Combustion chamber shape and valve angles in WWII aero engines

[z42]

I was thinking of the shape of combustion chambers in WWII aero engines, and how different they are from each other, and for that matter from post-war racing (and general purpose) engines. Going all the way from completely flat heads as in the Merlin to very highly domed (hemispherical) heads on many radials.

If we take a flat head with a flat piston as the base case, the hemispherical head was invented to bring a couple of advantages:

• To minimize heat loss to the walls you want to maximize the volume and minimize the surface area of the combustion chamber. So ideally it should be spherically shaped, with a half-sphere roof as the cylinder head, and another half-sphere cup in the piston. Now in practice for mechanical strength and minimum weight it's probably better to have a more or less flat piston crown. So in practice you then end up with a sort of half-sphere shaped combustion chamber.
• The domed cylinder head allows bigger valves than with a flat head.

The hemispherical head proved to be quite popular, being widely used also in race cars. And Chrysler successfully marketed their implementation under the "Hemi" brand name. Eventually this design largely faded away due to various disadvantages:

• It is often mentioned that hemispherical heads only work with two valves per cylinder. I haven't seen an explanation why it would be impossible to implement a four valve hemispherical head. Presumably because then the valves would be angled so much also in the longitudinal direction (if we take the longitudinal axle to be the line of the cylinder bank in an inline engine) that the valve mechanisms for neighboring cylinders would run into each other? So for a 4V head we necessarily end up with a much flatter head for packaging reasons. Or alternatively, if we'd angle the valves in the hypothetical 4V hemi head only in the perpendicular direction, there would probably be some pretty sharp corners at the longitudinal ends of the combustion chamber, which probably wouldn't be good either. Though in a way the modern "pent roof" combustion chamber is a sort-of hemi shape, just much shallower due to the need to fit four valves rather than two.
• With increasing automotive fuel quality and other ways of inhibiting knock it became possible to increase the compression ratio. Getting a high compression ratio, particularly with an oversquare style racing engine, required a domed piston crown. In practice destroying the "minimum surface area" benefit that was one of the motivating ideas of the hemispherical head in the first place, and leading to an inefficient crescent/boomerang shaped combustion chamber with high surface area?
• The highly angled valves took up a lot of space, and required a complex and heavy valvetrain, increasing cost and limiting max rpm.

For a look at the heads and valve angles on a somewhat modern Formula I engine, see F1 Cylinder Head Design and Pneumatics, a closer look - F1technical.net (apparently a V10 from 1999). The famous Cosworth DFV from 1967 apparently had a valve angle of 32 degrees (the angle between the valve stems, AFAICT) so pretty flat also.

Going back to WWII aero engines, it seems that while inlines had either completely flat (e.g. Merlin) or almost flat (e.g. DB 60x, Allison) roofs, presumably largely for the 4V packaging reasons mentioned above, the radials OTOH almost universally(?) had highly domed heads, often with valve angles close to 90 degrees (by eyeballing). I would guess for a WWII radial, being undersquare and having a (compared to engines today) low compression ratio, there was no need to go to domed pistons, and indeed googling for pictures seem to confirm they tend to be mostly flat surfaced (perhaps with indents for the valves). Also, a high valve angle might allow more finning and better airflow around the head, and slightly reduce the diameter of the engines compared to having them more or less vertical.

So in a way it seems although the combustion chamber shapes were quite different, both the 4V inlines and the radials had good reasons for being the way they were. However, I've seen it mentioned that one thing that distinguishes WWII engines from modern ones, and which to an extent explains being able to run modern engines with high compression ratios despite having lower octane fuel than WWII aero engines, is combustion chamber shape. So what is missing? Squish seems important these days, any other major thing WWII aero engines missed compared to contemporary practice?

 

[Simon Thomas]

I was hoping to hear from those who actually know, but in the meantime I will gladly give you the benefit of my ignorance.

The modern car engine has a high geometric compression ratio, however the valve timing is such that the effective compression ration is less.

The aero engines need to fit an additional spark plug into the head, which also imposes difficulties in making everything fit as well as everything needing to be operable and maintainable. IIRC, WWII aviation spark plugs are slightly larger than current automotive plugs.

I would also say that after the Merlin's ramp head debacle, it may have been a career limiting move for anyone at RR to propose a new head shape.

