Interview with Rob White
The photo above of Rob White, on the left, and Dennis Chevrier
appeared in the Formula One section of speedtv.com on September
14, 2006. Adam Cooper questioned Rob, ex-Cosworth now Engine
Technical Director at Renault F1, on the details of Fernando
Alanso's engine failure late in the race at Monza.
Adam Cooper: What are your first thoughts on the failure?
"I'd be a bit less grumpy if it hadn't happened. We've
had better weekends, and we'll have better weekends again. At
times it's a beautiful sport, at others it's an ugly sport! It
was a big, instantaneous mechanical failure with no warning from
any of the surveillance systems. We know Monza is a hard circuit
for the engine, but that in itself isn't sufficient to explain
it. It's best to reserve judgment until we get the engine in
pieces and have a look. From that we'll have to build an action
plan to make sure that we're not in the same position at the
Q: We know drivers have extra revs to use from time to time
during the race. Bearing in mind he had to come through the field,
had he used up his allocation?
"It was always going to be a long, hard race. But it
was inside the way in which we had expected to use the engine,
so we would want to look further than that to explain what happened.
My Interview with Rob
After I saw this piece I took a look at the interview I did
with Rob more than a decade ago and decided it's still interesting.
Today engine developers probably use more simulation software
but I'll bet most power gains still come from clever improvements
and running engines until they break.
This interview appeared in the Engine Chapter of Inside Racing
Technology published in 1995. I still have a few copies left.
If you'd like to have one send me an email using the link above
in the left column.
This is a very long interview but worth your time.
I visited Cosworth's Southern California facility several
times in the early 1990s. This is where they assembled engines
for the CART IndyCar series and I assume they now put engines
together for the current Champ Car teams. In October 1993 I went
again for an interview with Rob White who was the CART Track
At that time I was traveling to most of the CART races writing
articles for my newsletter, TV Motorsports and compiling material
for my first book, Inside Racing Technology. Racing technical
people aren't always forthcoming when questions are asked. Rob
was and, I assume still very helpful when asked a serious question.
He knows what he's doing well enough that he can provide useful
information without divulging proprietary stuff.
As I do before all my interviews I spent some time thinking
about what questions to ask and came up with the ones below.
I faxed them to Rob about a week before the trip.
1. Background, personal and professional.
2. Are developmental increases in power planned or serendipitous?
3. What are the sources of power increases? Mechanical improvements
4. What are the trends in materials?
Flame or plasma spray? Chemical vapor deposition?
Metals: titanium, metal matrix?
Carbon fiber rods?
5. How are fuel mileage gains being made?
6. What do multiple injection nozzles accomplish?
7. What are the current limiting factors? Flame travel? Component
8. What is the weak component at max rpm?
9. Piston trends. Materials? Ringless?
10. What changes are made to engines to maximize performance
at different tracks? Fuel and ignition maps? Mechanical parts?
11. Do you have simulation capabilities that work? Design? Dyno?
12. If you have the opportunity to influence chassis design,
what do you try to do?
13. What are the trade-offs on pneumatic valve control? Is rising
pressure a better choice than constant pressure?
14. Trends in overall packaging?
15. Most difficult/critical component to design/develop?
16. What do you do at a racetrack on race weekends? Technical
support? Sympathetic ear? Customer communications?
Interview with Rob White
When I arrived at the Cosworth shop in Torrance I was greeted
by Ian Bisco, the guy who ran the facility, and he and I sat
down for a talk about general Indy car/Cosworth stuff. Later
we found Rob and he took me into an office to talk. We sat down
across from each other at a desk, and I set up my tape recorder.
Rob had prepared for the talk by reading the questions I sent
and making some notes. He referred to the questions and his notes
throughout the talk.
Rob White is a very likable, university-educated professional
engineer in his thirties. Seen in an Indy car pit, Rob seems
intensely focused and a little stiff, but when he talks one-on-one
his eyes sparkle with enthusiasm and small smiles come and go.
Rob has that charmingly-British, disarming, almost-shy, self
depreciation that compliments his technical strength and helps
him perform what is a very, very difficult task, dealing with
Indy car owners, crew chiefs, and mechanics. They expect a lot
from a Cosworth engine.
After I got set up, Rob looked up from my question list, grinned
and started to talk.
RW: "I anticipated most of your questions, but not the
first one [laughs]. The one about how I came to be doing this.
It started a long time ago. When I was in school I used to race
karts and tinker about with Formula Ford engines in my spare
time. When I left school I worked at Jaguar Cars in Coventry,
England, and as a result of that they sponsored me to go to University,
and I did a bachelor's degree in meechanical Engineering. But
as soon as I got my degree I up and left to work for Cosworth.
I've worked for Cosworth since 1987, and most of my time at Cosworth
I've worked on the Indy car engine. In the early days I worked
for Steve Miller, then the chief engineer. It was essentially
just the two of us engineering the DFX, which would grow into
the DFS. At one point we decided we needed to do another engine,
and right from the beginning of the XB engine project I've been
in there in the thick of it. After the XB was laid out, designed,
and detailed, then I was working directly for Steve and responsible
for the performance development of the engine. At the beginning
of 1993 I came to Torrance to look after the servicing of the
engine and take care of the mechanical spec of the engine as
built at Torrance, which is a satellite plant. All the parts
of the engine are designed and built in England and serviced
PH: "Where did you go to school?"
RW: "The University of Southampton."
PH: "Was that a theoretical program? Did you get any
manufacturing exposure there?"
RW: "There was quite a bit of manufacturing content in
my degree courses, but in truth, nothing we could do at school
could compare with the practical experience you get when you're
doing things for money [laughs]. I learned a lot about that building
automobiles at Jaguar, and then, of course, one of Cosworth's
greatest strengths is its manufacturing. Even to this day, every
time you go into the machine shop you've got to be dumb not to
learn something. The place is still very close-knit, and you're
perpetually in contact with manufacturing people throughout your
working life at Cosworth. The manufacturing process and the discipline
it takes is crucial to making this whole thing work. We couldn't
work without it."
PH: "Were you a car nut?"
RW: "No, not really. I've never been a car nut. I'm a
working mechanisms sort of person. If you can take something
apart and put it back together and take out some of the clutter
inside, it wouldn't really matter if it was on a car or a boat
or an airplane bit. I'm still not particularly interested in
street cars. Engines always fascinated me; anything mechanical,
clever gizmos and stuff like that."
