Build the perfect water cooled engine

Ad: This forum contains affiliate links to products on Amazon and eBay. More information in Terms and rules

Interestingly, the air cooled air racers at Reno spray water onto the engine to keep their monsters cool. Liquid cooled racers spray water on their radiators - basically because in both instances they are operating beyond the cooling capacity of their installations.

I have seen a modification that sprays water onto an inter-cooler for supercharged car engines as well. Although this is performance enhancing rather than a cooling issue.
With the Reno racers, have the cowls been modified to make them more streamlined at the expense of cooling?
I'm a little surprised as cooling should be part of the racing cars design as overheating is as damaging as running too cool....
Cheers
John
 
When considering sleeve valve vs. poppet valve development, it is unfair to speak of "15 years of development and millions of dollars/pounds" when one can ask that how many years and how many millions were spent to developing poppet valves to the level they were in WW2?

As for air cooled vs liquid cooled issue, the JFC report has an interesting statement by "Mr. Steppe" of NAA. He stated that folks at NAA are convinced that using liquid cooled engines one can achieve similar range and speed performance on a much smaller and lighter airframe and pointed out the size difference between the P-51 and its R-2800 powered colleagues...

Personally, I would fit the radiators on the leading edge. I'd rule out belly scoop for its vulnerability when operating from unpaved runways. I would use coolant radiators on both wings in a way that if one of them is holed, one could isolate it and have still some remaining cooling capacity left to get home.

One more thing. When one compares e.g. 2000 hp engines to those projects with 4000 hp, one can note 2 things: the power/weight gets substantially poorer and the complexity goes disproportionally up.

So, in large bombers/transports wouldn't it been perhaps wiser to use e.g. 8 2000 hp engines engines instead of 4 4000 hp engines? Or e.g. in the case of the B-29, 6 x V-1710-89/111 instead of 4 x R-3350?
 
Last edited:
When considering sleeve valve vs. poppet valve development, it is unfair to speak of "15 years of development and millions of dollars/pounds" when one can ask that how many years and how many millions were spent to developing poppet valves to the level they were in WW2?

The entire cost of the development of the sleeve valve engine (at least for aircraft, there were a few luxury automobiles that used them) was borne by the British and basically two companies. The Development of the poppet valve engine was spread out over a number of countries and many companies. Development of the poppet valve engine was going to occur regardless of what other valve arrangement's were worked on. Some of the "improvements" benefited both engines. Better casting/forging techniques meant better/closer pitched fins for both types of air-cooled engines.

One more thing. When one compares e.g. 2000 hp engines to those projects with 4000 hp, one can note 2 things: the power/weight gets substantially poorer and the complexity goes disproportionally up.

This is because there is a limit to how big you can make a single cylinder and how much power you can get from a single cylinder. Yes you can run a smaller cylinder faster but the bearing loads go up with the square of the speed and the smaller cylinder has less cooling area for the power it is trying to make. Once you hit that limit per cylinder (which changes as technology moves) the only way to make more power is to use more cylinders per engine and that brings on the complexity and power to weight problems.
[/QUOTE]
So, in large bombers/transports wouldn't it been perhaps wiser to use e.g. 8 2000 hp engines engines instead of 4 4000 hp engines? Or e.g. in the case of the B-29, 6 x V-1710-89/111 instead of 4 x R-3350?[/QUOTE]

Now you are back to the complexity problem. You haven't cut the number of spark plugs that need changing by much. 8 x 18 cylinder engines or 4 X 28-36 cylinder engines. Valves to be checked/adjusted. Fuel and oil pumps, throttle linkages and other engine controls would actually require more maintenance. This is assuming that the big engines need contra-rotating props and the 2000hp ones don't. If the big engines can get by on a single propeller then the props may be cheaper and easier to maintain.

Wings with large numbers of propellers messing with the airflow don't seem to do as well in some flight conditions as wings operating with less disturbed airflow. It was also hoped that 4 big nacelles would have less drag than 8 smaller ones.
 
Flat engines can be arranged so that there are less vibrations, but that requires a more complicate dcrankshaft and a longer engine (one crank throw per cylinder, basically). In any case the 60° V12 is close to perfectly balanced.

The flat engine also has problems with rigidity. In WW2 aircraft the engine was usually a semi-stressed member of the airframe, in that the prop and thrust loads were transferred to the airframe through the engine and its mounting. The flat engine, especially one as long as a flat 12 aero engine, doesn't have that much rigidity in the vertical plane.
 
