Active Brake Balance Control in F1

Some time ago I designed an active brake balance control system for the Prost F1 car, from memory it was for AP04. The idea was that aerodynamic load builds up differently front to rear with respect to speed and therefore when a driver brakes into a corner and slows the car the load on the tyres and therefore the amount of grip front to rear changes as the car slows. The front wing is far more efficient than the rear, and so as the car slows the brake balance should shift progressively to the front.

Also what we found is that the difference in brake hose length between the front and rear mean't that the rear brakes take fractionally longer to come up to pressure on rapid application of the brake pedal, than the fronts do. So very briefly you have a lot more front bias than you would ideally want.

So our answer was to develop an active brake balance control system. It's hard to know what the other teams were doing at the time, but, we were pretty confident that we were the only ones using it. It never actually got raced because the rules were changed to ban it as soon as we were ready to race it. It seems that 2014's regulation might require this kind of system and that the rules may have been changed to allow it once more.

The heart of any active system is the Moog valve. This is a servo hydraulic control valve that F1 cars use to control the throttles, gear change, and in the early 90's active suspension systems. It's a very sensitive and rapidly responding valve with a built in feedback system that allows you to control the pressure in a system very accurately, and extremely quickly.

The active brake balance system was in essence very simple. All we had to do we 'T' off of the rear brake hose and put in a hydraulic piston. On one side of the piston was the brake fluid and on the other side was the hydraulic fluid controlled by the Moog valve. The hydraulic fluid circuit also included an accumulator and obviously a high pressure pump (the same pump is used for the gear change and throttles). The Moog valve was controlled by software that had the cars speed, front brake pressure and rear brake pressure as inputs. The software did the calculations to compare the front to rear brake pressure split to the nominal selected balance with respect to speed, so in essence we had a 2D map of what brake balance we wanted for each particular speed. The Moog valve then regulated the pressure of the rear brake line increasing or lowering the pressure many thousands of times a second to maintain the optimum balance.

For 2014 with the addition of extra KERS harvesting and therefore variation in the amount of braking that the rear of the car sees during this harvest, this active balance system would be able to reduce rear brake line pressure during harvesting to keep the balance stable.

The simplicity of this system mean't that if there was a failure in the active system the rear brakes would still function as normal because the separator piston had limited travel and if the hydraulic pressure were to fail then the piston would move back against a hard stop allowing the brake master cylinder to build pressure as in a normal system.

F1 cost capping

There's been a lot said recently about the proposal to apply budget restrictions in F1 because a lot of teams can't afford to compete anymore.

I think that's been part of F1 since it's inception and the highly competitive nature of sport will always drive costs up but there must be a limit to how much sponsors will put in before it's no longer cost effective.

From the inside as an engineer the scale of spending is even more apparent and shocking than the spectacle you see each weekend. For example at Prost we spent £15, 000 making two ratchet tools to remove the nose box. Why? Because we could it's as simple as that.  The tools were beautifully designed and made and no consideration to cost was ever given to it. Now that was in a back of grid team that went bust 6 months later,  so imagine the attitude that a top team would be taking.

I don't want to see a slow down in development ir a reduction in staff obviously because that would hurt people like me. I'd like to see a reduction in waste and therefore the cost in that. We made some carbon rear wishbones for Jaguar. 5 sets to a new design proven to be better in the wind tunnel.  They had to be done in a rush to get to the Australian GP. The day we finished them they updated the design because further wind tunnel testing had proven another design change was even better. So all 5 sets were scrapped and we started again to the new design. 

The process for making a carbon wishbone calls for lots of tooling specific to each design. You have to make aluminum patterns,  carbon moulds off if these patterns,  then wishone halves out of these moulds and then bond them and final machine them. The aluminum for the patterns is already many thousands of pounds and machining time is 90 per hour or more and they take days to machine. The waste of 5 sets of wishbones like this is possibly close on £100, 000 today.  And it's pure waste which is just accepted.

So my cost capping rules would first look at finding ways to ensure parts are homologated in such away as to prevent design and manufacture of endless new parts that subsequently don't get used.

