Questions part 3: Another variety pack.

[Simon Thomas]

I was talking about balancing the the two cranks onto the prop shaft. Simple gearing would not be accurate enough to ensure each crank was putting it's torque into the prop.

 

[Bretoal2]

I was talking about balancing the the two cranks onto the prop shaft. Simple gearing would not be accurate enough to ensure each crank was putting it's torque into the prop.

But Rolls-Royce abandonned the Halford balancing system and used direct gearing for "their" version of the Sabre (RR Eagle II). The direct drive was present on other Halford's H engines, Dagger and Rapier.

 

[Simon Thomas]

The balancing system gets more power to the prop than not having a balanced system.
A cheaper gearing system still works and is lighter, but is sub-optimal wrt power.

 

[don4331]

I am interested to see how these multi-component crankshafts are sectioned. Do you have any idea where I might find explanatory images?

 

[don4331]

The balancing system gets more power to the prop than not having a balanced system.
A cheaper gearing system still works and is lighter, but is sub-optimal wrt power.

The geared solution may also require the crankshafts to be further apart:

Not an issue with the air-cooled Dagger/Rapier, but in a liquid cooled (and nearly square i.e. stroke and stroke very equal) engine, i.e. Eagle II, the crankshaft needs sufficient vertical separation so the rods don't contact each other.​

 

[Bretoal2]

The geared solution may also require the crankshafts to be further apart:

Not an issue with the air-cooled Dagger/Rapier, but in a liquid cooled (and nearly square i.e. stroke and stroke very equal) engine, i.e. Eagle II, the crankshaft needs sufficient vertical separation so the rods don't contact each other.​

Rolls-Royce used its usual "floating coupling" between crankshaft and gears, in order to absorb some (very) minor angular variations and misalignments.

 

[Simon Thomas]

Of course the minor variations change as the cranks are in effect torsional springs, so no amount of fine tolerance Rolls Royce machining will ever eliminate the issue.

 

[Engineman]

Shortround, did this strategy of using rich mixture for cooling exist in the second world war as a holdover from the first, or was there some additional benefit to using fuel compared to water or a mixture such as MW 50? Other than only needing one reservoir and associated delivery components, none immediately come to mind. I am not aware when water injection (or anything-other-than-fuel-injection) was first developed, but I imagine it must have been prior to even the first world war.

Of course, there are different benefits to the use of fuel(s) or other adjuncts for cooling or power boosting purposes.

Eng

 

[AspiringHowardHughes]

On the topic of balancing crankshaft torque - this is not something I have previously considered, but in the absence of copying an existing design (which may need some modification to suit my hypothetical engine, such as accommodating counter-rotating cranks) my knee-jerk reaction is to do one of the following:

Option A: Isolate the cranks entirely, each driving their own half of a counter-rotating propeller.
This is probably unadvisable for countless reasons, not only because there will inevitably need to be some sort of synchronisation between the upper and lower halves for other purposes, like guns and auxiliary systems that would ideally not need to be duplicated.

Option B: Each crank communicates power to the unified propeller drive system through a sprung gear, such that the natural inclination of the propeller assembly is to resist the two drive inputs equally.
I imagine this is a far more reasonable way to go about it, and perhaps is being accomplished in the systems mentioned by others, albeit mechanically differently.

Would the issue of an "uncontributing" crank even arise if the two were geared together rigidly at each end of the engine? It would surely take a good deal of slack within this gear pair to allow for one crank to "trail" the other, and even then I would think that the unloaded condition of the "trailing" crank would cause it to immediately take back up the slack and resume contributing. In this way, the drive as a whole could be taken off one of the two cranks without issue. Of course, this could load the "main" crank needlessly, so in practice I think it would be more advisable to place the resistance of the propeller on a third gear that is connected to both the cranks in some way.

All this brings to mind another concern, which is the potential issue of the two cranks trying to synchronise through the auxiliary drives (oiling, gun timing, valve timing, etc) and in the process overloading them to the point of failure. Was this an issue that cropped up in service engines?

I quite enjoy these engineering challenges that might easily be "brushed under the rug" when considering a prospective aircraft as a whole, thank you for bringing them to my attention.