Were the planes with leading-edge radiators slow birds?

The Tempest I was 20mph faster than any other Tempest variant. Granted the MkI had more power (about 200hp) than normal Sabres at that time, but the aircraft was still faster than the other mks when they caught up in power.

The Mosquito was tested with chin radiators, using the R-R power egg from the Beaufighter, and also used on the Lancaster. I can't find the reference to what the performance difference was. However, the radiator design of the Mossie was said by de Havillands to be worth 5-10mph, depending on altitude and speed.
 
The entire cost of the development of the sleeve valve engine (at least for aircraft, there were a few luxury automobiles that used them) was borne by the British and basically two companies. The Development of the poppet valve engine was spread out over a number of countries and many companies.

Still the total amount of money and effort spent on perfecting the poppet valve train far exceeded the amounts spent on sleeve valves. Yet, I am 100% convinced that the sleeve valve is the better system. An exellent indication of the sleeve valve superiority is to compare the maximum allowable cylinder head temperatures of air cooled radials. E.g. the Centaurus 18 allowed 300 deg C as the maximum continuous (weak and rich mixture) CHT while e.g. a R-1820-56 (which has the famous W cooling fins) the weak mixture limit was 218 deg C with the 5 minute limit being 248 deg C.

But being better may not do, as was with the case of VHS vs. Beta.

[/QUOTE]This is because there is a limit to how big you can make a single cylinder and how much power you can get from a single cylinder. Yes you can run a smaller cylinder faster but the bearing loads go up with the square of the speed and the smaller cylinder has less cooling area for the power it is trying to make. Once you hit that limit per cylinder (which changes as technology moves) the only way to make more power is to use more cylinders per engine and that brings on the complexity and power to weight problems.
[/QUOTE]

My point exactly.

[/QUOTE]Now you are back to the complexity problem. You haven't cut the number of spark plugs that need changing by much. 8 x 18 cylinder engines or 4 X 28-36 cylinder engines. Valves to be checked/adjusted. Fuel and oil pumps, throttle linkages and other engine controls would actually require more maintenance. This is assuming that the big engines need contra-rotating props and the 2000hp ones don't. If the big engines can get by on a single propeller then the props may be cheaper and easier to maintain.[/QUOTE]

Perhaps, but you are leaving out accessibility. If one compares e.g. the R-2800 to the R-4360, one can note how the accessibility of the latter is disproportionally poorer. Plus, according to Graham White's R-4360 book, the R-4360 had significantly lower TBO than the R-2800.

[/QUOTE]Wings with large numbers of propellers messing with the airflow don't seem to do as well in some flight conditions as wings operating with less disturbed airflow. It was also hoped that 4 big nacelles would have less drag than 8 smaller ones.[/QUOTE]

That reminds me of another air cooled vs. liquid cooled thing. E.g. in the B-36 the engine installation had become extremely complicated to keep the engines cool. Again, liquid cooling offers better flexibility.
 
Still the total amount of money and effort spent on perfecting the poppet valve train far exceeded the amounts spent on sleeve valves. Yet, I am 100% convinced that the sleeve valve is the better system. An exellent indication of the sleeve valve superiority is to compare the maximum allowable cylinder head temperatures of air cooled radials. E.g. the Centaurus 18 allowed 300 deg C as the maximum continuous (weak and rich mixture) CHT while e.g. a R-1820-56 (which has the famous W cooling fins) the weak mixture limit was 218 deg C with the 5 minute limit being 248 deg C.

But being better may not do, as was with the case of VHS vs. Beta.

More money may have been spent on poppet valve systems but that may be because there are so many more of them. The basic poppet valve could be developed in increments. From basic steel to high temperature steel to sodium cooled and stelite facings. Valve seats could go from machined/ground in in an iron head to various insert materials and manufacturing techniques. Valve spring material and manufacture could also advance in steps.
The sleeve valve is more of an all or nothing proposition. Leaving out small scale production for a few luxury automobiles and the Early Bristol engines you can either manufacture working, interchangeable sleeve valves or you can't. There isn't much of a halfway point or a step in the sleeve valve progress where you can develop XX amount of power per cylinder but a small improvement gets you XX plus 10% or 20%. At least not in the sleeve valve mechanism itself. Different porting can and different cooling can.
I don't know about the Centaurus but the Hercules went through at least 5 (not the early 4) different cylinder head designs during it's history in a quest for more power/better cooling and the head had little to do with the sleeve valve. Again as technology improved more finning could be provided by improved foundry techniques in casting/forging and in machining. The Early Hercules (pre-war) had a sand cast head with 540 sq in of cooling, the first production versions had a die cast head with 581 sq in of cooling. The main wartime versions used a two piece die cast head with 728 sq in and were followed by a finer pitched head with 777 sq in that lowered CHT by 15 degrees C. The Post war engines used a head with a copper base fasted to a steel skirt faced with nickel on the combustion surface which lowered the CHT by another 25 degrees C.