 

[Bretoal2]

I was thinking of the shape of combustion chambers in WWII aero engines, and how different they are from each other, and for that matter from post-war racing (and general purpose) engines. Going all the way from completely flat heads as in the Merlin to very highly domed (hemispherical) heads on many radials.

If we take a flat head with a flat piston as the base case, the hemispherical head was invented to bring a couple of advantages:

• To minimize heat loss to the walls you want to maximize the volume and minimize the surface area of the combustion chamber. So ideally it should be spherically shaped, with a half-sphere roof as the cylinder head, and another half-sphere cup in the piston. Now in practice for mechanical strength and minimum weight it's probably better to have a more or less flat piston crown. So in practice you then end up with a sort of half-sphere shaped combustion chamber.
• The domed cylinder head allows bigger valves than with a flat head.

The hemispherical head proved to be quite popular, being widely used also in race cars. And Chrysler successfully marketed their implementation under the "Hemi" brand name. Eventually this design largely faded away due to various disadvantages:

• It is often mentioned that hemispherical heads only work with two valves per cylinder. I haven't seen an explanation why it would be impossible to implement a four valve hemispherical head. Presumably because then the valves would be angled so much also in the longitudinal direction (if we take the longitudinal axle to be the line of the cylinder bank in an inline engine) that the valve mechanisms for neighboring cylinders would run into each other? So for a 4V head we necessarily end up with a much flatter head for packaging reasons. Or alternatively, if we'd angle the valves in the hypothetical 4V hemi head only in the perpendicular direction, there would probably be some pretty sharp corners at the longitudinal ends of the combustion chamber, which probably wouldn't be good either. Though in a way the modern "pent roof" combustion chamber is a sort-of hemi shape, just much shallower due to the need to fit four valves rather than two.
• With increasing automotive fuel quality and other ways of inhibiting knock it became possible to increase the compression ratio. Getting a high compression ratio, particularly with an oversquare style racing engine, required a domed piston crown. In practice destroying the "minimum surface area" benefit that was one of the motivating ideas of the hemispherical head in the first place, and leading to an inefficient crescent/boomerang shaped combustion chamber with high surface area?
• The highly angled valves took up a lot of space, and required a complex and heavy valvetrain, increasing cost and limiting max rpm.

For a look at the heads and valve angles on a somewhat modern Formula I engine, see F1 Cylinder Head Design and Pneumatics, a closer look - F1technical.net (apparently a V10 from 1999). The famous Cosworth DFV from 1967 apparently had a valve angle of 32 degrees (the angle between the valve stems, AFAICT) so pretty flat also.

Going back to WWII aero engines, it seems that while inlines had either completely flat (e.g. Merlin) or almost flat (e.g. DB 60x, Allison) roofs, presumably largely for the 4V packaging reasons mentioned above, the radials OTOH almost universally(?) had highly domed heads, often with valve angles close to 90 degrees (by eyeballing). I would guess for a WWII radial, being undersquare and having a (compared to engines today) low compression ratio, there was no need to go to domed pistons, and indeed googling for pictures seem to confirm they tend to be mostly flat surfaced (perhaps with indents for the valves). Also, a high valve angle might allow more finning and better airflow around the head, and slightly reduce the diameter of the engines compared to having them more or less vertical.

So in a way it seems although the combustion chamber shapes were quite different, both the 4V inlines and the radials had good reasons for being the way they were. However, I've seen it mentioned that one thing that distinguishes WWII engines from modern ones, and which to an extent explains being able to run modern engines with high compression ratios despite having lower octane fuel than WWII aero engines, is combustion chamber shape. So what is missing? Squish seems important these days, any other major thing WWII aero engines missed compared to contemporary practice?

You forgot the Bristol Pegasus (and late Jupiter or Mercury) with four-valves pentroof combustion chamber - but with an awfull actuating system, unable to be enclosed in an oiltight case. Alfa-Romeo when trying to create a Jupiter-based twin row radial, returned to a 2-valves scheme.

A 4 valves hemispherical head with "concentric" valves was built in motorbike engines by Honda in the 80's.

