# How big should your stages be?

## Recommended Posts

It's not effective, either to have a single gigantic tank for Tylo and back, nor having two hundred separate stages for a trip to Mun. Is there a formula to calculate the minimum fuel to dead weight ratio for each stage in order to have a net delta-v gain for adding that stage?

##### Share on other sites

I want to say that number of stages is not amenable to an analytic solution, so you need to iterate over a few designs. For a two (or I suspect N) stage design you can figure out the optimal fuel distribution with some constraints.

There are some somewhat messy formulae in the stage sizing link in my signature that are hopefully related. (In general you want larger mass ratios in the high Isp stages, but can easily hit TWR limits)

Edited by UmbralRaptor
##### Share on other sites

"Is there a formula to calculate the minimum fuel to dead weight ratio for each stage in order to have a net delta-v gain for adding that stage?"

Generally, every stage you add will increase net dV, assuming of course that you are adding that stage to a prexisting craft. It may be far from optimal though...

It also depends on the configuration of the staging... are they serial stages (like a Saturn V), concurrent stages (Like the SRBs vs the SSMEs+ET on the STS), or crossfeeding designs. Is the stage simple drop tanks?

If going with drop tanks, in KSP it is very very very advantageous to have many staging events, as long as decoupler mass << jettisoned tank mass.

However, part count can be prohibitive there (and decouplers can be rather expensive). Not to mention that one may want to make re-usable/coverable craft.

In the case of reusable, stages must be landed or in space in order to not be deleted. In my designs, I've done:

A very high TWR, short burn boosters that jettison after a few seconds and can land on parachutes before they get outside of the active craft's physics bubble (wasn't vibale until the atmospheric physics bubble size was expanded)

A suborbital booster stage, which hangs around apoapsis long enough for the next stage to get to orbit so I can switch back to the booster (I suppose I could have multiple such booster stages, with different apoapses (apoapsii?), but managing them all would be too much of a pain)

A stage to put a payload at near escape velocity from kerbin.

A capture into an elongated orbit stage... etc.

Basically, when going reusable but maximizing efficiency, I aim to stage often, but leave the stages in stable orbits that don't cross SOI boundaries.

##### Share on other sites

My personal rule of thumb is 2000m/s dv per stage, with a ridiculously wide margin of +/- somewhere near 100%

##### Share on other sites

Mines are just casual things.
The first stage has the most fuel, second (and above) stages have less fuel, depending on the weight of the payload.
But for Jool missions, you need a lot of fuel, BUT if your stage has only less than 20 tons in weight, a few ion engines should do
A rocket with a payload, for instance, only has 2.1 tons and will go to Jool, less fuel is required.

The heavier = the more fuel

##### Share on other sites

I concur with SuperfluousJ; right about 2 km/sec DV per stage. This can be very flexible and still show good results,  so I alter it as necessary to give me convenient staging points in the mission.

Best,

-Slashy

##### Share on other sites

Rocket stages (sequential) with equal DV per stage form a roughly pyramid shape, where the bottom stage is x-times heavier than everything above it.

Now build a stage below it with a mass of 2t (the total rocket mass will now be 3t)
Next, build a stage below all of this. The "payload" for this stage, is 3t (everything above it), so this new bottom stage needs to be 6t
The total rocket mass is 9t at this point

This followed a 2x rule of thumb, it works alright at getting good dv per funds, but you can obviously use a different number

• NERV/Ion stages will definitely be better with a different number
• Tylo landers, Eve ascent vehicles will need way more engine/thrust

This makes designs really simple, but once you are in orbit, drop tanks will out perform anything else...unless you have mission requirements for re-usability
(i got to the 2x value by eyeballing some wet/dry mass vs DV graphs, i think anything around 2-3x seemed good for Terrier engines, but it is worth experimenting with different ratios and techniques)

##### Share on other sites

So, what do I do to save stage mass? I am at the point where I am using several Mastodon engines as mere RCS! And the rocket is at 32 kilotons, and it needs clamps lest the pad explode. Again, 200 clamps are there!

##### Share on other sites

2 minutes ago, Single stage to ocean said:

So, what do I do to save stage mass? I am at the point where I am using several Mastodon engines as mere RCS! And the rocket is at 32 kilotons, and it needs clamps lest the pad explode. Again, 200 clamps are there!

