there were a couple of costly unforced errors from vettel that really cost him dearly.

F1 is the best form of technology applied to a sport. If you like impressive machinery take a look at Lewis Hamilton hitting 5 to 6.5 Gs multiple times in an 82.188 second lap at the Australian GP in 2017:


That s-curve at turn 11 starting at the ~53 sec mark (video above) with 6.5 lateral G-force is taken at 300km/h, down to 250 at the apex. This is the same lap with the speedometer instead of g-force meter:

https://www.youtube.com/watch?v=4mtjC9DozXs (turn 11 at the ~59 sec mark)

I almost feel stupid but I never realized they take the corners THAT fast. Holy shit. Respect. That's insane.

>I almost feel stupid but I never realized they take the corners THAT fast. Holy shit. Respect. That's insane.

The camera angles Rarely do the cars justice in protraying their speed (though its gotten better). There's a great vid on HN with a Porsche 919 sportscar at the Nurbugring that portays the quick factor fairly realistically (you can quibble about the various diffrences between a Sportscar and an F1 car in terms of capabilities--both are other worldy quick, and fast). I like to post that vid when talk here turns to self-driving cars, but I digress; F1 pilots are simply amazing. The cars are too.


Old top gear put a formula something car on the test track, and it was insane how the car moved compared to everything else i'd seen travelling "fast" on the same camera angles.

Holy cow - I can’t even hit that speed in a freakin video game! Let alone someplace where I could actually crash in real life!

Those drivers can’t be human. Proof positive aliens live among us. Wow.

You need to remove the leading double quotes because it's being carried when clicked on mobile, and leading to a 404 on YouTube.

>You need to remove the leading double quotes

Sorry, pulled an all nighter, didn't notice. Too late to edit it now.


Always fun and andrenaline releasing to watch the Porsche 919 fly around the nurburgring!

He takes one of those corners at 350kph

I can't even. The fastest I've been was 185kph. I always think back to that experience when watching stuff like this and try and empathize with the feeling of speed. It was insanely exhilarating to go that fast and I felt like if I even make a tiny mistake I'm gonna end up in a flaming wreck. And that was just down a long straight. I simply cannot imagine going almost twice as fast and taking a corner.

It is and it is not!

It is a formula. The 13 inch wheel size is an example of where being a formula gets in the way of innovation. Plus the non-movable aero rule.

The rules are there to make it a sport with the drivers safe, close racing and the smaller teams not out-spent by the manufacturer teams.

At the moment there is a lot going on for the 2021 season to fix the 'close racing' problem. The aero will be changed to throw the 'wake' up and over cars following the leader. At the moment any car wanting to overtake has to battle against a tide of messy air that has been disrupted by the car in front.

The list of banned things is a list of technology that was not applied. Ten years ago Brawn F1 won with a 'double diffuser', 'spinner' wheel covers and a very different aero package than allowed today. They were bought out to become the Mercedes team.

The Williams team brought along innovations that were tantamount to magic in the 1980's and 1990's. Active suspension appeared and was banned. On the same car there was traction control - now banned. We don't hear the term 'driver aids' any more but things taken for granted today such as a semi-automatic flappy paddle gearbox were not introduced easily to the sport.

The engine is built to spec in F1. It has to have a certain amount of cylinders, the fuel rate is restricted as is the total fuel load. You would not be able to put a car together that didn't bother with the engine bit and had batteries to get through 200 miles of racing.

The mystique surrounding F1 engineering is cool but sometimes it is best to not meet your heroes.

If you go to a car meet such as at Goodwood you can see some fairly recent F1 cars and what is under the hood. Some of the engineering is quite shockingly primitive. There are reasons for this, for instance, if only having to get through a race and taking the car apart every weekend then bearings can be more basic than what you might find on a family mass market car. It is all a lot more bicycle workshop than NASA when you see these cars close up. This is not to deride what goes on, but, if engineering is bespoke, doesn't have to have customers fix it and only needs to last two hours of race distance then it has different production values.

I think that Formula E on 'proper circuits' rather than city streets would be the apex of technology in sport. I am not alone in this, Formula E has more sponsors and the Formula 1 people that have jumped ship to it are not 'has beens'. F1 really is in danger of getting stuck in the slow lane with an elderly audience who can remember when F1 was shown 'free to air'.

