3 row radials: pros & cons

[tomo pauk]

Apart from some (one?) experimental engines, such a development was as good as non-existent. I wonder if there were some/any plausible advantages vs. two-row 'classics', ditto for disadvantages

[Readie]

lack of cooling for the third row.
You would be better off with an inline liquid cooled motor.
John

[gjs238]

lack of cooling for the third row.
You would be better off with an inline liquid cooled motor.
John

Such as??????

[Shortround6]

With Readie do you really have to ask?

As far as the three row engine goes it has a lot of disadvantages and few real advantages.

Engines can be looked at several ways. One is to try and build on engine of a certain power class. Like a 1000hp engine, or a 1500hp engine or a 2000hp engine.
Another way is to say you have developed a cylinder of a certain size/power, now how can you arrange multiples of that cylinder to get the power you want? This last is especial true of air-cooled cylinders because of the cooling problem.
Let's say you have a 70hp cylinder (in 1937) and you are making both a 630hp nine and a 980hp fourteen. Were do you go from there? A 1260hp eighteen or a 1470hp twenty-one?
The 21 will be longer and heavier than the 18 with it's longer crankshaft and crankcase. Cooling the rear row will be harder. The frontal area will be almost the same if not exactly the same. Arranging the valve gear for the rear row is going to be a headache. Many two row radials used a common cam ring at the front and routed the push rods fo rthe rear cylinders in-between the front cylinders. Some used a cam ring both in front and in back. The 3 row engine is going to require two cam rings at a minimum. Or, as was done on the Deerhound, lining up the cylinders and using overhead cams with 7 different cam drives.
A two row 18 does weigh more than a two row 14 but the 21 only has 16.7% more displacement/power than the 18. Is it enough to make up for the extra weight, cost and complication?

[wuzak]

The Deerhound used vertical shaft drive to its overhead cams - like those employed in the Merlin, V-1710, etc. Not a great issue, I wouldn't have thought. Each of the cam drives was taken off a common gear ring - so in that sense is no different to the radial cam rings, in thet both needed to be driven by reduction gearing.

The Deerhound I was only 38l (2260ci), the II was increased in size to 41l (2510ci)and up to 1500hp. The redesigned III stayed at 41l and the same 44" in diameter as the II. The only on ewhich was built was tested at 1800hp. So, similar in capacity and performance to an R-2600 of the time, but much smaller diameter. It was only very early in the testing stages, however, so much work still had to be done.

Armstrong Siddeley engineers apparently originally wanted to make the engine liquid cooled but the MAP wanted competition for Bristol in air cooled engines.

The Deerhound was re-endineered to use reverse cooling early in the program. With hindsight, it may have been possible to provide the necessary cooling had they elected to use a system of baffles to guide the cooling air around the cylinders and heads - as was used in the R-4360.

[wuzak]

Just in case I didn't make the point clear, for a given capacity (and presumably power level) the diameter of a 3 row radial could be smaller than for a 2 row radial.

The Deerhound II/III had a 5" (127mm) stroke as compared to a Wright R-2600's 6.312" (160.2mm), the Hercules' 6.5" stroke (165.1mm) and the P&W R-2800's 6" (152.4mm). So, it may have been possible later in development to use more rpm than those - as it was the III tested at 1800hp at 2600rpm.

Also, comapre the diameters:
Deerhound 44"
R-2600 55"
R-2800 52.8"
Hercules 55"

So, 80% of the diameter of the R-2600 and Hercules, meaning 64% of the frontal area. 83% of the diameter of teh R-2800, 69% of frontal area.

[Shortround6]

You are quite right Wuzak. the 3 row radial will be smaller in diameter for the same power, that is about the only advantage.

The Deerhounds use of overhead cams was one way around the multiple rows of pushrod problems but introduced a few of it's own. A V engine uses two cam drives, a W engine uses 3 and the Deerhound used 7 cam drive units. While the parts are identical it is that many more parts that have to be made and fitted.
Overhead cam boxes also tend to limit some of the area that could be used for cooling fins. Perhaps not a big deal for an engine using 87 octane fuel but as the octane went up it becomes more important.

Does anybody have any information of the weight of the Deerhounds?

[wuzak]

No, I can't find any weight data.

[tomo pauk]

If I'm following this right, a 3-row radial would not made Tempest II/Sea Fury any wider in front vs. Tempest V? Or somewhere in-between?

[Shortround6]

The three row should be smaller than the Centaurus for the same power. Wither you can get it as small as the Sabre is a different question. Then you have the problem of the sleeve drives. easiest is to put the cylinders in line and run the shafts that control the sleeves between the rows, the inline configuration makes it harder to cool though, compared to staggered rows.

The question is if the 3 row is worth it. For a given power level it is smaller than a 2 row as Wuzak has shown. The engine will be longer, the power section will be at least 50% longer although the gear reduction section and the supercharger/accessories section will stay the same length. It will be heavier for the same power. It is more complicated and expensive to manufacture, It will be more difficult to develop. Look at the problems with mixture distribution in the R-3350. The R-4360 was no picnic either. Sneaking intake pipes forward to the front row can be done, it just takes a bit more work than a 2 row engine and the same for exhaust.

