hobbes154

My understanding is basically every tank had that problem once it was penetrated - might be better to say the later wet stowage Shermans were unusually safe.
 
I can't find a definitive source but see Table VIII here and the discussion here.
 
Edit: also this https://www.tankarchives.ca/2016/03/tank-crew-losses.html

My understanding is basically every tank had that problem once it was penetrated - might be better to say the later wet stowage Shermans were unusually safe.
 
I can't find a definitive source but see Table VIII here and the discussion here.
 
Edit: also this https://www.tankarchives.ca/2016/03/tank-crew-losses.html

There were many more fighter programs than tank programs, many of them producing dogs that never went into service. Of the ones that went into service, most were a disappointment in some way. Of the few that weren't, only one or two were outstanding. This gives you a good idea of the numbers involved: around 240 types used or tested, including foreign types, trainers, utility aircraft etc. Of those, maybe half were used in any great numbers in service. Of that 100-ish aircraft, perhaps two dozen rose above the level of mediocre. And of that two dozen, a handful are considered superlative in their class.
 
Aircraft design is very fiddly, and requires a mix of easily-ascertained factors (power-to-weight ratio, wing loading, armament etc.), hard-to-ascertain factors (top speed, turn times in various configurations, landing speeds) and factors which defied empirical modelling and could only be found by experiment (stability, stall characteristics, maintenance and service niggles, random engine/landing gear/aerodynamic bugs etc).
 
Making a good aircraft in WW2 was as much alchemy as science, and resulted in a lot of dead test pilots. Tanks were actually comparatively easier to design, and accordingly got designed by lesser talents on lower budgets (see, again, the example of British tank building in WW2, which was the product of a bare handful of second-tier engineers). Even today, the best mechanical engineers are mostly doing aviation and aerospace. 
 
 

Inspired by Collimatrix's excellent topic in the Aviation section, my attempt at doing the same for tanks. 

Tank design is often represented as a trade-off between firepower, armour, and speed. But this ignores many other, equally important variables. Furthermore, this formula doesn't explain why the trade-offs are there in the first place. So here is my attempt to make things a bit more complicated.

The Constraint: Compact, powerful and reliable engines – and the ability to use them
Firepower and armour are the obvious features that made a 1940 tank obsolete in 1945. Yet even a 1945 tank is trivial compared to a WWI battleship. Or more modestly, the German 88mm, Soviet 85mm and US 90mm AA guns were prewar designs. Why not put them on a tank from the beginning? Were people just stupid back then?

Up to a point, I would say ‘yes’. While it is unrealistic to think of 1945 combat aircraft, radar, or nuclear weapons in 1940 as the result of anything short of time travel, 1940 tanks could have been significantly better without anachronistic scientific or engineering breakthroughs. Like the assault rifle, this really was a case where the right people just didn't see the need or spend the money.
 
However, it's not the only reason. A tank with a big gun and thick armour that can't move is just a pillbox. Like aircraft, WW2 tanks were fundamentally limited by their engine power. This is less obvious because additional power was typically used to ‘buy’ weight rather than speed. Yet the trend is clear. Shown below are the improvements in engine power for the German and British (cruiser) tank lines, which were in the war the longest.

German Panzers
I
100 hp
II
138 hp
III, IV
250-300 hp    
Tiger, Panther
690 hp

British Cruisers
I, II
150 hp 
III, IV, Crusader
340 hp (Liberty)
Cromwell, Comet
600 hp (Meteor)
 
Arguably, more powerful tank engines could and should have been introduced much earlier (the Liberty was first used in a tank as early as 1918).  I will leave that aside, noting only that, relative to aircraft engines, tank engines were forced to use lower octane fuel for economic reasons (preventing tank use of the Napier Lion), and are harder to cool due to being inside a slow moving armoured box (this was a particular challenge with the Merlin's conversion to the Meteor). 
 
There's also the choice of diesel (compression ignition) vs. petrol/gasoline (spark ignition) engines. Diesels had higher torque and lower fuel consumption, but lower specific power, were heavier and cost more, and meant an extra type of fuel in your logistics train. Both Germany and the US decided against diesels for fear sufficient fuel would not be available.