> I am not alone in this, Formula E has more sponsors and the Formula 1 people that have jumped ship to it are not 'has beens'.

This is potentially a bit deceptive. The young Formula 1 guys who are driving in Formula E would rather have a seat in F1 and aspire to go back, and when it comes to money, sponsorship in Formula E costs approximately nothing compared to F1.

> I think that Formula E on 'proper circuits' rather than city streets would be the apex of technology in sport.

With some hypothetical, much faster, car that they haven't built yet. Is there some especially good technology that they're holding back from the sport?

Niki Lauda won a race for Brabham in the 80s driving a car fitted with a fan that sucked air from below the car. It got banned immediately, before the next race.

The 1978 Swedish GP in the BT46B.

Some say that he can make a brick fly using only a banana and a pair of socks and that he's been known to reverse the wind; all we know is he's called the Secret Aerodynamicist.

I really miss the Stig segments in the (now) old top gear with clarkson, may and hammond. Really unfortunate what happened to that show - it kept me going through some ugly periods in life.

The dynamic between the three of them is good but (and the ev supporters and musk drive me crazy too) what Clarkson did with Tesla was fundamentally dishonest. As was the Argentina thing. So.. clever and funny but also a bit suss.

The cool wall is every pub conversation ever.

The Argentina thing is the one thing that I feel like was actually just a mistake

If James may is to be believed, there were only two of that Porsche in the UK (And Jeremy was insisting on a(?) 928)

The Grand Tour is 95% the same show. In many ways I actually prefer it to Top Gear.

It’s so not. It’s conpletely missing what made the early seasons of new Top Gear so magical at times.

Now it is just typical 100% staged reality TV bullshit.

It is not really all that different from the later Top Gear Seasons though.

Having worked briefly on a unit attached to the show, huge amounts of Top Gear during the Clarkson era were also "staged". (I'm assuming by that you mean scripted).

Oh, I'm not saying that it wasn't, but in the early Clarkson years it was....less flagrant.. and the show was mostly (obviously scripted) studio segments. I have nothing against scripted television, I just don't like it when they pretend they're on some grand adventure when in reality everything was probably planned shot for shot in advance.

Everything became "oh, look how witty and clever and snarky we are" rather than humor developing organically.

> Everything became "oh, look how witty and clever and snarky we are" rather than humor developing organically.

Yes, agreed.

> Now it is just typical 100% staged reality TV bullshit.

Don't you feel they are mocking staged reality?

I honestly don’t like it, they are trying too hard to make it funny and while in top gear I sometimes wondered if the joke is scripted or a natural one, in grand tour I feel like every joke is scripted, and most of them are quite poor. Also they don’t spend as much time reviewing cars as they did in top gear, which I liked

I felt that the second season greatly improved from the first. I'd say the studio segments were still pretty subpar compared to the Top Gear days, but they seemed to get a better handle on the rest of it. I haven't watched the latest season yet, but hopefully they managed to improve again.

If you gave up after the first season, you might want to give the later ones a shot. Though I have to admit I never cared too much about the reviews, so that's worth keeping in mind.

s3 is really good - the car reviews are back, the panoramic views and cinematic genius of top gear is back.

Really? I honestly can't remember the last time I didn't think every single utterance in Top Gear was scripted. I can't remember the last time they reviewed a car either. Are you not thinking of Top Gear of 20 years ago?

The classic series of top gear (mid 2000s) are brilliant, organic and only scripted in terms of set pieces (No dialogue)

The script sounds like it was written by an AI trained on a combination of old top gear and an American dictionary (i.e. why is Clarkson giving temperatures in Fahrenheit when he used to shun all things American etc.)

Despite popular opinion, I like the new Top Gear (once they got rid of Chris Evans). The last two series with Matt LeBlanc was really good.

Will reserve judgement on the new-new line up with McGuinness and Flintoff [1]. McGuinness used to be funny, and Flintoff's podcast [2] is/was brilliant.

* [1] https://www.bbc.co.uk/news/entertainment-arts-45937945

* [2] https://www.bbc.co.uk/programmes/b08fr7t1/episodes/downloads | https://itunes.apple.com/gb/podcast/flintoff-savage-and-the-...

