Heavy rockets

[gorebello]

I'm trying to get to Jool with a 36 ton payload. But after the sixth keybodyne tank with mammoth attached, in line, with fuel crossfeed to central ones, the delta V added decreases significantly. So much that i can't pass the barrier of a certais (low) delta V with 600 ton of fuel. and I need 5000~6000 (based on a delta V map). That makes sense, but since I saw videos of ridiculously large rockets with many engines in line over many stages I don't get why. Can someone explain how am I supposed to get to Jool anda Eeloo?

Edited December 30, 2015 by gorebello

[StarManta]

It'll be tough to tell you what's going wrong without pictures of your launcher.

[gorebello]

Spoiler

Ugly rocket, I know.
This is one of the versions, but the concept is the same.

[Waxing_Kibbous]

Could be that KER isn't working quite right in 1.05, notice the readout shows S5, S6, and S8 yet you have a ton of rockets on S7.

[cantab]

A few factors:

There is a limit to the delta-V from a single stage, depending on the efficiency of the engine and the mass ratio of the fuel tanks.

Even with staging, each stage needs to either be larger than the one after it or delivers less delta-V. This is noticeable with asparagus designs, there comes a point where it's not worth adding more boosters the same size and you should instead go bigger.

You have KER in atmospheric mode. Rocket engines are more efficient in vacuum. Some are a LOT more efficient in vacuum, including the nuclear engine.

[Phil deCube]

5 hours ago, gorebello said:

I'm trying to get to Jool with a 36 ton payload. But after the sixth keybodyne tank with mammoth attached, in line, with fuel crossfeed to central ones, the delta V added decreases significantly. So much that i can't pass the barrier of a certais (low) delta V with 600 ton of fuel. and I need 5000~6000 (based on a delta V map). That makes sense, but since I saw videos of ridiculously large rockets with many engines in line over many stages I don't get why. Can someone explain how am I supposed to get to Jool anda Eeloo?

This is a Imgur gallery of the craft I took to Dres and Jool in one mission, Eeloo in another, and Moho in a third, in version 1.0.4 . I'm sending a 1.0.5 version of the same craft to Dres (then Jool) while I'm typing this.

The ship was assembled in LKO using three launches. Main capsule and science lab in one launch, Lander in another, and the fuel in a third.

Refuelling was carried out in high Minmus orbit (in the 1.0.4 version, but not the current one) then, after a landing on Dres, more fuel was taken on from an asteroid mining rig on one of Dres' asteroids.

This is the launcher for the capsule/science lab:

and this is the launcher for the fuel tanks. Note that tanks can be dumped as they become empty.

Edited December 31, 2015 by Phil deCube

[WuphonsReach]

You're in for a bit of fun getting 36t to Jool.  One solution is to send up the payload (the parts going to Jool) with empty tanks (reducing the mass that has to be lifted to orbit), then refuel the ship in orbit before sending it off to Jool.

[Norcalplanner]

Looks pretty overbuilt and unaerodynamic.  All the scanning satellites appear to have LV-Ns, which is overkill for the small size of the fuel tanks - try a Spark instead.

In terms of the lifter, your TWR is way too high.  Arrange it so you have one central stack, with both it and all the radial stacks having at least two or three white tanks on top of each Mammoth.

Here's a lifter I made a while back.  It gets two white tanks (160+ tons) into orbit, which ought to be plenty to get a 36 ton payload to Jool.  You can pick it up at KerbalX here: http://kerbalx.com/norcalplanner/200K-Fueler-LF-Mk11

 

 

Edited December 31, 2015 by Norcalplanner
Added photos and link

[Phil deCube]

Sigh!

[Norcalplanner]

53 minutes ago, Phil deCube said:

Sigh!

Not sure how to take that, but OK.

gorebello, if you want to largely keep the rocket design you have, the easiest improvement you can make is to put more fuel on each of the radial stacks.  Put an orange tank on each of the Twin Boars (with a nose cone on top) and then add a white tank and a nosecone on each of the radial Mammoth stages (after tweaking the struts a bit).  Delta V should go up a lot.

[gorebello]

Thank you all for the answers. Lets see:

23 hours ago, Waxing_Kibbous said:

Could be that KER isn't working quite right in 1.05, notice the readout shows S5, S6, and S8 yet you have a ton of rockets on S7.

I'm still using 1.0.4. But yes, I don't get that weirdness too. Looks like S6 and S8 added are the S7.

 

21 hours ago, cantab said:

A few factors:

There is a limit to the delta-V from a single stage, depending on the efficiency of the engine and the mass ratio of the fuel tanks.