While these junkhead changes allowed for more power they had little to do with the Sleeve valve itself and were paralleled by improvements in cylinder head design on air-cooled poppet valve engines, which again, had little to do with actual poppet valve mechanism.


Anther model of the R-1820 was rated at 205 deg C weak mixture 218 deg C rich climb and 232 deg C rich continuous (or reversed chart isn't quite clear) with a 260 deg C max for 5 minutes. P&W R-1830s (not the most sophisticated engine) use limits of 230 deg C for lean and 260 deg C for rich mixture limits as did early R-2800s

Some war time Hercules engines used limits of 270 C with 280 emergency , or in some cases 20 degrees higher if it was not possible to observe the lower limits.

The problem with these temperature limits is they tell us what was allowable by the manufacturer/government at the location of the thermal couple. Not what the actual thermal load was or how well it was dissipated under certain conditions.

Like the R-2800, it might have been held to the same 230 deg C limit lean and 260 deg C limit rich when it changed form the "B" series engine to the "C" series but the "C" series cylinder used an aluminium finned cooling muff around the steel barrel vs machined steel fins and it used a forged cylinder head with much more fining than the cast head on the "B" series so at the same temperature it was dissipating more heat.
 
With the computer design and simulation technology, and alloys that are available now to create a sleeve valve engine; would this engine type offer any advantages for light aircraft and automobiles that would be economically feasible?
 
I see what you're saying.
Rather than run too far off topic, i was simply taking notice that placement of the radiator intake would effect overall performance, and that leading edge vents would not present the best drag profile.
Whether its intended to be a ground attack/support fighter or not, having the radiator on the chin or near the engine would not make it more vulnerable, something that i'm sure was a consideration for war-time engineers.
Its also represented by the placement of the radiator on the P-40-XP to the first production model P-40B. Though that change was first implemented on the P-40(would've been A) and was said to have lowered performance. Obviously, there was more that went into the B that also accounted for its increase in speed from the XP model.


Bill

the P-40 was designed to be an " air superiority " (modern term) fighter for the American mainland.
it later served as an ground attack fighter/bomber ( hence its other nickname, the B-40) when more
advance fighters came to the war.

for the other comments on the P-39/P-63:

as far as the P-39.. Chuck Yeager said at a 100feet.. nothing could beat the P-39.
I don't think the USAAF ordered any P-63's. but the russians did.. alot of them.


as far as underwing radiators (ie: Bf109).. I'm pretty sure I read that they actually produced
to some degree thrust, which outweighted any loss in drag.
 
Last edited:
Were the planes with leading-edge radiators slow birds?
Compared to??
Its a simple observation of aerodynamics.
The boundary layer is disrupted by leading edge vents, and drag over the nose. Diminishing drag in these areas helps diminish drag over the wing. The air frame would therefore be cleaner and faster.
Could planes with leading edge vents be faster if they were located elsewhere?
Probably, but there needs to be room in the air frame for such a design.
Planes with leading edge slats, or under wing rigs are typically smaller, ie 109, Spitfire, 190, but they also showed up on the F4U.

In other words, Camm decided it would've been good to have draggier plane, rather than a 'clean' one?
Configurations might also have something to do with the size of the engine and the radiator to cool it.
The Chin radiator of the Tempest is draggy, but then the 24 cylinder Napier running at 4000RPM probably required such a rig.
Could a similar sized radiator have fit aft the pilot??
Yes, the plane would be faster if it could be achieved.
Danovan Berlin made the same observations about the later P-40s, something he did not directly design, (E and later) saying the P-40 was given a radiator chin/vent that was much bigger than required and did inhibit its performance.
 
With the computer design and simulation technology, and alloys that are available now to create a sleeve valve engine; would this engine type offer any advantages for light aircraft and automobiles that would be economically feasible?

In theory it will, in practice it may. The kicker is the economically feasible. How much a physical change in plant do you need? The sleeves's were forgings and needed specialized grinders. While a light plane engine or car wouldn't be as complicated as a 14cylinder radial airplane the Hercules used an enormous amount of gears in it's "valve train" which may be much easier to manufacture using today's technology. A plant to build sleeve valve engines would need a fair amount of retooling from a plant that makes poppet valve engines.
Could you sell enough to cover the cost of the R&D and retooling the plant?
 