Chris Harris is a legend. The segments of him racing would've been impossible with the old lot

I've worked a lot with Chris Harris on various projects with car manufacturers. He is indeed a very nice man, with a great sense of humour, and hasn't let his slightly bumpy road to televisual semi-stardom change him.

Curious fact about F1 aerodynamics: the regulations limit the computing power of the CFD clusters they use. The limit is about 30 teraflops but the teams can trade off less teraflops to more wind tunnel time.

As someone working with semiconductors, I think the limit should be watts instead of flops to encourage competition in this aspect too.

Wow, how can you enforce a limit on simulation time?

Can't they use a secret cloud somewhere?

They could, but they get audited a lot.

The fines for cheating (unless you are Ferrari, usually) are catastrophic (for a non-works team): Lots of money and probable disqualification from the WCC

Afaik they need to provide logs.

Of course they could cheat but the penalties of getting caught are severe.

The limit isn't very high, though. A single box with a few GPUs is enough to hit the limit.

fyi most cfd software is not using GPU, current cfd clusters are traditional cpu solving systems.

also 30 teraflops of RANS simulations is a lot. https://en.wikipedia.org/wiki/Reynolds-averaged_Navier%E2%80...

basic cfd with turbulence modelling can get you pretty close and then refine in wind tunnels.

the other techniques are LES https://en.wikipedia.org/wiki/Large_eddy_simulation

genetic algorithms and iterative optimization is leading to more organic and nature inspired shapes in the latest generations.

Are they limited to specific software? I wonder if they could "cheat" using the new neural network NS solvers that came out recently.

We're just a few hours away from qualifying at the Australian GP, so we'll know soon enough which approach is working better (at the moment they look _really_ close)

So it turns out Mercedes were sand bagging in testing, as usual. Beat Ferrari by 7 tenths of a second! Anyways, that's the same gap as last year's race, interestingly it turned out Ferrari had a car cappable to win the championship.

https://www.formula1.com/en/results.html/2018/races/979/aust... - 2018 Aus GP Quali result

https://www.formula1.com/en/results.html/2019/races/1000/aus... - 2019 Aus GP Quali result

But Ferrari fixed their floor for the next race which made the car much stronger. They don't seem to have any "problems" right now so may not be able to catchup like last year.

I think it's going to be another season of Mercedes dominating.

As a Ferrari fan this makes me sad.

I was _really_ hoping for Seb to win this year. It would have set up Lewis and Seb both going for their sixth driver's championship in the last season before the upcoming rule change...

I always wondered about the downforce. If one design is few % better than another, providing the key advantage, does that mean the drivers are able to drive so exactly that they can use the extra force without spinning out?

Is there some sort of feedback when you're really close? I've only had a couple of track days, never came close to understanding it.

There's an old saying for drivers (and it's usually trotted out when your favorite driver has had an off) that goes "In order to know where the limit is, one must occasionally cross it."

Usually drivers are going to test that limit in an area with less risk, for example a corner with lots of paved runoff. But small factors such as the ambient wind changing direction, or the sun coming out or going behind a cloud, can affect grip enough that the limit changes enough to catch a driver out. In these cases there's usually warning, that the car very slightly stops responding as exactly to your inputs as it was the lap before.

However the most common scenario for a single car shunt is when the driver is testing the track limits; for example riding 3 inches of his tire over the kerb instead of 2, or using up the edge of the track to where the tire is now on a painted line rather than bare asphalt. In these cases there is often no warning, the car just steps out.

Try pressing your hand against a flat surface and then slowly applying enough force in the way parallel to the wall so the hand starts to slide. After some practice you will feel the point of losing traction. The concept is the same with tyres. F1 as of 2018 actually doesn't allow drivers to push 100% all the time, otherwise the tyres will overheat and the strategy is screwed.

Plus, 4 tyres may have different traction depending on car balance (braking, accelerating or cornering), downforce (nearly zero at 50km/h but at 300km/h F1 car is on rails, and different for the front and the rear), track specifics (for example, most clock-wise tracks put a lot of pressure on the front left tyre) and many little other things. It's really complicated.

> Is there some sort of feedback when you're really close?

If you watch the practice for the race coming up shortly in Australia, you'll see one corner in particular where pretty much everyone including the fastest guys slid off the track. So a lot of it's a feel thing, but you've also got engineering data telling you all the measurable stuff about how the car is performing, which you can then feed into a computer simulator which knows the exact (one hopes) layout of the race track.