Cut down on dry-mass. This will reduce the amount of fuel + engines you need to lift that into orbit.

Making the final deployed payload - the satellite portion - 25% lighter will also make it possible to remove 25% of the rest of the rocket and still reach orbit. Unfortunately many kerbal designs are, well kerbal, and are not as simple as 3 stages and a satellite. At which point it helps more to simplify the mission instead.

##### Share on other sites

On 7/19/2021 at 2:24 AM, he77789 said:

Is there a formula to calculate the minimum fuel to dead weight ratio for each stage in order to have a net delta-v gain for adding that stage?

As has been said several times, every stage you add will increase overall dV unless it has a TWR <1 for its full burn while still on the ground.  For my part, when I'm designing rockets I have two key considerations in mind: the ultimate dV required to deliver the payload to all its destinations and the TWR required to get through each of the phases of that process efficiently.  In general, the more TWR you have in a given stage the less dV that stage will have, so I try to only have a high TWR when I really need it, which is for takeoff and landing, and use a lower TWR for all other phases of the mission.

So to illustrate,  let's say I'm doing a Tylo flag-planting mission with a Tylo orbit rendezvous design.  I'll start by building my lander ascent stage so that it has a TWR of around 1.4 (on Tylo) at takeoff, maintaining that or higher TWR for at least the first third of the  total dV required to reach orbit, which for Tylo  would be something like 800 m/s . For the second third, I would start at a TWR of around 1, topping out at maybe 1.4 again, and for the last third I might start with a TWR as low as 0.5, topping out around  0.8.  To this I would then add a descent stage that has the same overall dV as the ascent one, but starts at a TWR of at least 0.5 and ends at a TWR of at least 1.4. Once I have that built, I would assemble that package to a transfer stage that has enough dV  to take the full lander from LKO to LTO and then just the empty ascent stage from LTO back to a Kerbin encounter, with a TWR that remains somewhere between 0.2 and 0.4 for its whole journey.  Usually I'll do this by assembling my near-empty Tylo ascent stage to the 2.5m core of the transfer stage with enough tanks and Nervs on the bottom for a TWR of 0.25 and something like 2.8 km/s  dV in that configuration. Then I will fill the tanks and add my descent stage to that, and take note of the dV with those added.  I'll then start adding paired, asparagused side stacks of Mk1 fuselages and Nervs to this, until I've added another 2.8 km/s of dV to whatever the display showed before I started adding them. In this process, I'll try to keep the TWR between 0.2 and 0.4 for the whole burn time of each asparagus stage, setting it up so that each stage will run out of fuel right as the overall TWR reaches 0.4.  Once all that's built, I'll usually mount it on top of a 3.75m core stage with a Mammoth engine that has a starting TWR of right around 1 and maybe 1.8km/s total dV.  I will then put an asparagused pair of Mainsails on the sides of that, so that I'm adding another 700 m/s or so of dV at a starting TWR of maybe 1.2, and then finish it with a pair of giant SRBs to add another 700-800m/s of dV at a starting TWR of >=1.4.

Anyway, that was rather a complex narrative, but if you follow that sequence you should end up with a vessel that can get the job done with a reasonable if not perfectly optimal level of both efficiency and flyability, and without getting fancy about gravity assists.

Edited by herbal space program
##### Share on other sites

1 hour ago, Single stage to ocean said:

So, what do I do to save stage mass? I am at the point where I am using several Mastodon engines as mere RCS! And the rocket is at 32 kilotons, and it needs clamps lest the pad explode. Again, 200 clamps are there!

Uhhhm, don't build such a ridiculously gigantic ship? But seriously, the same rules of thumb wrt TWR and dV apply on all scales.

##### Share on other sites

From remembering my past readings in astronautics texts, to maximize performance, you want each stage to deliver about the same delta-V, which for engines of the same Isp would mean the same mass ratios between stage ignition and cutoff.   Other considerations often lead to the 1st stage having slightly greater mass ratio that the 2nd stage, primarily because even with the same engines, the Isp of the atmosphere flight of the 1st stage is lower than the mostly vacuum Isp of the 2nd stage.  Another productive change of a slightly bigger 1st stage is to move the staging event to farther along the ascent track.

