Open top menu
2 Dec 2015

This is reblogged from Willem's LinkedIn account with his permission: &

Power without control is worse than useless, although control without power is just frustrating (thanks Luca C).   Beautiful V12 Honda drawing.   I think racing engines are absolutely stunning works of art if you can ever see inside them.  This image of the V12 is not entirely accurate (you do not want to give away all your secrets) but it is a very nice example of a cutaway racing engine.

When you develop a racing engine, one of the things you try to do is maximise the power over the rev range a driver normally uses.  This depends on many things but one is how many gears you have.   Imagine developing an engine for a car with a perfect variable speed transmission.  You would be able to focus most of your effort over a very small rpm range.  If you have a 3-speed gearbox, as many road going cars had in the past, then you need a much wider rev range.  Modern road cars are tending towards having more gears, mainly so that the engine can be put into the best rpm range for fuel consumption even at higher (European motorway) speeds.

For the 1994 season the rules changed, removing active suspension and driver aids such as traction control.  This had an impact on engine requirements as well.

At Benetton the Cosworth “Ford” engine was used in 94.  It wasn’t the most powerful engine on the grid but it was really quite drivable and relatively simple to manage.   At Ferrari the V12 was a bit more of an “animal”.  No problem when you have traction control but as soon as it is turned off there could be issues.   What was the “problem”?   The drivers described it a bit like this.   “You come out of the corner and carefully increase throttle position until you can just go flat to the floor.  But then at 12,000 rpm wheelspin starts and I have to back off again – and I don’t know how much but it feels like a lot.”  Part of the reason for this is that tyre friction reduces with high levels of slip – so it does take a lot to reduce it again.  For an engine tuner it can be difficult to purposefully reduce power where you have peak torque in order to have a flat torque curve,  It can feel counter intuitive.  After all in the next gears the driver will want all the power he can get (even without downforce, as you go into higher speed gears, torque at the driven wheels is reduced due to the change in gearing).  After years of working with racing engines, both as a driver and as a member of the engineering team, I conclude that a driver likes to have a “torque pedal”.  He gets more acceleration with increasing throttle.   Perhaps obvious but it does not mean that it can be achieved with ease – in fact it is hard.  Of course, as soon as you are on full throttle you want more!
The 1994 Benetton Ford engine - picture taken in more recent times I believe.

In concept a 4-stroke engine sucks in charge air (1 of the 4 strokes – intake valve open), compresses it (2 - valves closed) the spark plug fires and the air fuel mixture pushes the piston down (3 – valves closed) and then, when the piston comes back up again (4 – exhaust valve open), the exhaust gas can escape.   When I first looked at camshaft timing back in the 1970’s I was shocked.  At this point engine builders will be laughing their heads off, I imagine!  They already know about valve timing.  The camshafts drive the valves which determine what can happen at which point in the engine’s cycle.   I found that there was a massive amount of overlap between exhaust closing and intake valve opening.   Then even more I was surprised at how early in the power stroke (3) the exhaust valve would start opening.   These timings are because of wave motion – which is fast – and I was clearly thinking in slow motion.

Despite valves being open together and being open “too early” compared to the view of a na├»ve young engineer, engines are remarkably efficient as pumps.   I reasoned that, with the best will in the world, an engine would pull in its theoretical capacity every two engine revolutions (4-stroke engine).   In practice, I thought, you’d never get there.  My life in F1 (measuring airflow into the airbox which feeds the engine with clean air) taught me that it swallows more than 120% of its theoretical capacity.   Pretty impressive and all due to wave motion.  The exhaust gas escaping from the pistons sucks some charge air out with them but, just before the exhaust valve closes, a pulse from the first joining of the exhaust pipes sends a pressure wave back which rams fresh charge back into the cylinder.   The inlet trumpets are also tuned to create a column of air that has momentum, which then pushes a bit more air in at the last millisecond.  Mere “conventional” aerodynamicists can only create about 4% of density with ram (due to car velocity) air but the acoustics make well over 20%.  The amount you can compress the air due to velocity increases with the square of speed so the percentage varies with the type of racing.