 

[AspiringHowardHughes]

Don, that seems an unnecessary number of parts. I am sure it is not, of course, but I would much rather a bit more of a complex machining and assembly process (tunnel crankcase) if it means I can use a superior crank (single-piece). Were such complex crankshafts exclusive to radials? I imagine that an inline engine, with each crank bearing less stressed, would be a little more conservative design-wise.

On the topic of potential altercations between the rods of the upper and lower banks, my hope by making the two cranks rotate opposite directions was to allow for them to "mesh" and thus have the operating space overlap. Unfortunately, the adoption of my sleeve-rotary valve system incentivised me to use a tunnel crankcase, which sadly prevents any overlapping of the two banks.

You have given me an idea, however. I could preserve my desired studless and headless block design while allowing the two cranks to "mesh" by installing the disk-webs not as part of the rotating assembly but as static elements of the block, installed by being heat-fitted or screwed into place. Unfortunately this would prevent using a disk-webbed crank, and is probably more trouble than it is worth. Hard to imagine the installation process, too.

 

[don4331]

Don, that seems an unnecessary number of parts. I am sure it is not, of course, but I would much rather a bit more of a complex machining and assembly process (tunnel crankcase) if it means I can use a superior crank (single-piece). Were such complex crankshafts exclusive to radials? I imagine that an inline engine, with each crank bearing less stressed, would be a little more conservative design-wise.

On the topic of potential altercations between the rods of the upper and lower banks, my hope by making the two cranks rotate opposite directions was to allow for them to "mesh" and thus have the operating space overlap. Unfortunately, the adoption of my sleeve-rotary valve system incentivized me to use a tunnel crankcase, which sadly prevents any overlapping of the two banks.

You have given me an idea, however. I could preserve my desired studless and headless block design while allowing the two cranks to "mesh" by installing the disk-webs not as part of the rotating assembly but as static elements of the block, installed by being heat-fitted or screwed into place. Unfortunately this would prevent using a disk-webbed crank, and is probably more trouble than it is worth. Hard to imagine the installation process, too.

How much extra weight is your tunnel crankshaft carting around? Not big issue in a Tiger tank, but much more in a Spitfire...

When your fuel source goes from 87 Octane to 100 Octane to 115/145, how easy is it to improve your one piece tunnel crank engine? What do you need for a press to manufacture your tunnel crank? The splitting of the cylinders from the head on the Merlin occurred to allow improvements to either head or cylinders independently without having to change all the production machinery.

I should have commented about S Shortround6 's piston speed. Piston speed largely stabilized because of limitations on the connecting rods. Before @Snowygrouch (Calum Douglas) left, we had a discussion about the Jumo 213 connecting rods and how they added a thin cut (as do current F1 rods) to ensure the rod bolts only see tension, not shear. It allowed the 213 to operate several hundred ft/min faster than its competitors: 3,700rpm on 165mm stroke vs 3400rpm on 152mm stroke for Allison.

While the production R-4360s used a 1 piece crankshaft and split master rods; P&W was planning a 5 piece crankshaft/1 piece master rods for future power development.

I was fortunate to grow up with mill and lathe on the family farm, so I understood the manufacturing as I was designing. But what the RR and P&W machinists achieved was incredible.​

I am still looking forward to how you manage to have your sleeve-rotary valve have equal/greater intake/exhaust ports while being both cooled and lubricated.

 

[AspiringHowardHughes]

Thank you for your interest, Don.

It is likely that weight would be a drawback here. While in theory it only needs to be as heavy as the strength required demands, you are correct that by virtue of being larger (and having more rotational inertia) there will be reinforcement needed beyond what a typical crank would have. I hope that the ability to dispense with head studs and gaskets would mean the rest of the block can be lighter for the same strength, as the one-piece nature of it would grant it immense rigidity. Additionally, more space could be made for cooling passages and the routing of intake and exhaust systems.

As for the ability to improve it over time, I think it would be no worse off than any other engine. Making the crank single-piece would definitely complicated improving it, but not to the point of impossibility. I should clarify that my hypothetical engine is not like the Offenhauser in that it does not fully enclose the cylinder within the block. Rather, the cylinder liner and the individual "head" for each cylinder are screwed down into the block past the BDC of the piston such that no associated seam is exposed to the pressure of combustion, and thus does not need any seal beyond a mechanical fit sufficient to resist the pressure within the crankcase. In this way, the "heads" could be swapped out (or carried over to a new and improved block) like any other engine. No part of my engine design mandates a single-piece crankshaft, but I would like it as a structural bonus nonetheless.