Simon Thomas points out another problem, the true C/R of heavily supercharged engines is far more than geometric figure... and spark plug location is very important owing to the great dimensions of aero engines - inducing a long flame travel. And a personal remark, the Merlin pancake head has no squish surface !

 

[Engineman]

I was thinking of the shape of combustion chambers in WWII aero engines, and how different they are from each other, and for that matter from post-war racing (and general purpose) engines. Going all the way from completely flat heads as in the Merlin to very highly domed (hemispherical) heads on many radials.

If we take a flat head with a flat piston as the base case, the hemispherical head was invented to bring a couple of advantages:

• To minimize heat loss to the walls you want to maximize the volume and minimize the surface area of the combustion chamber. So ideally it should be spherically shaped, with a half-sphere roof as the cylinder head, and another half-sphere cup in the piston. Now in practice for mechanical strength and minimum weight it's probably better to have a more or less flat piston crown. So in practice you then end up with a sort of half-sphere shaped combustion chamber.
• The domed cylinder head allows bigger valves than with a flat head.

The hemispherical head proved to be quite popular, being widely used also in race cars. And Chrysler successfully marketed their implementation under the "Hemi" brand name. Eventually this design largely faded away due to various disadvantages:

• It is often mentioned that hemispherical heads only work with two valves per cylinder. I haven't seen an explanation why it would be impossible to implement a four valve hemispherical head. Presumably because then the valves would be angled so much also in the longitudinal direction (if we take the longitudinal axle to be the line of the cylinder bank in an inline engine) that the valve mechanisms for neighboring cylinders would run into each other? So for a 4V head we necessarily end up with a much flatter head for packaging reasons. Or alternatively, if we'd angle the valves in the hypothetical 4V hemi head only in the perpendicular direction, there would probably be some pretty sharp corners at the longitudinal ends of the combustion chamber, which probably wouldn't be good either. Though in a way the modern "pent roof" combustion chamber is a sort-of hemi shape, just much shallower due to the need to fit four valves rather than two.
• With increasing automotive fuel quality and other ways of inhibiting knock it became possible to increase the compression ratio. Getting a high compression ratio, particularly with an oversquare style racing engine, required a domed piston crown. In practice destroying the "minimum surface area" benefit that was one of the motivating ideas of the hemispherical head in the first place, and leading to an inefficient crescent/boomerang shaped combustion chamber with high surface area?
• The highly angled valves took up a lot of space, and required a complex and heavy valvetrain, increasing cost and limiting max rpm.

For a look at the heads and valve angles on a somewhat modern Formula I engine, see F1 Cylinder Head Design and Pneumatics, a closer look - F1technical.net (apparently a V10 from 1999). The famous Cosworth DFV from 1967 apparently had a valve angle of 32 degrees (the angle between the valve stems, AFAICT) so pretty flat also.

Going back to WWII aero engines, it seems that while inlines had either completely flat (e.g. Merlin) or almost flat (e.g. DB 60x, Allison) roofs, presumably largely for the 4V packaging reasons mentioned above, the radials OTOH almost universally(?) had highly domed heads, often with valve angles close to 90 degrees (by eyeballing). I would guess for a WWII radial, being undersquare and having a (compared to engines today) low compression ratio, there was no need to go to domed pistons, and indeed googling for pictures seem to confirm they tend to be mostly flat surfaced (perhaps with indents for the valves). Also, a high valve angle might allow more finning and better airflow around the head, and slightly reduce the diameter of the engines compared to having them more or less vertical.

So in a way it seems although the combustion chamber shapes were quite different, both the 4V inlines and the radials had good reasons for being the way they were. However, I've seen it mentioned that one thing that distinguishes WWII engines from modern ones, and which to an extent explains being able to run modern engines with high compression ratios despite having lower octane fuel than WWII aero engines, is combustion chamber shape. So what is missing? Squish seems important these days, any other major thing WWII aero engines missed compared to contemporary practice?