I found for most Kerbal orbital launch vehicles, I could make a 2-stage design with about equal delta-V, with the mission vehicle itself being a lower delta-V 3rd stage mostly for parking orbit circularization.  This meant that the 2nd stage didn't end up in orbit as debris, which was handy.

Standard Kerbal stage dry masses are much larger that in our real world.  That means the fixed dry mass for each stage burn is worse in KSP.  These larger fixed costs makes a 3-stage design usually worse than a similar size/mass/cost 2-stage design.

##### Share on other sites

3 hours ago, Jacke said:

From remembering my past readings in astronautics texts, to maximize performance, you want each stage to deliver about the same delta-V, which for engines of the same Isp would mean the same mass ratios between stage ignition and cutoff.   Other considerations often lead to the 1st stage having slightly greater mass ratio that the 2nd stage, primarily because even with the same engines, the Isp of the atmosphere flight of the 1st stage is lower than the mostly vacuum Isp of the 2nd stage.  Another productive change of a slightly bigger 1st stage is to move the staging event to farther along the ascent track.

I found for most Kerbal orbital launch vehicles, I could make a 2-stage design with about equal delta-V, with the mission vehicle itself being a lower delta-V 3rd stage mostly for parking orbit circularization.  This meant that the 2nd stage didn't end up in orbit as debris, which was handy.

Standard Kerbal stage dry masses are much larger that in our real world.  That means the fixed dry mass for each stage burn is worse in KSP.  These larger fixed costs makes a 3-stage design usually worse than a similar size/mass/cost 2-stage design.

The first part of this may be roughly true for ascent stages, but I think transfer stages with their much lower TWR requirements should pretty clearly have more dV than the ascent stages for best performance.  As to the last part, it kind of depends on what you consider a stage. Asparagus staging is pretty clearly the most mass-efficient way to get to orbit, because it allows you to shed superfluous dry mass at the highest rate that your constraints of TWR will allow. My ascent stacks therefore typically have 4-6 side boosters that are dropped in pairs from a core stage so that the overall TWR starts pretty high and gets lower as the ascent progresses and gravity losses become less of a factor.

##### Share on other sites

10 hours ago, herbal space program said:

The first part of this may be roughly true for ascent stages, but I think transfer stages with their much lower TWR requirements should pretty clearly have more dV than the ascent stages for best performance.

i don't think i ever used a transfer stage; only drop tanks.

the difference between the two is that i can use a single engine for the whole rocket, while a transfer stage would require multiple engines.

on the other hand, the lander still needs its own, high thrust engine. I guess my standard architecture of using a mothership to carry around a lander does just that; a mothership is virtually the same thing as a transfer stage. except mothership implies you're using it for more than just a transfer. i always run missions to multiple planets - which is something else that improves efficiency, if you're going to jool you only have to add 500 m/s to also visit duna - so i call it a mothership. and when the mothership needs more than 4000 m/s for the multiple destinations, i achieve it with drop tanks.

##### Share on other sites

11 hours ago, herbal space program said:

The first part of this may be roughly true for ascent stages, but I think transfer stages with their much lower TWR requirements should pretty clearly have more dV than the ascent stages for best performance.  As to the last part, it kind of depends on what you consider a stage. Asparagus staging is pretty clearly the most mass-efficient way to get to orbit, because it allows you to shed superfluous dry mass at the highest rate that your constraints of TWR will allow. My ascent stacks therefore typically have 4-6 side boosters that are dropped in pairs from a core stage so that the overall TWR starts pretty high and gets lower as the ascent progresses and gravity losses become less of a factor.

If I'm getting it right what you mean by transfer stages, to me they're part of the mission spacecraft and their delta-V requirements are based upon what the mission needs.  There's going to be staging of a large enough mission spacecraft and for them, splitting that craft into roughly the same delta-V stages should maximize its performance.  However, again, KSP's large dry mass means staging is less advantageous and the size of a spacecraft that will actually benefit from being split into 2 stages is larger.  And just for mission simplicity, I prefer to adjust mission spacecraft stages to match the expected burn delta-V's events, so that staging happens between two of them, not during one, as moving off the ideal split isn't that much of a penalty.