Vehicle Dynamics - a layman's introduction

For those with a casual interest in vehicle dynamics but need an introduction that you can understand and doesn't involve huge calculations and derivations I hope you'll find this informative. If you watch motorsport and wish to understand a little more about why commentators bang on about tyres and aerodynamics etc hopefully this will help too.

Tyres

First thing you must know is that the tyres are the only part of the car that can transfer the loads to the ground and therefore are the most important part and limiting factor in a cars performance. So you must understand how they generate these forces and what the limits are to their performance in order to know what the overall limits to your cars performance are.

We can all probably understand that the "grippy" rubber interacts with the ground and the friction between the tyre and the road is what creates the available traction. Newton's equation for friction tells us that the Friction force is equal to the coefficient of friction multiplied by the force that's pushing the two surfaces together (referred to as the normal force, because it's the force normal to (i.e. right angles to) the direction of the friction force. The coefficient of friction is a value that is derived from how "sticky" or "smooth" two surfaces are that are in contact. For most materials this friction coefficient is independent of the normal force, but for a tyre it isn't. More on that later.

So from Newton's friction equation we can see that the more we push the tyre onto the road the more friction we will generate and therefore the more traction we will have for cornering or braking or accelerating.




The above graph shows the relationship between the vertical load pressing the tyre into the road and the amount of traction. Traction for a tyre basically means the amount of force it can generate overall, so you can use this all for cornering, all for braking/accelerating or a combination of both. When you try to go above this force then you have the situation where you lose grip and spin etc.

Slip Angle

The blog link below gives a nice explanation of slip angle. But essentially slip angle is the difference between the direction the wheel is going and the direction where the car is going. Might sound odd at first, as you would hope this is the same, but, to a small degree this is not the case. The wheel during cornering will be at a higher angle than the direction the car is turning. This is because the contact patch of the tyre, is being deflected, the angle difference is referred to as slip angle, and up to a point the greater this angle the more cornering force you can get from the tyre. So on a lot of tyre performance graphs you'll have multiple lines, so you can see the relationship between cornering force and both the slip angle and the vertical load.

http://www.pratperch.com/2011/05/tyre-side-slip-explained/

When you get above the maximum slip angle that the tyre can work at, or that the road conditions will allow then the tyre loses traction. Anyone who drives will likely have experienced this in bad weather conditions, maybe at a roundabout, you turn in and maybe accelerate and the steering feels lighter and less responsive and the car doesn't take the course that you though it would. This is understeer, and it's where you've taken the tyre above it's slip angle limit for those conditions.

Slip angle takes account of speed and steering angle. If you think of a F1 car, it can't take a hair pin bend like the Lowes bend at Monaco at 200 mph, but it can take large radius bends at Silverstone for example pretty much flat out. What will actually happen though is that the slip angle is the same whether it be 30 mph and very tight of 200 mph and very large corner radius. At the tight bend the steering angle will be 20+ degrees but because of the speed of the car the direction the car is taking is close to this steering anlge, and at a high speed corner the steering angle will be much less and the direction of the car also much less but equal as far as the "slip angle" or the tyres contact patch is concerned. So slip angle is important to keep in mind but it allows us to ignore how fast the car is going and the steering angle and equate everything to what the tyre is actually doing.



On a racing tyre like an F1 slick you don't get a lot of warning that you are about to exceed the maximum slip angle of the tyre, and when a driver does go over this limit the tyre's traction rapidly drops off and generally causes a dramatic incident, like a spin. On a road tyre they are much more forgiving and road cars and tyres are designed to give us less competent drivers plenty of warning of impending doom. So the drop off after maximum slip angle is less dramatic. Back to the example of driving a roundabout in wet weather what an inexperienced driver will do is to wind more lock onto the steering in the hope that this will make the car turn more, but, it just makes the situation worse. Big accidents are then caused by the inexperienced driver jumping on the brakes whilst turning more on the steering, and then all control is lost. What should happen is that you wind off some lock so you come back into the usable slip angle range, gently ease off the power until the steering begins to respond again and then you can take back control of the car.

