# Scaling an engine - performance

22 replies to this topic

### #1 Wuzak

Wuzak
• Member

• 4,539 posts
• Joined: September 00

Posted 31 May 2012 - 13:27

Let's say you were to make a scale version of an engine how well would it perform?

I was going to say scale an F1 engine, but as they already have stratospheric rpms maybe it isn't the best example.

Let's say we have a 400hp 4l V8. We want to make a half scale model - that is the bore and stroke would be half of the original, the physical dimesnions are half size.

The half scale model would have 1/8th the capacity of the original, but 1/4 the piston area.

Would the power be the same, or similar, to the hp/l of the original, or would it match the power per unit piston area (ie the same BMEP with twice the revs)?

Assuming you use the same materials, the scale version ought to be approximately 1/8th the weight of the original. Are there any areas where direct scaling would not provide the strength required, needing teh scale engine to have more material in certain areas?

### #2 Paolo

Paolo
• Member

• 1,673 posts
• Joined: May 00

Posted 31 May 2012 - 19:00

Let's say you were to make a scale version of an engine how well would it perform?

I was going to say scale an F1 engine, but as they already have stratospheric rpms maybe it isn't the best example.

Let's say we have a 400hp 4l V8. We want to make a half scale model - that is the bore and stroke would be half of the original, the physical dimesnions are half size.

The half scale model would have 1/8th the capacity of the original, but 1/4 the piston area.

Would the power be the same, or similar, to the hp/l of the original, or would it match the power per unit piston area (ie the same BMEP with twice the revs)?

Assuming you use the same materials, the scale version ought to be approximately 1/8th the weight of the original. Are there any areas where direct scaling would not provide the strength required, needing teh scale engine to have more material in certain areas?

Engine power is roughly proportional to piston area, for a given level of component stress.

### #3 Wolf

Wolf
• Member

• 7,881 posts
• Joined: June 00

Posted 31 May 2012 - 20:46

I think the maximum revs would change as well- remember how late sixties Honda RC116 (four-stroke, I think 2 cylinder, 50cc engine) had redline at 22,500rpm... That points out that valve springs are (as logic would imply) 'not exactly' scalaeble- smaller valves will require smaller/lighter springs and lighter springs will have higher resonant frequency, allowing the engine to rev higher... (That's just my 2c worth.)

### #4 saudoso

saudoso
• Member

• 5,404 posts
• Joined: March 04

Posted 31 May 2012 - 21:27

I guess this is a bit overexposed by now, but worth having a look under the circumstances:

Beautifully scalled from the sound of it.

### #5 Magoo

Magoo
• Member

• 2,701 posts
• Joined: October 10

Posted 31 May 2012 - 21:52

Let's say we have a 400hp 4l V8. We want to make a half scale model - that is the bore and stroke would be half of the original, the physical dimesnions are half size.

The half scale model would have 1/8th the capacity of the original, but 1/4 the piston area.

Would the power be the same, or similar, to the hp/l of the original, or would it match the power per unit piston area (ie the same BMEP with twice the revs)?

Assuming you use the same materials, the scale version ought to be approximately 1/8th the weight of the original. Are there any areas where direct scaling would not provide the strength required, needing teh scale engine to have more material in certain areas?

You probably noticed this, but you just built a 500cc Moto Guzzi V8.

### #6 MatsNorway

MatsNorway
• Member

• 2,239 posts
• Joined: December 09

Posted 01 June 2012 - 09:59

Not related to directly scaled down engines:

Smaller engines tends to make more hp pr liter.

They tend to handle and like rpms more.

Claim:
efficiency goes down compared to their hp output.

Reason: Due to the low volumes of air you can`t utilise the dynamics air got as good.

Edited by MatsNorway, 01 June 2012 - 14:37.

### #7 Wuzak

Wuzak
• Member

• 4,539 posts
• Joined: September 00

Posted 01 June 2012 - 11:23

I ask because on another forum someone suggested that P&W could have made a half size R-4360 for a lightweight fighter and have it ready before WW2 (ie late 1939).