RW: [starting on the list of questions] "Are development
increases in power planned or serendipitous? Bit of both I guess
is the answer. About 50-50 [grins]. In the development stage
of an engine you have to sort of organize your life. What happens,
I suppose, is you plan to go looking for some performance and
sometimes you find it and can put a tick in the box and say that
was a planned increase. Other times you don't find what you're
looking for. I guess it's important to be aware that even if
every time you make a change or do an experiment you don't find
a big lump of bolt-on power, oftentimes some learning takes place
along the way which you can sort of pigeonhole away for next
time. Undoubtedly though, there are times when you just get a
lucky break, and you stumble over something or you go looking
for something and it points the way to something that you haven't
anticipated. Probably the best examples of those you can't really
talk about, but there have been one or two of them.
"In general, to try and dispel the mysticism that there
is out there, most of the performance that you get out of an
engine after you move from putting the pieces together for a
new design is just a matter of continual optimization, and it
comes down to the age-old thing that people have known about
engines for probably 80 years ,which is if you want to make more
power, you've got to get it to swallow more air, make it breathe
better, and run more quickly. So you're into moving closer to
mechanical barriers which always exist, or you try to move the
mechanical barriers a little bit further away from where you
are at the time by making some changes. Everything that we do
in the way of performance development is focused on these two
areas. Sometimes you find that you're machining big pieces to
make a change to make it breathe better, and sometimes you're
doing what appears to be infinitesimally small work with, for
example, ports or valves or valve seats on a flow bench. Sometimes
you're working in thousands of an inch and other times you find
yourself sawing a couple of inches off a piece of steel pipe.
"Our goals are similar throughout all that though, and we're
always trying to make it use more air. That's because air is
the working medium for a reciprocating engine, and it's the stuff
that you don't have direct control over. You have direct control
over the amount of fuel that you put in and you can basically
put in as much fuel as you like. You have to pace that to the
amount of air that the engine will swallow. So, when we go to
the development phase of an engine, we go after making it swallow
more air first off. Then fuel it accordingly.
"The other big path that you follow is reducing mechanical
losses. The real big lumps of that come about when you redesign
your engine. Most of the mechanical losses that we're fighting
against are defined by the architecture of the engine, the layout
of the pieces, and based on the base dimensions that you choose
for the major components, and so are pretty much cast in concrete
as you move into the development phase of the engine.
"There are little things that you can do, and because
it's a game of optimization, there's some stuff you can do once
you've put your first engine together, things like scavenge pumps.
There's a balance between the power consumed in driving these
pumps and the power saved by running the crankcase in depression
[low pressure], and by completely evacuating the oil from the
engine. So you have to balance the two, and somewhere in between
there's an optimum speed for the scavenge pumps to run relative
to the engine speed."
PH: "When you're designing an engine, what do you work
with? Paper, computer, CAD [computer-aided design]?"
RW: "It's a mixture at Cosworth. It's still a very small
group of people who actually do the layout of a new engine. For
the Indy car engine right now that's Steve Miller, chief engineer;
Stewart Groves actually designed the XB engine, but he no longer
works for Cosworth. He works at Ferrari.
"The major items are the cylinder heads, block, sump,
crank, rods, pistons, and then the cam drive mechanism. The pumps
aren't trivial, but they're a packaging exercise most of the
time. This is one of the benefits that a company like Cosworth
has in that there is a lot of prior knowledge, a lot of information
available from previous designs, and, although it's important
to assess what you're doing each time you make a new engine,
it's equally important to make use of what's available to you
from previous times. There are some things which take place in
the pumps that make such good sense and get modified only after
a lot of soul-searching, and, as a consequence, the auxiliaries
themselves give us, typically, [knocks wood, grins] very little
trouble. I guess that the cylinder head is the biggest piece
in an engine. That's the piece that you tend to guard most jealously.
"Going on down the list, then?"
PH: "That's great."
RW: "Usual sources of power increases, mechanical or
software? Almost all are mechanical improvement and very little
software. In terms of raw horsepower there's almost nothing you
can do electronically or in software to improve the performance
of the engine."
PH: "So software has mainly improved the drivability
RW: "Yeah, and also the thing that's often overlooked
in the favor of the electronic stuff on the engine, is the facility
it gives you to do mechanical work so easily. Inevitably, when
you want to make a mechanical change, you have to make a change
to re-optimize fueling, ignition, or injector timing, and, with
electronics, you just twiddle a knob or edit a table and you're
there. Whereas, in the old days, if you wanted to change the
ignition timing, then somebody had to go in there while the engine
was running and crank a trigger backwards or forwards or, before
that even, crank the distributor backwards or forwards. If you
wanted to change the fueling of an engine while it was running,
with carburetors, you had to be hauling around choke tubes and
emulsion jets and other little pieces of the carburetor. More
recently than that and probably even more ghastly, you had to
tinker with cams and springs and orifices and restrictors and
such things in mechanical fuel injection metering units, all
of which would distract you from what you're trying to do, which
is investigate a mechanical change."
PH: "Now all that is just a huge waste of time that you
don't have to worry about anymore."
RW: "Sure. The electronics greatly increases the amount
of mechanical work you can do but, on their own, they're unlikely
to give you any power."
"So, I'll go on to your materials question. ['What are
the trends in materials?' was the question, and I listed some
possible subtopics: friction reduction, insulating--flame or
plasma spray; metals--titanium or metal matrix; ceramics; carbon
"All engineers are interested in materials, and we as
much as anyone else. Fortunately, most of the things you have
here are either specifically banned or their use is limited in
the rules for Indy cars. I guess it's still worth saying that,
in modern race engines, the most important materials are still
high-strength alloy steels and world-class aluminum castings.
There are other materials in there and you try to optimize the
materials you use with heat treatments and surface treatments,
but the real building blocks of the engines are still relatively
PH: "I've read that Cosworth has a casting process that's
better in some way?" [Since sold to Ford.]
RW: "That's true. I'm not sure of the history, but maybe
15 years ago or so we decided that we should have our own foundry
because of the difficulties you have in getting aluminum castings
of good dimensional accuracy and good structural integrity. Bottom
line now is we have our own foundry with our own unique process
that achieves this by a quite novel means.
