Three row radials

[swampyankee]

Why were there no three row radials?

There were one, two, and four row radials. Why not three?

Obviously, nobody thought they were worthwhile, but why not?

 

[wuzak]

Why were there no three row radials?

There were one, two, and four row radials. Why not three?

Obviously, nobody thought they were worthwhile, but why not?

The Armstrong Siddeley Deerhound was a 3 row radial, with the cylinders lined up in rows. It wasn't a success - it lacked power and was redesigned twice in its short life span.

 

[Shortround6]

It doesn't really offer a lot. For a practical standpoint you have a choice of 15 cylinders (3 rows of 5), 21 cylinders (3 rows of 7) or 27 cylinders ( 3 rows of 9).
For the first adding an extra crank throw and the longer heavier crankcase and the more convoluted intake and exhaust systems needed for one extra cylinder rules it out pretty quick.
For the 2nd it gets a bit closer, 21 cylinders in 3 rows vs 18 in two. It is a 16.6% increase in displacement. Is that enough to pay for the longer crank and crankcase the more troublesome intake and exhaust?
Please note the Deerhound was in serious trouble in an open market. It started as a 2260 cu in engine with 7 overhead camshafts. It may have been in competition with the 2360 cu in Hercules and 2360 cu in Gnome-Rhone engines. With it's short stroke of 5in (127mm) it had the potential to run at higher rpm. However it would also be in competition with the Wright 14 cylinder R-2600.
The lined up cylinders present a bigger challenge to air cooling than staggered cylinders. You basically need the airflow to do two 90 degree turns, while the staggered cylinders require only gentile turns in the airflow to reach the cylinders and exit.

"A Deerhound II engine installed in an Armstrong Whitworth A.W.38 Whitley bomber. Note the relatively small diameter of the engine compared to that of the firewall. The 44 in (1.12 m) diameter Deerhound replaced the 51 in (1.29m) diameter Armstrong Siddeley Tiger that was originally installed in the Whitley II."

from "Armstrong Siddeley 'Dog' Aircraft Engines"

However please note the area/volume of the overhead cam boxes and the area/volume NOT being used for fins/airflow to cool the cylinder heads (hottest part of the engine). upper and lower visible banks have cam covers removed.

Gnome Rhone engine

Note all push rods for valves come from 1 or 2 cam rings at the front of the engine. Trying to use pushrods from the front on a 3 rd row would be very difficult and adding cam ring/s at the back for a 3rd row increases cost/weight.

The three row arrangement is not as easy at it might appear.

 

[Jugman]

A two-row radial had a 25-33% labor cost and roughly a 10% weight gain over a single-row of similar displacement. P&W consider the smallest practical two-row air-cooled engine to be the same size as the largest practical single-row engine. A three or more row radial pretty much suffers the same problem in relation to two-row engines.

Other issues are:

1. Proper engine balance.
   
2. Engines that have cylinder counts per row that are not divisible by three require in line cylinders or an offset crankshaft for staggered rows. It's the opposite for cylinder counts that are divisible by three.
3. The points Shortround brought up.

 

[pbehn]

A two-row radial had a 25-33% labor cost and roughly a 10% weight gain over a single-row of similar displacement. P&W consider the smallest practical two-row air-cooled engine to be the same size as the largest practical single-row engine. A three or more row radial pretty much suffers the same problem in relation to two-row engines.

Other issues are:

1. Proper engine balance.
   
2. Engines that have cylinder counts per row that are not divisible by three require in line cylinders or an offset crankshaft for staggered rows. It's the opposite for cylinder counts that are divisible by three.
3. The points Shortround brought up.

In addition to the above, the four row radials only gained 10 cylinders more than a two row, based on 9 x 2 =18 to 7 x 4 =28. For any single engine use they were useless, to big too heavy and needing a massive fuel capacity. The "market" for these engines was mainly a stop gap between the two row radials and the turbo prop generation.

 

[wuzak]

Is that enough to pay for the longer crank and crankcase the more troublesome intake and exhaust?

I should think that the intake and exhaust were simplified over a staggered row 3 or 4 row radial.

The lined up cylinders present a bigger challenge to air cooling than staggered cylinders. You basically need the airflow to do two 90 degree turns, while the staggered cylinders require only gentile turns in the airflow to reach the cylinders and exit.

However please note the area/volume of the overhead cam boxes and the area/volume NOT being used for fins/airflow to cool the cylinder heads (hottest part of the engine). upper and lower visible banks have cam covers removed.

Note also that Armstrong-Siddeley may have desired the Deerhound to be liquid-cooled, but were asked by the Air Ministry to make it air cooled to compete with Bristol radials.

 

[kmccutcheon]

Below is a link to an article about Armstrong Siddeley three-row radials. Includes a good bibliography.
Armstrong Siddeley 'Dog' Aircraft Engines

 

[Shortround6]

I should think that the intake and exhaust were simplified over a staggered row 3 or 4 row radial.

Yes and no.
Using log manifolds is simpler as far as routing or fitting things in goes. But log manifolds are not as good at getting even mixture distribution.
Same with exhaust, can you get the 3 cylinders in one bank to fire at 240 degrees to each other (or close to it?)
Of course compared to the Bristol exhaust systems most anything was an improvement.

I am also not sure what running the intake and exhaust manifolds within a few inches of each other does.
None, (or few) of the pictures seem to show the exhausts in place and perhaps the baffling separated them when fully assembled.

as in this picture. Intakes are under the baffle.

 

[AMCKen]

Armstrong-Siddeley Boarhound and the Soviet (Gnome-Rhone) M-95 were 27 cylinder engines. Rumour has it no Boarhounds were actually built.
A-S also worked on the Hyena which was a 15 cylinder engine. AFAIK

 

[swampyankee]

Below is a link to an article about Armstrong Siddeley three-row radials. Includes a good bibliography.
Armstrong Siddeley 'Dog' Aircraft Engines

I am aware, now, about the Armstrong Siddeley "dog" engines.

Three row radials certainly seem to have been tried, but never produced.