Once you have an engine, you still need to put power to the wheels through what the British call a ‘transmission’ and Americans call a ‘drivetrain’, which also does the steering since almost all tanks turn by making one track spin faster than the other. And you need a track that won't fall apart and a suspension that will stop the occupants falling apart. These were big problems during WW1 – the first tanks had no suspension at all! – and into the 1920s, but by the 1930s you could more or less use the power of the available engines. Even in 1945, however, the Panther and Comet were deliberately speed limited to around 30 mph.

TRADE-OFFS
OK, we have an engine of a given horsepower, and the ability to turn that horsepower into forward motion off-road with an acceptable degree of unreliability and discomfort. What choices do we have to make now?

1. Weight vs. mobility
This is almost self-explanatory, but mobility is more than speed. Very roughly, multiplying the hp/ton ratio by two gives an approximate top road speed in mph, although looking at individual types this correlation is surprisingly loose. The 15 mph of the British infantry tanks was annoyingly slow, the 25 mph of most German tanks seemed good enough, and as mentioned above, anything over 30 mph arguably wore out the running gear and the occupants to little benefit. A high power-weight ratio was also useful to provide rapid acceleration to dash from cover to cover.
 
But weight has other penalties that are less amenable to increased engine power:
Reliability and maintenance time – pushing around more weight means more parts everywhere from the engine to the suspension will break. (US tracks lasted about 6000km on light tanks but only 2400km on medium tanks. See Exercise Dracula for the effect of maintenance downtime on overall mobility.) Fuel consumption – less range for the individual tank, more for the logistics train to haul. Bridging – if you can’t cross a bridge without breaking it, you may have to go a much longer long way round. Shipping – the M6 heavy tank was not adopted partly because it exceeded the 40-ton limit on many dockyard cranes.
Tanks that were kept in service a long time such as T-34 and Panzers III and IV tended to creep up in weight as bigger guns and frontal armour were added, but for new designs bigger engines and better transmission, steering and track technology (and bridge building) roughly kept pace. Except when they didn't.

2. For a given weight: armour vs. internal volume 
The basic choice is a smaller box with thicker armour or a bigger box with thinner armour. Similarly, sloped armour will give more protection for a given weight, but reduces the internal volume. At the start of the war, most countries tried to armour the front, sides, and even rear to a similar standard, using mostly vertical armour.  As (anti-)tank guns got more powerful, this became impractical, and focus shifted to improving the front armour, both by increasing thickness and sloping (sloping all round reduced the volume of the tank excessively, as in the pyramid shaped early T-34s). 

3a. For a given volume:  guns vs. crew vs. ammo vs. suspension...
Everyone wants a bigger gun, but you need to fit other things in too. For example, a three-man turret crew (commander, gunner, loader) worked better than one or two men, because everyone could focus on one job. But when the British upgunned their Crusader and Valentine tanks from 2-pdr to 6-pdr guns, there was not enough room in the turret for the third man.

Of course, if you run out of ammo, or your crew are bumping into something every time they move, your tank will not fight very well either. (A bigger gun has a double penalty: it reduces the room for other things, including ammo, and also makes each round bigger. The IS-2 looks stunning on paper, but remember that 122mm gun only has 28 rounds and has a slower rate of fire due to its separate loading ammunition.) The Soviets limited the height of their tank crews for this reason. 

If you just want to shoot an enemy soldier or two, the main gun is overkill, so almost all tanks have a ‘coaxial’ machine gun next to it in the turret. Is it worth having a second machine gun in the hull and someone to shoot it? This was nearly universal during the war but fell out of fashion soon afterwards, as bigger guns needed more room for ammunition. (I also imagine most people trying to sneak up on a tank didn't do so from the front.) A few designs even had little secondary MG turrets, a hangover from prewar, but these were quickly abandoned.

Tank suspension is a whole topic of its own. Broadly the choice was between types that allowed more independent wheel movement for a better ride but took up valuable room inside the tank (and were harder to repair in the field) like Christie and torsion bar, and types that gave a worse ride but were completely external (and easier to repair) like VVSS/HVSS, leaf spring and Horstmann. Also, lots of small wheels are better to spread weight evenly and not sink into mud or snow, but fewer bigger wheels are better if you want to drive fast over bumps. The Germans tried to have the best of both worlds with many overlapping large wheels, which was complicated and tended to freeze solid in the Russian winter.