"The basic difference is that instead of taking molten
aluminum from a vat and ladling it into molds, the stuff is pumped
out of the vat so you don't have a ladle passing through the
scum on the top of the vat or disturbing the stuff on the bottom.
You take pure, clean, molten aluminum from the center of the
vat and pump it through ceramic pipe with a type of pump that
has no moving parts. It's an electro-magnetic device of the type
used in nuclear power stations to pump liquid sodium.
"You haul this clean molten aluminum out of the middle
of the vat, and you pump it into the mold from underneath. So
the mold gets to vent at the top, and you don't have the gas
bubbles passing up through the whole melt. It has a clear passage
to the outside world. The other thing that's important is that
if you maintain a modest pressure on the bottom of the mold until
it's solidified, a number of things contribute to the improved
integrity of the casting. First is the cleanliness of the material.
Second, there is less porosity because there's a good route for
the air to get out of the mold, and the pressure helps too. There's
another important difference in the process, and that is the
sand used in the mold is not the same stuff that's used in a
typical molding application, which is usually silicon sand. We
use zircon sand. The key feature of zircon sand is that it doesn't
have a phase shift close to the temperature where aluminum melts,
whereas silica sand does have this phase shift. What happens
when silica sand undergoes this phase shift is there is a small
volume change which results in a dimensional change to the casting.
Fundamentally, the zircon sand gives you a casting which has
better dimensional control. The stuff is damned expensive, and
the only way you can do it is by recycling the sand. Another
part of the process is a novel way of reusing the zircon sand
so it's viable. I'm not an expert on the casting process and
that's about all I can remember.
"Our basic materials are high-strength steels, good castings,
and optimizing surface treatments. We use nitriding on crankshafts,
tufftriding on cams and tappets, and plating or vapor deposition
processes on titanium pieces.
"Titanium is an interesting material because it's light
and very strong. Typically it has yield strengths as good as
a reasonable piece of steel. It has several important disadvantages
compared to steel, though. First of all, it has terrible properties
when any sort of sliding friction is involved. Anytime a titanium
piece has to rub against anything else it's bad news. This means
that almost any application you can think of for titanium involves
a part that's moving, because it's used so you can reduce the
reciprocating mass of something. So anywhere you use titanium
there has to be a coating."
PH: "You're doing some chemical vapor deposition of materials
on titanium to solve this problem?"
RW: "Yes, but most of that is proprietary to a subcontractor.
Wherever you have titanium you have to do something to the surface.
We've been involved in research to try to improve the surface
properties of titanium, but in our engine there's not a single
component in the engine that doesn't have some sort of surface
treatment on it. And I can't think of a titanium component in
anyone else's engine I've ever seen that has been untreated either.
Even if it's only anodizing, or sometimes you have to apply a
clever surface treatment to the other piece. One example is titanium
fasteners. You'll typically use an anodized surface on the titanium
and a silver-plated nut for the female thread. Titanium fasteners
are not allowed on Indy cars, but if they were that's what we'd
have to do. I think I'm right in saying that, in Indy cars, titanium
is only allowed for valves and reciprocating parts in the valve
gear. Con rods are specifically excluded. Fasteners are specifically
excluded. Likewise carbon fiber and ceramics."
PH: "Just a speculation, is anybody using carbon fiber
rods now? In F1, for instance."
RW: "I can't imagine anybody who builds racing engines
in the fashion that we do, using them. The Mugens, Ilmors, Judds,
although I have no idea in truth, but I don't see them being
able to do that sort of work. Typically nobody has the resources
to devote to high-level research of that type. We'd like to make
use of novel materials as soon as humanly possible once they've
become established and proven technology."
PH: "So this is probably a much more conservative business
than you would think if you believed everything you read."
RW: "There's certainly some truth in that. It's conservative
in terms of the materials we use, but the design and the detail
design is very adventurous. The performance we achieve is very
adventurous despite these largely tried and tested, proven materials."
PH: "I was surprised to hear Eddie Cheever commenting
on ESPN's F1 broadcast speculating that the Renault engine might
be putting out 100 hp more than the other engines. One percent
seems a lot to me."
RW: "It's very difficult to speculate on that, but I
would have thought that, for one engine to be more than 3% or
4% up on another engine, is a real big step. Well, let me see,
1% is 7 horsepower so I guess you could be 5% ahead....or behind
[laughs]. Depending on your point of view. Yeah, you could be
5% behind in terms of raw horsepower. You're unlikely to be 5%
behind in terms of BMEP [Brake Mean Effective Pressure]. The
combination of BMEP and the speed you can run the thing, and
you might end up being 5% ahead or behind.
[Rob looked down again to the question list.] "How are
fuel mileage gains being made? Essentially the same mechanism
as for power. It's a matter of optimization. You've perhaps got
quite a long way to go in a methanol engine, because you're starting
from a position of less knowledge. You're starting with a fuel
that's less well understood. You're starting with a fuel that
has many endearing properties, but burning well isn't one of
them [grins]. I think where we are right now, we run methanol
engines far richer compared to gasoline engines of similar technology.
Some of this is linked to other elements in the rules package,
no intercooler, for example. I think it's fair to say that, certainly
in the past, everybody would run rich in order to preserve pistons.
"One of the things that has happened recently is that,
generally, people are running engines a little bit leaner and
there are mileage benefits as a result of this. At the moment,
very much without exception, the only way of increasing the efficiency
of the engine or reducing its specific fuel consumption is by
running it lean. None of the other measures that are taken by
teams at the track successfully improve the efficiency. People
who short-shift in order to get mileage or people who are altering
their driving style may be achieving better miles per gallon,
but they're not doing anything to improve the efficiency of the
engine. There are still measures we can take to improve the efficiency
of the engine which, of course, is reducing its specific fuel
"The next question, I think, is linked to the above.
What would multiple injection nozzles accomplish? Essentially
it's a question of mixture preparation. Particularly in Indy
car engines, where the amount of fuel you put into the engine
is horrific. With the type of speeds you do around the Speedway,
then you're using something of the order of 2 gallons a minute
of fuel. You can figure that easily by looking at the lap times
and how often the pits stops are. That's perhaps the rate at
which a bucket would fill from an ordinary water tap. It's a
lot of fuel, and passing it through a bunch of nozzles and trying
to prepare it for combustion, trying to get it into a highly
divided mist, it's better to pass it through a bunch of little
nozzles than trying to get it all through one big nozzle [laughs].