Another area of development for engines is the start.  Particularly true for very powerful cars.  To make a good start you need perfect control over the engine when it is producing quite low power – quite challenging to do.  With computer control it is easier to allow for things like the temperature of the clutch from pre-start burnouts but that’s not legal in F1 so the driver has to manage that himself.   A Formula 1 car goes from 0-100kph (62mph) in roughly 2.4 to 2.7 seconds depending on how good the start is.  Not bad but, as it only happens once per race, it isn’t the main design focus of the car.

I had the best (ex F1) engine in British Hillclimbing - a very special 4 litre Judd EV, and I sold it - to get more power.  Photo is of me in the Pilbeam Judd at Loton Park in April 2005.   Here I'm doing some pre-start burn outs.  650 bhp with truly delightful drivability.  This car allowed me to win a number of rounds of the British Hillclimb Championship and set an outright hill record (for a short time).  Image Derek Hibbert.

When a racing engine is “mapped”, starts are one of the areas that need to be considered.   The ex-F1 Judd I had in my Pilbeam hillclimb car was fantastic for starts.   It had wheel spin capable, but fully controllable, torque from 2500 rpm to 12500 rpm .  With a 145kph (90mph) first gear in the car I would hold the engine at 3800 rpm on the start line, take up the driveline slack and then drop the clutch (slide sideways off it) when it was time to go.  The rpm would drop to about 2500 and I’d just play with the throttle to control the wheelspin.  When it was slippery, of course, the start rpm would be even lower.   0-100kph was about 1.9 to 2.0 seconds.   Felt “hooked up”.  With the turbo engine, the starts went out to nearly 3 seconds.  Pretty poor by comparison and most of it due to lack of control.

At the end of 2005, having come second in the championship and wanting more, I looked around for what we could do to make the car faster.  Three obvious ways stuck me: increase aero, reduce weight and increase power.   To get the aero I needed to do some research and probably build a new car.  I rejected that as being too expensive and needing too much work/time from me personally – the day job in F1 really didn’t allow me time for that.  I looked at all the available options and chose a new power plant.  Not all the decisions you make in life are the best ones – but hopefully we learn from our mistakes!   The new power plant was an ex-Indycar engine of about 3.2 litres which would be rebuilt and tuned with twin rally turbochargers.  In the end we could not get any power down low and also could not reduce turbo lag to a reasonable level.  I went from about 650 bhp to nearly 900 but the car was much harder to drive and ultimately much slower coming out of corners (not to mention off the line).  As a driver the feeling was nothing, nothing, nothing, wheelspin (and sometimes “sideways”).  In truth the power change was probably from 200 bhp (feels like nothing) to 700 (spins the wheels and feels like it is out of control).   The project also ate up some money of course, so I could not “go back” and go for the new car with the old engine, or even an aerodynamic update on the old car, both of which would have been better choices.
Judd EV 4.0 in my Pilbeam.  Tyres wrapped in security wrap to keep them clean.
About the Judd EV 4.0.  It originally came out of a 1989 Leyton House F1 car and was a 3.5 litre V8.  For hillclimbing it had been “stroked” (more piston travel via a different crankshaft and conrods) out to 4 litres.  It had milder intake camshafts which gave it great torque.  You could idle it around the paddock under complete control.  Throttle response was fast – faster than anything I’d driven before.  After some experimentation I put about a 90 mph first gear into the car for most hills.  145 kph (90 mph) at 12,500 rpm.  Hairpins were down to about 40 kph so they would take the revs down to about 3,500 rpm.   With this engine more than enough to thrust the car forward (and sideways).   My competitors used to bring their engine tuners to the startline at hillclimbs and ask them to watch and listen.  I imagine they were asking why their engine didn’t get off the line like this one.  Start line marshals too would be warned – you won’t know when this one is about to go – be careful.

For a lot of hillclimbs I’d put the car into second once off the line and leave it there.  That left me with a go pedal and a stop pedal.  Nothing could be easier.  Just so you know I’m not a great driver – just very enthusiastic!