The R-4360 was a 3000hp engine in 1944/45 from its 4360 cubic inches. Making a half scale version would give 545 cubic inches and, using the piston area, 750hp. His idea was for a 77mm bore (why I don't know) down from 5.75" - for a scale factor of 0.527, giving a total of 639 cubic inches and 834hp based on piston area.

The R-4360 is air cooled, so theoretically the surface area of the finning should go down by the square of the scale, which would keep it in proportion with the power. However, the ability of the manufacturers to make greater fin area improved throughout the war, and the cooling capacity of a 1944 model R-2800, or the R-4360, was far better than the R-2800 of 1939.

For this power to be made the mini R-4360 would need to turn at 5300rpm. Surely that would be difficult for a radial engine at the time, even if it was so small?

### #8 saudoso

saudoso
• Member

• 5,404 posts
• Joined: March 04

Posted 01 June 2012 - 12:11

The RPM increase is easily explained by the square/cube phenomenon. When you scale down the weights go down by the cube, the cross sections go down by the squares. So you have a scaled down version that is in fact stronger built.

### #9 Kelpiecross

Kelpiecross
• Member

• 1,043 posts
• Joined: October 10

Posted 01 June 2012 - 14:10

I ask because on another forum someone suggested that P&W could have made a half size R-4360 for a lightweight fighter and have it ready before WW2 (ie late 1939).

The R-4360 was a 3000hp engine in 1944/45 from its 4360 cubic inches. Making a half scale version would give 545 cubic inches and, using the piston area, 750hp. His idea was for a 77mm bore (why I don't know) down from 5.75" - for a scale factor of 0.527, giving a total of 639 cubic inches and 834hp based on piston area.

The R-4360 is air cooled, so theoretically the surface area of the finning should go down by the square of the scale, which would keep it in proportion with the power. However, the ability of the manufacturers to make greater fin area improved throughout the war, and the cooling capacity of a 1944 model R-2800, or the R-4360, was far better than the R-2800 of 1939.

For this power to be made the mini R-4360 would need to turn at 5300rpm. Surely that would be difficult for a radial engine at the time, even if it was so small?

I would have thought by "half-size" they meant 2180 cubic inches not physically "half-size".

### #10 Magoo

Magoo
• Member

• 2,701 posts
• Joined: October 10

Posted 01 June 2012 - 21:18

I ask because on another forum someone suggested that P&W could have made a half size R-4360 for a lightweight fighter and have it ready before WW2 (ie late 1939).

Because there was such a desperate need among the Allies for fighter aircraft with limited range, poor armor, and inadequate armament.

### #11 Lee Nicolle

Lee Nicolle
• Member

• 7,481 posts
• Joined: July 08

Posted 03 June 2012 - 08:13

Manufacturers have done it regularly for decades in useable parameters. Scaled down V8s into V6 from Buick, Chevy, Ford. Nissan did scaled up withj the L series engine and Holden did scaled down with the blue 6 to the Starfire.
Seldom particularly succesful but useable.
But to do a small scale engine of the original will run but probably be little more than a toy.
Scale down a V8 Chev, it has been done a a model and it runs, and even sounds the part but I doubt remotely near as efficiently as a say 300cc purpose built engine. And they are a simple engine.
To try and do a DFV would probably be impossible yet alone a modern F1 engine of say 300cc, yet alone 30. The originals already are very minute and finicky to start with.

### #12 mariner

mariner
• Member

• 1,542 posts
• Joined: January 07

Posted 03 June 2012 - 08:40

BRM V - 16 1.5 litres and 16 cylinders ??

One scaling issue is the ignition process , making tiny plugs with enough air gap is probably very hard . I don,t know what the famous Ferrari 312 guy did but very small plane IC engines go "diesel" , i.e Nitro based fuel and a glow plug only.

### #13 Wuzak

Wuzak
• Member

• 4,539 posts
• Joined: September 00

Posted 03 June 2012 - 09:06

Manufacturers have done it regularly for decades in useable parameters. Scaled down V8s into V6 from Buick, Chevy, Ford. Nissan did scaled up withj the L series engine and Holden did scaled down with the blue 6 to the Starfire.

But not scaled 50%.

As for the Starfire, starting with a 3.3l 6 and ending up with a 2l 4 isn't a lot of scaling.