It's pretty straightforward. The best you'll get through a big
nozzle is a jet, but you want a fine spray.
"And then the position of the nozzles is the same thing,
trying to prepare the mixture the best possible way. Less fundamentally,
but of equal practical importance, is that you tend to use proprietary
pieces for the injectors on the engine and, at the type of flow
rates that we use on the engine, there's a very limited number
of parts to choose from. Unless you're fortunate enough to have
an injector manufacturer in the palm of your hand, you probably
need to use something that's already in existence. It may well
be that this forces you, in an Indy car engine, to use two injectors
per cylinder even if you wanted to use just one. As it happens,
every single Indy engine in the last five years has two injectors
per cylinder. The DFX did, the Chevy A did, the Porsche did,
Alfa did, our engine does.
"The other thing that it can do for you is give you the
opportunity to improve the accuracy of the fuel at part throttle,
because you have the opportunity to use only one injector at
very low load. Clearly, if you have this huge fuel flow going
through the engine at full throttle, then at idle you have more
or less no fuel going through the engine, so you have to have
a solenoid device which can turn itself on and off with sufficient
accuracy to meter the fuel at both flow extremes. Having multiple
nozzles gives you the opportunity to turn some of them off completely
and double the flow rate through the other one that stays switched
on, therefore reducing the dynamic range that the injector needs
to have. That's important to fuel consumption and the accuracy
of measuring fuel consumption.
"Current limits in the engine? Flame travel and component
fatigue strength? 'Yes' to both of those. The flame travel thing
is really interesting. Obviously if you have infinite flame speed,
then you could fire the engine at top dead center and would need
to do no negative work. Typically you'll find that you have to
fire the engine 30 or 40 degrees before top dead center, and
all that time there's a flame growing and the pressure is rising
in the combustion chamber at a faster rate than it would normally
rise just from compression, and negative work is power that's
being stolen from the engine. If you can improve the flame speed
you could get a lot back.
"But it's difficult to improve the flame speed and so, for
most of us, it comes down to the layout for the combustion chamber
and the detail design for the top of the piston and the cylinder
head around the valves."
PH: "I've read numbers for Formula 1 piston bores and
strokes that indicate they're going to bigger and bigger piston
bores and shorter strokes. Are their flame travel problems worse
than yours because of this?"
RW: "I think the trend toward bigger bores is being paced
more by the old air consumption problem. Bigger bores definitely
get more air in and then there's the reduction of mechanical
loads in the con rod and big end bolts with shorter stroke. The
flame speed issue would tend to push you in the opposite direction,
and the fact that we're still moving toward bigger bores means
that we haven't got to that limit yet. Or rather that the balance
still weighs in favor of bigger bores.
"It's good that you've described it as component fatigue
strength rather than material fatigue strength. If you had infinite
fatigue strength we could run the car infinitely fast and get
more power. We can't do that because the bits break. It's important
to remember, though, that most of the parts that break in fatigue
break of a result of their detail design rather than for a lack
of material strength. Almost every piece that suffers a fatigue
failure in the engine will have some form of stress raiser on
it as a consequence of the way it's designed or manufactured.
You have to get rid of any of that, and the most powerful tool
is detail design."
PH: "When I was in college I bought a Mk. VII Jaguar
sedan and drove it for several years. One day the gear shift
lever came off in my hand. The break was a classic fatigue failure.
The crack had been growing for years, and there was only a thin
shiny line across the center where it finally broke. There was
a sharp corner, no radius at all, where the threaded part on
the end changed to a larger diameter. The material was strong.
It was the design that was at fault. If they had put a small
radius on it, it would have lasted forever. You guys must be
looking at things like that trying to improve the fatigue strength
of critical parts."
RW: "Yeah, every piece. Even the most benign component
can break as a result of a mistake or lack of attention to the
design. Even in the case of the big bits that you really pay
attention to, you can usually get a much bigger improvement of
the life of the piece by detail design, addressing the concerns
about where it's breaking, than you can from a big change in
materials. There is no wonder material that can fix all our problems.
"Weak components at max rpm? [Another question on my
list.] Mainly valve gear pieces. It's always been the case that,
sort of jockeying for position in a racing engine, you have valve
gear pieces and con rod pieces, either the rod itself or the
fasteners that hold it on the crank, or the piston. The layout
of the current Indy car engine is such that the valve gear is
the weakest link in everybody's chain. I think part of that is
a result of improvements in piston design and changes in the
way people have laid out their engines. Bigger bore, shorter
stroke, lighter pieces have tended to reduce the loads on the
bottom of the engine and refocused the attention on the top end
of the engine. Particularly because of the way the rules are
that exclude pneumatic valve train, solutions to this problem
that have been learned elsewhere, like Formula 1, can't be applied
directly. So we're still fighting a battle with Indy car engines
that F1 has turned their back on for the time being. F1 is almost
certainly fighting the bottom end of the engine more than the
"The key to the valve gear thing is understanding valve
spring design and the things that upset the valve spring. It's
no secret that every single valve spring is measured before assembly
and it's individually shimmed as it goes into the engine, so
they're very close to coil bind. They're installed in such a
way that they're always a very highly stressed piece, even when
the engine is just turned over at hand speed. The thing that
really kills them is spring surge, which is the resonance of
the valve spring under the influence of the camshaft.
"There are two big elements to the design of the valve
gear. One is to make sure that the cam is designed so the valve
spring can keep the valve gear in contact with the cam so your
reciprocating bits are in control with respect to the rotating
bits. We depend on the valve springs to keep the valve in contact
with the cam profile. The second and less clearly understood
part is making the valve spring survive under the action of the
cam irrespective of what the rest of the stuff is doing. You
require that the valve spring can conform to what the cam is
making it do. If you can imagine the spring in between two surfaces
both moving backwards and forwards [Rob uses his hands to illustrate],
the spring has to accommodate that movement without getting excited.
Typically the spring will resonate, and that's when you get valve
gear problems. That's the limit that wire springs still face.
It's possible that another way to understand and improve valve
gear design will allow us to move valve gear back down to second
or third place on the list. Then we'll be limited by con rods
or pistons again."
PH: "So when incremental developments improve something,
it just shifts the emphasis to the next weakest link. You just
improve yourself to a new problem."
