26 replies to this topic

[VAR1016]

I know that this one will get those-in-the-know-and-who-are-not-talking laughing, but I am curious about materials used in the manufacture of F1 engines/

For example, I heard a while back that the inlet valves on the Renault F1 engine had 40mm heads, were about 100mm long and had 4mm stems. what the hell are they made of?

Similarly I heard that carbon crankshafts have been tried.

Does any one have more information on this sort of thing?

VAR1016

[12.9:1]

The rules stipulate what can be used
FIA-regs here




.

[VAR1016]

Thanks for that.

I looked at the relevant section and it is amazingly vague. I had always understood that only poppet valves may be used, but there is no mention of this.

Anyway the materials mentioned are simply steel for crankshafts and cast iron or steel for camshafts. So obvioulsy composite cranks are out. A pity really as this could have useful production possibilities.

This all means still that pistons could be say ceramic, and again what are the valves made from? Titanium alloys perhaps? But the exhaust valves?

A friend has always posited "unobtanium" of course...

VAR1016

[12.9:1]

Though the seventh, of the seven veils of F1 secrecy covers the combustion chamber, the valves do stick up a bit, and are said to be titanium.

Unobtanium wen found, is promptly (after kicking and screaming) banned.
McLaren/Mercedes are only now showing signs of recovery after loosing their Beryllium/aluminum pistons in 2000.



.

[desmo]

Until recently the vast majority of materials that made up an F1 engine were remarkably prosaic e.g.:

Pistons- RR55Al alloy, originally developed by Rolls-Royce (hence the RR) for the Merlin WWII aero engine!

Rods and Valves- Ti alloys

Block and Heads- Cast Al alloy

In fact, as recently as 5 years ago the only materials employed that weren't readily available- at least for aerospace applications- 40 years ago, were the multi-phase steel bolts and the PTFE (Teflon) pneumatic valve seals!

CF and aramid fiber reinforcements are prohibited for engine internals. And of course the crank and cams must have a "basic structure" made of steel or cast iron. What "basic structure" exactly means is undefined. Perhaps steel-matrix MMCs might usefully raise the elastic modulus and hence the harmonic frequency of the cranks enough to be useful, and still meet the letter of the regulation. At what mass fraction do MMC reinforcements exceed the "basic structure" limits?

Also, the specific modulus limit seems to limit the use of super strong and light materials such as AlBeMet for engine internals, but a closer reading of the regulations reveals that that reg only applies to "metallic materials" and ceramic reinforcements appear to potentially skirt the issue as long as the metallic matrix material is within the specific modulus limit.

Rumors have been about for several years now that pistons have gone to Al-matrix MMC with ceramic preforms reinforcing high-stress areas of the piston. Weight taken off the pistons will raise the critical torsional harmonic frequency of the crank, widely believed to be the single largest factor currently limiting rpm. Liners are likely similar MMCs.

Of course, no one knows what all the teams are using and where, even those inside the sport. It's still fun to speculate.

[Cory Padfield]

Around 1995, Del West Engineering (the major titanium valve supplier for racing teams) was using Ti-6Al-4V for intake valves and Ti-6Al-2Sn-4Zr-2Mo-0.2Si for exhaust valves. I believe these are most likely the same today, although I read that exhaust temperatures are higher now due to traction control and more unburned fuel, so perhaps the exhaust valve is something new. Ti-6242S was near its limits before, so it may not be capable.

I read recently that head temperatures are above the limits of the standard Al casting alloy A357 (due to the desire to run engines hotter and use smaller radiators, higher rpm, etc.), so now they may be using hypereutectic alloys (maybe 319?).

One interesting development may be the removal of main bearings in engine blocks - it is speculated that BMW run the crank directly against the aluminium block, which must have a suitable coating such as one of the new microarc oxidation types (e.g. Keronite).

[H. Eckener]

One interesting development may be the removal of main bearings in engine blocks

I am curious as to what advantage that would offer. Perhaps the plane bearing alloys didn't have the required strength and durability to stand up to the loads of lubricating the crank? Although, instinctively I don't care for that answer.