Even with staging, each stage needs to either be larger than the one after it or delivers less delta-V. This is noticeable with asparagus designs, there comes a point where it's not worth adding more boosters the same size and you should instead go bigger.

You have KER in atmospheric mode. Rocket engines are more efficient in vacuum. Some are a LOT more efficient in vacuum, including the nuclear engine.

Didn't knew that what I was doing was calles Aspargus. That was instructive.

So those ultra big rockets can't be made with keybodine? It bugs me because I can lift different parts of the ship and assemble it in space, so theoreticaly (for me at last) I should be abble to offer the same power to a huge rocket and it would go to the destination. Are my physics wrong? Because this guy did this LOL thing. https://www.youtube.com/watch?v=MA1ILgysnYE

---

@Phil deCube

Liked your tank dumping system. What mod do you use for mining?

 

19 hours ago, WuphonsReach said:

You're in for a bit of fun getting 36t to Jool.  One solution is to send up the payload (the parts going to Jool) with empty tanks (reducing the mass that has to be lifted to orbit), then refuel the ship in orbit before sending it off to Jool.

There are 6 satellites, one for each moon, that why it's so heavy. Refueling is an idea, but I'm too lazy for that haha.

 

18 hours ago, Norcalplanner said:

Looks pretty overbuilt and unaerodynamic.  All the scanning satellites appear to have LV-Ns, which is overkill for the small size of the fuel tanks - try a Spark instead.

In terms of the lifter, your TWR is way too high.  Arrange it so you have one central stack, with both it and all the radial stacks having at least two or three white tanks on top of each Mammoth.

Here's a lifter I made a while back.  It gets two white tanks (160+ tons) into orbit, which ought to be plenty to get a 36 ton payload to Jool.  You can pick it up at KerbalX here: http://kerbalx.com/norcalplanner/200K-Fueler-LF-Mk11

gorebello, if you want to largely keep the rocket design you have, the easiest improvement you can make is to put more fuel on each of the radial stacks.  Put an orange tank on each of the Twin Boars (with a nose cone on top) and then add a white tank and a nosecone on each of the radial Mammoth stages (after tweaking the struts a bit).  Delta V should go up a lot.

It got huge as I was trying to fit those huge things with enough fuel. Aerodynamics play that much of a role? I read that I don't even need nose cones, only for aesthetics.

Good idea with the Spark, I didn't saw engine. That brings me to another doubt. How do I know how much fuel should I use depending of the engine? Too much and it doesn't move, too few and the engine weights too much and it doesn't move too.

You said too much TWR. Isn't it the more the better?

Will try the ideal too. EDIT: I tried it, got only 600 more delta V.

 

Edited December 31, 2015 by gorebello

[luizopiloto]

this kind of mission I usually build an Interplanetary Freighter on orbit... so, once I have one in orbit...  I only need to send payload and fuel for the missions...

Edited December 31, 2015 by luizopiloto

[Phil deCube]

On 12/31/2015 at 4:47 AM, gorebello said:

Liked your tank dumping system. What mod do you use for mining?

No mods. Mining equipment is stock in KSP. The only mods visible in the picture of the asteroid miner below is the liquid only fuel tanks from the Fuel Tanks Plus mod.

[Geschosskopf]

On ‎12‎/‎30‎/‎2015 at 3:58 PM, gorebello said:

I'm trying to get to Jool with a 36 ton payload. But after the sixth keybodyne tank with mammoth attached, in line, with fuel crossfeed to central ones, the delta V added decreases significantly. So much that i can't pass the barrier of a certais (low) delta V with 600 ton of fuel. and I need 5000~6000 (based on a delta V map). That makes sense, but since I saw videos of ridiculously large rockets with many engines in line over many stages I don't get why. Can someone explain how am I supposed to get to Jool anda Eeloo?

A 36-ton payload isn't very big as things go, about average for me actually.  This is totally and easily doable in stock.  Which part of the Jool system are you going to?  I'll assume Laythe and no aerobraking so that's 4370 m/s from LKO, to make this sort of a worst-case.  Dunno where you're getting the 5000-6000 from.  Anyway, call it 4400-4500 in the transfer stage to have a slight margin.  With this, worst case, you can capture into an elliptcal orbit at Laythe and then do some gentle aerobraking if you're running low on fuel when you get there.