RW: "Exactly. Piston trends [the next question on the
list] is a similar story. It comes down to detail design. Again,
we're very fortunate at Cosworth that we have complete control
over the piston. We can design and manufacture every piston for
every racing engine that we build. We make our own forgings,
and we completely machine them ourselves. The trends are really
in detail design. The pieces are getting smaller and lighter
the whole time. We're learning things about the details of the
piston construction that lead to better fatigue life of the component
although we use, at the moment, the same material that we have
done for some time. We use a forging alloy that was developed
for fighter engines during the second World War.
"Although we're interested in new materials, we haven't
yet found one that has the same combination of properties as
the one we're using today. In the piston you're looking for a
material that combines light weight, high strength, and good
fatigue properties, particularly at elevated temperatures. It
has to be hot all over, but it's also got to cope with a big
temperature distribution. You've probably got 300 or 400 degrees
on the piston crown; that's degrees Celsius [400 degrees C is
about 750 degrees F], and 150 degrees C [300 F] on the piston
skirt. The thing has to be able to cope with rapidly fluctuating
changes in temperature. The temperature on the inlet stroke is
obviously greatly different than the temperature on the power
stroke. It's really detail design that sets a good piston apart
from a bad piston.
"You only have to have a look at a Periodic Table of
the Elements to decide where you might get a few further improvements.
You can look at the light elements up on the left-hand side and
start thinking about aluminum alloys that have lithium or magnesium
in them, or look at magnesium alloys. We could make pistons from
some of these alloys, but I don't think we could use them seriously
yet. There is potential for changes in piston material, but I
don't think it's imminent.
"We're looking at ways of reducing piston friction. I
don't think getting rid of the piston rings is possible at the
moment. Most of the work on reducing the friction is getting
the skirt shaped right and using the right combination of piston
rings and liner materials. Piston rings and liner materials is
one place where we will confess to a bit of witchcraft. It's
one of the areas of the engine where it's typical to say we're
using trial and error techniques and experience rather than first
principles. Finding these combinations is sometimes easy and
sometimes not so easy.
[Again, Rob looks at the next question on the list.] "What
changes do you make to the engines to maximize performance at
different tracks? In this series we try to maintain the mechanical
spec of the engine as uniform as we can. We try not to have a
road course spec or an oval spec. The reasons for this are mainly
that nobody has enough engines to be able to manage that type
of approach. At the moment, the character of the engine is such
that the benefits of doing that would be quite small in any case.
You probably wouldn't want to do it.
"Typically, we would be a bit more conservative about
the way we use the engine at the Speedway [Indianapolis Motor
Speedway]. To go 500 miles, we run the engine a little less quickly
and would pay more attention at rebuild to 500-mile race engines.
These engines get more new pieces. We have a schedule of replacement,
and some parts get replaced every build and others every third
build or whatever. A lot of those intermittently replaced pieces
are refreshed for a 500-mile race. But in terms of the performance
spec of the engine, they're essentially the same for all tracks.
"In terms of software changes, nothing changes. These are
not usually driven by a desire to optimize performance for different
tracks. They're more driven by a necessity to get the necessary
mileage. Fuel consumption is incredibly important in Indy car
racing. It's the thing that has the greatest bearing on how you
run the race."
PH: "I hate that. I think it ruins the whole thing watching
those guys short shifting and waiting until after the last pit
stop to start to race. Is that just the alcohol? Would that change
if they used gasoline?"
RW: "They could just increase the fuel allowance. That
would fix the problem if you thought there was one. You wouldn't
have to increase it very much. Of course it's not an issue at
the Speedway. Everybody can get all they need there."
PH: "Why do they limit it? I thought it was because people
would dump a bunch of fuel in for the cooling effect. Is that
RW: "In our case, it wouldn't do much good. It's probably
wrong to have a completely unlimited supply of fuel, but the
balance could be shifted in favor of more."
PH: "I guess if the pressure is on and you guys get the
baseline fuel mileage up a little bit, then you'd have some to
play with. So maybe the pressure for fuel mileage is good?"
RW: "At the moment, you end up running the engines very
lean compared with what they need at maximum power. The drivers
don't like it. It's kind of worrying because you have to pay
phenomenal attention to the accuracy of how you measure the fuel
consumption as well as the amount of fuel you have to put into
the engine. It's interesting. I quite like it how it is, but
I've kind of gotten used to it."
[Rob looks back at the list again.] "Simulation capabilities
that work? Yeah, it starts pretty early on and it continues through
the development of the engine. Most of the drafting for the engine,
most of the layouts for the engine are done on paper, not on
CAD. But some of the parts are laid out on a computer. Particularly
pieces that change a lot or mechanism sorts of pieces. It can
be useful to be able to plot the movement of pieces, and the
computer is good at that. It doesn't come under the realm of
simulation, but it gives you the data to run in other models
and other programs which are simulation. I'm thinking about finite
element analysis [FEA] and valve train dynamic models that we
have, programs for designing camshafts and examining the influence
of valve gear changes on what we do.
"The finite element thing is important to us, but it
would be wrong of me to paint a picture of a big finite element
model of the whole engine. It's not the case. What we use finite
element for is looking at the sensitivity to change of parts
that we know to be critical to us. FEA is a great way of looking
at fatigue-type problems. It's best suited to the type of problem
that doesn't involve plastic deformation; it doesn't involve
absolute failure. The type of stresses that lead to fatigue failures
are pretty well modeled by even FEA programs that run on office-based
personal computers rather than huge mainframe computers. So what
we use them for is evaluating different types of detail design
on highly stressed bits that we know by experience are prone
to breaking. It sort of quantifies the seat-of-the-pants feeling
that everyone has when you're putting in a fillet radius or smoothing
out a change in section or any of these classic textbook problems.
We all know that putting in the fillet radius will make it better.
But sometimes the FEA programs are a good way of finding out
how much better. And whether it's a change that's going to go
far enough to fix the problem, or whether it's only going to
stall the real solution while you're doing the experiment to
"FEA is typically done on a piece you're having a problem
with or a piece you've had a problem with that you're about to
redo fresh. We can see how to improve the part incrementally,
quantify that increment, and estimate the cost and benefit of
"We have programs that can more or less design cams on
their own, if you know what you want."
PH: "Internally developed?"