[Colin]

The advantage of running without main bearings is the ability to lower the crank slightly and thus make the whole engine slightly lower, with a lower CG.

I think Cory you meant A390 hypereutectic casting alloy; A319 is a hypoeutectic alloy of much less strength. The problem with hypereutectic alloys is their brittleness with all of the hard silicon in the alloy; the ductility of A357 is about three times that of A390.

www.matweb.com <-- good site.

[H. Eckener]

Cory,

Once you have lowered the crankshaft so the counterweights are touching the pavement, the counterweights, not the bearings are limiting your ability to push CG even lower. Then reducing the stroke will also help the CG, rotational inertia, and weight of the crankshaft, but those probably aren't the best reasons to go and play around with the stroke.
Installable bearings have some very handy properties/features. And probably the only reason to do away with them is if you start exceeding their mechanical/physical properties.

[VAR1016]

[QUOTE]Originally posted by Cory Padfield
[B]Around 1995, Del West Engineering (the major titanium valve supplier for racing teams) was using Ti-6Al-4V for intake valves and Ti-6Al-2Sn-4Zr-2Mo-0.2Si for exhaust valves. I believe these are most likely the same today, although I read that exhaust temperatures are higher now due to traction control and more unburned fuel, so perhaps the exhaust valve is something new. Ti-6242S was near its limits before, so it may not be capable.


I have been fascinated to read these reponses to my original post. And I am also fascinated that titanium alloys are used for exhaust valves. Given the likely temperatures I had assumed that a development of one of the Nimonics would have been selected - perhaps even with a little cobalt for toughness? (so much for my metallurgical knowledge )

Meanwhile, on the subject of valves - does anyone give credence to the stories about some sort of solenoid operation?

VAR1016

[MRC]

On the subject of valves, I believe most of the turbo rally cars use Iconel (or some other high heat alloy) for the exhaust valves due to the very excessive ignition ****** used for anti-lag.

VAR1016, I think most anyone on this forum will emphatically say that no one in F1 is using solenoid operated valves. Is anyone in F1 doing some R&D with solenoid operated valves of some sort? Probably.

[desmo]

Speaking of valve materials, at least one team has extensively tested Al MMC intakes. I can't say if they've made it to the grid.

[VAR1016]

Originally posted by MRC
On the subject of valves, I believe most of the turbo rally cars use Iconel (or some other high heat alloy) for the exhaust valves due to the very excessive ignition ****** used for anti-lag.

VAR1016, I think most anyone on this forum will emphatically say that no one in F1 is using solenoid operated valves. Is anyone in F1 doing some R&D with solenoid operated valves of some sort? Probably.

MRC:

Thanks

VAR1016

[Cory Padfield]

Colin,

You are correct about the alloys, although 390 is a high pressure die casting alloy and heads are usually produced by gravity die casting, sand casting, or investment casting. I suspect new alloys are being used here, but I have no actual information.

Regarding exhaust valves, the problem with Nimonic alloys is high density. Stainless steels are competitive here as well. One reason that titanium can be used is that a majority of the heat in the area goes through the seat rather than through the valve, so the valve temperature is not quite as high as the gas temperature.

[H. Eckener]

...will raise the critical torsional harmonic frequency of the crank, widely believed to be the single largest factor currently limiting rpm

Ah hell I've got nothing going on at the moment, so I might as well throw this out there. Anyone like the notion of using radial ball bearings instead of plain bearings on a crankshaft? Like the old Porsche and Maybach engines, not to mention a few other old names.

[Top Fuel F1]

Originally posted by VAR1016

This all means still that pistons could be say ceramic,

I suppose there is solid ceramic parts used some where; but it seems to me that when they are talking about ceramic in regard to F1 they are talking strictly about metal parts with various types of ceramic coatings. For instance Nikasil would be one type of coating. It is a Nickel/Silicon galvanizing process applied to AL parts

[klthomas]

Ah hell I've got nothing going on at the moment, so I might as well throw this out there. Anyone like the notion of using radial ball bearings instead of plain bearings on a crankshaft? Like the old Porsche and Maybach engines, not to mention a few other old names.