And obviously you need fuel, and a radio or two, a boiler for tea if you're British, compressed air tanks for cold weather starts if you're Russian, and other stuff I haven't mentioned or thought of...

3b. For a given volume: height vs. width
A taller tank is easier for the enemy to spot. A wider tank lets you have a bigger turret ring and therefore a bigger gun. So it seems like a low, wide tank is the ideal. But make your tank too wide and it can't fit on railways – the British were particularly constrained with a narrow railway loading gauge, but even the Germans had to put narrower tracks on their Tigers and Panthers to transport them by rail – or narrow roads and bridges.  Also, height allows more ground clearance to get over obstacles and greater gun depression to shoot at the enemy while hull down.

Finally, height can actually substitute for width to some extent in fitting a bigger gun: a tall hull as in the Sherman allows the turret ring to be extended over the tracks, or a tall turret as in the Challenger (and later Strv 74) allows the gun to recoil (and the crew to squeeze in) above the turret ring. 

Length tends to be the residual in the equation, within limits - too long and you can't steer, too short and the crew gets motion sickness. The Sherman was stretched as required to go from short radial to longer inline engines. Similarly the Challenger was basically a stretched Cromwell.

4. For a given sized turret/gun: AP vs. HE
Tanks sometimes shoot at enemy tanks, but mostly at other, softer things. If you need to punch through armour with AP rounds, you want a high velocity gun (penetration increases with roughly the square of the velocity, but only linearly in calibre). If you want to blow things up, it's all about calibre (HE capacity increases with the cube of calibre, or even a bit more when you consider the minimum size of the fuze and thickness of the shell wall). So for a gun that will fit in a given sized turret, you can have a smaller calibre high velocity hole puncher or a larger calibre, low velocity HE lobber. While you can build specialised guns for each job (and even different tanks to put them in, as the Germans did, which is going a bit far), it's better to have one kind of gun that can do both reasonably well in most of your tanks, since you never know what they will run into.

Conveniently, while tank armour increased throughout the war, common building materials and the human body stayed the same. Therefore, while AP rounds needed ever greater performance, HE didn't. Around 3 inch calibre, with increasing velocity as the war went on, proved a good compromise. The Americans and British both picked the medium velocity 75mm over the high velocity 6-pdr, as did the Soviets with the 76.2mm over their own 57mm AT gun. The higher velocity 76mm, 17-pdr and 77mm then gave the needed AP upgrade while the Soviets went for 85mm (probably because they already had the AA gun rather than any ideal calibre calculation). The Germans also ended up with high velocity 75mm guns on most of their late war tanks (except the 88s on the Tigers, again copied from the AA calibre).

5. For a given budget: quality vs. quantity
Obviously, a bigger tank uses more steel and other resources, and fancy gadgets like better radios, optics and steering systems have a cost. The Tiger was hugely expensive compared to the German mediums (and, more speculatively, other countries' tanks). The Panther was surprisingly cheap for its size, but partly by skimping on the final drive, which crueled its reliability. 

This trade-off applies to distribution as well as production. If you need to move your tanks across the ocean like the Americans, or even by rail across the steppe like the Germans and Soviets, a bigger and better tank at the factory gate meant fewer delivered to the battlefield for the same freight tonnage. So we are back where we started with weight vs. mobility, except in terms of numbers rather than the individual tank's capability.

Inspired by Collimatrix's excellent topic in the Aviation section, my attempt at doing the same for tanks. 

Tank design is often represented as a trade-off between firepower, armour, and speed. But this ignores many other, equally important variables. Furthermore, this formula doesn't explain why the trade-offs are there in the first place. So here is my attempt to make things a bit more complicated.

The Constraint: Compact, powerful and reliable engines – and the ability to use them
Firepower and armour are the obvious features that made a 1940 tank obsolete in 1945. Yet even a 1945 tank is trivial compared to a WWI battleship. Or more modestly, the German 88mm, Soviet 85mm and US 90mm AA guns were prewar designs. Why not put them on a tank from the beginning? Were people just stupid back then?