But to do a small scale engine of the original will run but probably be little more than a toy.
Scale down a V8 Chev, it has been done a a model and it runs, and even sounds the part but I doubt remotely near as efficiently as a say 300cc purpose built engine. And they are a simple engine.

However, scaling a 4000+ cubic inch down to a 500 cubic engine shouldn't make it too toy like.

I was wondering about cooling. For a liquid cooled engine the volume in the cooling passages will be 1/8th, but if the power scales with the piston area the power will be 1/4.

### #14 Tony Matthews

Tony Matthews
• Member

• 17,503 posts
• Joined: September 08

Posted 03 June 2012 - 11:22

To try and do a DFV would probably be impossible yet alone a modern F1 engine of say 300cc, yet alone 30. The originals already are very minute and finicky to start with.

A friend of mine is making a 1/4 scale Ferrari V10, complete with fuel injection and pneumatic valves...

Edited by Tony Matthews, 03 June 2012 - 11:26.

### #15 Bloggsworth

Bloggsworth
• Member

• 7,681 posts
• Joined: April 07

Posted 03 June 2012 - 14:03

Windage. You can't scale the clearances. Revs would be no problem at all, the rotating forces would reduce, so 1/4 scale would be about 1/64th the weight, so the rotating forces should be no problem, small glo-plug engines can run at 40,000RPM with no stress at all - granted, scale con-rods would be a lot heavier than those in small CI engines, and that might be the limiting factor - All that said, you can't scale the bearing and bore clearances, so the engines may rattle. I have now finished talking twaddle, so I'll go...

### #16 TDIMeister

TDIMeister
• Member

• 265 posts
• Joined: January 06

Posted 04 June 2012 - 20:46

IMEP and mean piston speed will remain fairly consistent if other factors remain fixed; scale for power and RPM using the downsizing factor from there.

Edited by TDIMeister, 04 June 2012 - 20:46.

### #17 Greg Locock

Greg Locock
• Member

• 4,907 posts
• Joined: March 03

Posted 04 June 2012 - 23:17

The cynic in me says that if this were a good way to design an engine then people would already do it. By and large they don't, because different things sclae differently with changes in scale, so what is optimal at one size is less so at another.

### #18 bigleagueslider

bigleagueslider
• Member

• 1,117 posts
• Joined: March 11

Posted 05 June 2012 - 01:53

The RPM increase is easily explained by the square/cube phenomenon. When you scale down the weights go down by the cube, the cross sections go down by the squares. So you have a scaled down version that is in fact stronger built.

When scaling a recip engine linearly by bore and stroke dimensions, the square/cube relationship affects more than weights. The things most affected are heat transfer, dynamic loads and friction losses. Just consider the change in cylinder surface area/volume ratio when reducing the bore and stroke dimensions by half.

### #19 Kelpiecross

Kelpiecross
• Member

• 1,043 posts
• Joined: October 10

Posted 05 June 2012 - 04:49

However, scaling a 4000+ cubic inch down to a 500 cubic engine shouldn't make it too toy like.

I was wondering about cooling. For a liquid cooled engine the volume in the cooling passages will be 1/8th, but if the power scales with the piston area the power will be 1/4.

I can't see why anbody in their right mind would consider building a 500 or 600 cu engine based on the P and W 4360 - a four-row radial of 28 cylinders and pushrod as well? A V-8 or V-12 would be more sensible.

And what possible use could it have even in a "lightweight" fighter? As I think Magoo was alluding to, even a small fighter will need six machine guns, big fuel tanks, armour plate etc. and will not weigh less than 6000lbs or so - a 600cu engine is just too small.

Wiki tells me that P and W did build a "half-size" 4360 - the R-2180-E - a 14 cylinder twin-row engine.

### #20 bigleagueslider

bigleagueslider
• Member

• 1,117 posts
• Joined: March 11

Posted 05 June 2012 - 05:25

I can't see why anbody in their right mind would consider building a 500 or 600 cu engine based on the P and W 4360 - a four-row radial of 28 cylinders and pushrod as well? A V-8 or V-12 would be more sensible.

And what possible use could it have even in a "lightweight" fighter? As I think Magoo was alluding to, even a small fighter will need six machine guns, big fuel tanks, armour plate etc. and will not weigh less than 6000lbs or so - a 600cu engine is just too small.