That would require a built up crankshaft or split-shell bearing races and I don't think they would last at current rpm's.

[MRC]

I believe their have been roller cranks around before that had split bearings (old Carrera's). I don't think the rpms would be of any issue with hybrid ceramic bearings. I think the main downside would be the larger packaging envelople required of the roller bearings as compared to plain bearings. Higher friction may or may not be another downside, too.

[H. Eckener]

Hello,

To expand on what MRC said, the roller crank Porsche's of 1950's routinely ran up to 10,000 rpm I believe without much issue. The German fighter plane BF-109 of WWII utilized three rowed roller bearings in the big end of their connecting rods in the DB600 engine series, that was a supercharged engine with a fairly high output for its time from my understanding. Japanese efforts to produce that engine for their own war effort ran into reliability issues mainly because Japanese roller bearing quality was significantly inferior to German roller bearings.

Maybach's extensive fatigue testing and R&D in the 1960's, on their high-speed high-output diesel engines, found that for a given mass their roller cranks were superior to their plain bearing cranks in terms of stiffness, which reduced the effects of vibrations and improved durability. They determined that for a given horsepower output, to improve durability and the life of the bearings, it was significantly better to increase RPMs rather then increase torque. Sounds like F1 in some respects.

From what I gather ceramic hybrid bearings represent almost a revolution in technology over their steel counterparts, whereas plain bearings are more or less an evolution of the same practice.

Packaging maybe an issue, but improvements in stiffness and reduction in oiling requirements and possibly friction (I've heard it both ways), may out weigh the packaging.

Fun speculation either way.

[VAR1016]

Originally posted by H. Eckener


Ah hell I've got nothing going on at the moment, so I might as well throw this out there. Anyone like the notion of using radial ball bearings instead of plain bearings on a crankshaft? Like the old Porsche and Maybach engines, not to mention a few other old names.

Certainly I have always understood that plain bearings offer lower friction that roller bearings - ever since I asked someone why roller bearings were not used in F1. I imagine this must have been known by engineers for years which of course begs the question: Why did Mercedes-Benz use a Hirth (built-up) crank in the W196?" Obviously money was no object so what was it? Tradition? I doubt that Uhlenhaut & Co were sentimentalists in that way!

There is also the weight and packaging aspect; perhaps the dreaded aerodynamics outweigh everything else?

VAR1016

[klthomas]

To expand on what MRC said, the roller crank Porsche's of 1950's routinely ran up to 10,000 rpm I believe without much issue. The German fighter plane BF-109 of WWII utilized three rowed roller bearings in the big end of their connecting rods in the DB600 engine series, that was a supercharged engine with a fairly high output for its time from my understanding. Japanese efforts to produce that engine for their own war effort ran into reliability issues mainly because Japanese roller bearing quality was significantly inferior to German roller bearings.

Sorry, I don't agree. There's a world of difference here. A flat 6 has good primary balance, a 90 + degree v10 doesn't. I seem to recall Renault having alot of problems trying to get their 111 degree engine to not tear itself apart AND make enough power. Also there's a big difference between 10,000 and 19,000 rpm. Forces go up exponentially with revs. I seriously doubt anybody could make a multi piece crank that would work at 19,000 revs and not be way too heavy. Also roller bearings require more oil flow which could cause windage problems.

As far as the Daimler aircraft engines. Yeah, they had roller big-end on the crank, but look at the size of those parts. Also they only ran at around 3500 rpm.

[H. Eckener]

Var1016,

I am not so sure that it is an established fact that plain bearings offer the lowest coefficent of friction out (COF) there. Looking through some books I see ball bearings around 0.001, rollers at 0.002, and needles at 0.004 (these numbers predate ultra high quality steels and things like DLC coatings and cermics). Whereas for a plain bearings the number is around 0.005 to 0.001.