Up to a point, I would say ‘yes’. While it is unrealistic to think of 1945 combat aircraft, radar, or nuclear weapons in 1940 as the result of anything short of time travel, 1940 tanks could have been significantly better without anachronistic scientific or engineering breakthroughs. Like the assault rifle, this really was a case where the right people just didn't see the need or spend the money.
 
However, it's not the only reason. A tank with a big gun and thick armour that can't move is just a pillbox. Like aircraft, WW2 tanks were fundamentally limited by their engine power. This is less obvious because additional power was typically used to ‘buy’ weight rather than speed. Yet the trend is clear. Shown below are the improvements in engine power for the German and British (cruiser) tank lines, which were in the war the longest.

German Panzers
I
100 hp
II
138 hp
III, IV
250-300 hp    
Tiger, Panther
690 hp

British Cruisers
I, II
150 hp 
III, IV, Crusader
340 hp (Liberty)
Cromwell, Comet
600 hp (Meteor)
 
Arguably, more powerful tank engines could and should have been introduced much earlier (the Liberty was first used in a tank as early as 1918).  I will leave that aside, noting only that, relative to aircraft engines, tank engines were forced to use lower octane fuel for economic reasons (preventing tank use of the Napier Lion), and are harder to cool due to being inside a slow moving armoured box (this was a particular challenge with the Merlin's conversion to the Meteor). 
 
There's also the choice of diesel (compression ignition) vs. petrol/gasoline (spark ignition) engines. Diesels had higher torque and lower fuel consumption, but lower specific power, were heavier and cost more, and meant an extra type of fuel in your logistics train. Both Germany and the US decided against diesels for fear sufficient fuel would not be available.

Once you have an engine, you still need to put power to the wheels through what the British call a ‘transmission’ and Americans call a ‘drivetrain’, which also does the steering since almost all tanks turn by making one track spin faster than the other. And you need a track that won't fall apart and a suspension that will stop the occupants falling apart. These were big problems during WW1 – the first tanks had no suspension at all! – and into the 1920s, but by the 1930s you could more or less use the power of the available engines. Even in 1945, however, the Panther and Comet were deliberately speed limited to around 30 mph.

TRADE-OFFS
OK, we have an engine of a given horsepower, and the ability to turn that horsepower into forward motion off-road with an acceptable degree of unreliability and discomfort. What choices do we have to make now?

1. Weight vs. mobility
This is almost self-explanatory, but mobility is more than speed. Very roughly, multiplying the hp/ton ratio by two gives an approximate top road speed in mph, although looking at individual types this correlation is surprisingly loose. The 15 mph of the British infantry tanks was annoyingly slow, the 25 mph of most German tanks seemed good enough, and as mentioned above, anything over 30 mph arguably wore out the running gear and the occupants to little benefit. A high power-weight ratio was also useful to provide rapid acceleration to dash from cover to cover.
 
But weight has other penalties that are less amenable to increased engine power:
Reliability and maintenance time – pushing around more weight means more parts everywhere from the engine to the suspension will break. (US tracks lasted about 6000km on light tanks but only 2400km on medium tanks. See Exercise Dracula for the effect of maintenance downtime on overall mobility.) Fuel consumption – less range for the individual tank, more for the logistics train to haul. Bridging – if you can’t cross a bridge without breaking it, you may have to go a much longer long way round. Shipping – the M6 heavy tank was not adopted partly because it exceeded the 40-ton limit on many dockyard cranes.
Tanks that were kept in service a long time such as T-34 and Panzers III and IV tended to creep up in weight as bigger guns and frontal armour were added, but for new designs bigger engines and better transmission, steering and track technology (and bridge building) roughly kept pace. Except when they didn't.

2. For a given weight: armour vs. internal volume 
The basic choice is a smaller box with thicker armour or a bigger box with thinner armour. Similarly, sloped armour will give more protection for a given weight, but reduces the internal volume. At the start of the war, most countries tried to armour the front, sides, and even rear to a similar standard, using mostly vertical armour.  As (anti-)tank guns got more powerful, this became impractical, and focus shifted to improving the front armour, both by increasing thickness and sloping (sloping all round reduced the volume of the tank excessively, as in the pyramid shaped early T-34s). 

