Explaining burnout asymetry

[MKI]

I am still at awe of the Air intake and burnout symmetry on my SSTO planes.

Does anyone have a good explanation as to how air intakes affect my engine performance in low oxygen environments? (such as the edge of space)

For those a little confused, my 2 jet engine Single Stage To Orbit will always lose control to one engine, leaving the other at a different thrust level and thus creating a flat spin type scenario which is NOT good at high speeds.

So how do you prevent these things on your own SSTO's and exactly why does this stuff still happen, i thought they fixed it a few patches ago. (i took a break from SSTO's)

Oh and don't say use one engine lol, i already know that one haha

Edited June 27, 2014 by MKI

[Red Iron Crown]

Kasuha discovered the details of intake air flow and posted them in his fuel flow rules thread, along with some demonstrative experiments.

Surprisingly, the order of placement of parts determines which engines get fed by which intakes. If you have two engines and eight intakes, construct your ship by placing 4 intakes, then an engine, then the next four intakes, then the other engine. They will then get the same amount of intake air.

[FenrirWolf]

If I recall correctly, the last engine that you place will be the first one to burn out. And when using symmetry to attach multiple engines, the "last engine placed" is whichever one you actively attached, not the duplicate created by symmetry.

In which case a good way to reduce flat spins is to have an odd number of engines, whether it be 1 or 3 or 5 or whatever, and have the center engine be the last one that you place.

EDIT: Looks like I got ninja'd by a more thorough response.

[LethalDose]

So how do you prevent these things on your own SSTO's and exactly why does this stuff still happen, i thought they fixed it a few patches ago. (i took a break from SSTO's)

Oh and don't say use one engine lol, i already know that one haha

Most of my SSTOs have two RAPIERs, and my solution has been to use action groups. I have one action group that either shut's down my turbo jets, or switches engine mode for my RAPIERs. When I start noticing unequal thrust, I hit that action group. It doesn't prevent the vessel having these goofy fuel feed issues, but it does handle the issue well.

After I hit ~ 22 km altitude, I activate my RCS to keep the vessel flying straight. The size of the RCS blocks' exhaust also give me a good warning for when to switch modes/shut-down the air-breathers.

Kashua did a great job documenting the effects that the placement order of the engines has on fuel flow. However, trying to translate that into practical vessel design other than "Place lateral engines at the same time" makes my brain hurt.

[Claw]

If I recall correctly, the last engine that you place will be the first one to burn out. And when using symmetry to attach multiple engines, the "last engine placed" is whichever one you actively attached, not the duplicate created by symmetry.

The last engine placed is the first to burn out. But when placing via symmetry, the last one placed is the one that is mirrored, not the one that's under the mouse cursor. For symmetry higher than 2x in the VAB, it seems to place symmetry clockwise when viewed from above, but I haven't confirmed that absolutely.

[Red Iron Crown]

Kasuha did a great job documenting the effects that the placement order of the engines has on fuel flow. However, trying to translate that into practical vessel design other than "Place lateral engines at the same time" makes my brain hurt.

I think I have a grasp of how it works based on Kasuha's explanation, it works like this:

When doing intake air calculations, the program runs through the .craft description in order; when it hits an intake it adds its air to the IntakeAir pool, when it hits an air-breathing engine it subtracts its requirements from the pool. Since parts are added to the .craft description in order of placement, it matters what order you place the parts on the ship.

In my example of the two engine, eight intake design, the way most of us would put it together is to place all the intakes at once, then place all the engines. So the .craft description would be (simplified, of course):

Intake

Intake

Intake

Intake

Intake

Intake

Intake

Intake

Engine

Engine

So the first engine finds the IntakeAir pool filled by eight intakes and takes its full requirement; the second engine is left with whatever the first doesn't take, so it will starve for air first. This produces asymmetrical flameout.

If instead you place half the intakes, then an engine, then the rest of the intakes, then the other engine, the .craft description looks like this:

Intake

Intake

Intake

Intake

Engine

Intake

Intake

Intake

Intake

Engine

So when the program runs the IntakeAir calculation it gets to the first engine with four intake's worth of air, which the engine then takes. The next four intakes fill the pool again for the second engine to clear. This is the recipe for symmetrical flameout.

To put this knowledge to work: Build your ship by placing half the intakes, then an engine, then the remaining intakes, then the other engine. Alternately, build an assembly with one engine and half the intakes and duplicate it with symmetry. Never place engines alone with symmetry.

This is not very intuitive or obvious, kudos to Kasuha for figuring it out.

Edited June 27, 2014 by Red Iron Crown

[Red Iron Crown]

The last engine placed is the first to burn out. But when placing via symmetry, the last one placed is the one that is mirrored, not the one that's under the mouse cursor. For symmetry higher than 2x in the VAB, it seems to place symmetry clockwise when viewed from above, but I haven't confirmed that absolutely.

It's not necessarily the last engine placed that flames out first, though that is common. Imagine you place two intakes, then an engine, then four intakes, then an engine. The first engine placed will flame out first.

To avoid asymmetrical flameout, it is important to never place air-breathing engines alone with symmetry, only use symmetry when they are part of an assembly including intakes and an engine.

[MaverickSawyer]

I use SmartParts, and if it senses flameout on one engine or the other, regardless of which engine it is, it'll trigger an action group to change engine cycles. Whether that's switching modes on a RAPIER or shutting down the turbojets and switching to rockets, it still runs on the same action group system. It's hassle free once it's set up. Just point the nose where you want to go and start the air-breathing engine(s), then just keep going until you make orbital velocity.

[Claw]

It's not necessarily the last engine placed that flames out first, though that is common. Imagine you place two intakes, then an engine, then four intakes, then an engine. The first engine placed will flame out first.

Yeah, sorry. I wasn't trying to copy and paste all the nuances. I was just addressing this specific statement, which was the one I quoted.

And when using symmetry to attach multiple engines, the "last engine placed" is whichever one you actively attached, not the duplicate created by symmetry.

So my statement was meant to be aimed at placing symmetric type parts. But you're right, what I said isn't fully correct.

[MKI]

Alright im glad there are defined principles of how it all goes down.

So sticking with the rule of placing one sides air intakes and the engine its feeding, THEN the other sides air intakes and the engine they are feeding should balance out the air intakes as a whole.

Good to know and it should help refine some of my bigger more demanding SSTO designs that have failed for this exact reasoning.

With this knowledge looking back it does make sense as to why this happens. The ordering of the engines in the file really explains it all. Thanks for helping me figure it all out.

[Laie]

So sticking with the rule of placing one sides air intakes and the engine its feeding, THEN the other sides air intakes and the engine they are feeding should balance out the air intakes as a whole.

Yes. Just be careful with symmetry: first placing two intakes, with symmetry, then two engines, will lead to one engine having two intakes and the other having none.

However, you can prepare matched sets (say, a nacelle with n intakes and one engine) and then symmetrically place these subassemblies. No need to save them as subassembly first, or to actually use the "nacelle" part; but I call them nacelles and subassemblies, if you don't mind.

How symmetry works:

let's assume this is your plane:
---------------|    |---------------
                \__/

you're placing nacelles under the wings, with symmetry, inner engines first:
---------------|    |---------------
7   5   3   1   \__/   2   4   6   8


preparing an entire wing, nacelles and all, then placing it with symmetry:
---------------|    |---------------
4   3   2   1   \__/   5   6   7   8

I much prefer wing-wise symmetry. I allows me to switch off 
every other engine, or all but two, and still get symmetric thrust.


Be aware that the inner engines don't become #1 automatically.
If your pre-assembled wing had the engines placed in this order:
---------------|    |
4   1   2   3   \__/ 

you'd end up with this:
---------------|    |---------------
4   1   2   3   \__/   7   6   5   8

No matter how you went about construction, the last action counts. Symmetrically placing the nacelles lets you see how the CoM shifts as you go. Then rip off and re-attach the wing, and behold: wing-wise symmetry.

[Pecan]

Not completely on-topic but I think it's worth repeating that you can usually tell when your jets are getting close to flameout. Throttling-back reduces their demand so can keep them thrusting longer.

[Geschosskopf]

Not completely on-topic but I think it's worth repeating that you can usually tell when your jets are getting close to flameout. Throttling-back reduces their demand so can keep them thrusting longer.

This really isn't necessary these days. The RAPIER (and IIRC the Turbojet) both automatically throttle themselves down as the amount of air available decreases. This was an under-the-hood change introduced at the same time the RAPIER was released. Although the throttle indicator next to the navball doesn't move, you'll see the engine thrust decreasing gradually if you right-click on it. That's the automatic throttle-down in action.

[xcorps]

Very interesting comments about engine placement.

As for flameouts, Mechjeb has a pair of settings in the Utility tab:

1)Manage air intakes - automatically opens and closes intakes based on demand, manages drag very well.

2)Prevent jet flameout - reduces throttle as available air intake lowers, with a customizable safety margin. This usually takes a test flight to get right, but it works.

Red Iron Crown:

Never place engines alone with symmetry.

You mean placing them alone without air intakes?

I like to use the manifold adapters (bi, tri, quad) for engine mounts. I usually have to place the engines on each node individually, then if I want more engines I use symmetry on the adapter. Does that break the rule I quoted?

Edited June 27, 2014 by xcorps

[Red Iron Crown]

You mean placing them alone without air intakes?

I like to use the manifold adapters (bi, tri, quad) for engine mounts. I usually have to place the engines on each node individually, then if I want more engines I use symmetry on the adapter. Does that break the rule I quoted?

It does break that rule. If you place engines, or an assembly with engines and no intakes, using symmetry they will be adjacent to each other in the .craft description with no intakes between them, which is what causes the asymmetric flameout.

[xcorps]

So for best intake efficiency I want to place say a triadapter, then an engine, then an intake, then an engine, then an intake, engine, intake.

So if I want two assemblies of triadapter engines, I can place the adapters with symmetry on then place all six engines manually alternating with intakes. Right?

[Red Iron Crown]

Yes, assuming one intake per engine. Structural parts like the tricoupler have no effect, whether placed with symmetry or not.

More generally: If you have I intakes and E engines, place I/E intakes, then place one engine. Repeat until all are placed.

If I is not evenly divisible by E, you will have asymmetric flameout.

Edited June 29, 2014 by Red Iron Crown
More detail.

[Kasuha]

Notice that with two engines, you will still have asymmetric flameouts even if you place intakes and engines 'properly'. Just not as intense as before. At some point, your intakes for first engine will become unable to provide enough air to keep the engine on, it will flameout, but twice the intakes will still keep the other engine alive.

If you're working with four engines, it's good idea to place them left-right-right-left so when this kind of flameout occurs, one engine on each side will flameout. Then you can throttle down to keep them all on without ill effect on your ship stability.

[O-Doc]

Asymmetry is quite useful for maximising your engine thrust. Just pay attention to your yaw indicator for the early signs of possible flame-out. Knowledge of engine placement is good for small SSTOs and keeping your flame-out engine in the center of your engine stack for large SSTOs. Don't get hung up on counting intakes with engine placement, on many designs it is either not practical, or, simply not possible.