I find it easiest to design rockets by the burns required for the various phases of its flight, starting in reverse order.  So you've got your mission payload of 36 tons which I'm assuming includes whatever fuel it needs to do it's job once it gets there.  So, you need 1 or more transfer stages to get it from LKO to LLO.  Depending on how annoying you find long transfer burns leaving Kerbin, you can either do this as 1 big LV-N stage (which I do in this example) or you can break this up.  Have a smaller LV-N stage right under the payload with about 2200m/s to do the burns near Jool and Laythe, and a 2300m/s LFO/Rhino stage for a short burn out from Kerbin.  But that will require a bigger lifter so probably the Nuclear Lightbulb is your best bet. 

Anyway, let's assume stock here.  4500m/s pushing 36 tons will need 4x LV-Ns, each under a 3/4 of a 1.25m long tank radially attached to a big orange tank.  Put nosecones on the 1.25m tanks, add 2 large folding radiators to keep the LV-Ns happy, and add 2x 2.5m reaction wheels.  That's your payload and transfer stage with a total weight of 101 tons.  Now, with 4x LV-Ns, you'll only have a TWR of 0.21 leaving Kerbin, which will require doing the transfer burn over 2 or 3 orbits instead of all at once so if it was me, I 'd the Nuclear Lightbulb engine from the Atomic Age mod instead of the LV-Ns.  Or I'd break this into a 2200m/s OMS stage and a 2300m/s chemical stage

So now for the lifter itself.  You only need about 3400-3500m/s to reach LKO these days.  This is EXTREMELY easy to do these days with stock parts, without even needing crossfeed, let alone asparagus.  Make a stack of the 2.5m-to-3.75m adaptor, 2 of the biggest 3.75m tanks, and a Mammoth engine.  This will give you above 2600m/s and a TWR of 1.46, low but good enough.  So now you need another 1000m/s.  For this, make 4 radial stacks of LFBs with the 1/2 size 2.5m tank (the "oil drum") with a nosecone on each.  No fuel lines.  Set these up as the 1st stage (1347m/s and 1.56 TWR) with the central Mammoth being the 2nd stage.  You now have 3570m/s in these 2 stages combined.  Now you could use Mainsails instead of LFBs, and you could use crossfeed or even asparagus to get by with less booster tankage, but not doing that just shows how doable this rocket is just being simple.  Of course, the LFB stacks will need a pair of centrally mounted Septratrons each, and the central stack should have some Sepratrons, too.  Plus you need 4 Vernor engines and some small tail feathers at the bottom of the central stack.  But that's really all there is to it.  If I was building this for msyelf, I'd have the Space-Y mod and would use some of its mondo SRBs instead of the LFB 1st stage, and maybe have a smaller central stack as a result, but again, stock works fine.  This lifter will get the 101-ton payload + transfer stage to a 100km orbit with about 140m/s left in the central tank.

Here's a link to the craft file if you want to take a look.

[ForScience6686]

Work to reduce payload mass, this will make huge savings as the dv will increase for every stage.  Nukes are heavy and should only be used on the interplanetary voyage.  Smallest engine possible otherwise.  Twr is not as important as dv.  Dv is king as you won't be going anywhere without enough.  Empty tanks should be discarded to improve both dv and Twr.  36t is not that large so I would focus on improving your launch and interplanetary stages.

[Vegetal]

Your rocket has quite a lot wrong, but I can tell you this: You are basing your design on old tutorials and videos. Your rocket is very un-aerodynamic and your payload is too heavy for it's objective. It's a monster of a ship that is failing to fulfill it's mission.

First thing, those huge engines are only good for lifting your ship into orbit, after that, small and highly efficient engines are the best for interplanetary burns. There's also engine weight. Nuke engines are very efficient for heavy ships, but too heavy for small satellites or probes.

I would reduce the payload size considerably and use other engines instead. Just by doing this you will be able to go to Jool with a ship maybe a third the size.

[Plusck]

Wait a second. There are three things that your design needs to do:

- get to LKO

- burn to a Jool encounter

- capture around Jool.

For the first thing, you should be able to get 36t up to LKO without any real problem. It's going to be a big build but if you're ready for Jool I'd assume it is not a problem at all.

For the second, it is a significant burn but you know that the only real solution is the LV-N. For such a huge payload, you also know that you'll be better off sending an LV-N (or two) + fuel into orbit separately and docking before the transfer burn. LfOx propulsion for 3200 dV in vacuum is going to be much, much heavier.

For the third, encountering Tylo can guarantee a capture to virtually any level of orbit around Jool. You shouldn't need any burn at all to get to an eliptical orbit within the Jool system that corresponds to where you're going.