Sadly, a book that says "if you had an F1 budget and engineers with any material/coating available this is the best COF for the roller bearing you could get is ____ and with the same skys the limit budget, engineerings, materials, and oils the best ... plain bearing is____", does not exsist. Its probably close, but my feeling is that the quantum leap in materials may help the antifriction bearing eek out.

The plain bearing coefficent of friction is also not as constant as an antifriction bearing in its operating envelope, ie the plain bearing isn't always at optimum. I have seen it stated that the COF in a plain bearing is related to (viscosity*rpm)/pressure.
Viscosity is dependent on oil chemistry and temperture.
Rpm is dependent on the driver, but we know the range is roughly 0 to 19,000rpm.
Pressure is dependent on journal dimensions and combustion loads.

A F1 engine's bearings probably operate at their optimum COF the same time and place the rest of the engine operates around optimum, but how badly the bearings properties drop off from that optimum I don't know.

An antifriction bearing on the other hand once broken in pretty much operates at a constant COF.

In terms of weight, packaging, and aerodynamics, you have several things going on. In the weight case according to Maybach's R&D you could go with a lighter crank and still have the same stiffness. You would also need less oil, which means smaller oil lines, oil pumps, and oil coolers. In regard to packaging the space the crankshaft sweeps out is pretty much wasted space, you couldn't store MS's hash stash there anyways, because on the first revolution it would be knocked apart by the crank, so long as you just use that space up you should be fine. On the issue of aerodynamics, I don't see the engine getting that bigger and with the oil cooler/radiator getting smaller, aerodynamics may improve.

[H. Eckener]

klthomas,

I thought of all the 6 clyinder arrangements only the inline 6 engine offered perfect primary and secondary balance? Perhaps the sign of the 111 degree Renault failure is a sign of bearing failures. It was plain bearing failures that first inspired a move to antifriction bearings in Grand Prix engines of the 1920's. Also from Maybach perspective, stiffness and consquently vibrations are lessened in a roller crank over that of a plain bearing crank. Either way the V-10 impacts plain bearings and roller bearing equally.

Generally engine bearing loads are depedent on three things fixed loads (ie gravity, piston ring tension, etc.), gas loads (ie normal and abnormal combustion forces), and inertia loads (which is related to the mass*piston speed^2). The first two are probably at a minimum, what with gravity being constant (roughly), piston ring tension as low as possible. And the gas loads probably are not that high with this being a naturally aspirated engine. With the third basis for bearing load mass is as low as possible. In the case of piston speed, you have something that is dependent on both the engine's stroke and rpm. Its true that rpm has nearly doubled, but the piston speed has not increased significantly.

Honda starting in 1964 and 1967 with their 1.5L and 3L F1 engines used roller bearing cranks as far as I know, and I don't think there was much issue with them.

Maybach cranks were not mutlipiece, I would think a billet crank could be designed. Maybach prefered high speed diesels over lower speed diesels, partly because it was easier to forge and heat treat a one piece small crankshaft over a large multipiece crankshaft.

Antifriction bearing do not require much if any oil to work properly, where as a plain bearing relys on oil for its life blood in order to function properly. If you have any doubts look at a two stroke motorcycle engine.

Sure the Daimler reved low, but I suspect its piston speed was high, the components were also over built for field use. Lets also not forget that this was some 60 years ago, and that Germany was fighting a two front war with constant aerial bombardment, and that all exotic alloys were scarce.

[Billy Gunn]

The biggest single problem with rolling element bearings is just that! It's a rolling element.