Suspension

So above we have described a little about the performance and limits of the tyres, now the suspension systems job (as far as vehicle performance is concerned) is to keep the vertical load on the tyres consistent. If you were to hit a bump in a very stiffly sprung car the tyre would see a rapid increase in vertical load (and therefore traction) and then the car would bounce on the tyres (because the tyres are very springy) as the tyre oscillates the load will go from very high to very low (low traction) to very high again until after some time the oscillations are damped out and the tyre load goes back to the normal static load from the weight of the car. Obviously if you are cornering, braking or accelerating whilst the tyre is cycling from high traction to low traction the cornering force will cycle from high to low and the cars cornering ability will suffer.

So the job of the suspension is to minimise the effect that bumps etc have on the normal load on the tyre. Just keep this in mind for one moment, whilst I explain about the effect of load transfer as the effect on efficiency is the same.

Load Transfer

When cornering the car will lean (or roll) over to the outside. What happens here is some of the weight of the car stops acting equally side to side on the cars tyres and puts more on the outside tyres and less on the inside tyres. If you were to stand with one foot on one set of bathroom scales and the other on another set, when you stand up straight both sets of scales will read half of your weight. If you then leaned your torso to one side one set of scales will increase and the other decreases by the amount the others increased. Your overall weight hasn't changed just the distribution from one leg to the other has changed. So the same is true with the car, what the outside tyre gains the inside tyre loses.

No problem you would think, we know that the tyres gain traction with increased vertical load, the problem is that it's not a linear relationship. When you look at the first graph above you'll see that as you increase the vertical load, the cornering force goes up, but not as much as it did for the previous increment in load. i.e. tyres are more efficient the less vertical load they have on them. From the first graph at 100 kg's of vertical load we get something like 140 kg's of traction, between 500-600 kg's of vertical load we only add about 50 kg's of traction. So in the case where we are transferring load from one side of the car to another, what happens is we lose more traction on the inside tyre than we gain on the outside tyre and thus overall the amount of traction the cars tyres produce is reduced.

Back to the case where the stiffly sprung car is bouncing on it's tyres and cycling between high and low vertical forces, what we get is less gain in traction at the high load points of the cycle than the loss of traction at the low load points, and therefore overall the tyre is less efficient and less traction is available.

Aerodynamic Load

When you load the tyres with aerodynamic downforce then as above we see that the traction force will increase. With racing slick tyres what you get is tyres that produce more cornering force than the vertical force acting on them. This is where tyres start to defy Newton because Newton says that you can't have a coefficient of friction greater than 1. Whereas racing tyres have a friction coefficient of around 1.5.

'G' we heard talked about a lot on racing commentary and this referred to the number of times heavier an object is whilst cornering or braking than it would be at rest. So a drivers head seems heavier to the driver whilst cornering than it would at rest. As far as the tyres and this blog is concerned "G" refers to the efficiency of the tyres. So a typical Formula Ford on slick tyres and no aerodynamic loads can pull about 1.5 G in the corners. So this means that the cornering force the tyres are generating must be 1.5 times greater than the force of the weight of the car acting on the tyres. Now with an F1 car and it's huge levels of downforce the cars can pull about 4 G in the corners meaning the tyres cornering force is 4 times greater than the normal force caused by the cars weight.

In the last two sentences I've made reference to the vertical force caused by the vehicles weight, because in terms of efficiency and therefore the G pulled during the corner we use the downforce + vehicle weight to determine the tyres cornering force, then divide that by the force caused by the vehicles weight alone, to determine the efficiency and G. So aerodynamic load is free efficiency if you like. I'll totally understand if that's a bit confusing, you might have to just accept it.

Centre of Gravity Height

Weight transfer as mentioned above is not good for a car tyres performance efficiency. The amount of weight transfer you will see is proportional to the cars centre of gravity height (CofG). If you imagine a car has a spot in the middle of it about which the entire cars weight can be considered to act and then when you roll the car over a few degrees this point will move over to outside of the roll. Therefore the weight of the car is now no longer acting perfectly in the middle of the two tyres, and is slightly close to one tyre than the other. Therefore more weight will be acting on that tyre, in proportion to how far this CofG point has moved from the centre point of the tyres.

Now if you lower the CofG, you'll notice when the you roll the car the distance off centre that the CofG moves is reduced, and therefore the amount of weight transfer is reduced.