I believe I am familiar with the Jumo 213 rod design you describe, I have a copy of "The Secret Horsepower Race", within which there is a diagram depicting exactly that. What a forward-thinking way of tackling that problem.

On the topic of my valve design:

Because the ports in the sleeve are not assigned to intake or exhaust, they must be the same size. Any desired difference in effective size must be adjusted by varying the size of the ports in the aforementioned cylinder liner.
Duration of each cycle can be easily adjusted by varying the size of the ports in the sleeve and/or liner, in fact they are adjustable in two "dimensions" just like a poppet valve (though with more freedom). I will happily expound upon this if desired.

Cooling would be superior to a conventional sleeve valve for two reasons.
1: There is no "hot side" of the cylinder, because the same ports are exposed to intake and exhaust in turn. This effectively eliminates thermal cycling of the sleeves as the engine starts and stops, as well as warping.
2: There is no "edge" of the sleeve, as rather than terminating within a "junk head", the rotating sleeve also occupies the entirety of the cylinder roof, save for a central spark plug position. This means that not only is there less sleeve perimeter from which oil can intrude into the combustion chamber (or be contaminated by combustion gas), but the entirety of the combustion chamber is surrounded by an oil film. In addition to this, sleeve friction would be reduced (no junk head to recede into) as well as sleeve inertia (movement is monodirectional). I think all this would produce very good cooling (and lubrication) properties. If anything, a larger oil cooler or more oil-coolant heat exchange within the block would be necessary.

 

[AspiringHowardHughes]

I have just now realized that by "equal/greater" you may have meant compared to a poppet or conventional sleeve engine.

The advantage of my sleeve-rotary design is that the shape of the ports does not need to be constrained so heavily as a typical sleeve valve.
An entire "slice" of the sleeve side (and the sleeve roof, if desired) can be dedicated to gas flow.
I imagine leveraging the full ability of this design would not be necessary, as other factors would impair breathing before the limit of port area could be reached.

 

[Bretoal2]

I have just now realized that by "equal/greater" you may have meant compared to a poppet or conventional sleeve engine.

The advantage of my sleeve-rotary design is that the shape of the ports does not need to be constrained so heavily as a typical sleeve valve.
An entire "slice" of the sleeve side (and the sleeve roof, if desired) can be dedicated to gas flow.
I imagine leveraging the full ability of this design would not be necessary, as other factors would impair breathing before the limit of port area could be reached.

The description of "your" sleeve-rotary design seems to correspond to the processes patented just before the war by Roland Chilton (US Patents 2137390 of November 22, 1938, 2146428 of February 7, 1939, and 2150541 of March 14, 1939).

The fact that no engine designed in this way ever materialized and that Chilton moved on to other projects suggests that the idea had far fewer advantages than assumed, and that, on the contrary, it was a source of unexpected difficulties.

Ultimately, the rotating "roof" solution seems to have never worked well; among many other ones, the Aspin motorcycles met with only very limited success. And yet, their engine had a bore much smaller than that required for an aircraft engine.

In Germany, engine manufacturers were being pressured by the RLM to abandon poppet valves, but the studies made by the DVL (German Development Laboratory) led Daimler-Benz and Junkers towards the over head rotary flat disc solution – see Callum Douglas's excellent book on the subject. I will be publishing some completely new information on the DB 611/613 in a few months.

 

[AspiringHowardHughes]

I believe you greatly misunderstand my described design, it does not overlap sufficiently with the patents you have provided to be considered infringement.
The vast majority (if not all) of the port area in my design lies in the wall of the cylinder, exposed and concealed in turn by the rotating sleeve.
When I describe the sleeve also covering the roof of the combustion chamber (save for the spark plug, as I discussed earlier) this is to achieve the goal of reducing the "edge" area of the sleeve, and thus improving oil sealing alongside the related benefits of cooling and lubrication.