The inertia of 20 odd balls or rollers whistling around in a cage, being accelerrated and decelerated at engine speed (300 times a second) beggers belief. Admittedly the oil requirements will be less (and different!) as it will not now be needed as a hydrastatic film but more of a heat transfer medium. Manx Norton's back in the 50's used a roller big end, and even at modest power out puts of around 100 HP/litre at 7-8,000rpm were suffering roller failure even with silver plated Titanium cages, due to the scuffing action of the constant accel/decel regime inherrent in a roller. Also the inner race and outer race peripheral speeds are dissimilar adding to the scuffing problem. Norton's recognized the limitations of the rolling element bearing and were developing plain big ends by the mid to late 50's. There is currently a 500cc 4 valve Manx run in the Classic Racing Club in the UK which uses the works '54 350 cc plain bearing bottom end with a BDG Cosworth piston and DFV valve gear. It revs to over 10,000 and is reputed to give over 65HP.

The biggest problem with the FIA regs is as always their wording. The specific strength per volume indicator is presumably tested at room temperature????????? Another problem with MMC's is how much 'seeding' is performed in the preparation of the 'mix'; and this argument can be taken through to casting techniques as well. I know of a steel alloy which could be used in fabricated cylinder blocks - which would weigh less than a current Al casting for the same or even greater strength. Problem is the only way of shaping it currently is 'Explosion forming'.

Titanium valves are common commodities nowadays - even the Roush Nascar's have run them for a long time in both inlet and exhaust.

BTW, Desmo - I think you mean RR58 (a 1.25% ti aluminum alloy) RR55 was about 1938, RR56 was circa 1941 and RR57 and 58 were circa 1942. These were also known by a DTD number as well (Director of Technical Developments ~ a wartime office of the MAP - Ministry of Aircraft Production). One of the wonder materials to come out of this age was Nimonic 80 and its derivatives, and the ubiquiteous DTD 5212.

'yerduntgetowtfernowt'

[desmo]

BG- Thanks for the correction on the alloy nomenclature, I knew it was one of the two and I just went from memory (might work for some people!). I believe that during the Bernard era at Ferrari, steel crankcases were employed. I can't find whether these were cast or fabricated or perhaps both, or formed by some other method.

Plain bearings are nearly magical when they are working as designed, hard to imagine a better solution for the applications that suit them. Rollers will always have metal to metal (or ceramic )contact to some extent while running won't they?

I'd love to hear that Manx-Cossie at full song, 65bhp from 500cc with a '54 bottom as its basis is an imperessive feat indeed! I have often wondered what a 300cc single based on a reasonably modern F1 engine might do in a road racing chassis.

[Billy Gunn]

Desmo - Believe me it sounds great!

To digress a little there was a 500cc single cylinder championship set up in he UK a few years ago ~ simple Formula racing! - 2 wheels and a single cylinder 4 stroke not exceeding 500cc, it was sponsored by Honda UK. The 4 valve Manx was competing against lightweight aluminium framed Honda RFVC's and Yamaha bikes. It went into the last race needing only to finish 4 th or better if the next best won the race. It was an 8 lap race and 'Horror of Horror's' it refused to fire up on the push start off the grid. When it eventually started the rest of the field had a 3/4 lap advantage! By the 6th lap the rider of the Manx was lying second (smashing the lap record on the way), and had the Honda mounted leader (and championship contender) in his sights, swept into the lead on the 7th lap only for the motor to let go at the start of the 8th lap. Boy were the Honda BigWigs relieved. When the engine was stripped in the paddock after the race to see what the damage was all the Honda reps were peering over the guys shoulder to see if they could learn something!!!!!

I askedthe rider why he didn't back off when he was in 3rd place ~ typical racers answer "...well I was so angry it wouldn't start , an once I'd got it going I just couldn't back off...."

Coworth did a bike GP engine for Norton in the '70's ... it was a hopeless flop, no torque or 'driveability' (or should that be'rideability')

Billy G

[Top Fuel F1]

Originally posted by desmo


Rumors have been about for several years now that pistons have gone to Al-matrix MMC with ceramic preforms reinforcing high-stress areas of the piston. Weight taken off the pistons will raise the critical torsional harmonic frequency of the crank, widely believed to be the single largest factor currently limiting rpm. Liners are likely similar MMCs.

Dr. Theissen (F1/BMW) commented last year that he found the MMC pistons "not ready for prime time". Too bad the article did not elaborate further.