Is the extra work and weight worth the smaller diameter?

When designers reached a plateau in engine power per cylinder they fiddled a bit more with odd ball combinations to use more cylinders, Wright tried making a 22 cylinder two row radial but didn't get very far. an R-4090 or 4100 of pretty much the same diameter as the R-3350.

[tomo pauk]

I was looking at the in-line cylinders and poppet valves, ie. a layout of the Deerhound, but circa 3000 cu in displacement (for mid-war and after that). The 3-row is bound to be longer than a 2-row, but also shorter than an in-line; the complete powerplant section (from supercharger end to tip of spinner) shuld be somewhere in between.
About the extra work: the all mighty 'depends' is the keyword. If RAF (or LW) can have 2000-2500 HP engine of the layout & size like I've proposed in 1943-44, than it compares rather favorably with what they had historically at the disposal at same time frame.

[wuzak]

The three row should be smaller than the Centaurus for the same power. Wither you can get it as small as the Sabre is a different question. Then you have the problem of the sleeve drives. easiest is to put the cylinders in line and run the shafts that control the sleeves between the rows, the inline configuration makes it harder to cool though, compared to staggered rows.

A 3 row radial will be more difficult to cool than a 2 row radial, just as the two row radial is harder to cool than the single row, and the 4 row harder than the 3 row. Not sure that the in-line configuration makes it any harder to cool than a staggered arrangement. Without the baffles the R-4360 wouldn't have worked at all, and I am sure the 3 row radial, be it staggered or lined up, could be made to work with similar arrangements.

I figure the Sabre could fit in a round cowling of approximately 48-50" in diameter, similar to that required by the Vulture. The Vulture installations, however, could be even more tightly packaged with bulges covering the front of the cam covers. Not sure if that could be done or was desirable for the Sabre.

Tomo never specified sleeve valves, but having the cylinders in line should simplify the sleeve drive arrangement.

The question is if the 3 row is worth it. For a given power level it is smaller than a 2 row as Wuzak has shown. The engine will be longer, the power section will be at least 50% longer although the gear reduction section and the supercharger/accessories section will stay the same length. It will be heavier for the same power. It is more complicated and expensive to manufacture, It will be more difficult to develop. Look at the problems with mixture distribution in the R-3350. The R-4360 was no picnic either. Sneaking intake pipes forward to the front row can be done, it just takes a bit more work than a 2 row engine and the same for exhaust.

For the same capacity I'm sure the power section will not be 50% longer, since the bore will be smaller, and each cylinder will have less heat rejection requirements and thus require smaller fins - so the bore centre distance could possibly be reduced.

Mixture distribution could very well prove to be problematic. It was for the engines you mentioned, as well as many others, including at least one V12 (the V-1710 had this problem before the "ram's horns" intae manifolds were developed).

Certainly it will be more expensive to build and develop, and probably heavier.

Is the extra work and weight worth the smaller diameter?

That's a question for aerodynamcists everywhere.

Certainly there is no benefit in using an engine as an alternative in some circumstances - like the Tempest, for example. It was deisgned for use with the Centaurus, so its fuselage was probably bigger than it needed to be for the Sabre, and using the smaller diameter Deerhound would probably give negligible aeordynamic benefit.

However, an aircraft designed around the more compact engine may gain more benefit from the drag reduction than the cost of the extra weight.

Perhaps a liquid cooled 3 row radial would be more viable? It would allow for single piece blocks, strengthening the whole unit. Cooling would be less of an issue, the engine will be shorter, though its installed weight would go up.

[wuzak]

I was looking at the in-line cylinders and poppet valves, ie. a layout of the Deerhound, but circa 3000 cu in displacement (for mid-war and after that). The 3-row is bound to be longer than a 2-row, but also shorter than an in-line; the complete powerplant section (from supercharger end to tip of spinner) shuld be somewhere in between.
About the extra work: the all mighty 'depends' is the keyword. If RAF (or LW) can have 2000-2500 HP engine of the layout & size like I've proposed in 1943-44, than it compares rather favorably with what they had historically at the disposal at same time frame.

That may have happened had the Deerhound been less troublesome or Armstrong-Siddeley hadn't been advised by the MAP to concentrate on axial flow gas turbines.

[tyrodtom]

Isn't there cooling drag with a air cooled engine ? The more complicated the cooling airflow path is under the cowling, the more cooling drag there'll be.

While the smaller diameter of the engine will mean less drag, it would increase the cooling drag. You've got to get air thru there someway, a engine driven fan, like the Fw190, takes engine HP, or cowling flaps, increase drag.

Reducing the diameter would decrease drag, but it's not a free ride, it introduces it's own problems with drag also.

[tomo pauk]

Indeed, we could use some data re. cooling drag.
Anyway, I'd really love to see a Spitfire with a 2000 HP in small diameter radial in 1943*, or, some early, yet not wide Tempest II/Sea fury.

*a plane somewhere in between Yak-3U and F2G (the Super Corsair)