How things go wrong in an F1 windtunnel

McLaren's problems have showed no signs of going away, and the French Grand Prix weekend was its least competitive performance of the season so far, with both drivers eliminated in Q1.

After qualifying, it was very interesting to hear team boss Zak Brown explaining that McLaren could not identify its problem in the windtunnel even though it can see the problem on track.

Windtunnels are just a simple tool to prove your ideas. You can't see anything, so you have to believe they represent reality as closely as possible. The Toyota Motorsport facility in Cologne that McLaren uses for aerodynamic testing is one of the best, and the team isn't blaming the windtunnel. After all, a windtunnel is simply equipment that produces a vast set of numbers.

It's what you do with those numbers, and what you do to create them, that really matters. Aerodynamicists have to sit down to decipher them and draw conclusions on what's happening and what direction needs to be taken.

You can use a windtunnel in many different ways. That includes everything from just having the model sitting there in one position, to a fully transient model simulating the car going around the track and all the dynamic changes that involves.

Most tests will be done using a sequence of varying ride heights. This simulates the car's changing ride height because of different speeds - the faster you go the lower it is thanks to the downforce pushing it down, and when you brake it rises up again.

This is a critical area, because the diffuser will stall as it gets close to the ground and during braking the airflow needs to reattach very quickly as the car starts to rise up. If not, the initial braking can be tricky. Note that McLaren has been working on solving rear-end instability into lower-speed corners all season.

After an initial ride height survey to see if the new components are working as predicted, you would then do a 'roll and yaw' sweep to simulate the car mid-corner to see if it still performs. After all, it's in the corners that you want to improve the car's grip levels, so this test is critical.

That test then needs to be combined with a steering survey to see how the downforce, and more importantly the centre of pressure, changes with varying steering lock.

Different radius corners require different amounts of steering lock. So if, for example, the centre of pressure is moving rearward with increased steering lock then the car will just have more and more understeer as the driver applies more steering lock.

Doing the whole lot together is the best way, but this can be a very time consuming for every test component. So, with the restricted windtunnel running time already used, compromises have to be made.

The airflow separation characteristics of a car in the windtunnel and on the circuit will never be the same

It's those compromises that can get you in trouble. It is critical that any new components are fully tested, otherwise they can just be a waste of time and money and, more importantly, add to the confusion of which development direction is the correct one.

Sometimes you have to read between the lines. The airflow separation characteristics of a car in the windtunnel and on the circuit will never be the same, because the car is never in a steady state condition.

Any aerodynamic stall will be much more aggressive on the track and more aggressive when the ambient and track temperatures are higher. This is because the car is never sitting still, it's always on the move at some frequency or another.

In the tunnel, the car will go from one ride height to another and settle there before any readings are taken, but it's not like that on the track.

So, on the circuit, it will always be more critical in terms of airflow separation. This is an area that has caught me out on more than one occasion and really needs to be analysed in very close detail. The aerodynamic hysteresis of any aerodynamic stall must be fully researched and understood.

One of the main problems is how the airflow reattaches. For example, if you have a diffuser stall - which all the cars will do - let's say that happens at 24mm rear ride height when the car is being compressed by the increasing aerodynamic loads going down the straight.

Then the driver hits the brakes. Does the airflow reattach at the same 24mm on the way back up again, or is it slower? Let's say it doesn't fully recover until the rear ride height reaches 30mm; if this is so and the airflow stall is of reasonable magnitude, then the braking stability will be very inconsistent.

In that instance you could then adjust the aerodynamic balance to give the driver more braking confidence, but that will just lead to mid-corner and exit understeer - a spiral to nowhere.

Very seldom have we heard from either of McLaren's drivers that the car is a bit of a pig, or that it has too much understeer or oversteer. Normally they are reasonably happy with the balance, but it is basically too slow. I wouldn't be surprised if what the drivers say in public is all McLaren-speak rather than reality.

But the car on track will never be as the windtunnel figures suggest. The tunnel is a tool to help you understand and you use your time to explore the areas that are questionable and those that are a good crossover.

Even the surface finish on a 60% model will never be the same as on the real car. That alone can lead to correlation problems.

If McLaren knows what the problems are, then it's up to the team to instigate research to understand them away from the track too. It's much easier to do so away from the track rather than relying on what you do on a race weekend.

McLaren also has computational fluid dynamics (CFD) technology available, but it is a fairly new tool in the box. It's very good for some things, such as designing ducts where the airflow is contained by the component's surfaces and also for understanding airflow away from the surfaces of the car.

But to use it to optimise a diffuser will be incredibly difficult. With the car's ride height constantly changing and the car's underbody surface working against the track, things get very complicated.

If you are a bit lost, and McLaren is, then there is no point in doing more of the same

We see all the teams using airflow rakes during Friday practice sessions to try to correlate their CFD data to actual track data, but as of yet I haven't seen one under the car. Yes, the underfloors can be fitted with pressure sensors, but this is only one of the surfaces.