The Chilton patent not only requires a "head" for the engine (unlike mine), but the port area is small compared to what my design allows, and the combustion chamber shape is horrid as a result of what appears to be an attempt at directing gas flow more beneficially.

In comparison, my design allows the designer complete freedom over the shape of the combustion chamber, provided it is rotationally symmetrical. Hardly a downside.

The overhead rotary valves are poor for similar reasons, many of which Calum Douglas explains in "The Secret Horsepower Race".

I will produce a simple sketch so that I might be better understood.
I do appreciate the efforts at fact-checking the originality of my design, though.

 

[Bretoal2]

I believe you greatly misunderstand my described design, it does not overlap sufficiently with the patents you have provided to be considered infringement.
The vast majority (if not all) of the port area in my design lies in the wall of the cylinder, exposed and concealed in turn by the rotating sleeve.
When I describe the sleeve also covering the roof of the combustion chamber (save for the spark plug, as I discussed earlier) this is to achieve the goal of reducing the "edge" area of the sleeve, and thus improving oil sealing alongside the related benefits of cooling and lubrication.

The Chilton patent not only requires a "head" for the engine (unlike mine), but the port area is small compared to what my design allows, and the combustion chamber shape is horrid as a result of what appears to be an attempt at directing gas flow more beneficially.

In comparison, my design allows the designer complete freedom over the shape of the combustion chamber, provided it is rotationally symmetrical. Hardly a downside.

The overhead rotary valves are poor for similar reasons, many of which Calum Douglas explains in "The Secret Horsepower Race".

I will produce a simple sketch so that I might be better understood.
I do appreciate the efforts at fact-checking the originality of my design, though.

There's no idea of copyright infringement in my comment (especially for patents that have likely been in public domain for decades ! ); I'm simply discussing the viability of the concept from a historian's perspective.

A sketch would certainly be welcome and would prevent any misunderstanding.

 

[AspiringHowardHughes]

Courtesy of Google Drawings, a simple cross-sectional diagram.
-Green is the piston
-Red is the sleeve
-Black is the cylinder liner and manifolds

The upper diagram shows the cylinder during compression, with the sleeve closing off access to the ports.
The lower diagram shows the cylinder during intake and scavenging, with the ports available.
In this case I have chosen to demonstrate an implementation of my sleeve using a two-stroke operating cycle, but I am sure it can easily be visualized for a four-stroke cycle. One could think of this image as presenting a moment of valve overlap in a four-stroke cycle, I suppose.

I have not taken any care to present the piston or cylinder as being at all optimised, But I am sure it does not impair comprehension.
The ports may seem in this diagram to cause areas of great weakness in the combustion chamber, this is an obstacle that would be designed around like any sleeve-valve engine.
It is possible that this design would emerge with many ports of a smaller size to remedy this.

 

[AspiringHowardHughes]

It is possible the valve cutout in the sleeve could take the form of a thin vertical slot, allowing the engine to move from sealed to fully open as quickly as possible without weakening the combustion chamber excessively. This might also introduce some beneficial turbulence.
I have not illustrated the spark plug in this diagram, it would be placed directly in the centre of the cylinder roof, protruding through a small hole in the sleeve, which rotates around it.
If the engine is fuel-injected, this would be done through another small cutout in the sleeve, such that the injector head is exposed only as it actuates, and shrouded from the strain of combustion otherwise.

 

[AspiringHowardHughes]

For clarity I have produced a corresponding pair of top views.

Again, few details here are critical, the intent is only to communicate the general principle of the design.
For example, the intake and exhaust manifolds might in a real installation be located somewhere other than 180 degrees apart, because this is how the valve timing is adjusted.
Similarly, the appropriate ratio of open space to sleeve wall is up for debate.
Suggestions for what these values ought to be and factors that contribute to these conclusions are very much welcome.

 

[z42]

If one spark plug is in the center of the head, where is the other one?

For two stroke designs, loop scavenging is usually considered superior to cross flow scavenging. Then again, for a two stroke design you can just use piston porting and dispense with the sleeves, no?

For valve opening and closing, what you're looking for is known as a "valve timing diagram". You'll find plenty of them online as well as discussion. Obviously almost entirely assuming poppet valves, but the same principle applies to other valve types as well.