The pictures above are an extreme view, but, they do clearly illustrate how the position of the centre of gravity moves to one side with roll angle, and how this is reduced the lower the centre of gravity.

There's a lot more to this of course, than I've written above. I'll collect my thoughts together and write more soon on how this relates to suspension geometry, and how CofG relates to other aspects of suspension geometry and vehicle dynamics.

F1 Carbon Pedals

In 2001 for the Prost AP03 car we were charged as always with the job of pushing forward the extremes of engineering, in particular weight saving on the mechanical aspects of the car.

I was looking after a lot of the brakes and cockpit components that year, and we fancied taking a look at the pedal arrangement. In F1 you have a funny situation of spending a lot of time designing parts to be a few grams lighter than the year before, and the same or better stiffness. Every year you believe that you have pushed it to the boundary of what is possible and yet the very next year you some how manage to rearrange things to pinch another few grams. In order to really make a sizeable weight saving you have to take chances and go to the extreme. At the time no one on the grid was prepared to take the chance and use carbon for the brake pedal. It's hard to explain why really when most of the rest of the car is made from the stuff.

The pedals on all of our cars had always been fabricated Titanium pieces. If you read my last blog you'll know the processes involved in making fabrications from Titanium. It's very labour intensive and tricky stuff to work with. Carbon provides you with a lot of ways of putting material exactly where you need it, without having to carry around material where you don't need it. You can vary the thickness through the part, you can orientate it differently to get different characteristics, and you can use different cloths and resin systems to get different material properties. The base fibres and resins are also stiff and very light, so ideal for making lightweight high stiffness parts so long as you can design the shape in such away as to physically make it in a mould(s).

Each driver has the cockpit laid out to their preference. So for each driver the pedals were different, shapes, positioned differently in the cockpit, so had side fences on to trap the foot in the pedal, others had a fence on one side but not the other, or like Alesi (picture above) no side fences. In this era the two pedal arrangement was just becoming popular, with the clutch on the steering wheel. Alesi preferred the three pedal arrangement pictured above, and so the brake pedal has no side fences because he needed to move his foot from one pedal to the next. Most other drivers on the grid used left foot braking and so there feet were held into the pedals to prevent them slipping off.

For the spare car you needed to be able to quickly change the cockpit to suit which ever driver was about to jump into it, so we had removable pedal pads that could be swapped over. At Prost the money was an issue and so in the later years we had a lot of different drivers paying for their seat, and each driver needed a new design of pedal arrangement, which drove me mad.

Mid season we decided to experiment and make a carbon brake pedal to retro fit into the car if it was successful. So the pedal set you see in the picture was born. The brake pedal is made in two halves, with a split line in the middle of the front and back faces. The pedal pad was a bolt on arrangement. The layup was varied to give a strong/stiff pedal face and basic neck of the pedal, and thinner in the less well loaded parts. The bottom was solid carbon, which we then machined and sleeved with Titanium, into which the pivot bearings were fitted. There was also a Titanium plate and insert where the master cylinder push rods attached either side. The two halves of the pedal were made so they socketed together and were bonded together.

Like all Carbon parts, you have a lot of things to make before you get to make your final component. You have first to make a pattern, which in this case the two halves of the pedal shape made in aluminium. Then these patterns get laid up with a tooling Carbon fibre pre-preg material, and "cooked" in an autoclave to produce the moulds. Now aluminium expands a fair amount with the heat of the curing process so the patterns have to be made smaller by the exact amount that they will expand. The contraction after curing does help release the patterns from the moulds. One you have the carbon mould tools, you can then make your final brake pedal, by laying up the carbon pre-preg in the moulds, and curing them in the autoclave. These mouldings then go to be machined which often meant making mould shaped jigs so they component could be held correctly whilst being machined. The final part of assembly is then to bond in the Titanium fittings and the two halves together.

Once we'd made the first one and tested it by applying loads to the pad face with an enormous hydraulic test rig we had, we knew that it was both considerably lighter and very much stiffer than the previous Titanium one. On subsequent pedals we were a lot less cautious with our layup reducing weight further. The unnerving thing about Carbon in these applications is that when you first apply the considerable load, the component makes large cracking sounds, as the fibres start to align with the direction of the load. You think it's going to snap, but, after repeated loading cycles it goes quiet. McLaren at the time had very sophisticated sonic measuring equipment and kept track of these noises on all components, and they used it to "life" each part.

