Asparagus vs regular

[CosmicCharlie]

I have the same craft, asparagus-style and regular-style. With the regular one, I can get an apoapsis of 80km with less fuel than I can the asparagus.  I only have 4 tanks connected asparagus-style because otherwise the stacks of 3 tanks run out before the last stacks of 2 tanks.  Still, the asparagus should be more efficient, shouldn't it?

Discuss.

https://drive.google.com/file/d/0B6Dffk6hHIF3SWR4LTBGZ3g3dVU/view?usp=sharing

 

https://drive.google.com/file/d/0B6Dffk6hHIF3Nk1NNkpuc0YzZjg/view?usp=sharing

 

[ToukieToucan]

18 minutes ago, CosmicCharlie said:

I have the same craft, asparagus-style and regular-style. With the regular one, I can get an apoapsis of 80km with less fuel than I can the asparagus.  I only have 4 tanks connected asparagus-style because otherwise the stacks of 3 tanks run out before the last stacks of 2 tanks.  Still, the asparagus should be more efficient, shouldn't it?

Discuss.

https://drive.google.com/file/d/0B6Dffk6hHIF3SWR4LTBGZ3g3dVU/view?usp=sharing

 

https://drive.google.com/file/d/0B6Dffk6hHIF3Nk1NNkpuc0YzZjg/view?usp=sharing

 

Asparagus is more effiecent if you link up everything to a 'layer' down, this way a middle tank will never be empty and the outer tanks will always be empty first.

[Snark]

29 minutes ago, CosmicCharlie said:

I have the same craft, asparagus-style and regular-style. With the regular one, I can get an apoapsis of 80km with less fuel than I can the asparagus.

Discuss.

https://drive.google.com/file/d/0B6Dffk6hHIF3Nk1NNkpuc0YzZjg/view?usp=sharing

That's not asparagus.  The staging's not set up correctly, nor are there enough fuel ducts.

A correctly set up bilateral asparagus will have the following properties:

• ALL the engines fire upon takeoff.
• The radial decouplers are always in pairs.  That is, each stage will fire exactly two decouplers.
• The order that the decouplers fire in is the same order that the tanks are drained.
• The pattern is:  the "outermost" (according to fuel duct arrangement) tanks drain 100% and are then jettisoned.  Then the next pair drains, and so forth.

12 minutes ago, ToukieToucan said:

I only have 4 tanks connected asparagus-style because otherwise the stacks of 3 tanks run out before the last stacks of 2 tanks.

Except that that will never happen with properly set-up asparagus.  Only one pair of stacks will drain at a time, and you jettison the empties, so that each time you decouple, you're left with a ship that has 100% full fuel tanks.

33 minutes ago, CosmicCharlie said:

Still, the asparagus should be more efficient, shouldn't it?

Not necessarily.  The way you have it set up, you have some very odd staging (lots of engines that don't come online until later), there may be some TWR issues there.

[CosmicCharlie]

OK, thanks guys.

Hey Snark, you've been helping me a lot. A whole lot.

Maybe you should just give me your phone number so I don't clutter up the boards...

...

...

(Kidding.)

[Snark]

Okay, I tried grabbing your "asparagus" craft, tinkering with it to fix the staging of decouplers and engines, and adding appropriate fuel ducts.

Here's the craft file.  Changes from the original:

• removed MechJeb part ('coz I don't have that installed)
• updated engine and decoupler staging
• added fuel ducts
• added some angled fins on the top of some stages to help them peel away when jettisoning, so they don't hit the central core

 

Here's what I ended up with in 80 km circular orbit.

Here's the staging there.  Note that the apparently not-completely-full status for the various engines' fuel bars is misleading, due to the aspargus setup.  Only the two emptiest-looking stacks (i.e. the ones that show about a quarter-full) are really accurate.  All the other stacks are 100% full.

Here's the resource count.

...So, that's 1117 tons of LFO delivered into 80 kilometer circular orbit.  That's as hand-flown (I don't do MechJeb).  Not necessarily an optimal flight path, but not a terrible one, dunno if MechJeb would have done better.  Maybe you could squeeze another couple of percent out of it with an optimal path.

In your non-asparagus setup, I don't know how much LFO you got to orbit, but is it safe to assume it was less than 1117 tons?

[Plusck]

Snark's results come extremely close to my own with the smaller (two-ringed) version I was tinkering with earlier (final version: Refuel Asparagus 3). I got that one to to orbit with 74,101 liquid fuel left out of a takeoff total of 275,400 (just under 27%), while this one gets 100,600 to orbit out of 379,080 (again, close to 27%).

I didn't add fins though - but simply resorted to cutting throttle when staging the inner circle. The outer circle was fine on the two-ringed craft.

Edited January 19, 2016 by Plusck

[Snark]

13 minutes ago, Plusck said:

I didn't add fins though - but simply resorted to cutting throttle when staging the inner circle.

Yah, I ended up having to cut throttle to avoid collisions, too, once I got above 20 km or so and the fins had less bite.  Mainly wanted to avoid thrust loss during that initial ascent.

14 minutes ago, Plusck said:

the smaller (two-ringed) version

My vote for naming the thing that I just flew:

"The Three-Ring Circus"

[Jimbodiah]

Asparagus is the most efficient way to use boosters, but you have to do it properly. It's only efficient in KSP as the dry weight of everything it so high that is is actually helpfull to drop off unused weight in favor of losing a little thrust (you need less and less the higher you get anyway).

[GoSlash27]

1 hour ago, Jimbodiah said:

Asparagus is the most efficient way to use boosters, but you have to do it properly. It's only efficient in KSP as the dry weight of everything it so high that is is actually helpfull to drop off unused weight in favor of losing a little thrust (you need less and less the higher you get anyway).

 Weeelll...

 It is in theory, but in practice... it's more complicated than that.

Merely getting something to orbit ( approx. 3500 m/sec DV) doesn't take enough DV to justify more than 1 staging event with most boosters. If you experiment, you will find that a 3 stage booster winds up being heavier than a 2 stage design for most applications, and the more stages you add (dividing the DV budget), the heavier the lifter becomes.

 Since most asparagus stacks consist of 3 or more stages, applying asparagus staging usually winds up creating a slightly- more efficient version of a less- efficient stack design. End result is you wind up with a stack that weighs about the same as a generic 2 stager at best, and costs a lot more.

 I used to be big into asparagus back when the lighter engines had better Isp and orbit took another km/sec DV, but nowadays they don't get much use.

Best,

-Slashy

[Temstar]

17 minutes ago, GoSlash27 said:

Merely getting something to orbit ( approx. 3500 m/sec DV) doesn't take enough DV to justify more than 1 staging event with most boosters. If you experiment, you will find that a 3 stage booster winds up being heavier than a 2 stage design for most applications, and the more stages you add (dividing the DV budget), the heavier the lifter becomes.

I rather doubt that. In the stock payload fraction challenge the pure vertical climb rocket with the highest payload fraction is a 7 core asparagus. If you can hit more than 26% with one or two stages then you should enter a design.

Even if that was true, asparagus staging is still the better design because a lower proportion of the rocket's dry mass is engines and more of the wet mass is fuel. And we know that's good because engines are expensive and fuel is cheap.

[Snark]

Asparagus has its place.  I adore it, but in the post-1.0 world, aerodynamics favors a tall skinny design rather than a wide one, so asparagus ain't quite what it used to be.

Asparagus designs are worse for aero, but better for eagerly dumping dead weight.  Which effect predominates depends on a lot of things, including how overall streamlined the ship is, overall ship scale, etc.

For example, the square-cube law favors big ships for aerodynamics.  If you're making ginormous ships like the one in this thread, the aero is a much lesser consideration, because aero matters less as you scale up.  If you double all the linear dimensions, you're moving eight times the mass (and presumably build it with eight times the thrust), but it has only four times the cross-sectional area.  Really big ships hardly even notice aero drag (as long as they're not stupid about it).  Therefore, I'd expect bigger designs to do better with asparagus.

Then there are practical considerations.  This ship weighs thousands of tons on the pad.  Unless you resort to something like KJR, I suspect you'd have real trouble trying to make this fella tall and skinny. 

[GoSlash27]

6 minutes ago, Temstar said:

I rather doubt that. In the stock payload fraction challenge the pure vertical climb rocket with the highest payload fraction is a 7 core asparagus. If you can hit more than 26% with one or two stages then you should enter a design.

Temstar,

 I have seen that challenge, and I'm not interested in participating in it. I used to design for max payload fraction back in the day, but these days I design for economy.

9 minutes ago, Temstar said:

Even if that was true, asparagus staging is still the better design because a lower proportion of the rocket's dry mass is engines and more of the wet mass is fuel. And we know that's good because engines are expensive and fuel is cheap.

 Actually, I'd say this is precisely why it's *not* good. Unless you're using an add- on like stage recovery, anything you drop during a staging event is not recovered. I'd rather not drop a bunch of engines I don't need to drop during a launch if I don't have to. *Especially* not Vectors.

 More importantly, there is an economy of scale when you get into really big launchers that doesn't hold when working in the range of lifters that don't need to be clustered. Asparagus is nearly always better when you *have* to cluster, but it doesn't compete when you don't have to cluster.

Best,

-Slashy

 

[Rocket In My Pocket]

I tried Asparagus once but I didn't care for it.

Made my pee smell funny.

[Jimbodiah]

Big ships will do better as the dry weight to fuel ratio on smaller craft is higher, so you are carrying a lot of extra dead weight vs a bigger version.

I just like how they stage, but for my career mode I can do pretty much everything with a two stage setup. For really heavy stuff (60+ tons) I use a Delta IV Heavy setup I have which uses two LF boosters, and for 200+ tons I have a Jupiter III which is just awesome to see launching. But 99% goes up in a sleek 2-stage or an Ares I style setup with SRB 1st stage and LF second (so also 2 stage).

Then again, I don't think I use a lot of stock parts, maybe landing legs and the 1.25/2.5m probe cores. Stock parts are too small and limited (and part-count goes through the roof). I really think they should add more 3.75 and 5m parts in the next update.

Edited January 20, 2016 by Jimbodiah

[Temstar]

6 minutes ago, GoSlash27 said:

 I have seen that challenge, and I'm not interested in participating in it. I used to design for max payload fraction back in the day, but these days I design for economy.

Well I can definitely agree with that point of view. Someone need to do a "Stock cheapest payload per ton challenge".

That said the economy doesn't seem that bad. I manage about $800 per ton to orbit with my large asparagus launcher and I can see development paths that will further decrease this to somewhere in the region of $600 per ton with recovery of the Mammoth ULA Vulcan style. This doesn't seem all that bad next to the best reusable rocket which is in the region of $400 per ton to orbit.

I suspect the reason why asparagus cost per ton is decent (besides the obvious it uses fewer engines for a given payload) has to do with the size. By their nature they are big rockets and so there's more room for fine tuning various aspects like TWR curve, engine combination, fraction of the dry mass of the guidance package and so on.

[GoSlash27]

2 minutes ago, Temstar said:

Well I can definitely agree with that point of view. Someone need to do a "Stock cheapest payload per ton challenge".

That said the economy doesn't seem that bad. I manage about $800 per ton to orbit with my large asparagus launcher and I can see development paths that will further decrease this to somewhere in the region of $600 per ton with recovery of the Mammoth ULA Vulcan style. This doesn't seem all that bad next to the best reusable rocket which is in the region of $400 per ton to orbit.

I suspect the reason why asparagus cost per ton is decent (besides the obvious it uses fewer engines for a given payload) has to do with the size. By their nature they are big rockets and so there's more room for fine tuning various aspects like TWR curve, engine combination, fraction of the dry mass of the guidance package and so on.

Temstar,

 Yeah, I agree with this. *my* point is that asparagus only works with large rockets. The benefit you see is due to the fact that they are large rockets, not that they are asparagus staged.

If you try to make an asparagus staged lifter in the 30t class (for example), you will find that you can make a lighter and cheaper 2 stager that does the same thing.

Best,

-Slashy

[Jimbodiah]

There is economical efficiency, and there is fuel efficiency. Asparagus optimizes the latter, not the first.

[mcirish3]

4 hours ago, Snark said:

Not necessarily an optimal flight path, but not a terrible one, dunno if MechJeb would have done better.  Maybe you could squeeze another couple of percent out of it with an optimal path.

Base on everything I have been reading the current best way to get to orbit with maximum efficiency is with [1.0.5] GravityTurn version 1.2.2 - Automated Efficient Launches which has MJ integration for the final circularization.  I will let you know when I solve the Full Godard problem for KSP.  Also you are very much correct for large ships and Asparagus staging.

[numerobis]

3 hours ago, Temstar said:

Well I can definitely agree with that point of view. Someone need to do a "Stock cheapest payload per ton challenge".

Right over here! I did it as passengers instead of payload to avoid the usual bickering over the definition of payload.

[GoSlash27]

On 1/20/2016 at 6:29 PM, GoSlash27 said:

Temstar,

 Yeah, I agree with this. *my* point is that asparagus only works with large rockets. The benefit you see is due to the fact that they are large rockets, not that they are asparagus staged.

If you try to make an asparagus staged lifter in the 30t class (for example), you will find that you can make a lighter and cheaper 2 stager that does the same thing.

Best,

-Slashy

After thinking about this and experimenting, I have to retract this statement. I was able to come up with a 24t asparagus lifter that was both lighter *and* cheaper than it's 2- stage analogue. This only worked due to the Reliant's excellent cost- effectiveness as compared to the Mainsail and it wasn't economical compared to a SRB lifter, but it *did* outperform a 2 stager in small scale.

Best,

-Slashy

[Temstar]

2 Stage? As in Falcon Heavy style?

[suicidejunkie]

For absolute economy, a SSTO spaceplane with a small mining package can get the remaining payload mass to orbit for negative money per ton.

However that's only because Jeb hasn't noticed that we aren't actually paying him.

[Norcalplanner]

On 1/20/2016 at 4:22 PM, Temstar said:

Well I can definitely agree with that point of view. Someone need to do a "Stock cheapest payload per ton challenge".

This had been discussed a little bit in the tutorial thread linked in my signature.  I have the rules almost done, but can't spend as much time on KSP these days as I'd like.  If someone is willing to co-judge with me (since there will be frequently be a few days at a time when I don't do anything KSP), or if people are OK with a delay in the scoring, then I can get it up and running by the tail end of this weekend.

And yes, it will be called "The Cheap and Cheerful Rocket Payload Challenge".