3a. For a given volume:  guns vs. crew vs. ammo vs. suspension...
Everyone wants a bigger gun, but you need to fit other things in too. For example, a three-man turret crew (commander, gunner, loader) worked better than one or two men, because everyone could focus on one job. But when the British upgunned their Crusader and Valentine tanks from 2-pdr to 6-pdr guns, there was not enough room in the turret for the third man.

Of course, if you run out of ammo, or your crew are bumping into something every time they move, your tank will not fight very well either. (A bigger gun has a double penalty: it reduces the room for other things, including ammo, and also makes each round bigger. The IS-2 looks stunning on paper, but remember that 122mm gun only has 28 rounds and has a slower rate of fire due to its separate loading ammunition.) The Soviets limited the height of their tank crews for this reason. 

If you just want to shoot an enemy soldier or two, the main gun is overkill, so almost all tanks have a ‘coaxial’ machine gun next to it in the turret. Is it worth having a second machine gun in the hull and someone to shoot it? This was nearly universal during the war but fell out of fashion soon afterwards, as bigger guns needed more room for ammunition. (I also imagine most people trying to sneak up on a tank didn't do so from the front.) A few designs even had little secondary MG turrets, a hangover from prewar, but these were quickly abandoned.

Tank suspension is a whole topic of its own. Broadly the choice was between types that allowed more independent wheel movement for a better ride but took up valuable room inside the tank (and were harder to repair in the field) like Christie and torsion bar, and types that gave a worse ride but were completely external (and easier to repair) like VVSS/HVSS, leaf spring and Horstmann. Also, lots of small wheels are better to spread weight evenly and not sink into mud or snow, but fewer bigger wheels are better if you want to drive fast over bumps. The Germans tried to have the best of both worlds with many overlapping large wheels, which was complicated and tended to freeze solid in the Russian winter.

And obviously you need fuel, and a radio or two, a boiler for tea if you're British, compressed air tanks for cold weather starts if you're Russian, and other stuff I haven't mentioned or thought of...

3b. For a given volume: height vs. width
A taller tank is easier for the enemy to spot. A wider tank lets you have a bigger turret ring and therefore a bigger gun. So it seems like a low, wide tank is the ideal. But make your tank too wide and it can't fit on railways – the British were particularly constrained with a narrow railway loading gauge, but even the Germans had to put narrower tracks on their Tigers and Panthers to transport them by rail – or narrow roads and bridges.  Also, height allows more ground clearance to get over obstacles and greater gun depression to shoot at the enemy while hull down.

Finally, height can actually substitute for width to some extent in fitting a bigger gun: a tall hull as in the Sherman allows the turret ring to be extended over the tracks, or a tall turret as in the Challenger (and later Strv 74) allows the gun to recoil (and the crew to squeeze in) above the turret ring. 

Length tends to be the residual in the equation, within limits - too long and you can't steer, too short and the crew gets motion sickness. The Sherman was stretched as required to go from short radial to longer inline engines. Similarly the Challenger was basically a stretched Cromwell.

4. For a given sized turret/gun: AP vs. HE
Tanks sometimes shoot at enemy tanks, but mostly at other, softer things. If you need to punch through armour with AP rounds, you want a high velocity gun (penetration increases with roughly the square of the velocity, but only linearly in calibre). If you want to blow things up, it's all about calibre (HE capacity increases with the cube of calibre, or even a bit more when you consider the minimum size of the fuze and thickness of the shell wall). So for a gun that will fit in a given sized turret, you can have a smaller calibre high velocity hole puncher or a larger calibre, low velocity HE lobber. While you can build specialised guns for each job (and even different tanks to put them in, as the Germans did, which is going a bit far), it's better to have one kind of gun that can do both reasonably well in most of your tanks, since you never know what they will run into.