Below is a picture of a Coulthard pedal that we went on to design and make for that years McLaren. David was very unsure about using a Carbon brake pedal and apparently took a lot of convincing. Like all things these have become a bog standard item now. In the picture below the two recesses in the pad face take gripper tape, which is essentially like a sticky backed sand paper to give the driver added grip.

Titanium upright design and manufacture

In what seems a previous life now I used to design Formula 1 cars. I thought you might be interested in some of the details of these extreme parts that don't come to light normally.

This picture is of a rear upright from the 2001 Prost AP03 car. It's fabricated from Titanium and a pair of them cost around £15,000 to produce.

The centre of the upright is made from a solid billet of Titanium 6Al4V, the first stage in the process is to wire erode all the triangular shape cut outs in the billet. These cut outs are used to pass cooling air from the brake duct through to the brake disc. The wiring eroding took about 20 hours from memory. This upright has a steel axle in it made in California by Metalore from a special steel that undergoes a special surface treatment. You can probably just make out that the tracks for the CV joint are machined directly into the axle.

The upright has a pair of pretty slender bearings that on this car featured ceramic balls. The ceramic balls reduced the rolling friction by an incredible amount and they are also considered self healing in that the balls are so hard if there's any damage to the bearing races the balls will smooth it out.

The outer parts of the upright were fabricated from Titanium sheet, normally 1.2mm thick, often formed over knocking blocks to get the shapes that we required. Brake caliper pickups, lower wishbone pickups, top wishbone and track rod pick ups are all Titanium machined pieces that are welded into the sheet.

Titanium has to be welded in an inert atmosphere or else is gets contaminated and can fail so we built special chambers where the Titanium parts were placed and the air removed and replaced with pure Argon. The welders had to weld them like you see Biologists in labs, buy placing their arms through in-built rubber gloves and looking through a window equipped with auto-darkening filters.

After welding the uprights were heat treated to remove any internal stresses.

The accuracy required means that any distortion due to welding would be unacceptable, so each critical part of the upright will have a stock of material on that after welding gets machined back to where it should be. The whole upright assembly would therefore be put on a lathe first and have the bearing features machined and then go onto one of our 4 axis milling machines to have all the other pickup points and mounting points machined in exactly the right places.

After machining these uprights were then shot peened and where then inspected and finally fitted to the car. The design of the rear upright system, including the upright itself, axle, wheel, brake package, carbon brake duct and tooling and all the data logging sensors for things like brake temperature took approx. 12 weeks each year and a pair of these uprights could just about be made in a week if two fabricators worked on them. I think we made 5 sets for the first Grand Prix each year and then another 5 sets during the season dependent on damage etc.

Below is a picture of the AP04 front upright, that shows the Titanium axle, Aluminium wheel nut and the full front brake duct assembly.

Design of monocoque chassis

I've been asked to explain how monocoque chassis are designed. It's a bit of a challenger because the details of how a chassis is designed is dependent on the materials used and the application.

However, the overall basic principle is the same. A ship, an aeroplane and most race and road cars could all be described as monocoques. They are all made from relatively thin walls formed into shapes, or supported by other elements to provide strength and stiffness.

A cardboard box is a form of monocoque, in that it had thin walls, and is only a rigid structure when the ends are folded to close the box off. What happens in the case of the box is that you have a shear plane on each face of the box and therefore you have a shear panel in each of the principle directions so that the box can't easily deform. In vehicles you will have a floor, sides, front and rear bulkheads and in a lot of cases a roof to form the enclosed stiff structure.

The problem comes when parts of the system feed loads directly onto these thin walls, obviously there is no stiffness there, so in order to stiffen the panels at the points that these loads are applied, by adding in bulkheads (to form an extra shear web). These don't always need to be full bulkheads, but they can be box section elements to help stiffen the panel, and direct the loads into surrounding panels.