So, in large bombers/transports wouldn't it been perhaps wiser to use e.g. 8 2000 hp engines engines instead of 4 4000 hp engines? Or e.g. in the case of the B-29, 6 x V-1710-89/111 instead of 4 x R-3350?

Consider the fuel economy and range of the bomber.
I could be wrong, but radials were seen as more reliable for that role, and didn't require the plumbing for a cooling system.
There's also the weight issue. 8 engines on one wing might actually make up more gross weight than the potential of 4 engines with similar net output.

Am i mistaken in thinking that large displacement/more cylinders is also more capable of running on leaner mixtures compared to less cylinders/smaller displacement?

It would also be interesting to have seen further development of Chrysler's first hemi, first applied to the P-60 and P-47H. It may have had limited applications but could make a good duel prop engine. The V16 would be slender and thinner than a bulky X design and output was a considerable 2400+ hp.
Again, a proper cooling system design could more than make up for the added length.
The best out of the box V engine of that era, IMO, would be the DB-600s engines.
They had decent power to weight and could be applied to a variety of applications.
They also tried or attempted various duel V designs, and turbo/supercharged arrangements. Its no wonder they appeared on fighters and light bombers.
 
How would the P-51 have performed (and appeared) with a DB601/DB605?
For that matter, how would any Merlin powered craft have performed appeared with Teutonic power?
 
How would the P-51 have performed (and appeared) with a DB601/DB605?
For that matter, how would any Merlin powered craft have performed appeared with Teutonic power?

The Germans, clever though they were, never bested the Merlin/Griffon. There are other engineering examples where British design has proven superior. Motorcycles being one.
Cheers
John
 
well I can't answer those questions, but the P-40F used the Merlin engine ( built by Packard to exacting british specs)
and ground crews pilots complained the Merlin was tempermental, unreliable, had mediocre performance.

the next version, the P-40K went back to the Allison V-1710, which was more powerful, more reliable, and faster.

I think the myth of the Merlin stems from its use in the P-51, which wasn't that great of an aircraft IMO.
 
Last edited:
Am i mistaken in thinking that large displacement/more cylinders is also more capable of running on leaner mixtures compared to less cylinders/smaller displacement?

I believe you are mistaken. Large cylinders are harder to cool than small cylinders, ratio of cylinder wall/head surface to volume. More smaller cylinders may be able to run leaner than fewer large cylinders of the same displacement. Think R-1820 vs R-1830. But the 14 cylinder engine weighed more and had more friction so I don't know where the total economy wound up.
It would also be interesting to have seen further development of Chrysler's first hemi, first applied to the P-60 and P-47H. It may have had limited applications but could make a good duel prop engine. The V16 would be slender and thinner than a bulky X design and output was a considerable 2400+ hp.

While was thinner than some X designs it was actually wider than a Griffon which was almost the exact displacement. It was also quite a bit longer and almost 400lbs heavier (than a two stage Griffon). Weight to the Chrysler does not include turbo. There were a few questionable constructional choices but the decision to take the drive from the middle of the crankshaft points to the problem with V-16 engines.
The best out of the box V engine of that era, IMO, would be the DB-600s engines.
They had decent power to weight and could be applied to a variety of applications.
They also tried or attempted various duel V designs, and turbo/supercharged arrangements. Its no wonder they appeared on fighters and light bombers.

The power to weight of the early 601s was no better (and no worse) than the Allison or Merlin. And not much different than the Hispano Y or Soviet M-105.
The latter two had darn little development potential though.

The Germans used the liquid cooled engines because they only had ONE viable radial engine, the BMW 801 and it was sort of a late comer.
It can take 4-6 years to take an aircraft engine from drawing board to service.
Trying to use existing bits and pieces is a way to try to shorten up that time span.
 
How would the P-51 have performed (and appeared) with a DB601/DB605?
For that matter, how would any Merlin powered craft have performed appeared with Teutonic power?

Very much the same. Comperable V-1650s and 605s had very similiar power curves, weights and dimensions. The plus side would be that the DB engines had better fuel effiency, so either the P-51 would possess even more range, or could do away with the rear fuselage tank. That would lighten the aircraft..
The downside was that the 605AS engines appeared in late 1943, so there would be a month or two delay getting them to the frontlines.
 

Users who are viewing this thread

Back