RW: "Yeah. Not much you can say about those except if
you know what you want, they work. They're always limited by
the amount of understanding you had of the problem in the first
place when you wrote the program.
"Once you get them up and running, dyno simulations of
a racetrack are very valuable. They basically come into two categories,
one being performance work and the second being durability work.
Performance testing of the engines goes hand in hand with having
done a lot of engines for a lot of years. We know a lot about
testing racing engines. Although we use typical, standard Heenan
and Froude dynos, we just use the big mechanical pieces and the
installation and control pieces are designed and built internally.
This enables us to make pretty good performance measurements
of the engine--steady state performance within 0.5%. That would
be a repeatable 0.5%. It's difficult to ascribe an absolute accuracy
to it. The absolute accuracy is not really that important as
long as you can be honest with yourself when you analyze the
results. It's not that often that something happens to change
the absolute accuracy of the measurement."
PH: "Are dynamometers accurate one to another? Is there
variation from one piece of equipment to another?"
RW: "They should be very close."
PH: "That same 0.5% number?"
RW: "Maybe 1%. There's a lot more difference in test
results coming about from dishonesty or different corrections
than there is in actual changes in the measured performance of
PH: "Are all power measurements corrected to standard
RW: "Yeah. You can get quite big changes if you change
what you ascribe to be standard, though. We correct our numbers
to 25 degrees C and 66% relative humidity and 45 inches of mercury,
which is the nominal boost allowed in the rules. We'll adjust
the waste gate so the absolute boost is 45 inches each time and
correct the observed power according to the condition of the
PH: "In an engine on a track, how close does the boost
get to the allowed 45 inches of mercury? How much of the time
does the engine run close to the max boost?"
RW: "It depends on the type of track. It's everybody's
goal to have the engine at 45 inches of mercury the whole time
the driver has his foot to the boards. The whole time the engine's
at full throttle, it should be running at 45 inches."
PH: "I guess I don't know how that's controlled."
RW: "There's a pop-off valve that's supplied by CART.
That has no influence on boost control. It's the penalty for
screwing up the boost control. The rules say we use waste gates
that are controlled by the plenum pressure. Essentially, the
current arrangement for all the cars that race is that there
are two waste gates. The waste gate is a device that measures
the boost pressure in the plenum and, according to that pressure,
it either spills exhaust gas bypassing the turbo or it forces
exhaust gas to go through the turbo. Basically, if the waste
gates are open, then there is a direct passage out to atmosphere
for the exhaust gases. If the waste gates are shut, the exhaust
has to go through the turbo and the turbo will be increasing
the pressure in the plenum. When the waste gates are open the
turbo will be slowing down and the pressure in the plenum falls.
"Effectively, you match waste gate performance to engine
performance using a spring. It can be open or shut or anything
in between. Typically, around a superspeedway, the plenum will
be at 45 inches all the time. Around a short oval the plenum
will be at 45 inches the whole time the driver is on the throttle.
He will lift briefly in the corners, but the boost doesn't fall
sufficiently far that there is a significant response time when
he gets back on the throttle. It's instantly back at 45 inches.
On street courses and road courses it's a bit more difficult.
There's substantial part-throttle content at a road course. You
wouldn't want it to get to 45 inches during the time the driver
is at part throttle. At part throttle the turbo slows down and
cools down, and when the driver goes back to full throttle there's
a response delay before you achieve full boost."
PH: "I hear the drivers blipping the throttle in slow
corners. I assume that's to keep the turbocharger spinning?"
RW: "Yeah. There'll be some of that taking place with
changing gears and making the car change direction, which some
of them don't want to do this year [smile].
"Boost control is a subject all to itself. It's a big
part of making the engine work in the car.
"Going back to the dyno, the important thing is the dyno
measures the steady-state performance of the engine. It doesn't
measure the performance as the driver nails the throttle coming
out of a turn. That'll be related to the boost control. There
are a number of ways of going after this. We have a performance
development dyno in England that's capable of measuring the transient
performance of the engine in addition to steady-state performance.
There's a big flywheel mounted on the back of the dyno. It sounds
simple and looks very scary, with big pieces of metal rotating
at engine speed. It's got big railway engine-type disc brakes
on it to slow it down. But what it enables you to do is add the
inertial load to the hydraulic load on the dyno so you can examine
the performance of the engine with realistic rates of change
of engine speed which you cannot normally do. Transient testing
on a dyno is useful for performance work and durability work.
"The nature of superspeedways is such that you can do
a very good simulation with just a regular dyno control. You
have to be able to run an engine flat out for a number of hours."
PH: "That'd be scary too."
RW: "You've got to do it. [Laughs with pained expression.]
We've learned a lot doing that; bolting the engine down and power
testing it in the normal fashion and then doing a superspeedway
simulation. You can choose the range of engine speed and cycle
it up and down at full throttle. If you have something you're
particularly worried about you might have to stop prematurely,
but we have one engine which for the last two and a half years
has been continuously built and run in this endurance cycle and
then rebuilt. This is to find which pieces break."
PH: "So you've got a continuous durability program?"
RW: "It's about the only thing you can do to be able
to go into one of these big 500-mile races with the belief that
you can finish the race. Sometimes you can have that belief and
it still doesn't happen.
"Going back to the questions. If you have the opportunity
to influence chassis design, what do you do? The main thing here
is making sure the bits that affect the engine work. There's
a divided responsibility in putting an Indy car together. Our
biggest interest when the chassis is being built is in making
sure the systems that interact with the engine do what they have
to do without causing performance or durability problems. What
we're thinking about here are big mechanical bits. We want to
see that the engine mounts are sensible so that the mechanical
loads from the car that need to be carried by the engine are
being put into the engine in a sensible fashion. And we need
to see that all the plumbing is as it needs to be, in particular
the oil tanks and the breather systems. They're crucial to making
the engine last.
"I guess it's not immediately obvious to someone watching
the race on TV that if you're pulling 4 Gs around a turn at a
short oval, then the oil that would normally be sitting on the
bottom of the tank will be sort of splurged up the side of the
tank at an angle of about [Rob pulls out his pocket calculator
again.] 75 degrees or something, making it very difficult for
the oil tank to cope with the oil being settled on the bottom
at one point in the lap and up the side at another. Of course
on a road course it can slosh to the other side on another part
of a lap. Getting an oil system that works reliably is real important
to us all [Rob grins].