Conveniently, while tank armour increased throughout the war, common building materials and the human body stayed the same. Therefore, while AP rounds needed ever greater performance, HE didn't. Around 3 inch calibre, with increasing velocity as the war went on, proved a good compromise. The Americans and British both picked the medium velocity 75mm over the high velocity 6-pdr, as did the Soviets with the 76.2mm over their own 57mm AT gun. The higher velocity 76mm, 17-pdr and 77mm then gave the needed AP upgrade while the Soviets went for 85mm (probably because they already had the AA gun rather than any ideal calibre calculation). The Germans also ended up with high velocity 75mm guns on most of their late war tanks (except the 88s on the Tigers, again copied from the AA calibre).

5. For a given budget: quality vs. quantity
Obviously, a bigger tank uses more steel and other resources, and fancy gadgets like better radios, optics and steering systems have a cost. The Tiger was hugely expensive compared to the German mediums (and, more speculatively, other countries' tanks). The Panther was surprisingly cheap for its size, but partly by skimping on the final drive, which crueled its reliability. 

This trade-off applies to distribution as well as production. If you need to move your tanks across the ocean like the Americans, or even by rail across the steppe like the Germans and Soviets, a bigger and better tank at the factory gate meant fewer delivered to the battlefield for the same freight tonnage. So we are back where we started with weight vs. mobility, except in terms of numbers rather than the individual tank's capability.

A few years late, but have to say this was a great OP I wish I'd stumbled on sooner (plus a pretty epic flamewar). Fascinating for those of us who don't work on or aspire to design aircraft but want to understand the choices designers made. Hope it is not too late to add something to the conversation.

1. Some more fighters with outer wing fuel tanks:
(some) Corsairs (some) Spitfires (VIII, XI, XIV-) TA 152H Ki-84 Yak-7 Possibly LaGG-3 & La-9? 2. Some more points worth mentioning:
Airframes that were arguably heavier and stronger than optimal for pure air-to-air or economy of production (most US fighters, Fw 190, Typhoon/Tempest) had an advantage when converted to fighter-bombers, since you could hang more bombs on them without breaking. (Of course if you were optimising for load carrying you would also make the wings bigger for more lift, with a drag penalty, which is one reason bombers and transports tend to be slower than fighters even unloaded.) The distinction between maximum speed and initial acceleration is important in a dive as well as level flight. E.g. the Spitfire had AFAIK the best critical Mach number of any WW2 prop plane, but didn't pick up speed very quickly in a dive due to its lightness. Re the earlier civilised discussion on the P-38, the following critical Mach numbers from Wings on My Sleeve by Eric Brown may be of interest:
P-38 0.68
P-47 0.71
Me 109 0.75
Fw 190 0.75
P-51 0.78
Spitfire 0.83 (?) If you want to use your plane on a carrier, you need extra weight (strength for basically controlled crash landings), bigger wings (more lift for short takeoff runs & slower landing/stall speeds), and better visibility illustrated e.g. in the "hump" on the Sea Fury at the cost of some drag (similar to the move to from razorback to bubble canopies in general). The poor forward visibility mentioned on the Fw 190 would have been a problem. 3. And finally push back a little on a few statements:
This isn't obvious from the data: leaving aside the Lightning which is granted as an exception, the Me 110 was competitive in top speed (it was the manouverability that killed it as a day fighter), the Whirlwind was as fast as the Spitfire with less powerful engines and the Mosquito and DH Hornet were faster with the the same power. Same for the Tigercat vs the Bearcat. Geometrically, too, the pilot can't hide completely behind the engine in a single engine fighter that needs a minimum level of visibility.
Theoretically yes but practically didn't all a/c (not just US ones) develop excessive control forces, hit critical Mach, or simply break apart before that point? AKA "why dive bombers needed dive brakes".
Again, not obvious as a major problem: aircraft with good power-weight and moderate wing loadings (no I don't have E-M charts for them all!) like the Spitfire, Zero, Oscar, early 109s were pretty good at both, even if there is a marginal tradeoff. Even the J2M given as an example of a fighter completely optimised for climb was not that bad a turn fighter: "the Raiden, despite not being designed for maneuverability, still had a lower stall speed than the Hellcat, and could turn tighter." There is a much more obvious tradeoff between climb and turn performance on the one hand, and dive on the other, based on weight, seen most notably in Japanese or really anyone else vs most US fighters.