Modelling Torque Discussion

From reading the tank files in datamine, the engine’s energy output, minimum and maximum RPM are specified, but there are no individual elements to set the torque value, torque curve, or engine type.

Therefore, it is likely that the torque value corresponding to the RPM value is specified by some constant common to all vehicles, and maximum torque is probably generated at high RPMs even if the engine mechanism is different.

When observing the behavior in manual transmission mode, the amount of acceleration is different at low and high RPMs, so it seems that there is probably a torque curve.

Also, some wheeled vehicles are equipped with a system that causes a delay between the displayed RPM and the actual RPM, in which case the vehicle speed and displayed RPM will not be synchronized.

Indeed, either power is constant, or it linearly decreases with speed, both aren’t correct to any engine other than an electric motor.

To add some color to this thread, I will write down the information I know about torque converters.

This text has been machine translated into English, and I’m not very good at writing, so it’s in bullet points. Please feel free to criticize or point out any mistakes.

A typical engine produces the highest output (horsepower) at high revolutions.
However, if the engine is mechanically connected to the axle, the engine speed has to depend on the vehicle speed, and ideal horsepower can only be achieved in a few situations.
This is where the torque converter comes in. By using this, the engine speed and axle speed are freed from the dependency, and horsepower can be freely adjusted over a wide range.

With a simple fluid coupling that is not a torque converter, if there is a difference between the engine speed and the axle speed, the horsepower that can be delivered is reduced. The ideal horsepower can only be delivered when the axle speed begins to synchronize with the engine speed.
In contrast, when the revolution speed is out of sync with the torque converter, the torque converter converts the energy of the revolution speed into “torque” and transports it. This is “torque amplification.”
This allows the engine’s horsepower to be transmitted to the axle without loss at all speeds.

It is important to note that the total amount of energy output by the torque converter does not exceed the engine’s output energy. The torque amplified by the torque converter is the energy of rotation speed in a different form.

Therefore, the torque converter can be seen as a kind of “infinitely variable reduction gear”.

The torque converter realizes a wide ratio transmission. Originally, a transmission with a wide gear ratio interval has a strong drop in engine speed after shifting and has uneven horsepower, making it difficult to use in vehicles that are always underpowered, such as tanks.

If a torque converter is inserted here, it prevents the horsepower drop immediately after shifting. This is because the engine can always maintain a high rotation range.

If a wide ratio can be achieved, the theoretical maximum speed can be increased even with a small number of gears. Transmissions with a small number of gears are generally small, which is advantageous. Many modern NATO MBTs have this design.

However, torque converters also have their weaknesses. When a torque converter amplifies torque, losses always occur due to fluid resistance. In other words, a few percent of the engine’s horsepower is lost before it is transmitted to the axle.

The horsepower that could be increased by always having the engine in the high RPM range is also partially lost in this loss. If the design is incorrect, it will become a completely inefficient system.

For this reason, modern torque converters are equipped with a “lock-up mechanism” that directly connects the output shaft and the input shaft and fixes them in situations where amplifying the torque is not very meaningful.

Torque converters have another useful function. They can be installed as a substitute for the clutch between the transmission and the engine. This is very useful for automating the transmission.

Among the vehicles implemented in WT, there are surprisingly many vehicles that are equipped with torque converters in the real world.

The M18 self-propelled gun also uses a torque converter to achieve a wide ratio transmission. The transmission is supposed to be a three-speed, but because the game is unable to replicate a torque converter, it is treated as a six-speed, which is exactly what Fireball_2020 has pointed out as a problem with the game.

3 Likes

Was a full suggestion posted for this?

There is a Suggestion for CVT;

2 Likes

Yes it was:

1 Like

Thank you both. Fingers crossed it’s implemented, would make a huge difference in game.

2 Likes

Ever day I pray for my CVT suggestion to get added into WT
:D

2 Likes

This, CVT and regenerative steering might make the CH2 less ass to drive

2 Likes

I think the only tank that would benefit from a CVT (HMT) change would be the type 10. But any changes to increase mobility are welcome changes.

I would support better torque implementation as right now a lot of US tanks that use the 810HP gasoline engine are better than the ones using the 750HP diesel engine designed to replace it.

2 Likes

In the current version of Warthunder, power losses caused by friction and fluid resistance are not taken into account at all, although this is intentional. As a result, the catalog spec value of the engine output is transmitted directly to the drive wheels.

This causes problems when modeling torque converters and CVTs. Although these devices are convenient, their biggest drawback is that they cause large power losses. If these devices are implemented in the game without reproducing the power losses, the performance will be far removed from reality.

The power losses are broadly irrelevant, what’s useful is the torque multiplication effect, since the converter is locked up when not in use, its loss is broadly 0%.