Wiki tells me that P and W did build a "half-size" 4360 - the R-2180-E - a 14 cylinder twin-row engine.

The whole principle behind the Pratt & Whitney R-4360 engine was scaling. While other engine manufacturers were taking the approach of using lots of small displacement cylinders operating at high rpm's, with liquid cooling, overhead cams, and 4 valves per cylinder in order to maximize power-to-weight, Pratt & Whitney took the opposite approach. Pratt designed the R-4360 with massive air-cooled cylinders, using 2 valves per cylinder with pushrods/rocker arms, turning at low rpm's, and using lots of boost. The result being that the R-4360 achieved fantastic BSFC rates, great reliability, and excellent power-to-weight.

As for WWII fighter engines, the air cooled, 2-valve pushrod Pratt & Whitney R-2800 was arguably one of the best piston engines produced. It was extremely reliable, durable, cost effective, powerful, fuel efficient, and built in far larger numbers than almost any other engine. The R-2800 was designed using the same "low tech" principles as the R-4360. While they weren't sexy, the big Pratt & Whitney air cooled radial designs definitely made sense.

slider

### #21 Wuzak

Wuzak
• Member

• 4,539 posts
• Joined: September 00

Posted 05 June 2012 - 07:28

The whole principle behind the Pratt & Whitney R-4360 engine was scaling. While other engine manufacturers were taking the approach of using lots of small displacement cylinders operating at high rpm's, with liquid cooling, overhead cams, and 4 valves per cylinder in order to maximize power-to-weight, Pratt & Whitney took the opposite approach. Pratt designed the R-4360 with massive air-cooled cylinders, using 2 valves per cylinder with pushrods/rocker arms, turning at low rpm's, and using lots of boost. The result being that the R-4360 achieved fantastic BSFC rates, great reliability, and excellent power-to-weight.

As for WWII fighter engines, the air cooled, 2-valve pushrod Pratt & Whitney R-2800 was arguably one of the best piston engines produced. It was extremely reliable, durable, cost effective, powerful, fuel efficient, and built in far larger numbers than almost any other engine. The R-2800 was designed using the same "low tech" principles as the R-4360. While they weren't sexy, the big Pratt & Whitney air cooled radial designs definitely made sense.

slider

The P&W R-4360 was basically 28 R-2800 cylinder barrels put together in one engine. The prototype even used the then current R-2800 parts. But some things had to change - like the orientation of the push rods.

### #22 Wuzak

Wuzak
• Member

• 4,539 posts
• Joined: September 00

Posted 05 June 2012 - 08:22

Personally, I don't see adding or subtracting the same size cylinders as scaling.

That is, a 3.0l V8 may be three quarters of a 4.0l V8, but it isn't a three quarter scale model of a V8.

There is an example of scaling with Rolls-Royce. In the mid 1920s they built the F (Kestrel) in response to the Curtiss D12. It had a 5" bore x 5.5" stroke.

Deciding there was a market for a larger, more powerful engine, for transports, bombers and the like, Rolls-Royce built the H (Buzzard) as a 6/5 scale version of the F. It had 6" bore x 6.6" stroke. Some details had to be changed for strength reasons.

The Buzzard would be modified to become the R.

### #23 bigleagueslider

bigleagueslider
• Member

• 1,117 posts
• Joined: March 11

Posted 08 June 2012 - 23:56

Personally, I don't see adding or subtracting the same size cylinders as scaling.

Wuzak,

I would agree. Simply adding or subtracting similar sized cylinders scales the number of cylinders, but it doesn't scale the number of other components such as oil/fuel/coolant pumps, cam drives, filters, etc.

The type of combustion cycle used can also have a big impact on how well the engine responds to scaling (up or down) of cylinder dimensions. SI engines seem to perform better at small to medium bore diameters. The practical upper limit for SI engines seems to be about 6 inches (or 152 mm). On the other hand, CI engines seem to perform better at medium to large bore diameters. With bore diameters below about 3 inches (or 76 mm) CIDI engines can be difficult to start and tend to have reduced efficiency. But CIDI engines with bore diameters as large as 36 inches or more work very well and are super efficient.