For example an average car is made from sheet steel about 1mm thick, and if you ever try to handle a 1mm thick sheet of steel you'll see it's incredibly hard to handle. But fold a simple return flange around the edge even only a few mm's thick and it's stiffness increases dramatically. The floor of a car is essentially a sheet of steel, so it's not at all stiff until they fold it around the edges to form the sills, and they will form rib shapes into the footwell areas, and normally form or spot weld on a tunnel section where the gearstick and hand brake lever will fit. Normally they'll be a boxy section under the rear seats. There are plenty of pictures of floors and other car chassis parts being stamped online and no doubt on YouTube to give you a clearer idea of the sort of shapes and forms they add to the floor in order to make it stiffer. The rest of the cars chassis is similarly designed so that panels, folds and stamped forms add stiffness to the basic panel, before it all comes together to add stiffness and strength to each other sub-structure.

To simplify it further take a look at the common beam bending equations. Far and away the most important aspect of how much a beam bends or twists is the sectional properties of the shape. In bending that's I and in torsion J. If a car is to be stiff in beaming, then the general cross section needs to have a large Ixx value, and large Ixx values come from having the walls a long way the neutral axis, and from having lots of walls that are in the shear plane of the load. So for a car, you have a large approx. square shape, with a two extra shear webs from the tunnel in the middle and another couple of small shear webs formed by the hollow sills on the outside edges of the floor.

In a F1 car you have the same principles at play, except the Carbon fibre material used gives you a greater degree of control of the material properties. An F1 car is not actually all that stiff in comparison to a road car, and racing car with a roll cage in a production shell will often exceed the torsional stiffness figures of an F1 monocoque. This is primarily because of the Ixx value that you can achieve from the much larger cross section of a typical road car. However where an F1 monocoque wins is in it's stiffness to weight ratio.

Carbon monoques are have walls made of two skins of carbon fibre sandwiching a honeycomb material, which is usually aluminium. The aluminium honeycomb is nothing more than a thin foil maybe less than 0.5mm thick formed into hexagonal cells. The honeycomb comes in sheet form and typically F1 cars use about 10mm thick honey and 2-2.5mm thick carbon walls. Carbon fibre cloth is approx 0.2mm thick so you have to lay up around 10 sheet of carbon (termed plies) to form one wall. This sandwich structure forms exceptionally light and yet stiff walls for the monocoque. Far stiffer than the thin sheets of steel, and so you don't always need to make use of shape and bulkheads and other supporting structures in the same way as you would with steel. However, like any material if you try to put point loads directly into the walls of the carbon monocoque you'll have problems. Carbon is very brittle and you'll crack the skin of the monocoque unless the skin and wall is properly supported.

In a F1 monocoque in areas where they need to attached components like the suspension directly to the outside of the monocoque they be a solid carbon or metal insert replacing the honeycomb in the wall structure. In the case of suspension parts the walls will need added support and you'll find some bulkhead structure behind the inside wall to provide support. There is a great YouTube clip showing the layout of a Sauber F1 monocoque where the chassis is cut in half so you can see everything inside. This clearly shows where the bulkheads and other shapes and features are that support the outer walls when there are loads that require the added support.

A company called Hexcell composites have a great website where you can download a lot of details about working with their pre-made honeycomb sandwich boards. A lot of the techniques they describe in that documentation are used in vehicle composite structures.

With carbon composite monocoques you have the advantage that you can vary the material properties at any point on the chassis to aid stiffness, strength or support for other elements. You can thicken the area by adding more plies of carbon cloth, you can vary the orientation of the cloth so that more of the carbon fibres are aligned with the direction of the resulting load thus stiffening that area, or you can use Uni-direction cloth which has all the fibres runnin in a single direction, thus meaning that you can put all of the carbon to work in the direction of the load. Usually a carbon monocoque will be made of a mixture of cloth and uni-directional (UD) fibres. The UD plies are often wrapped around the chassis at approx 45 degree angles to help provide extra stiffness in torsion, and the woven cloth will be orientated differently through each ply and sandwiching the UD in order to support the UD fibres and provide a more universal load carrying capability.

Saying all of that, everything basically boils back down to the simple cardboard box concept, of thin walls supporting other thin walls.