Probably the "best" head shape for a conventional poppet valve engine is a four valve slightly pent roof with some squish at the edges from a flat topped piston. This can give good airflow because the double intake and exhaust valves can facilitate greater airflow potential than a single larger valve in the same limiting dimensions. They are also lighter and that makes a whole world of difference, and they facilitate heat flow into the head/cooling. If you think four-valves can easy go into a real Hemi head, try drawing it out.
Additionally, modern inlet design has mastered the airflow into the cylinder through twin inlet valves to get great scavenging. Paired exhaust valves can be advantageous for flow/scavenging compared to single big ones in the same package space.
The WW2 engines are really a bit different to modern automotive, and the DB's did have a touch of valve angle, the V-1710 more so. However, although at no expense spared redesign, they might possibly gain a little from just a bit of valve angle, I don't think it would be huge. The real question is how would they look if you needed to build something to do that performance today?

Eng

 

[Simon Thomas]

The later Bristol poppet valve system was exposed by choice. The stream of cold air at cruising speed was reasonably effective at keeping the exhaust valve stems cool. Fedden mentioned in one of his papers that he could have enclosed them, but the benefit of enclosing them wasn't worth the effort.

I think the temperature compensating actuating system was a rather clever solution to a non-existent problem. As the Wright and P&W system showed, a fixed valve gap setting was adequate.

 

[z42]

The modern car engine has a high geometric compression ratio, however the valve timing is such that the effective compression ration is less.

Are you referring to Atkinson cycle engines? Aren't those mostly used on hybrids, whereas most "normal" car engines are still running on the "traditional" Otto cycle?

The aero engines need to fit an additional spark plug into the head, which also imposes difficulties in making everything fit as well as everything needing to be operable and maintainable. IIRC, WWII aviation spark plugs are slightly larger than current automotive plugs.

Yes. E.g. the DB 601/605 have the two spark plugs at the side of the combustion chamber, fairly close to each other. Probably not a very optimal placement, although easier for maintenance as both of the plugs are on the outside of the banks. For the 603, one of the spark plugs moved to the opposite side, inside the V.

I would also say that after the Merlin's ramp head debacle, it may have been a career limiting move for anyone at RR to propose a new head shape.

Hehe, true. Although TBF after wasting so much time with the ramp head, they were in a huge hurry to get anything that worked. A R&D project to investigate a better shape would have to wait for another day.

 

[Deleted member 68059]

Its all about squish, but, its very difficult to get just right. Because if you make the gap too small you`ll smash the pistons to bits if you rev up the engine when cold, and if you make the squish distance in a certain "twilight zone" clearance range you get severe detonation.

To do a big engine like that with the a really modern chamber would be hard, and would certainly need very rigidly imposed operating procedures to deal with the collossal thermal expansion changes in head vs crank centreline (no revs about a certain level until hot, etc). So no more emergency take offs in winter when cold !

In this respect something like a BMW 801 might fare well as the crankcase is steel like the crank and rods, hence slightly less expansion issue to look after (but it runs hotter than water cooled so.... marginal perhaps)

 

[z42]

Probably the "best" head shape for a conventional poppet valve engine is a four valve slightly pent roof with some squish at the edges from a flat topped piston.

Yes, that was the tentative conclusion I came to when trying to find some information about the topic online.

If you think four-valves can easy go into a real Hemi head, try drawing it out.

No, I don't think it would be easy, or maybe even possible at all, at least not without substantially increasing bore spacing (which of course has a number of negative side effects which probably makes the whole thing not worth it). Together with increasing compression ratio in modern engines necessitating a flatter head to avoid the need for domed pistons, the question is why would this even be an interesting exercise anymore?

The WW2 engines are really a bit different to modern automotive, and the DB's did have a touch of valve angle, the V-1710 more so. However, although at no expense spared redesign, they might possibly gain a little from just a bit of valve angle, I don't think it would be huge. The real question is how would they look if you needed to build something to do that performance today?

Indeed, that's a very interesting question!

 

[J.A.W.]

Check out a Honda XR 400 RFVC head, it is an example of a radially disposed 4V design.

https://farm8.staticflickr.com/7351/12996811113_698db6d5df_z.jpg

Of course, the best head designs do not have poppet valves crowding in & spoiling the shape/heat-transfer & etc...

https://thecrankshopvt.com/pub/media/catalog/product/1/0/1025trdome2.JPG

 

[Deleted member 68059]

Check out a Honda XR 400 RFVC head, it is an example of a radially disposed 4V design.

https://farm8.staticflickr.com/7351/12996811113_698db6d5df_z.jpg

Of course, the best head designs do not have poppet valves crowding in & spoiling the shape/heat-transfer & etc...

https://thecrankshopvt.com/pub/media/catalog/product/1/0/1025trdome2.JPG

Sadly this is all utter nonsense because a hemi chamber has a far larger surface area than a narrow pent roof head with 4 valves,
which means it has lower themal efficiency (also very low compression ratio) this can only be corrected by pushing the piston crown up into the head
volume giving a terrible chamber with awful burn paths and even higher surface area.