It is true that lockup mechanisms reduce power loss, but they are a relatively new idea and are not typically found in torque converters used in WWII vehicles.

Kinda why torque converters aren’t as present on older vehicles.

Ground vehicles already simulate powerbands. It’s very clearly obvious on things like the stock M41. However, it does absolutely need an overhaul since power is not applied to the wheels, but rather a thrust effect that pushes the entire vehicle along like a plane. This is why tanks sort of just glide forwards like butter on a hot pan, instead of pitching and rocking like irl. The most extreme and visible example would be the Ha-Go. In addition, transmissions shift instantly and without any loss in power, which makes many tanks a lot faster than they were in real life. If you accelerate forwards and see you went straight from 1st to 8th gear, then that’s a problem lol. There are a few exceptions, such as American tanks which used hydraulic transmissions, and modern ones, but overall the system is incredibly outdated and desperately needs a rework

This is a great idea. War thunder needs to decide if it’s a simulator or a game with cool ballistic mechanics.
Right now it’s not even close to a simulator at least on the ground side.
There are so many of these “hidden statistics” that dumb the game down. That would effect actual performance so much. It’s way more important than ray tracing. WT does such a great job with airplane models. But with ground as cool as war thunder is, it’s a very incomplete.

I’ve been asking for torque curves, BHN and material hardness and quality calculations, firecontrol pulse update times, thermal flir power and a huge huge thing, ping times and fcs throuput numbers. Generally a real life equipment quality check.

As cool as torque is, they shouldn’t use any engine statistics for tracked vehicle. There’s just too big of a gap in performance waste between the varied approaches.
How many drive lines? How many gears? How efficient is the torque converter not just It’s a ratio. Does it have limited slip? Does it have traction control? Does the computer slow one track up when it detects track speed alignment issues?
Also, temperature play a massive roll. In cold air, the engine that has the most boost will create more power.

DU acts the same as tungsten ingame for god sakes, and very similar to steel in game. Which is insane. DU self sharpens, reacts and pre-melts armor while the molecules hold onto to each other and keep their shape 15% better. Darts are all about length, material and velocity vs what materials and what order at what angle they are going through.

Stabilizers are all treated the same, which is hilarious. Gyros and ballistic computers are not made equal.
The US for a long time and now Germany and France use independent co-processors each with their own designated memory. So the laser rangefinder., the barrel bend sensor, the windage and elevation and tracking and GPS and thermals are not running through the same processor… leopard 2a4 down and every t80 72 ingame used a single processor to deal with all of those inputs…
For God sake’s our gaming routers and smart tvs nowadays use tons of co prosessors because shading and lighting and the smart function is too much.
Now imagine you’re tracking a moving target and eighth the size of your pinky nail, firing at 2000 yards with two changes of wind, from a moving platform and need to hit it with a dart… do you see why 500 ms lag spikes could be a problem?
And when do spikes occur? Oh that’s right, when you’re actually sighting and ranging your target. Essentially the other fire controls have to do one job at a time didn’t pass the data to the other. Us and now nato do it at once.

Thermal scopes are pretty much treated the same and don’t use actual heat levels. It’s just a fancy filter and they get a certain resolution. What generation one thermals are not created equally. The Hughes has 18 top quality thermal cameras all going to a processor.
Many other gen ones are going to 4 and we’re piped in to the fire control.

Without these mechanics actually flushed out, countries that lie or overstate can claim whatever they want and it doesn’t match reality at all. Then it’s not a simulator. It’s a fun tank game. Gaijin can use its bias to prop up their favorite ew premium or country with these vague systems in plac. It’s actually a major problem.
When it becomes a real issue is when they artificially nerf or buff certain tech and then the country, knowing it was an issue actually changed the design for the better or it was such a good technology multiple vehicle vehicles then got it.

When they do things like over buff a single dart for their new premium tank because it’s in between tiers or it wasn’t that great in real life, or they don’t have the subsystems and micro gameplay to make it good in game. Then the over buffed dart comes out on the next tank in which the country added a ton of resources to up its armor or its reload to make up for it.
Now you have an overpowered or underpowered piece of technology that gets worse and worse and worse.

When a steel, short Dart that is machined 3 times looser that has been RL tested using NATO standards comes in 30% different than WT numbers… that’s an issue.

It’s already bad enough when they remove things that they said was a test round like the 40mm on the m247
Making it go from 130/90 pen to 36. Meanwhile, the SU 57 that never carried APHE in rl is better at killing than anything. Or when the 45 mm Russian APHE payload gets buffed in tanks, which fixes a small problem in a mainly unused tank. But it creates planes that carry these that now one click anything from the sky.

If it’s a simulator, please show me where the Russians were getting thousands of heavy kills with 45 mm. If real life was anywhere close to war thunder, the US and Britain and France would’ve all bought yak 9s and not run any other plane. It could kill fighters and bombers and King Tigers while being fast and completely stable.