"Car designers have these things in their minds but there's
other things in their minds as well. Like aerodynamics, they've
got packaging goals. They want everything tiny, tiny, tiny. They
want it to sit low down. Sometimes these things come into conflict.
The engine people just want their engine to live [laughs]. You
have to pay attention to a new car when you first test it. If
you have a new car and the oil tank doesn't work, you blow some
PH: "I guess the clutch and gearbox comes in there too."
RW: "The transmission and the way the transmission is
attached to the car is not often open to us to have a big influence.
In the case with a car that's being laid out and there's fairly
close cooperation with the engine manufacturer, then there will
be dialog. If there's anything that strongly offends the engine
guy, then he's going to tell the car guy what's going to happen
because he's interested in the car doing well.
"The other things that are important to us are exhaust
layouts and installation and turbo position and installation.
The turbo position determines what the exhaust pipe ducting looks
like and what the ducting from the outside world into the compressor
looks like, and the compressor up to the plenum. All of these
factors can affect the performance of the engine so we like to
know what's happening there early in the proceedings so we can
"Likewise the boost control system is an important part
of getting the car performance sorted. We like to do more than
influence the design of the boost control; we like to control
it. We make the waste gates, and we make the boost controller.
Although in the past this has been the car builder's territory,
we build them. It's one of the things we do to reduce the amount
of grief we get at the racetrack. Historically, the boost control
has been a gray area, a bit of divided responsibility. The pieces
previously would have come with the car, but now they bear more
on the engine performance than car performance. The engine person
is more likely to understand these pieces than the chassis guy.
"Trends in packaging? [Next question on the list.] The
trend in racing the world over is toward smaller, lighter pieces,
and to engines that are more greatly integrated with the chassis
design. That's become more and more important."
PH: "More integrated in both structure and aerodynamics?"
RW: "Exactly. Specifically, engines have got narrower,
particularly in the sump area, under pressure from car designers
to give them more space for the diffusers underneath the car.
That wouldn't apply to an Indy car because the underside is controlled,
but it would apply to a Formula 1 car. Engines have got narrower
and could get lower. A recent rule change in Indy cars has perhaps
taken some impetus out of repackaging engines, but I don't think
they can completely detune us. A new rule controls the shape
of the engine covers and is set up around the existing dimension
of the Cosworth XB engine.
"The packaging thing is interesting. There's no doubt
that if a car designer could convince the engine people of a
needed change it would get done, but not if it affected engine
performance. You'd only make a change that didn't influence the
power because the aerodynamic things can't be measured with the
same precision you can measure engine performance, so it would
be a fairly risky activity to give away some engine performance
in exchange for an aerodynamic hypothesis."
PH: "That's a very interesting point. So the package
is important, but you would never sacrifice engine power."
RW: "You'd never give up horsepower. When it comes to
horsepower, more is definitely better than less.
"The other thing that's increasingly important to the
car people is the cleanliness of the engine. It's becoming the
case that they want fewer components external to the engine.
So the auxiliary things which used to sprout from engines, things
like fuel pumps and alternators that used to be external pieces,
are now more integral with the engine. The fuel pump is another
example of a chassis bit which has always been built by the engine
manufacturer. You can see that engines of today have a much cleaner
appearance than engines of 10 or 15 years ago. You can't see
through them any more. The space in the valley between the banks
is better occupied and there are fewer and fewer external parts.
[Rob again refers to my list of questions.] "The most
critical piece to design and develop? Probably all of them [laughs].
It goes, really, back to the bits that are the weakest link.
So that it comes down again to the valve gear. The big mechanical
bits that people think about when they think about engines are
the most difficult to design. The valve gear, the cylinder heads,
the block, the crankshaft, the rods, and the piston. The bits
that are relatively easy to design and develop are the ancillaries,
the pumps and the bolt-on bits on the top. The inlet arrangement
might be quite hard, but the design is relatively straightforward.
You just do it."
PH: "What do you do on a race weekend? At the Indy car
race at Vancouver, Ian said I should try to catch you Saturday
afternoon because you might have some time to talk. I hung around,
and I watched you in one of the team areas sitting on an electric
cart, and there were never less than three or four people there
talking to you. What do you do there?"
RW: "It's incredibly difficult to explain sensibly. [Rob
shakes his head and we both laugh.] I cannot give an adequate
answer to this question no matter how many people ask me, whether
it be socially or at work. I have to try to describe what happens
at a typical race weekend, and that's very hard.
"Essentially, I'm there to take care of the engines.
I'm there to do all of the things that you've suggested in the
question here, to give technical support, to be a sympathetic
ear, and to communicate with the customer. But the balance or
the level of dominance of each of those things varies according
to circumstance. Being relatively fortunate this year that the
engine has been, on the whole, reliable; we haven't had too many
repetitive problems. We haven't had too many common threads running
"A big part of my job is to be there to gather information
about problems or failures--as early as possible. And try and
get information from the teams while it's fresh in everybody's
mind and while they're wanting to talk about it. If you don't
get the information right away, then it's often lost for all
"It's kind of difficult to know how to support the engine
in service. We try to maintain very close links between the customers
who use the engine and Cosworth, who builds and services the
engine. We want to make sure that there's always a channel for
communication in both directions just to make sure that if anyone
has a problem at a race weekend that they won't think that it's
being glossed over. We want them to have confidence that we know
about it, for starters, and we want them to have confidence that
once we find out about it we're going to get someone to try and
fix it. So a big part of my job is information gathering, particularly
on problems that exist. If we go to a race weekend and we experience
engine failure, then I like to be in there as soon as possible
after the fact to try and establish exactly what went wrong.
"Some of the tools that you have in your armory here
are the diagnostic equipment that's carried on the engine itself.
The ECU [engine control unit] that controls the engine has some
diagnostic capability. All of our customers use very good proprietary
data logging equipment and we have access to that in order to
diagnose what went wrong if there's a failure--temperatures and
pressures and the exact sequence in which the traces changed
during a failure.
"The failure analysis is important because it enables
people to continue through the weekend with some knowledge of
what happened rather than continuing in ignorance. If you have
to take an engine out of a car, it's good to know if there's
a chance of a similar failure with the engine you're about to
put in or if it's a random thing that's not likely to recur or
if it's a mistake that's happened, which means that it's almost
certain not to recur.