 

[J.A.W.]

Sadly this is all utter nonsense because a hemi chamber has a far larger surface area than a narrow pent roof head with 4 valves,
which means it has lower themal efficiency, this can only be corrected by pushing the piston crown up into the head
volume giving a terrible chamber with awful burn paths and even higher surface area.

Not necessarily, since if it is boosted, no bulging piston crown is needed for combustion events, & the splayed valves allow for the flow
to be aimed away from the cylinder walls. The venerable Norton Commando was designed with decades of GP racing experience backing
it up, & those Engineers eschewed the lumpy piston top altogether see below:

http://www.britcycle.com/Products/06/06_4048_001.jpg

 

[Deleted member 68059]

Not necessarily, since if it is boosted, no bulging piston crown is needed for combustion events, & the splayed valves allow for the flow
to be aimed away from the cylinder walls. The venerable Norton Commando was designed with decades of GP racing experience backing
it up, & those Engineers eschewed the lumpy piston top altogether see below:

http://www.britcycle.com/Products/06/06_4048_001.jpg

Thermodynamics doesnt matter because Norton (who went bust) raced an obsolete 2-valve hemi engine decades ago ?

I`m completely lost as to what it is you are trying to say about WW2 aero engines, or indeed anything else related to what
a high technology engine looks like.

 

[J.A.W.]

Thermodynamics doesnt matter because Norton (who went bust) raced an obsolete 2-valve hemi engine decades ago ?

I`m completely lost as to what it is you are trying to say about WW2 aero engines, or indeed anything else related to what
a high technology engine looks like.

No, that is a brand new Commando head, its a hemi, like an R-2800.

 

[Deleted member 68059]

No, that is a brand new Commando head, its a hemi, like an R-2800.

Pouring metal into moulds designed half a century ago is not a "brand new" cylinder head design.

Its hopelessly archaic and you would never, ever design anything like this now. The chamber area is very high
the surface area is very high, the compression ratio is far too low, the spark plug location is wrong (which is
of course always the case in a 2 valve head as you cant get it in the middle), and the single large inlet valve will be so heavy
its probably limited to very lethargic open/shut profiles as otherwise it will never follow the cam.

 

[J.A.W.]

10 to 1 isn't bad for an air-cooled SI 2V, only fairly recently have average cars run higher than that.

If I can find it to post, there is an excellent article written by Mercedes-Benz Engineers explaining how adoption of
the Cosworth DOHC 4V head changed the characteristics of the engine previously running their SOHC 2V hemi.

Really, only in the 5,000-to-7000rpm range was the Cosworth design superior, 'for the sporting driver' as they put it.

As for sparkplug location, the hemi R-2800 did locate its plugs more usefully than SOHC designs like the Merlin,
even if not nearly as advantageously - as the Napier Sabre - though.

 

[J.A.W.]

Pouring metal into moulds designed half a century ago is not a "brand new" cylinder head design.

Its hopelessly archaic and you would never, ever design anything like this now. The chamber area is very high
the surface area is very high, the compression ratio is far too low, the spark plug location is wrong (which is
of course always the case in a 2 valve head as you cant get it in the middle), and the single large inlet valve will be so heavy
its probably limited to very lethargic open/shut profiles as otherwise it will never follow the cam.

Arcane maybe, but punchy even so, 150hp out of that old twin mill - on Nitro-methane - reliably.


View: https://www.youtube.com/watch?v=tV92j51lUrU

 

[Deleted member 68059]

Arcane maybe, but punchy even so, 150hp out of that old twin mill - on Nitro-methane - reliably.


View: https://www.youtube.com/watch?v=tV92j51lUrU

Yes, quite true, you need to be very clever to get good results from old geometry, NASCAR engine people are very clever - they get great power from
absoutely horrific basic engine geometry, in some ways their work is more impressive than F1 for this reason.