When they simply accept RHA estimates, (vs actually modeling the layers) add weight and velocity, that leaves out about 1000 different factors. It allows manufacturers or countries to not invest in tons of real life, quality control, testing, and pad their numbers as much as possible.

The perfect example of the problems ingame is Abrams, DU and composite in General.

Gaijin did dmg control during the leaks, and they were very clever. Rather than addressing the fact that 99% of sources on earth agreed the m1 had 350-400mm lower hull protection which was increased to 550 to 600 with m1a1. They instead focused on a small part of claiming DU was in the lower hull and proved it wasn’t.

The reason they fought back so hard and won’t just make it the listed stats, it’s because they’re NERA in X-ray
Has a single value. US, German, French Brit’s all just say Nera and thickness and it’s all the same. On the Russian tanks, they add plates and smaller Nera so they are much stronger.

The point is, the US has never sold its composite armor scheme. They do not use the same materials as other countries. Most the other countries simply license Chobham armor and buy composite screens.
But Chobham is simply the sandwich of whatever ceramic, rubber and fiberglass you use and the order to put them in.

The US used simple high bhn armor with chobham blocks on the xm1. But the Abrams and especially the m1a1 used US high hardened steel with vulcanized, rubber and fiberglassdipped in 1 inch ceramic coating for both the front of the sandwich that was 4 inches thick and the back plate which was 2 4 inch plates back to back. All the plates are opposed to each other.
Now the nera block come in units and they do look like what’s modeled in game but each 4 inch block has a 3.8mm plate.

This is not what the other NATO countries do. They use old school Chobham design with various thickness and materials. They don’t do this because it’s much more expensive to do ceramic coatings, high hardened steel and opposed plate positions.

I tested weapons for six years. All kinds of composite armors in different variations. Simply switching your hardest material from the back to the front (if the armor is comparable) will defeat 15% more often. If you take 5 inches of RHA vs 1 inch face hardened 3inches in 3 plates each facing 15 degrees different, Putting fiberglass, vulcanizedrubber and tile between each layer you save 15% weight and defeat 25% more, have 0 spall and is basically immune to HEAT up to a reasonable point.

RHA with a 600 bhn at 60 degrees acts completely differently than its equivalent 1200 bhn FHA at 60.
The steeper the angle the higher, the BHN the more drastic.
There was a test with 25 mm of armor at 65 degrees. The cast steel at 450 bhn was penned 100%. It’s soft so it catches the round.
They changed it to 1200 BHN and it deflected 85% and only took slight damage 15%.
They went to 18 mm and went to 68 degrees and once again the cast steel was penetrated. The 1200 BHN was immune all the way to 15mm.
The high hardness would simply not allow the round to gain any traction.

So when you see a 38 mm plate on the Abrams at 83 degrees It would be possible to penetrate it if it was soft RHA. It’s not. It’s high hardened and it’s so hard the frontal plates sits on small solid bushings so it won’t crack. But the Abrams has never been penetrated there after they shot it 5000 times with darts.

All these things, they were doing in 1954 in the American Silacious Core Armor experiments. They used every combination we had. Some people confuse this with the chobham armor. But it’s not the same thing.
Chobham simply ceramic rubber and fiberglass in precast blocks. Nera blocks.

The conclusion of the salacious, core armor test was, materials are not all created equal, the order in which you put those materials can completely change what you’re facing and it cost a ton of money to make plates that are good against most situations.

The US Army learned to spend a lot of money on its armor materials and QC. It’s all fracture and stress tested under electron scan microscopes, it’s all triple CNCd to make sure it’s uniform so it won’t take loads unequally.
More importantly we make our own plates.
We learned it would be very easy for a country to sell us bad armor.

When you see 1500 hp or horsepower per ton, it really means very little. At what RPM? Is that a reasonable RPM to be running at? (car companies often times Dyno test 20% over red line to advertise way more horsepower) how fast the engine revs based on stroke length piston and crank materials, cam profiles etc.

Horsepower is actually a very lazy way to measure speed. I can show you 1000 hp import that will get crushed every single time by a 800 hp muscle car unless they have 5 miles to race. Torque is king. Horsepower is just torque divided by RPM. So if you can make a car, do double the RPMs or over rev or outside it’s useful power band, you can appear fast.

You can argue the same for the current clutch implementation. Alot of those questions would also relate to regenerative steering being modelled, currently all vehicles use clutch and brake to steer directly via driver input(sometimes damped using “driver assistance mode”)

Typical Brinnell hardness for modern plates would be 300 ±40, sometimes higher for high hardness plates.

All of this is irrelevent because we can produce or study the engines torque curve, how that torque is achieved is not useful to us.