##### Share on other sites

3 hours ago, Jacke said:

If I'm getting it right what you mean by transfer stages, to me they're part of the mission spacecraft and their delta-V requirements are based upon what the mission needs.  There's going to be staging of a large enough mission spacecraft and for them, splitting that craft into roughly the same delta-V stages should maximize its performance.  However, again, KSP's large dry mass means staging is less advantageous and the size of a spacecraft that will actually benefit from being split into 2 stages is larger.  And just for mission simplicity, I prefer to adjust mission spacecraft stages to match the expected burn delta-V's events, so that staging happens between two of them, not during one, as moving off the ideal split isn't that much of a penalty.

What I call a transfer stage is the whole part of the ship that does all the pushing from LKO to low orbit of whatever the target body or bodies are.   It can have any number of engines, but they will always all be firing together, as that is clearly the most efficient way to do things.  As my TWR gets higher than what I need, I'll generally drop them in pairs to trade that excess TWR for more dV.  So if I have a core stack surrounded by three pairs of asparagused side boosters, is that one stage or four?   I think what you are talking about only really applies to stages that are stacked on top of each other, which is quite clearly not the most efficient way to do things in general because the upper stages are all dead weight until they fire. Which raises the question of what exactly do you mean by maximizing performance anyway? What I am talking about is optimizing the tradeoff between having enough TWR to  get off the ground and do your other burns relatively efficiently, and also having enough dV to get where you need to go.  Any other measure of "performance" is pretty much only of academic interest in my book.

##### Share on other sites

14 minutes ago, herbal space program said:

What I call a transfer stage is the whole part of the ship that does all the pushing from LKO to low orbit of whatever the target body or bodies are.

At least in KSP, I think of that as part of the mission spacecraft, as that's the thing that can change atop a more standardized launch vehicle.  I really don't bother to try to do any staging optimization of it, as mission step subdivision of the spacecraft is usually more important.  But calling the 1st stage of most of the spacecraft the transfer stage is a good label.

Example:

For Mun I've done things like put a return reentry body carrying experiments I want to return, like Mystery Goo, on top of a Mun orbital stage with other experiments I can transmit (this was using Better Than Starting Manned, where there was more orbital science that could be transmitted for 100% return).  Launch it into LKO, final orbital circularization using a bit of burn from the mission spacecraft.  Then burn for a free-return trajectory to Mun.  Approaching Mun, I'd separate the return body from the orbiter.  Return body would run the Mystery Goo experiments, then I'd jump back to the orbiter, burn to switch to an orbital pass, then burn into Mun orbit.  The return body would be headed back to Kerbin and I'd just have to do a course correction; it usually only had RCS but with enough delta-V to cover that maneuver, aiming for a good re-entry pericenter altitude; it would often dump the RCS blocks and tank revealing the heatshield to minimize final landing mass (because BTSM used Deadly Reentry).  The orbiter would take Mun orbital science and transmit it back to Kerbin.

A common Mun lander design mission spacecraft would have its transfer stage cover the trans-Mun injection burn and the Mun-orbit insertion burn, with any reserve used to start the descent to landing.  The lander descent stage would be used down to Mun landing.  The ascent stage would burn to Mun orbit, including circularization, then the Mun escape burn to return to Kerbin.  After course correction and just before re-entry, like the previous flyby return body, I'd stage to drop the engine and tanks as well as RCS to minimize the final landing mass.

##### Share on other sites

7 hours ago, Jacke said:

At least in KSP, I think of that as part of the mission spacecraft, as that's the thing that can change atop a more standardized launch vehicle.  I really don't bother to try to do any staging optimization of it, as mission step subdivision of the spacecraft is usually more important.  But calling the 1st stage of most of the spacecraft the transfer stage is a good label.