After two seasons of competition I had the engine rebuilt.  JF Engines I was told was the place to go.  Now, John and Fred (J & F) used to build these engines when they worked for Judd back in their F1 days.   They had retired but built engines they enjoyed building for people they liked – they were retired so they did it more for fun than anything else.  Happily we got on well and they agreed to do a rebuild.  I gently asked what it might cost.  Don’t know boy, depends what it needs but something near 2,000 pounds plus whatever special parts it needs.   Two thousand pounds!   Hell we’d spent 4 times that rebuilding the engine in the Peugeot 2005 hillclimber.  I didn’t believe them, but hoped for the best.  Well it needed some valves and a few other parts so it came to a bit under 2,500 pounds.   I could literally not believe it.   They were such lovely old boys too.   They found me a bit entertaining I think because I asked lots of questions about crack detection and how they did things.   Good powers of observation and experience told them more than most young people would learn with the best crack detection equipment on the planet.   They used it but only on a few parts.

When I put the turbo engine into the Pilbeam at the end of 2005 I sold the Judd.   Man, that was a daft thing to do!   I dare say if you went anywhere else to have your ex-Formula 1 engine rebuilt you can start adding zeros on the end of the quote!   I really like the song “Let Her Go” by Passenger.  A comment about how blokes sometimes have no emotional intelligence.  Not quite the same thing but...........  The great thing about life is that you’re never too old to learn – and arrogance is eventually knocked out of most people!   I hope I have learned.   A number of good friends did their best to warn me beforehand not to go there!

There is a short compilation video of some of the climbs from 2005 available via the link below.   At Shelsley the commentators talk about a start time – that is the time to complete the first 64 feet – at one G (acceleration) it would take 2 seconds.  It’s quicker than that – I think my best ever was 1.60 at Gurston Down.  If you watch the video. listen to the start rpm.....  Notice too how wide and accommodating the UK hillclimb tracks are (not)!

The image below – used in an earlier post is not a fake – this really happens with some cars at this place on Gurston Down.  Combine super sticky tyres with too much (really) weight on the rear axle and great torque and the car just comes up (almost gracefully at first) as the bump in the hill flattens.  You have to actively come out of the throttle.   There is one other place on the championship calendar where the car would lift front wheels and that is at Loton Park after the last corner if you’re having a really good run (lots of grip).  More dangerous there as it is higher speed and you absolutely have to come out of the throttle if it happens.
Image Gerry Marshall?  Pilbeam at Gurston Down.

For UK followers of the blog the following opportunities exist for anyone interested in Willem's work...

2 Dec  18:30 to 20:30 plus Oxford - Free public lecture - F1 Performance, Design and (maily) Aerodynamics see   There will be some entertaining stories and time for questions.  Free, book early.  This is the one to aim for if you work as it starts at 6:30 pm for 7 pm talk start point.   Organised by the IMechE.  Some refreshments available from 6:30 pm.   Questions and discussion due to finish at 20:30 but I'm happy to discuss any questions you may have for a bit longer.  I will bring additional material so we have the potential to illustrate answers to questions.

4 Dec - 13:00 - 17:00 approx.  Southampton University (Building 45 room 0045 which should be on the ground floor and is a large lecture theater - see site map here   Guests welcome and free.   First lecture is similar to the one on the 2nd - Formula 1 performance, design, & aerodynamics.  This does not require specialist knowledge.   Would be suitable for higher school pupils, motorsport enthusiasts, engineering students and engineers.  After the first lecture we then focus more on the use of CFD to develop a race car (aimed at university students but anyone using CFD may find it interesting) and how new aerodynamic testing restrictions (in the FiA regulations) are changing the approach F1 teams are taking to aerodynamic development.   There will also be discussions with the Formula Student team which are probably not open to all.

1 comment:

  1. Back in the late 80s and early 90s when I was much more interested in F1 than I am today (wonder why?) it occurred to me that the answer to some of the power delivery problems would be to have a drive-by-wire throttle linked by a fairly simple program to other parameters - like gear ratios. Of course you could add any number of additional reference values, and I'm sure that today they certainly do.
    It would then have been possible to have non-linear throttle response with the amount of pedal travel related to whatever gave best control sensitivity for the current mode the car was in - ie more travel with smaller increments where the possibility of wheelspin was most likely - and vice versa.
    I've often wondered if I was ahead of the game with this idea?


Total Pageviews