Efficient Ratio Booster/Payload Weight for LKO ?

[KerbService]

I have seen and abused formulas for payload mass but that is not what I want to know.
I want your best guesses for an efficient booster weight to payload weight ratio for :

a) Rocket - no wings
b) Payload reaches LKO at 85k
c) any combination of STOCK engines/fuel - single stage/multi -stage no matter

Please specify Booster Weight, Payload Weight, Atmospheric deltaV to reach 85k LKO, beginning TWR, and any comments you care to make.

Example: 18.0 ton Booster with 1.5 ton Payload dV 3491 using LV-TX87 single stage. TWR 2.02

This is just for RULE OF THUMB - BACK OF THE BEER MAT guesses at general efficiency of a design.

STOCK PARTS please -- Thanks!

[Nich]

All of my numbers are old but 16% is a good target.  I have seen 26% in challenges.  Problem is if you want mass efficiency it will cost you extra funds.

Design is very dependent on payload but generally

First stage 1.2 TWR AIr breathing first stage gets AP to 40k with 1600 m/s

Second stage LFO vacuum 1 TWR 2200-2300 dv get you 85k

For cost I generally like

Generally Liquid core with .8 TWR on the pad

Add solids until 1.25ish TWR on the pad

3200-3600 vacuum dv depending on aerodynamics and path to orbit

[KerbService]

3 minutes ago, Nich said:

All of my numbers are old but 16% is a good target.  I have seen 26% in challenges.  Problem is if you want mass efficiency it will cost you extra funds.

3200-3600 vacuum dv depending on aerodynamics and path to orbit

6.25 to 1 sounds good. Do you find that lighter rockets need a higher ratio? And when you say expensive, where comes the cost?

[Nich]

Yes the smaller rockets are more draggy because of the volume law I believe.

If you play career mode you have to pay for the airbreathing engines you are staging off and they are way more expensive then the LFO engines.

Also if you are really Min/Maxing decouplers and nose cones can add 10-15% to the cost of your rocket

[KerbService]

 

35 minutes ago, Nich said:

Yes the smaller rockets are more draggy because of the volume law I believe.

If you play career mode you have to pay for the airbreathing engines you are staging off and they are way more expensive then the LFO engines.

Also if you are really Min/Maxing decouplers and nose cones can add 10-15% to the cost of your rocket

Maybe I should add that cost is not something I concern myself about. I have countless slave planets supporting my sandbox.

But maybe part count for the rocket would be a good addition to the replies.

I will try to compile the info I get from this topic into a general build guide. If people would like that.

Here on IMGUR is a 5.75 to 1 -- 69 ton booster of 46 parts for a 12 ton payload dV of 3384 and on the pad TWR of 1.22 -- This I built in 1.3.1

 

 

[Pecan]

Aim for 20% payload ratio.  Settle for teens.
Move-on hurriedly from 1.5m <15t launch vehicles and pretend it never happened (many will be single-figures, possibly even low single-figures, depending on payload mass!).

Obviously your launch engines need to provide over 1 TWR for launch; I find 1.3 about the minimum that's comfortable.  Choose those for ASL ISP.  Less obviously, second-stage engines will fire/become-mains high enough that you should almost certainly choose them for Vacc ISP and TWR as low as 0.8-0.9, depending on start altitude.  Without making the construction too complex about the best you can do for mass-efficiency is three-stage asparagus (two booster-pairs around core) for LFO engines.

Edited November 14, 2018 by Pecan

[Pecan]

Er, I forgot how OP the Vector engines are for larger payloads.  I just checked, without any huge attempt to minimax:

238.25 tonne launch vehicle for 72.5 tonne payload (launch mass 310.75), dV 3,198m/s (atm)/3,415m/s (vac), launch TWR 1.64, 33 parts, 133,100 funds.

Core = jumbo-64, (inverted) Rocomax adapter, Vector engine (TWR 0.9 with 72.5 payload)
2nd booster stage (pair of) = jumbo-64 and X200-32, (inverted) Rocomanx adapter and Vector engine, fuel lines to core (TWR 1.33)
1st booster stage (pair of) = jumbo-64, (inverted) Rocomax adapter, Vector engine, fuel lines to 2nd stage (TWR 1.64 launch)
TT-38K radial decouplers, protective rocket nose cones and sepratron 1s on each booster, autostrut to grandparent (ie; core via decoupler)

Once again I've been playing games all night so I might be calculating this all wrong but as far as I can see, according to NASA (https://spaceflightsystems.grc.nasa.gov/education/rocket/rktwtp.html), that gives a payload ratio of better than 30%!  Even on a simple payload/launch-mass it's over 23%.  I've built batteries, SAS and probe core into the payload while I was chucking it together so it's not a properly inert payload but then again, there's a lot of purely structural stuff here (nosecones & adapters) that may not be necessary.  (Oops, I had struts as well, the masses above are slightly higher than you should get in practice).

Edited November 14, 2018 by Pecan

[KerbService]

 

@Pecan Great link to the NASA page. Thanks! Just what I like to work with to see the progress I am making building boosters. My main rulers for progress so far have been explosions count and frame rate. And I have moved on to much larger boosters. My largest reliable is for up to 2000 tons and my best looking (maybe) is for 1000 tons. The most difficult problem for the heavy loads has been design of the booster core of the lowest stage. It has to still carry at low altitude the payload and maybe the multiple stages above it so requires many engines for the needed TWR it alone has to support during ascent.

Here on IMGUR is a 5.00 to 1 -- 5000 ton booster of 419 parts for a 1000 ton payload dV of 3817 and on the pad TWR of 1.17 -- This I built in 1.4.1 + MH

It lifts to 100k LKO and almost never blows up! Thanks again for the NASA link.

P.S. All of that is auto-struted to Heaviest - one of my posts no one agrees with. Still, seems to work for me.

Edited November 14, 2018 by KerbService

[Pecan]

"The most difficult problem for the heavy loads has been design of the booster core of the lowest stage. It has to still carry at low altitude the payload and maybe the multiple stages above it so requires many engines for the needed TWR it alone has to support during ascent."

Meh, I hardly ever serial-stage except for the last.  Everything else is onion/asparagus so all the engines are working for their pay right from the start.  If I can't lift it with 4 asparagus stages (core plus 3 booster-pairs, which is the most that fit conveniently around a same-size core) I'm not interested.  Hehe, at that I've never wanted to lift anything more than around 80t so I've never really tried ^^.

That NASA site is great for everything about rocketry.  It's one of the places I learnt from when I started with KSP.

[Pecan]

And another quick check:-

3-stage asparagus using Kerbodyne tanks, mammoth+mainsail cluster engines gives you better than 20% for a 615 tonne payload but that's about your max payload for high-efficiency and a (comparatively) simple build I think.

[KerbService]

@Pecan I have never built anything I wanted to lift in a single go that was 1000 tons. Most of the craft that I have built that I think are good are 100 tons or less or more. So this is just a game. How far can I take it before it breaks and I kill my favorite Kerbal and have to shill out the money for a bucket of flowers for the funeral. It is a challenge that is not going to get me laid anyway I win.

Asparagus (bad vegetable)((Idon't like vegetables)) staging depends on the core stage delivering the required TWR for it's part of the ascent. But my dream of a space program is to lift BIG parts to  LKO (LEO) orbit and bolt that together there then go somewhere fun. That initial climb deep in the atmosphere seems to to be better conquered by brute force. 

So.. (bad vegetable) staging is complex and kills frame rate and not used in real life for the same reason. You did 73 tons to LKO with 33 parts. How can that happen without the complexity of (bad vegetable) staging, how can it be made simple?

 

[Pecan]

" So.. (bad vegetable) staging is complex and kills frame rate and not used in real life for the same reason. You did 73 tons to LKO with 33 parts. How can that happen without the complexity of (bad vegetable) staging, how can it be made simple?"

I'm sorry, I don't understand what you're asking.  Are you really saying 33 parts is a lot and kills frame rate?  Do you really think 3-stage asparagus is complex?   Do you think many Kerbal things happen in real life?  (There has never been a real-life SSTO for a start BUT, haha, we invented the ladder before the rocket!).

However you arrange your staging you're going to depend on your last one to deliver the required TWR - so what?  We're talking about a core stage with four boosters here.  The only difference between asparagus and radial is that there are fuel lines between the tanks.  You could presumably lose a couple of them by using fuel-crossfeed instead.  That's it.

If you really mean how can you get the same payload ratios using radial or serial staging the answer is simple - you can't.  With a three- stage serial design you'd save - exactly 4 fuel-lines and 2 decouplers*.  You still need 2 stack decouplers and 3 sets of engines. Of course your upper-stage engines have to deliver the required TWR but now all the lower-stage engines have to as well, because they aren't helping each other.   So your serial-staged design will have massive engines and tanks to feed them so a lousy payload ratio.  With radial you can do a bit better because all the booster-engines can fire at once but you won't save anything except the fuel lines.  Onion is close - and you get a Korolev cross, like radial - but you won't save any parts (the fuel lines just run straight into the core from all 4 boosters).

I note "Here on IMGUR is a 5.75 to 1 -- 69 ton booster of 46 parts for a 12 ton payload dV of 3384 and on the pad TWR of 1.22 -- This I built in 1.3.1" - is 13 parts more and seems to have a lot of fuel lines.  My nearest equivalent is 17 tonnes lighter, 14 parts fewer and designed for 13.5 tonnes.  I do not include parachutes for stage recovery though so you'll have to think about how to standardise that.

(*Probably fewer, larger tanks as well actually)

Ha!  There is an even more efficient (theoretically) staging system than asparagus called 'twisted candle'.  If you think asparagus is complicated though you'd probably hate that - even though it's stack-staged.

Edited November 15, 2018 by Pecan

[Curveball Anders]

42 minutes ago, Pecan said:

" So.. (bad vegetable) staging is complex and kills frame rate and not used in real life for the same reason.

There's also the minor detail that the glorified garden hose used to transfer insane amount of both LF and oxidiser hasn't shown up in RL yet ...

[Pecan]

1 hour ago, Curveball Anders said:

There's also the minor detail that the glorified garden hose used to transfer insane amount of both LF and oxidiser hasn't shown up in RL yet ...

Whereas the spaceplanes have, flying throughout the solar system, mine and processing ore on other bodies to refuel as they go?  Everyone happy with the years-long missions never leaving the seat and surviving on snacks?  You read that part as well, I take it?

[KerbService]

@Pecan I looked at twisted candle... I am having nightmares now.

[Pecan]

Yes, twisted candle means twisted nightmares.  In case you didn't find it when trawling through the archives I think this:

Is about the best overview of staging strategies.  Be warned though that it's four and a half years old so any mention of specific engines/fuel tanks/other parts is almost certainly irrelevant.  Analysis of the different strategies is permanently sound though.

BTW: The reason twisted candle doesn't usually compete with asparagus (or even onion) is that it relies on radial-attached engines, which means limited choice.  Using normal node-attached engines with the strategy instead requires building girder frameworks for them, which adds a lot of mass and drag.  Same thing for the fuel-lines which you have to run up the stack (although this may be changed by the fuel-priority system, I haven't tried it).

[FinalFan]

On 11/15/2018 at 4:46 PM, Pecan said:

Yes, twisted candle means twisted nightmares.  In case you didn't find it when trawling through the archives I think this:

Is about the best overview of staging strategies.  Be warned though that it's four and a half years old so any mention of specific engines/fuel tanks/other parts is almost certainly irrelevant.  Analysis of the different strategies is permanently sound though.

BTW: The reason twisted candle doesn't usually compete with asparagus (or even onion) is that it relies on radial-attached engines, which means limited choice.  Using normal node-attached engines with the strategy instead requires building girder frameworks for them, which adds a lot of mass and drag.  Same thing for the fuel-lines which you have to run up the stack (although this may be changed by the fuel-priority system, I haven't tried it).

Aren't the aerodynamic issues an even bigger reason?  It's true there aren't all that many radial engines but you can get reasonable quality out of the ones that are there:  the Cub from MH has comparable performance to the Skipper in TWR and Isp. 

[Pecan]

5 hours ago, FinalFan said:

Aren't the aerodynamic issues an even bigger reason?  It's true there aren't all that many radial engines but you can get reasonable quality out of the ones that are there:  the Cub from MH has comparable performance to the Skipper in TWR and Isp. 

You're asking a hard question because I don't  think I've built a twisted candle for about 4 years!

Having said that, aerodynamics should favour the design as thinner than radial/onion/asparagus, not having the side-boosters that they do.  The drawback, as I noted, is the fuel-lines & supporting structure you have to build up the stack, in order to ensure tanks empty and can be dropped from the bottom up.  Also note I said "usually" and "limited choice" regarding the engines.  At certain mass-points the engines available may be good choices and twisted candle a viable option.  At others there's simply better choices for normal engines, so the comparison is not just between the staging strategies.

I might have a play with the candle to see what I can do ..

[magnemoe]

On 11/15/2018 at 9:24 PM, KerbService said:

@Pecan I looked at twisted candle... I am having nightmares now.

this uses serial staging as first stage, then an simplified twisted candle 9 LV-N in 2rd and 16 in 3rd stage, 3rd stage is also used to circulate and to push it to Minmus for refueling and drop down to LKO. 

One of the payloads, an base for Vall and an tylo capable lander who is SSTO with mining. 

[magnemoe]

On 11/15/2018 at 4:12 PM, Curveball Anders said:

There's also the minor detail that the glorified garden hose used to transfer insane amount of both LF and oxidiser hasn't shown up in RL yet ...

Don't use it anymore myself, stuff like 5 meter stacks, vectors and the shuttle srb made it redundant, you can cluster the SRB even reduce trust on one stack so they last longer. 
Still it has its uses, like rescuing kerbals midway between Moho and the sun.

Booster was refueled in orbit and used to eject this out of the solar system, a bit above 6km/s burn to match orbit then 6K for return. 

[Pecan]

9 hours ago, FinalFan said:

Aren't the aerodynamic issues an even bigger reason?  It's true there aren't all that many radial engines but you can get reasonable quality out of the ones that are there:  the Cub from MH has comparable performance to the Skipper in TWR and Isp. 

It turns out, as far as my testing went, that aerodynamics are only a problem regarding where to fit fins for drag/control.  That said, the best I got is only 21.48% payload ratio (which isn't at all bad but not as good as asparagus):

13.88 tonne payload, 64.6 tonne launch vehicle, total launch mass 78.484 tonne, 47 parts, 32,300 funds
Stage 1: Rocomax x200-16 fuel tank, twin thud engines (the cub is nothing like a Skipper, by the way! - 33.75kN vs 568.75) TWR 0.98, 1,048m/s dV
Stage 2: same as stage 1, TWR 1.36, 680m/s
Stage 3: same as stage 1 plus 2x AV-R8 Winglets, TWR 1.56, 501m/s
Stage 4: same as stage 3 plus X200-8 fuel tank, TWR 1.56, 571m/s
Stage 5: same as stage 4 without winglets, TWR 1.56, 447m/s

Stages 2 - 5 all have fuel line to cubic octagonal strut and then another to tank in stage above (unless there's a better way to drain tanks from the bottom).  I found the minimum twist was 6x shift-E rotations, by the way (hope that makes sense) - narrower engines (eg; spider) might allow for 5, I didn't try.

(3-stage asparagus for comparable 13.77 tonne payload is 23.35%, being a 52.25 tonne vehicle and is also fewer parts and cheaper)

As noted before, I really don't have much experience with twisted candle so I'd be interested to see any examples that do it better than this.

Edited November 17, 2018 by Pecan

[FinalFan]

14 hours ago, Pecan said:

It turns out, as far as my testing went, that aerodynamics are only a problem regarding where to fit fins for drag/control.  That said, the best I got is only 21.48% payload ratio (which isn't at all bad but not as good as asparagus):

13.88 tonne payload, 64.6 tonne launch vehicle, total launch mass 78.484 tonne, 47 parts, 32,300 funds
Stage 1: Rocomax x200-16 fuel tank, twin thud engines (the cub is nothing like a Skipper, by the way! - 33.75kN vs 568.75) TWR 0.98, 1,048m/s dV
Stage 2: same as stage 1, TWR 1.36, 680m/s
Stage 3: same as stage 1 plus 2x AV-R8 Winglets, TWR 1.56, 501m/s
Stage 4: same as stage 3 plus X200-8 fuel tank, TWR 1.56, 571m/s
Stage 5: same as stage 4 without winglets, TWR 1.56, 447m/s

Stages 2 - 5 all have fuel line to cubic octagonal strut and then another to tank in stage above (unless there's a better way to drain tanks from the bottom).  I found the minimum twist was 6x shift-E rotations, by the way (hope that makes sense) - narrower engines (eg; spider) might allow for 5, I didn't try.

(3-stage asparagus for comparable 13.77 tonne payload is 23.35%, being a 52.25 tonne vehicle and is also fewer parts and cheaper)

As noted before, I really don't have much experience with twisted candle so I'd be interested to see any examples that do it better than this.

I have a lot of trouble picturing your craft.  My attempt to recreate it did not have the right tonnage. 

In any case, I think you may have misunderstood my comment about the Skipper.  Clearly the Cub has much less raw thrust, but their performance relative to size is pretty similar: 
—Engine TWR ASL/VAC is 19.33/22.09 versus 19.12/22.66
—Isp ASL/VAC is 280/320 versus 270/320

In fact, you can slap 16 of them around a 2.5m stack and have almost the equivalent of a Skipper (with extreme vectoring) ... for 16,000 funds instead of 5,300, but c'est la vie. 

If cost is no object, I believe Cubs would be better than Thuds outside of the bottom stages.  There are two main difference between the Cub and the Thud: you need 3 Cubs to be as strong as 1 Thud; and the Cub has decent vacuum Isp.  The Cub also has much higher engine TWR.  3 Cubs also have the same VAC thrust but about 7 less ASL; and Isp that is a bit worse ASL but evens out at just 2400m. 

[Pecan]

2 hours ago, FinalFan said:

I have a lot of trouble picturing your craft.  My attempt to recreate it did not have the right tonnage.

Yer 'tis:

What tonnage do you get?  I tried 3x engines and things but couldn't find a comfortable match between payload, thrust and tankage that you need for payload ratio.  Plus, of course, more engines on each stage make it much harder to avoid exhaust hitting something on the lower stages.

I'd have preferred to use 1.5m tanks for this sort of payload but the thud's thrust is so low and so many stages required that that became ridiculously - and unflyably - long.  (It's also unflyable if you make the payload 2.5m)

Edited November 18, 2018 by Pecan

[FinalFan]

10 hours ago, Pecan said:

Yer 'tis:  [PICTURE NOT PICTURED]

What tonnage do you get?  I tried 3x engines and things but couldn't find a comfortable match between payload, thrust and tankage that you need for payload ratio.  Plus, of course, more engines on each stage make it much harder to avoid exhaust hitting something on the lower stages.

I'd have preferred to use 1.5m tanks for this sort of payload but the thud's thrust is so low and so many stages required that that became ridiculously - and unflyably - long.  (It's also unflyable if you make the payload 2.5m)

Thank you very much.  That makes much more sense than what I was imagining.  I was thinking of the old "hang things off the side" construction even though it's obviously not necessary with radially attachable engines. 

Three tries were not enough for me to replicate your success, no doubt due to pilot error.  I also had terrible issues with wobble from the upper portion.  I tried a version using a 1.875m payload and it seemed to be flyable, but I still did not immediately succeed in orbiting without help from the payload. 

My "what about Cubs instead?" version was easily able to do it with 1.875m and 2.5m payloads, replacing the 3rd, 4th, and 5th stage Thumpers with groups of 3 Cubs per Thumper.  Fins were not necessary; these tests produced some of the smoothest no-control gravity turns I've ever done.  (I discovered this when trying to counteract SAS oversteering; the Cubs have really outrageous gimbal for this vessel.)  Unfortunately uncontrolled gravity turns tended to not be shallow enough for best fuel efficiency so my successful tests were manually controlled.  In the end I set all the Cubs to 50% gimbal. 

Initial tests made orbit with between 14% (1.875m) and 19% (2.5m) fuel remaining in the last launch stage.  However, there was a problem:  My 13.88t payload vessel had less launch weight than your vessel, biasing the test in favor of the Cubs.  I then adjusted it for a 76.323t total weight (14.683t payload), with 2.16t difference due to the Cubs weighing less than the Thumpers they replaced and 0.001t lost due to inaccuracy.  It made orbit with 13.5% fuel remaining.  The Cubs' weight advantage may have helped, but the biggest difference was probably the Isp advantage. 

[Pecan]

Congratulations.  Presumably you had to twist the candle more to fit 3 engines in place of each of mine - which would have had the exhaust from the topmost thuds hitting the bottom ones.

If your launch mass is less for the same payload then that's perfect, after all the point is to maximise payload ratio.  I'm convinced though that, like multi-layer asparagus, any theoretical efficiency with twisted candle is lost to practical construction restrictions and difficulty. 

Edited November 20, 2018 by Pecan