Example:

For Mun I've done things like put a return reentry body carrying experiments I want to return, like Mystery Goo, on top of a Mun orbital stage with other experiments I can transmit (this was using Better Than Starting Manned, where there was more orbital science that could be transmitted for 100% return).  Launch it into LKO, final orbital circularization using a bit of burn from the mission spacecraft.  Then burn for a free-return trajectory to Mun.  Approaching Mun, I'd separate the return body from the orbiter.  Return body would run the Mystery Goo experiments, then I'd jump back to the orbiter, burn to switch to an orbital pass, then burn into Mun orbit.  The return body would be headed back to Kerbin and I'd just have to do a course correction; it usually only had RCS but with enough delta-V to cover that maneuver, aiming for a good re-entry pericenter altitude; it would often dump the RCS blocks and tank revealing the heatshield to minimize final landing mass (because BTSM used Deadly Reentry).  The orbiter would take Mun orbital science and transmit it back to Kerbin.

A common Mun lander design mission spacecraft would have its transfer stage cover the trans-Mun injection burn and the Mun-orbit insertion burn, with any reserve used to start the descent to landing.  The lander descent stage would be used down to Mun landing.  The ascent stage would burn to Mun orbit, including circularization, then the Mun escape burn to return to Kerbin.  After course correction and just before re-entry, like the previous flyby return body, I'd stage to drop the engine and tanks as well as RCS to minimize the final landing mass.

It all really depends on what your constraints are. Early in the (career) game, part count is a major limitation, so landers that can make it all the way from LMO to landing to home are the best. Later in the game, having autonomous transfer stages, aka "motherships", that drop off and pick up landers at various target bodies is clearly the most efficient way to go. When I was milking Mun for all its science in my current career game, I had an orbiting  transfer stage/mothership module that refueled the lander module four or five times so that it could collect surface science from different biomes.

##### Share on other sites

I don't know if this answers your question exactly, but when I come up with the staging used for my rockets, I usually try to target a similar initial TWR for each stage (1.0-1.5 for lower stages, 0.6-1.0 for upper stages, and >0.3 for vacuum stages), while trying to ensure each stage has a similar amount of Delta-V (for stock parts, around 2500-3000 m/s per stage does the trick for me). Note that the provided numbers are for RSS, where lots of vacuum delta V is needed, but the general strategy remains - ignore mass, prioritize TWR and Delta-V.

Edited by Beriev
##### Share on other sites

12 hours ago, Beriev said:

I don't know if this answers your question exactly, but when I come up with the staging used for my rockets, I usually try to target a similar initial TWR for each stage (1.0-1.5 for lower stages <snip>

I don't think it's ever good for a rocket to have a TWR of less than around 1.4 at takeoff. You'll just be throwing a bunch of  dV down the gravity hole. In fact, I'd say my most efficient lifters generally start at around 1.8. With the aero model as it is, what you shave off in gravity losses that way far outweighs what you'll lose to drag in the lower atmo.  If you've hit it right, your limiting factors should be staying in control through max Q, followed immediately by not burning up before you get above 25km.  At least in Stock. I have no idea how RSS might alter that calculation.

Edited by herbal space program
##### Share on other sites

On 7/22/2021 at 1:12 PM, herbal space program said:

I don't think it's ever good for a rocket to have a TWR of less than around 1.4 at takeoff. You'll just be throwing a bunch of  dV down the gravity hole. In fact, I'd say my most efficient lifters generally start at around 1.8. With the aero model as it is, what you shave off in gravity losses that way far outweighs what you'll lose to drag in the lower atmo.  If you've hit it right, your limiting factors should be staying in control through max Q, followed immediately by not burning up before you get above 25km.  At least in Stock. I have no idea how RSS might alter that calculation.

It is occasionally good. When using SRBs, the thrust tends to run away as they get close to empty. My SRB designs typically leave the pad at 1.2g to compensate for this. My liquid fuel boosters typically leave the pad at 1.4G in order to minimize overacceleration in the lower atmosphere, which leads to aerodynamic instability and aggravated cosine losses. I find these to be bigger dV wasters than gravity losses, but YMMV.

Best,

-Slashy

##### Share on other sites

On 7/22/2021 at 6:12 PM, herbal space program said:

I don't think it's ever good for a rocket to have a TWR of less than around 1.4 at takeoff. You'll just be throwing a bunch of  dV down the gravity hole. In fact, I'd say my most efficient lifters generally start at around 1.8. With the aero model as it is, what you shave off in gravity losses that way far outweighs what you'll lose to drag in the lower atmo.  If you've hit it right, your limiting factors should be staying in control through max Q, followed immediately by not burning up before you get above 25km.  At least in Stock. I have no idea how RSS might alter that calculation.