"There are times at some of the racetracks where you
can influence the performance of the car, where you can work
with the customers, making just small changes to things. You're
never going to reinvent the wheel on a race weekend, but you
can get a little further down the optimization road. You can
maybe do something to some elements of the mapping of the engine
to contribute to drivability concerns. You can maybe do a small
experiment looking for fuel mileage gains. You can't really do
any experiment that has a lot of risk attached to it. If you
judge the change carefully, then you can go in with an experiment
that might get you something back on drivability or fuel mileage
or something like that.
"There's a lot of learning to be done about the car systems
that we mentioned earlier. The only real experience that we gain,
the real-life performance of oil tanks and waste gates and turbochargers,
is during race weekends. For all the work you can do back at
the factory on dynos and stuff, you just can't cover the same
amount of ground that even one or two cars can cover in a race
weekend. Learning the character of these subsystems is very important.
"It's important to keep the customers up to date with
what's going on back here, to relay to them the status of any
engines they have in work here at the shop, to keep them abreast
of any changes in the performance spec of their engine. It's
the nature of racing engines that the spec changes quite rapidly.
It's a bit of a juggling act. Inevitably you end up releasing
parts still with some degree of risk attached to them. So there's
an unfolding story really, as you learn about the engines you
have in service.
"If you judge it right, what happens is you introduce
a change in spec and you try to go in very conservatively. So,
shortly after you've changed the spec of the pieces, you can
turn the wick up the whole way. When you introduce new pieces,
you'd like to think you don't go whole hog when there's still
some risk attached to them. For example, you might choose not
to run a big rev limiter when you have some new pieces in the
valve gear or something like that. Then later on you might learn
that all looks good in the engines that you have out in service
at the lower limiter, so you can maybe go to the limiter that
you thought you should have had all along.
"Alternatively, you might have made a mistake; you might
have not been conservative enough. As a result of running some
engines with some new pieces in, you might have to do some swift
reactionary stuff in order to stop the thing growing into a crisis.
It's important to communicate that sort of information as well
as gather it. It's got to come back in from the team on one hand,
but you've got to be sure you're giving feedback on the other
hand or the sources of information will dry up.
"It's particularly the case in Indy cars that even the
biggest teams don't employ engine specialists. So you need to
be there in order to advise and comfort and just to see that
they don't do something inadvertently that will upset the engine."
PH: "That's a big point that's new to me. I didn't know
that the teams don't have an engine guy. That's changed. Twenty
years ago all they really knew to fiddle with was the engine.
They didn't know what the shocks did and not much about the suspension.
The engine, now, is a black box that they plug in and out and
they don't want it to break and they expect increases in power
and they expect someone like you to be right there to make sure
all that happens. That's an amazing change in some of the racing
series in the United States. Is it the same in Formula 1?"
RW: "Not to the same extent. In Indy cars, Cosworth has
one technician with each team, the exception being our two lead
teams, Newman-Haas and Ganassi, who have one technician per car.
During the race weekend they report to me. They look over each
engine every time it goes in the car. They make sure there's
no obvious problems. They're there to work closely with the team.
They're the primary contact for educating the guys on the team
who are maybe dressing the engines. They would deal with most
of the mechanical things that would crop up on the engines. The
background of these people is that, typically, they've been engine
builders. They will also prepare the electronics on the engine
for use on the track.
"There are several different features that need to be
programmed into the engine, into the ECU. They take care of all
that and they will liase with the engineer on the team and decide
how lean they're going to run the engine depending on where they
are on the weekend. They just check that the car systems are
as they should be, and they encourage the teams to pay attention
as well. They sort of poke and prod around and ask questions
about how much oil the thing's using, and keep abreast of the
temperatures the thing runs at. They've only got two feet and
they can only get around so far, but the basic idea is they try
to maintain an awareness of the needs of the engine. If there's
a problem or crisis, then I get wheeled in to find out what's
"It's very difficult to make sense of. Some weekends
you'll find yourself lying on your back under a racecar looking
at a hole in the side of an engine. Thankfully that doesn't happen
very often. Other times you'll be taking somebody's turbocharger
or waste gate apart and putting it back together, or maybe you'll
spend hours giving yourself a headache over a set of data logger
results trying to understand why it's different from the last
time at this place or trying to achieve some incremental improvement
in some feature like the boost response or something like that.
Very, very diverse. The problems of each track are very different.
Some places you'll be scratching around the whole weekend looking
for how you're going to make it round the track on the necessary
fuel allocation. The whole thing is very difficult to describe."
PH: "What parts of your job do you like the most and
least? Do you like the race weekends?"
RW: "I like the uncertainty of it. I like the fact that
you don't know going in what a weekend's going to hold in store
for you. I like the fact that, by our responding quickly to problems,
then you can make life better for a lot of people. It forces
you to think on your feet. It forces you to assimilate facts
very quickly. Hopefully, you don't make too many mistakes. If
you react to a problem that you've discovered and you get a fix
in place, then you can maybe stop some people from falling out
of a race and save some people some money at rebuild. Even if
it isn't that clear cut or even if the news isn't good, sometimes
you can tell a team they have to change an engine early on and
they get it done sooner and they go to dinner earlier.
"The thing I don't like about it? I don't like the feeling
of impotence you sometimes have. The feeling you have that the
die is cast and you're powerless to do anything about it. I guess
the best example of that was during 1992, when we had that auxiliary
drive belt problem. We thought we didn't know how to fix it but
we knew that until we did fix it we were certain to experience
failures. That's a terrible feeling.
"But, likewise, unless you understand the thing, you
can't fix it. You have to focus on understanding the thing rather
than making up a random change. You couldn't hope it would go
away. Although we didn't understand the problem we knew it was
something to do with the car and it turns out it probably was,
but we still needed to make the engine resilient to this and
the change we ended up making was huge. It involved a new front
on all our engines. It made half our parts list obsolete overnight.
It was a big change and certainly not a change we could entertain
without making sure we grasped what was taking place."
PH: "When did you join Cosworth?"
RW: "September 1987."
PH: "How much longer will you be in the United States.?"
RW: "I'm still not sure what I'll do next year. I'm not
going to England right after the last race. I'll be here till
Christmas time. I'm interested to see what happens the next few
months. We'll be testing a new car, the Reynard. Once it gets
back over here we'll find out if it's a sinker or a swimmer.