Building a rocket bottom up and the how to's?

[LN400]

After countless repetitions of the way I always built rockets (Payload then custom built rockets for each payload) I decided to try it somewhat differently for the challenge:

Build the rocket to maximize the capabilities of each engine in the game (+ parts mods like KW, Space Y and MRS), then when I do have the payload, pick the pre-built launch vehicle that will get the job done with as little fuel left as possible when it was done doing its thing.

For testing, I choose a 2 stage rocket that will put the payload short of orbit so the 2nd stage will fall back to under 20km to burn up.

Before building I decided to try to vary these parameters for each stage: Starting TWR and dv.

 

Once the TWR has been decided and the engine is picked, I an work out the starting (wet) mass of the whole circus:

m_w = thrust_ASL / (TWR_start * 9.80665) (eq.1)

 

To figure out the fuel for the desired dv for that stage, I find the ratio wet/dry mass using the ASL value of I_sp

r = e^{dv/(I_sp * 9,80665)} (eq.2)

The dry mass for the whole rocket is then

m_d = m_w / r (eq. 3)

and the fuel mass is then

m_f = m_w - m_d (eq. 4)

9/20th of the fuel mass is liquid fuel and since a unit of fuel has the mass of 0.005 t, I find the number of units liquid fuel (l) by

l = 9/(20 * 0.005) m_f = 90 m_f (eq. 5)

Using procedural tanks, I custom build a tank to hold as close to l as possible, rounded up.

 

The first stage is: 1 decoupler, 1 tank, 1 engine, 4 fins. After building the stage, I use KER to quickly find the mass of the stage (m_s).

By subtracting m_s from m_w I find what the stage can lift (m_p) to get the desired TWR and dv.

m_p = m_w - m_s (eq.6)

 

m_p is in turn the new wet mass for the 2nd stage its payload.

The process is repeated to build the 2nd stage to find the payload mass that will give both stages the designed TWR and dv except now I use the vac values for thrust and I_sp.

After repeating the calculations, the 2nd m_p will be the actual payload's mass.

 

One thing that complicates things is that each payload has a different drag profile unless under fairings but then again, even the fairings have different drag profiles (modders, how about a mod for fairings of fixed sizes and shapes in addition to the already existing proc. fairings?), so for the time being, the payload will be a capsule + parachute + weights to reach the target wet mass, trying to keep the drag profile the same for all.

 

For my first 1.25m rocket using a T45 for the lifter and a 909 for the 2nd stage, 1700 m/s ASL to around 25km altitude and 1700 m/s vac did not quite make it to space but I am still working on the ascent technique and a good TWR.

 

Now, how about you? Are you like me in that you always have built the rocket around the payload, or are you building rockets for the stable and pick the right one for the job? If so, how do you go about it?

 

 

[Geschosskopf]

1 hour ago, LN400 said:

Now, how about you? Are you like me in that you always have built the rocket around the payload, or are you building rockets for the stable and pick the right one for the job? If so, how do you go about it?

I always build top-down.  However, I'm also now using BARIS so it pays to at least keep re-using the same booster parts over and over.  Thus, although my lifters are all built for the specific payload, they're often similar enough for the same general size of payload that I might as well start naming them.

When I was 1st starting in KSP, I tried to build standardized launch vehicles but decided it wasn't worth the trouble (especially as that was before subassemblies were really a thing).  

[LN400]

4 minutes ago, Geschosskopf said:

[...]I'm also now using BARIS[...]

Nice! Never heard of it, looked it up and it does appear at first look to be something I have fantasized over the last couple of weeks! Thanks tons for the headsup!

[Aegolius13]

I have a set of expendable launch rigs for getting particular mass ranges to LKO.  Also various classes of reusable SSTO rockets for different payload ranges. 

But I still mostly build top down, v especially for one-off mission profiles.   Even though I've done it a zillion times at this point, I still find it oddly satisfying to pick out just the right engine and tankage for each stage. 

[Geschosskopf]

1 hour ago, LN400 said:

Nice! Never heard of it, looked it up and it does appear at first look to be something I have fantasized over the last couple of weeks! Thanks tons for the headsup!

BARIS has a HUGE amount of customization.  It can be everything from a minor annoyance to a total nightmare, and everything in between.  Choose wisely  

[Spricigo]

I find more easy to design my rockets in the order I use the stages, last to first. And that don't prevent the possibility of an assortment of pre-built LV for a variety of different payloads.

 

3 hours ago, LN400 said:

One thing that complicates things is that each payload has a different drag profile unless under fairings but then again, even the fairings have different drag profiles (modders, how about a mod for fairings of fixed sizes and shapes in addition to the already existing proc. fairings?), so for the time being, the payload will be a capsule + parachute + weights to reach the target wet mass, trying to keep the drag profile the same for all.

You may try making the fairing part of each LV, as in: Lifter I is to payload up to 1t that fit in the fairing, for more bulky payload there is Lifter I-B (which may, of course,  have other differences  not only the fairing). The fairing may be 'too big' for some payloads but having always the same aerodynamic profile may avoid those slight flight profile corrections and make the LV more easy to use in the long run.

Edited October 30, 2017 by Spricigo

[OhioBob]

5 hours ago, LN400 said:

Now, how about you? Are you like me in that you always have built the rocket around the payload, or are you building rockets for the stable and pick the right one for the job? If so, how do you go about it?

My techniques have evolved with time.  Back during the time of v0.90 I did things much the way that you describe.  I pre-build rockets around certain engines and/or clusters of engines and saved them as subassemblies according to lifting capacity.  If a wanted to launch a payload of, say, 10 tons, I would just grab the smallest rocket capable of lifting 10 tons.

I've since gone away from that.  I now custom build each launcher for each payload, but I adhere as closely as possible to a specific set of guidelines.  By following the guidelines I can slap together a launcher pretty quickly, and I get something that's going to give pretty consistent performance.

As far as my guidelines are concerned, the launcher is just the part that gets the rest of the rocket/payload to low orbit.  So any upper stage used to eject the spacecraft into another orbit is classified as payload for the purposes of design.  I also throw in the mass of any fairings, decouplers, etc.  So "payload" is everything the sits on top of the launcher at liftoff.  The guidelines are pretty simple:

• 2nd (upper) stage:  Propellant mass = mass of payload; TWR = 1.0 to 1.2.
• 1st (bottom) stage:  Propellant mass = 2 * mass of payload; TWR = 1.2 to 1.5.

It's not always possible to hit those marks exactly because there are only a finite number of parts to work with.  I just get it as close as I can and then make whatever adjustments are necessary to make it work.  If there is no 1st stage engine available that will hit within the target TWR range, I'll usually go small and then add a couple SRBs to augment liftoff thrust.

[LN400]

30 minutes ago, OhioBob said:

It's not always possible to hit those marks exactly because there are only a finite number of parts to work with.  I just get it as close as I can and then make whatever adjustments are necessary to make it work.  If there is no 1st stage engine available that will hit within the target TWR range, I'll usually go small and then add a couple SRBs to augment liftoff thrust.

This is one reason I use procedural tanks and not the standard sized ones. It does mean the rocket won't be the cheapest possible but here the entire project is to break up the monotony, and have some fun with maths while I'm at it. I have the door open for SRB for those extra kilonewtons and m/s but since I need to learn a lot here, I think I'll keep it simple until I get the hang of these single engine rockets.

Quote

• 2nd (upper) stage:  Propellant mass = mass of payload; TWR = 1.0 to 1.2.
• 1st (bottom) stage:  Propellant mass = 2 * mass of payload; TWR = 1.2 to 1.5.

This is quite an interesting bit of info there. I'll have to play around with that and see how that goes.

[OhioBob]

8 minutes ago, LN400 said:

This is quite an interesting bit of info there. I'll have to play around with that and see how that goes.

It would be interesting to see how closely my simply rules of thumb line up with your calculations.  The guidelines aren't just random numbers that I pulled out of the air.  I did quite a few calculations and optimization tests.  From that I tried to boil it down to the simplest set of rules that I could easily remember.

 

Edited October 30, 2017 by OhioBob

[OhioBob]

@LN400, it's pretty easy to do a back of the envelope check of my design guidelines.  If we assume each stage is 75% propellant and 25% inert mass, then we can break the whole launch vehicle down into simple proportions:

• Payload = 3 parts
• 2nd stage = 4 parts (3 parts fuel)
• 1st stage = 8 parts (6 parts fuel)
• TOTAL = 15 parts

Specific impulse varies, but let's assume an average of 320 seconds.  Then the delta-v is:

• 1st stage = 320 * 9.81 * LN(15/9) = 1604
• 2nd stage = 320 * 9.81 * LN(7/4) = 1757
• TOTAL = 1604 + 1757 = 3,361 m/s (vacuum)

Which is, or course, just about what it takes to get to orbit.
 

Edited October 31, 2017 by OhioBob

[LN400]

Using my equations, for the T45 engine, TWR = 1.5 and dv first stage = 1700 m/s ASL value, I do get a wet to fuel ratio of pretty close to 2, like really close down to the 4th decimal. For the 2nd stage using a 909, TWR = 1.79, dv = 1700 m/s vacuum value, I get a wet to fuel ratio of 2.53-ish to maintain the wet mass, TWR and dv for the first stage.

To punch in the figures:

I want a starting TWR of 1.5 and a first stage dv of 1700 m/s ASL. At sea level, the T45 has the following specs:

Thrust: 167.969 kN, Isp 250s.

That gives a starting wet mass of the whole tabernacle of 11.4187 t.

To get a dv of 1700, I need a wet to dry ratio of 2 + a tiiiny bit  which gives a dry mass of 5.7079 t which in turn gives me the fuel mass of 5,7108 meaning 514,0 units of LF and 628.2 units of LO.

Decoupler, engine, 4 fins, fuel and the tank has a combined mass of 8.009 t which means it can carry 3.4097 t. That will be the wet mass for the 2nd stage.

The 909 has the vacuum specs of

60 kN, 345s.

To get 1700 m/s, I need 1.3467 t of fuel. Decoupler, engine, fuel and tank has a combined mass of 2.058t which gives the payload mass of 1.3517t or about 11.8% of the full mass of the entire thing.

 

Edited October 31, 2017 by LN400

[OhioBob]

13 minutes ago, LN400 said:

Using my equations, for the T45 engine, TWR = 1.5 and dv first stage = 1700 m/s ASL value, I do get a wet to fuel ratio of pretty close to 2, like really close down to the 4th decimal. For the 2nd stage using a 909, TWR = 1.79, dv = 1700 m/s vacuum value, I get a wet to fuel ratio of 2.53-ish to maintain the wet mass, TWR and dv for the first stage.

I find that my rules generally fall a little short for 1.25-meter rockets.  I usually have to bump up the proportion a propellant just a bit.  But my rules generally hit right about on the money for 2.5-meter rockets.

[IncongruousGoat]

It depends. If I'm playing stock, it's easy enough to slap together a launcher that I mostly use custom launchers. If I'm playing RSS/RO/RP-0, however, I start building launcher families, just because of how much work it is to design a functional LV.

[Superfluous J]

I tried this a while ago when I was using the Routine Mission Manager mod. I made multiple launchers for getting different sized payloads to orbit, and then launched them to prove them out, and then used them to skip that whole "getting to orbit" part of getting to orbit.

It was fine, I suppose, but I had this terrible aversion to launching a 9 ton payload into orbit with a launcher rated for 10 tons (for example). Seemed I was wasting money, even though money is never really a problem in career. So I ended up making custom launchers all the time and the act of proving them turned into more of a chore than just launching the darn thing into orbit manually.

Edited October 31, 2017 by 5thHorseman

[Tyko]

15 hours ago, 5thHorseman said:

It was fine, I suppose, but I had this terrible aversion to launching a 9 ton payload into orbit with a launcher rated for 10 tons (for example). Seemed I was wasting money, even though money is never really a problem in career. So I ended up making custom launchers all the time and the act of proving them turned into more of a chore than just launching the darn thing into orbit manually.

A big way KSP deviates from real world is that there aren't huge up front R&D costs to building a successful launcher. A large percentage of the cost of the first launch of the rocket is the development. Also reliability tends to improve with multiple iterations. So, in RL it makes total sense to use a 10 ton lifter to lift a 9 ton payload. Launchers like the Atlas V save money by using a common core and then varying the number of SRBs depending on payload.

In KSP, without BARIS or KCT/UPFM or any other failure mods, I tend to somewhat standardize my launchers because my limiting factor is my gaming time - repeatedly building the same launcher is not very exciting. I'll build a standardized launcher but use at least one procedural tank in each lifter stage so I can easily vary the fuel capacity/weight to account for different payloads. This seems to work pretty well and makes me feel like I'm somewhat following RL practices without needing the extra mods.

Edited October 31, 2017 by Tyko