If you are a bit lost, and I think McLaren is, then there is no point in doing more of the same. You need to stand back and think of what you are not doing and increase your research in other areas. This may appear to be wasting potential development time, but sometimes you need to bite the bullet because the reality is what you are doing now is already a waste of time.

As the drivers will seldom criticise the car, I assume that its balance is not a major problem. That means it's basically the level of overall downforce.

From the numbers produced, any car with a nice stable balance will look good in the windtunnel, but McLaren will need to be finding 8-10% more downforce with no drag penalty. That's around 150kg of load at 240km/h and is roughly the difference between its performance and that of fellow Renault-powered team Red Bull.

Process is everything when it comes to windtunnel testing. You can go through the motions and blindly pursue one or two kilos of extra downforce, or you can slow everything down a bit and put more effort into trying to understand where and why you lose or gain that downforce.

When a team gets into a situation like McLaren is in now, sometimes the upper management - who have no idea of how to go about windtunnel testing - starts pushing the aerodynamicists for improvements in downforce levels. They are bosses, so the employees normally do what they are asked to do and, again, that normally leads to a downward spiral.

It's been said many times that McLaren has no hiding place this year. This problem with the aerodynamic testing is one area where it has been exposed. The one positive is that, as I always say, you learn a lot more when things aren't going well.

The question is whether McLaren can get on top of its many difficulties, which is surely what all F1 fans want to see.

Thin-gauge Pirellis back in the spotlight at French Grand Prix

During the French Grand Prix, we again saw the thin-gauge Pirelli tyres not doing much for the performance of Red Bull and Ferrari but certainly not hurting Mercedes. It was a strange race, fairly exciting all the way through but that's largely down to the dodgem-car tactics of some of the drivers on the first lap.

If it hadn't been for that, it would just have been another one-stop procession. Kimi Raikkonen was able to go to lap 34 on the ultrasofts, which was the softest available compound.

I'll say it again: hypersofts, supersofts and mediums should be Pirelli's standard tyre range for all circuits. Drop the others and simply call them soft, medium and hard. It would just be so much easier for everyone to understand.

If Ferrari really believes it could have won in France that then this is the attitude that could very easily lose it the championship

As for the overtaking, that was all down to a very long DRS zone combined with the headwind on the straight, which increased both the tow and the overall drag reduction.

In my opinion, DRS overtaking just makes F1 look silly. It's a bit like playing rugby against a team in which each player has one hand tied behind their back.

After the race, Ferrari said it believed it had the pace to challenge Mercedes for the win. I'm afraid if the team really believes that, then this is the attitude that could very easily lose it the championship, as I don't think anyone outside of Ferrari saw its performance at the same level of Lewis Hamilton in his Mercedes. And when you consider that Hamilton probably had a fair amount in reserve, he really didn't have any close competition.

Haas, Renault and Sauber survived to get some more points, and a special mention has to go to Charles Leclerc for qualifying eighth and finishing 10th. It was a professional weekend for the new kid on the block.

Coming back to McLaren's terrible weekend, Fernando Alonso wanted to give up at the end of lap 39. He came on the radio saying he had no brakes, no tyres and there was no point carrying on. At the end of lap 47 he came in and the team fitted a set of hypersofts so he could try for fastest lap.

This is the mentality of a small team trying to achieve something to sell to the sponsors, not a team of McLaren's standing. In the end, McLaren finished up 12th and 16th (last) with a broken rear suspension for Alonso on the last lap. This is not what McLaren expected from this season after dumping Honda for Renault. Never mind the huge financial hit, it's just not doing any better.

Comparing its improvement over the season to Sauber doesn't make for positive reading. Sauber started out at the back of the field and just over a third of the way into the season it qualified a car ahead of both Haas, Toro Rosso, McLaren and Williams machines.

Sauber has shown that it can be done, but you need to define the development direction and sometimes that has to come from seeing past the hype of the day.

It's a reminder that while McLaren faces a big task, it is possible to climb up the grid with clear direction and understanding.

There's also an interesting comparison to be made between Sauber and another fallen giant, Williams. Watching third practice from Paul Ricard in the wet, I noticed something intriguing.

Williams and Sauber were among the few teams that ran in those conditions. A car's airflow picks up the water, and although it increases the air density, it still takes it in the same direction.

Looking at the Williams, its outwash from the trailing edge of the bargeboards was very uniform. This, combined with the floor ducts in front of the rear tyres, formed a flow regime that took it out around the outside of the rear tyre. On the Sauber, the airflow was much more disturbed and much less consistent.

So, why was the Sauber able to qualify eighth with the Williams cars propping up the rear of the grid? Well, it's actually possible to do things too well.

The Williams team believes that it has a diffuser airflow stall problem. Well, if the bargeboards and the floor ducts are working as well as I think they are, it is then sealing the underfloor very well. This alone can lead to a diffuser stall.

On the Sauber, if it isn't sealed as well then the diffuser might produce less downforce, but it will be consistent downforce that gives the driver the opportunity to use his skills without ending up in the kitty litter - or in the case of Paul Ricard, the stripy run-off areas.

It's never as simple as making something work better. There are consequences with every development, some positive and some negative.