5 hours ago, GoSlash27 said:

It is occasionally good. When using SRBs, the thrust tends to run away as they get close to empty. My SRB designs typically leave the pad at 1.2g to compensate for this. My liquid fuel boosters typically leave the pad at 1.4G in order to minimize overacceleration in the lower atmosphere, which leads to aerodynamic instability and aggravated cosine losses. I find these to be bigger dV wasters than gravity losses, but YMMV.

Certainly in the real world, rockets like the Saturn V lifted off with as low an initial TWR of 1.2.  It had sufficient control authority with 4 of the F-1 engines gimballing (the fins would have little control at low speed).

I think the problem is that it's harder to control rockets with a low TWR in KSP, although not impossible.  Using an ascent autopilot with finer control shows that lower initial TWR down to 1.2 is quite possible.

I think gravity losses are taken far too seriously.  Drag losses are greater but for orbital craft, both are small compared to the delta-V just needed to get up to orbital speed, even in stock KSP.  I find a greater problem is to get the ascent trajectory right.  It's very easy to either push the apoapsis too high when at too low an altitude or overspeed at too low an altitude.  The high dry mass of KSP stages makes it difficult to go with more than 2 stage without losing performance.

I've found my launch TWR going with about 1.2 to 1.4 (though it's been a while).  Sometimes the right size of SRB burning at the same time as the main stage helps push that TWR higher until the propellants burn off.

##### Share on other sites

12 hours ago, Jacke said:

Certainly in the real world, rockets like the Saturn V lifted off with as low an initial TWR of 1.2.  It had sufficient control authority with 4 of the F-1 engines gimballing (the fins would have little control at low speed).

Well I looked it up just now and it was actually 1.25 at launch, but that was for reasons that have nothing to do with the most mass or energy-efficient way to get to orbit. The F1 engines  could not be throttled, so they were like SRBs are in this game. If you put that together with the totally un-KSP mass ratios in real life, you've got a rocket for which the chief limiting factor is too much acceleration at the end of the burn.  They dealt with this on the Saturn V by shutting off the center engine when they reached 4g, since they couldn't throttle back.  As @GoSlash27 pointed out (and I acknowledge), that sort of situation is pretty much the one case where it  actually makes sense to have rocket take off with a TWR as low as 1.2. The STS took off at a much higher TWR of ~1.75 1.6, because that is more efficient  and the higher payload fraction and the ability to throttle the RS-25 engines mitigated against too much acceleration at the end of the burn.

12 hours ago, Jacke said:

I think the problem is that it's harder to control rockets with a low TWR in KSP, although not impossible.  Using an ascent autopilot with finer control shows that lower initial TWR down to 1.2 is quite possible.

I've taken off just fine with rockets that barely lifted off, so that is definitely not an issue for me.

12 hours ago, Jacke said:

I think gravity losses are taken far too seriously.  Drag losses are greater but for orbital craft, both are small compared to the delta-V just needed to get up to orbital speed, even in stock KSP.  I find a greater problem is to get the ascent trajectory right.  It's very easy to either push the apoapsis too high when at too low an altitude or overspeed at too low an altitude.  The high dry mass of KSP stages makes it difficult to go with more than 2 stage without losing performance.

I have tested this pretty exhaustively, both on Kerbin and on Eve, and based on that and a number of discussions in this forum, I think what you say here is just not true. Unless you push your TWR to implausible extremes or fly an excessively flat ascent profile,  gravity losses far outweigh drag losses for a well-designed craft. You just have to keep it pointed prograde the whole time.

12 hours ago, Jacke said:

I've found my launch TWR going with about 1.2 to 1.4 (though it's been a while).  Sometimes the right size of SRB burning at the same time as the main stage helps push that TWR higher until the propellants burn off.

Pretty much the only way I ever use SRBs is to provide that extra oomph at launch. I start with my core stage at full throttle and a TWR of ~1.8, getting my speed to around 200m/s as quickly as ever I can, then throttle back to keep things from getting out of control as I do my gravity turn. Both STS and now SLS pretty much do the same thing.