One of the younger guys who qualified surprisingly well supposedly spent hundreds of hours on a PC simulator before this latest race.

Yes it's a big advantage specially on fast corners (where you don't lift or hit the brakes to take it). You can maintain or increase your speed if you have more downforce than your competitor.

I’ve never driven a single-seater, but to some degree it’s intuition, and to another it’s just getting to know the limits through practice. I think Gasly spinning in FP2 while trying to find the limit is an example of that. You’ll be familiar with the twitching feeling when traction starts to give.

I don’t know if they have any on-board telemetry to measure this.

The article makes it seem like all teams add more rear downforce as the season progresses.

I don't follow F1 - is this true about the increasing downforce and if so, why is that so?

Is it simply that the drivers get more comfortable pushing the limits?

I’m by no means an expert—or even an expert amateur—but I’ll try until someone with more knowledge comes in!

All of the technical regulations are set in advance of the season (this year’s is https://www.fia.com/file/78015/download/26184). The teams are only allowed to test their cars during a fixed number of “test days”, so they do a _ton_ of simulation pretty much all the time (https://arstechnica.com/cars/2017/04/formula-1-technology/).

The chassis (the tub that the driver & engine sit in) doesn’t change much over a season, so the time will be occupied by the team finding ways to move and re-shape things like front and rear wings, so as to increase downforce.

Downforce is king because it allows you to make sharper turns at higher speed without understeeting (where the front refuses to turn) or oversteering (where the back refuses to stop turning). And the technical regulations are really finicky about many areas of the car. And in many ways, the three “free practice” sessions before qualifying are as much about in-season development as they are about getting used to the track.

Yes, as drivers get time in the real car, they begin to understand where they can push. But also the engineers are seeing how the car is responding during real-world conditions, to feed that into the simulation.

> is this true about the increasing downforce and if so, why is that so?

The regulations have changed this year so the teams can't use the same aerodynamic designs they had last year. The idea of the new aero regulations is that they will promote overtaking, but it also means the cars will lose downforce:



They'll regain some (and maybe even all) of the lost downforce through new aero approaches that are allowed under the new regulations. The designs will be developed and refined as the season progresses.

My understanding is that the wash from old designs would reduce downforce for cars that follow, which in turn would slow them down and make it harder to overtake. This made for less interesting racing.

The new regulations are intended to promote overtaking. Teams of course are trying to maximize downforce within the new regulations.

> My understanding is that the wash from old designs would reduce downforce for cars that follow, which in turn would slow them down and make it harder to overtake. This made for less interesting racing.

My understanding is that to reduce drag, the intakes for cooling air (engine and possibly intercooler) have been minimised to the point where they are sufficient for clean air but insufficient for turbulent air. The engine loses power.

No changes have been made to the intake, it's the aero package. From the horse's mouth:


"What’s the change: A wider, higher – and much simplified – front wing

Why has it been made: To help chasing drivers follow the car in front more closely - and in turn increase the possibility of overtaking."

What about the underside of the vehicle? I remember hearing it was a super big tightly guarded secret. Was that ever true? Is it still?

Chain Bear F1 on YouTube has great video from an incident last season which helps explain the floor setup.


>What about the underside of the vehicle?

A very astute question. Yes, as much effort goes into the design of the underside as does the shiny side.

The most recent article I found via a quick G search is from 2013. For sure the rules have changed since then, but it's still a fair benchmark to illustrate the complexity of overall design with regards to the aero.


More effort probably, most of the downforce is from the floor AFAIK

It used to generate a big fraction of the downforce, so much that drivers pressed the accelerator when going into corners.


Right now this ground effect has been banned and it is mostly flat, except for some things in the tail.

Floors in F1 have been flat for a while with the exception of the FIA mandated plank down the middle.

The rear diffuser is sometimes blocked from cameras when the car is parked. For example 2009 with the double diffuser.

There isn't that much teams can do on the underside of the car. A lot of effort goes into preventing air from "leaking" under the car in the first place.

While you don't see the underside of the car much (for obvious reasons), the occasional roll-over crash (or crane lift after a 'regular' crash) gives us a glimpse from time to time. Check out Nico Hulkenberg's crash at Abu Dhabi pay year for a good example.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact