Engine placement and centre of thrust on planes

[baldamundo]

I've mostly always been quite boring in my plane designs so far, and always kept my engines layouts fairly simple, symmetrically around the central axis of the craft and all facing directly backwards, such that the centre of thrust is neatly aligned directly behind and level with the centre of mass.

A lot of real world planes and space shuttles and such have more complicated arrangements, sometimes with all the engines higher than the centre of mass, or e.g. like the space shuttle launching with it's engines angled diagonally, and I'd really appreciate it if someone could either point me to a guide, or just explain some of the basic considerations about, I guess, (relatively) more advanced engine placement - what are the advantages of having engines in different alignments, what are the effects on stability, how do you balance these things out, etc Thanks!

Edited September 27, 2015 by baldamundo

[The_Rocketeer]

Don't be under any illusions, the combined CoT of the OMS engines on the Space Shuttle, and any other space-faring aircraft, point straight through the CoM. Off-centre thrust would just cause the spacecraft to tumble.

Think of the line between the CoM and the CoT as a lever-arm like a tyre-iron. The CoM is like the centre of the wheel-nut (the pivot or fulcrum), and CoT is the opposite end of the lever where you push with your hand. The direction of the arrow coming out of the CoT is the opposite of the direction you're pushing on the lever (think of it like the engine pushing on the craft, rather than the craft pushing on the air).

Now if you push the tyre-iron towards the wheel nut, nothing turns - if the handbrake is off you might accidentally push the car off the jack doing this - this is like the engine pushing the plane forwards thru the air. It's not what you want for the car, but it's exactly what you want for a spacecraft.

If you push the tyre-iron up or down though, you use your leverage to turn the wheelnut. This is what happens if the CoT isn't aligned with the CoM, you push the CoT around the CoM. This would cause the whole plane to rotate or tumble.

Although the space shuttle has big boosters that point in odd directions, and tho their CoT is off-centre of the CoM of the Orbiter, they're only used during ascent when the big external tank is altering the CoM. The OMS engines, the little ones high on the back, are angled to point thru the CoM of the orbiter, so although they look like they're too high, they're not, because the orbiter is designed to thrust in a nose-up attitude without tumbling.

Now, commercial jets and sub-sonic aircraft, like jumbo-jets with the engines slung under the wings, has a CoT that's off centre, but this is made up for by the CoL also being off centre, beings somewhat behind the CoM. These planes fly by doing a balancing act of thrust keeping the nose up and lift keeping the tail up. The advantage of this is it helps keep much of the attitude control on the throttle - high throttle to climb, low throttle to cruise - and avoids having to use the control surfaces too much, which would increase drag. Really this is all about fuel efficiency at sub-sonic speeds.

Having the engines nearer the front of the aircraft is also inherently more stable than having them near the back - it's like the difference between balancing on a unicycle vs hanging from a crane. You can easily tip over the unicycle, but the crane? Not so much. The difficulty is you then have to get the thrust out behind the craft without it hitting any part of the vehicle like tail-wings.

Edited September 26, 2015 by The_Rocketeer

[surge]

In short, there is no advantage to having wierd thrust angles. It always loses some energy. It's only ever done to compensate for poor design and/or materials limitations.

As The_Rocketeer said, the big engines on the NASA shuttles are only for the huge tank on the bottom. Of note is since they carry solid boosters on that axis, these engines are the widest gimbaling ones ever made - during SRB boost, they are pointing 'up', but when it drops them, they must point 'down'. KSP does not have engines that can gimbal so much, so although it's possible to build a shuttle-like craft, it's very difficult.

Also, commercial aircraft wings have dihedral (look it up) which, among other things, puts the underslung engine thrust closer to the centre of mass.

If you design all your craft like you say, you are doing great! Maybe you should apply for a job at NASA, and show them how to do it

[baldamundo]

Thank you! This is really interesting and useful stuff, exactly what I was looking for

Don't be under any illusions, the combined CoT of the OMS engines on the Space Shuttle, and any other space-faring aircraft, point straight through the CoM. Off-centre thrust would just cause the spacecraft to tumble.

I figured this had to be the case - that was part of why the space shuttle had me so confused

Although the space shuttle has big boosters that point in odd directions, and tho their CoT is off-centre of the CoM of the Orbiter, they're only used during ascent when the big external tank is altering the CoM. The OMS engines, the little ones high on the back, are angled to point thru the CoM of the orbiter, so although they look like they're too high, they're not, because the orbiter is designed to thrust in a nose-up attitude without tumbling.

Ahhh, I didn't even realise the space shuttle had two sets of engines.

I'm trying to make sure here that I understand the practical design implications of this stuff and also struggling to remember my high school physics:

What I think is confusing me now about the space shuttle, is that I'm not sure I understand how to work out the net result of the shuttle's multiple sources of thrust during launch - I know to conceive of the boosters' thrust (parallel to the centre of gravity) as a turning moment around the centre of mass, and the shuttle engine's thrust (diagonal to both) as a vector that can be resolved into horizontal and vertical components...but what exactly is the interaction between a moment (in Newton-metres, or Joules) and a vector (presumably just in Newtons)? I guess the vertical component of the shuttle engines' thrust would produce a moment in the opposite direction that balances part of the boosters'...but surely that would still leave it pivoting upside down, albeit more slowly, while the horizontal component of the thrust just accelerates it...uhh...towards the orbiter's underside?

Am I barking up the wrong tree here?

EDIT: I am an idiot talking gibberish. The above is nonsense. See posts further down for less nonsensical exposition.

Now, commercial jets and sub-sonic aircraft, like jumbo-jets with the engines slung under the wings, has a CoT that's off centre, but this is made up for by the CoL also being off centre, beings somewhat behind the CoM. These planes fly by doing a balancing act of thrust keeping the nose up and lift keeping the tail up. The advantage of this is it helps keep much of the attitude control on the throttle - high throttle to climb, low throttle to cruise - and avoids having to use the control surfaces too much, which would increase drag. Really this is all about fuel efficiency at sub-sonic speeds.

Having the engines nearer the front of the aircraft is also inherently more stable than having them near the back - it's like the difference between balancing on a unicycle vs hanging from a crane. You can easily tip over the unicycle, but the crane? Not so much. The difficulty is you then have to get the thrust out behind the craft without it hitting any part of the vehicle like tail-wings.

Do you know what's going on on something like the A-10 Warthog that looks as if the centre of thrust is above the centre of mass? i.e. http://justrcmodels.com/content/images/thumbs/0002238_12-ch-lx-super-a-10-warthog-thunderbolt-ii-rc-edf-jet-kit-only.jpeg

- - - Updated - - -

It's only ever done to compensate for poor design and/or materials limitations.

I suppose perhaps this is what I'm really asking then - in what circumstances and what limitations can make weird thrust angles necessary?

Edited September 27, 2015 by baldamundo

[surge]

I suppose perhaps this is what I'm really asking then - in what circumstances and what limitations can make weird thrust angles necessary?

In KSP, none! Or when materials are preventative. If you dont need to make stupid inefficient designs, don't! And I expect to see you land and return a couple of Kerbals from Eeloo soon!

It's that simple.

That's why both professional aerodynamists and yokel idiots love this game.

Edited September 27, 2015 by surge

[The_Rocketeer]

Ahhh, I didn't even realise the space shuttle had two sets of engines.

Yes, the OMS engines were essentially big monoprop thrusters, and they actually weren't fired up for the ascent until after the external tank had been staged.

I'm trying to make sure here that I understand the practical design implications of this stuff and also struggling to remember my high school physics:

What I think is confusing me now about the space shuttle, is that I'm not sure I understand how to work out the net result of the shuttle's multiple sources of thrust during launch - I know to conceive of the boosters' thrust (parallel to the centre of gravity) as a turning moment around the centre of mass, and the shuttle engine's thrust (diagonal to both) as a vector that can be resolved into horizontal and vertical components...but what exactly is the interaction between a moment (in Newton-metres, or Joules) and a vector (presumably just in Newtons)? I guess the vertical component of the shuttle engines' thrust would produce a moment in the opposite direction that balances part of the boosters'...but surely that would still leave it pivoting upside down, albeit more slowly, while the horizontal component of the thrust just accelerates it...uhh...towards the orbiter's underside?

Physics is one subject I didn't take beyond GCSE, so in terms of the mechanics you'd be better of with a real physicist (terms like 'moment' make me glaze over)! But imagine the situation where the orbiter's engines are parallel to the SRBs and not pitched - in this situation, the TWR of the orbiter alone is very high, and the ET/SRB assembly very low, because the SRBs are bearing their own fuel mass plus the mass of the ET. The effect of launching like this would be for the whole assembly to rapidly pitch down (from the orbiter's orientation) and crash. So, the orbiter's thrust vector is angled off parallel, toward the CoM, until the vertical thrust contribution of both sides is balanced, resulting in launch without rotation.

However, if you watch a video of a shuttle take-off, you'll notice that it's movement at launch isn't vertical. The whole craft actually moves laterally in a ventral direction for some time. This is because of a slight imbalance of horizontal thrust 'y axis' thrust caused by the angle of the orbiter ascent engines. However, acceleration in this direction is slight compared with vertical acceleration, its resisted by aerodynamic forces (lift AND drag), and since the shuttle rolls 90 degrees shortly after launch and then makes a dorsal-down gravity turn, this is insignificant. It's also necessary to avoid the staging of the SRBs causing an immediate tumble, since net velocity has to be just right at this point to transition to the orbiter engines alone.

Do you know what's going on on something like the A-10 Warthog that looks as if the centre of thrust is above the centre of mass?

I don't know for sure, but if I had to guess I'd say the reverse situation. From what I understand the A-10 is a ground-attack aircraft, so it could be that it's convenient for high-thrust to accelerate the aircraft into a strafing dive, i.e. causing a slight pitch-down rather than pitch-up. If so, then the relative positions of the CoL and CoT are probably just reverse, with the CoT high behind the CoM, and the CoL somewhat in front of the CoM.

I suppose perhaps this is what I'm really asking then - in what circumstances and what limitations can make weird thrust angles necessary?

Well, in as far as anything in KSP is necessary, I'd say if you enjoy testing your understanding and engineering skills to make it work that's a good enough reason. Otherwise, there are almost always other engineering solutions that are simpler. I've built a few STS replica shuttles where the off-line CoM made angled thrust necessary, but otherwise I find it much more convenient to keep everything symmetrical on the x and y axes - it just saves a lot of trouble!

[Alshain]

I suppose perhaps this is what I'm really asking then - in what circumstances and what limitations can make weird thrust angles necessary?

They are only necessary when design consideration require them, and even then... NEVER in space (or even high atmosphere). Without a countering force it will tumble. This plane is an older one of mine. It demonstrates the offset thrust principle that The_Rocketeer was describing, but when it switches to rocket engines, they are perfectly balanced. In this case it uses the little orange radial engines to balance it since the main engine (the aerospike) doesn't have gimbal to help either. Overall, it was very temperamental anyway, and I don't recommend it unless you are like me and just want to do it to prove you can (I tend to be a glutton for punishment).

[Kryxal]

I don't know for sure, but if I had to guess I'd say the reverse situation. From what I understand the A-10 is a ground-attack aircraft, so it could be that it's convenient for high-thrust to accelerate the aircraft into a strafing dive, i.e. causing a slight pitch-down rather than pitch-up. If so, then the relative positions of the CoL and CoT are probably just reverse, with the CoT high behind the CoM, and the CoL somewhat in front of the CoM.

There's actually some discussion of this very issue on Wikipedia's "Talk" page about the Warthog: Talk:Fairchild_Republic_A-10_Thunderbolt_II

[baldamundo]

There's actually some discussion of this very issue on Wikipedia's "Talk" page about the Warthog: Talk:Fairchild_Republic_A-10_Thunderbolt_II

I have a feeling the commenters there are working with different understandings of what it means for the engines to be "tilted upwards" (i.e. which end of the engine is tilted upwards), because otherwise the comments there seem to be completely contradictory - it's not at all clear, to my eye at least, whether the exhaust is tilted at all, although the intake certainly seems to be

[cantab]

An atmospheric aircraft is built so that aerodynamics make it fly straight. The torque from off-axis thrust is readily counteracted by the aerodynamics. In KSP that means you can put your jet engines where you like, within reason, and be able to make a flyable plane. That said you will still get changes in pitch when you throttle up or down with off-axis engines, which you'll need to counter with trim or let an assistant tool compensate for you. This drone of mine showed that effect quite strongly, and with a V tail pitch trimming adds extra drag, but nonetheless it was perfectly flyable and I took it quite some distance around Kerbin for survey contracts.

PS: To get an idea of how strong aerodynamic forces can be, pitch up hard and many KSP planes can easily pull 10 g. For engines to produce the same acceleration the aircraft would require a Kerbin TWR of 10, around ten times what most planes actually have.

Edited September 27, 2015 by cantab

[surge]

https://www.grc.nasa.gov/www/K-12/airplane/bga.html

Rockets that generally pass mach 2 are not relevent of course, except for 'ol baldamundo there, who seems to know whats up.

And not down. Do you try to ditch stages in atmosphere thus avoiding space junk? Or if a long mission, fire directly at the planet, dump the boost stage so it falls into it and then orbit?

Those are the important KSP expert traits.

Edited September 27, 2015 by surge

[baldamundo]

Yes, the OMS engines were essentially big monoprop thrusters, and they actually weren't fired up for the ascent until after the external tank had been staged.

Physics is one subject I didn't take beyond GCSE, so in terms of the mechanics you'd be better of with a real physicist (terms like 'moment' make me glaze over)! But imagine the situation where the orbiter's engines are parallel to the SRBs and not pitched - in this situation, the TWR of the orbiter alone is very high, and the ET/SRB assembly very low, because the SRBs are bearing their own fuel mass plus the mass of the ET. The effect of launching like this would be for the whole assembly to rapidly pitch down (from the orbiter's orientation) and crash. So, the orbiter's thrust vector is angled off parallel, toward the CoM, until the vertical thrust contribution of both sides is balanced, resulting in launch without rotation.

However, if you watch a video of a shuttle take-off, you'll notice that it's movement at launch isn't vertical. The whole craft actually moves laterally in a ventral direction for some time. This is because of a slight imbalance of horizontal thrust 'y axis' thrust caused by the angle of the orbiter ascent engines. However, acceleration in this direction is slight compared with vertical acceleration, its resisted by aerodynamic forces (lift AND drag), and since the shuttle rolls 90 degrees shortly after launch and then makes a dorsal-down gravity turn, this is insignificant. It's also necessary to avoid the staging of the SRBs causing an immediate tumble, since net velocity has to be just right at this point to transition to the orbiter engines alone.

I realise now, in my previous post, I was being an idiot and talking nonsense about vectors and moments, the solution's actually a lot simpler.

The key thing I think I wasn't taking into account, was that the fuel tank/booster assembly is MUCH heavier than the shuttle itself (I've just looked it up, and I make it about 17x heavier). Which means that the thrust of the boosters is much closer to the centre of mass (the fulcrum), which means that even though the shuttle's engines are weaker than the boosters combined (I think?), they're probably orders of magnitude further away from the centre of mass, meaning the turning moment will be that much stronger. I'd just assumed that the booster engines were a lot stronger and that the centre of mass was closer to the mid point, and such that the tendency would be to flip over backwards, in which case angling the engines in that way would be totally counterproductive. But actually the problem is the opposite - that the shuttle engines are giving it a tendency to tip over forwards, in which case the tilt makes perfect sense. I think this also is just a way of rephrasing what you've said, though. And yeah, the ventral movement would make sense

By the way, your eyes shouldn't glaze over when you see the word 'moments' - it's actually a really simple concept (if I'm remembering my high school physics right): it's basically just a way of calculating how much something's going to rotate - and it's done by multiplying the force applied (I.e. how hard you're pushing down on the lever) by the distance the force is applied from the centre of mass (i.e. how long your crowbar is). And if you've got something that's having different forces try to rotate it in opposite directions (like the space shuttle with it's two sources of thrust), you can just calculate the moment produced by each of them, and if one's bigger it tells you what direction it's going to turn in, and how fast it'll accelerate.

I was just pointlessly overcomplicating it and confusing myself when I started trying to talk about vectors though.

Edited September 27, 2015 by baldamundo

[surge]

hmm, maybe not. you know there are several buttons to display centre of mass, thrust and lift, right?

Left bottom in assembly.

There's nothing wrong with you talking about moments. The game just doesn't show it... and some countries are never taught about torque.

http://https://www.youtube.com/watch?v=1hwX854eG8E

Let this thread be answered, please.

Edited September 27, 2015 by surge

[baldamundo]

http://https://www.youtube.com/watch?v=1hwX854eG8E

Cheers for those!

By the way, does anybody know how Buran coped with this balance problem? Did the Energia engines just gimbal in the opposite direction to the American shuttle orbiter's?

[surge]

No, why would they do that? They are still 'above' the mass when carrying the main tank, and they still need to point 'down' through the centre of it. Because they're not so high up, they don't need to so much though.

Also, some images I've seen suggest they had a small 3rd rocket pointing very nearly 'down' (some 50-80 degrees from from flight path) which I assume gets turned off when they drop the boosters. I further assume the big control flap below that was used to control that torque as the fuel runs out.

Hmm, I wonder if I can do that in KSP...

[DancesWithSquirrels]

Tilting a high engine down to point more towards the CoG and CoL greatly reduces any pitching moment from off-center thrust.

And having raised engines on the A-10 keeps the engines from breathing nothing more than smoke from the 30mm canon the plane is built around...