Edited by herbal space program
##### Share on other sites

5 hours ago, herbal space program said:

Well I looked it up just now and it was actually 1.25 at launch, but that was for reasons that have nothing to do with the most mass or energy-efficient way to get to orbit. The F1 engines  could not be throttled, so they were like SRBs are in this game. If you put that together with the totally un-KSP mass ratios in real life, you've got a rocket for which the chief limiting factor is too much acceleration at the end of the burn.  They dealt with this on the Saturn V by shutting off the center engine when they reached 4g, since they couldn't throttle back.  As @GoSlash27 pointed out (and I acknowledge), that sort of situation is pretty much the one case where it  actually makes sense to have rocket take off with a TWR as low as 1.2. The STS took off at a much higher TWR of ~1.75 1.6, because that is more efficient  and the higher payload fraction and the ability to throttle the RS-25 engines mitigated against too much acceleration at the end of the burn.

I've taken off just fine with rockets that barely lifted off, so that is definitely not an issue for me.

I have tested this pretty exhaustively, both on Kerbin and on Eve, and based on that and a number of discussions in this forum, I think what you say here is just not true. Unless you push your TWR to implausible extremes or fly an excessively flat ascent profile,  gravity losses far outweigh drag losses for a well-designed craft. You just have to keep it pointed prograde the whole time.

Pretty much the only way I ever use SRBs is to provide that extra oomph at launch. I start with my core stage at full throttle and a TWR of ~1.8, getting my speed to around 200m/s as quickly as ever I can, then throttle back to keep things from getting out of control as I do my gravity turn. Both STS and now SLS pretty much do the same thing.

herbal,

I agree with most of what you're saying, especially AFA it applies to the real world. However... I would argue that the priorities are a bit different in KSP, not only due to physical differences in the environments, but priorities for the end-users.

While you state (correctly) that you can make orbit with a minimum dv at high thrust, minimum dv to orbit isn't the ultimate benchmark in KSP. What matters most is 1) Minimum *cost* per tonne to orbit, 2) repeatability, and 3) ease of use.

We have done extensive testing and competitions around these criteria, and have found that the rockets that were able to deliver the most 'bang for the buck' operated at the lower end of the TWR spectrum, were also the easiest to fly and achieved more uniformly repeatable gravity turns than their higher powered counterparts.

This also holds true in the exploration of high payload fraction launchers; the lower thrust ones have consistently outperformed the higher thrust ones.

Best,

-Slashy

Edited by GoSlash27
##### Share on other sites

17 hours ago, GoSlash27 said:

While you state (correctly) that you can make orbit with a minimum dv at high thrust, minimum dv to orbit isn't the ultimate benchmark in KSP. What matters most is 1) Minimum *cost* per tonne to orbit, 2) repeatability, and 3) ease of use.

What I was talking about was indeed what gets a given payload to orbit with the minimum total impulse and not the other things. If launch cost is what you're all about, then SRBs are the way to go, except on Eve where their bad ISP makes the Twin Boar more economical.  If nonrecoverable cost is your criterion, then reusable space planes will win the day every time.  Fraction of launch mass to orbit OTOH I'm not so sure about, at least if you exclude using air-breathing engines. Rapiers, with a TWR of up to 23.74 and a flat 3200 ISP, are so out of line with everything else in the game that the gravity penalty of barely taking off with their weak low-speed thrust is made negligible by comparison. For ordinary rocket engines however, with much lower vacuum ISPs and even worse atmospheric ones, that penalty becomes a lot more significant, especially if you're talking about Eve rather than Kerbin. There was a contest a while back to get ore from Eve sea level to Gilly for the lowest cost, which (as usual) was won by @ManEatingApe.  I'd be interested to know what the TWR at launch was of that vessel. If you can show me an example of a non-Rapier based high-payload fraction winner that takes off from Kerbin at a TWR of only 1.2, I'd be very interested to see that too. It's kind of hard for me to fathom how taking 30-50 seconds to get to 100m/s rate of climb can be the most mass-efficient solution even on Kerbin.

Edited by herbal space program

## Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

×   Pasted as rich text.   Paste as plain text instead

Only 75 emoji are allowed.