Edited January 3, 2016 by Plusck

[technion]

On ‎2‎/‎01‎/‎2016 at 0:45 PM, Geschosskopf said:

Anyway, let's assume stock here.  4500m/s pushing 36 tons will need 4x LV-Ns, each under a 3/4 of a 1.25m long tank radially attached to a big orange tank. 

Don't forget that an orange tank, by default, has a lot of fuel that the LV-N won't be using

[Wanderfound]

On 30 December 2015 at 8:58 AM, gorebello said:

I'm trying to get to Jool with a 36 ton payload. But after the sixth keybodyne tank with mammoth attached, in line, with fuel crossfeed to central ones, the delta V added decreases significantly. So much that i can't pass the barrier of a certais (low) delta V with 600 ton of fuel. and I need 5000~6000 (based on a delta V map). That makes sense, but since I saw videos of ridiculously large rockets with many engines in line over many stages I don't get why. Can someone explain how am I supposed to get to Jool anda Eeloo?

What everybody else said, basically. Getting to Jool isn't too hard, even if you're taking 36 ton (although you don't need to). But your design has some issues that need fixing if it's to work.

Think of it in four stages: lift, transfer, mission, return.

Lift is the boost to LKO. This needs high TWR engines (aim for a TWR of 1.5-2.5 throughout the lift) and about 3,500m/s of ΔV. It also requires smooth aerodynamics and decent control authority.

Transfer is the shift from Kerbin to your destination planet (Jool, in this case). For this, you don't need high TWR; this is where the Nukes come in. One nuke if you can tolerate the long burn, 2-4 if you're impatient. Jool transfer is about 2,000m/s if you're aerobraking (take a heatshield, it's toasty), doublt that if you're doing a non-aero capture.

Mission is what you're doing while you're there. This is where the LV-909 Terriers come in; small vacuum landers don't need big heavy nukes. A small Terrier ship usually has the ΔV to hit several biomes on a low-g moon, though.

Return is the trip home, if it isn't a one-way mission. Transfer back is generally the same as the transfer there; 2,000m/s in this case if we're assuming a Kerbin aerocapture at the end. Again, this is a low-TWR high efficiency stage calling for LV-Ns or similar.

Big doesn't mean crude; you still need to pay attention to aerodynamics. But if you build it right you can launch huge payloads with fairly simple launchers:

 

Edited January 3, 2016 by Wanderfound

[Wanderfound]

High ΔV is more about efficiency than size:

 

Don't carry any more engine than you need; try to keep your TWR around 2 while getting to LKO, and as low as you can tolerate after that. When you can, make the same engine work for multiple stages. Asparagus staging, lateral drop tanks, etc. Never carry two of something when one will do.

Edited January 3, 2016 by Wanderfound

[AeroGav]

Not to trivialize your problem Gorebello,  but this sounds like a really fun challenge.   36 Tons - is that the gross weight of the interplanetary stage, ie what departs low Kerbin orbit, or are you really bringing 36 tons of Kerbals, instruments and probes out that far?      Jeez, my heaviest rocket weighed that much on that pad.    It flips out of control 75% of the time and kills its crew.  Let's see how far i get with this. 

[Temstar]

On 2/1/2016 at 0:45 PM, Geschosskopf said:

4500m/s pushing 36 tons will need 4x LV-Ns

How did you come up with that 4 x LV-N number? Is it a personal minimum TWR requirement?

[Wanderfound]

BTW: if you're determined to carry multiple satellites and things (why not one and refuel it?), I'd probably use a nuke-powered mothership with a Mk3 cargo bay packed full of satellites and landers.

Give it a Mk3 LF fuselage and strap a bunch of 1.25m LF tanks on the outside as drop tanks. Four to six LV-N's should do for thrust; spread 'em out so they don't heat each other up. Use Mk3-2.5m adaptors to streamline each end; that'll provide some oxidiser storage for refilling the landers. A 2.5-1.25m adaptor and a shielded docking port should give a nice pointy nose. Don't forget nosecones on the drop tanks.

With enough effort you could lift it loaded, but it'll be much easier if you hoik it up empty and fuel it in orbit. A quartet of asparagus-staged 2.5m boosters should see it up easy enough; add some SRBs if it needs a bit of launchpad kick.

 

[Geschosskopf]

5 hours ago, technion said:

Don't forget that an orange tank, by default, has a lot of fuel that the LV-N won't be using

D'oh!  In the relatively short time that the LV-N has burned just LF instead of the LFO I grew up with, I have become so used to this that I forgot that having a fuel switch on the tanks (to change their entire volume from LFO to LF) isn't technically stock (even if one of the devs, @Porkjet did release a "mod" called "Stock Fuel Switch" which says it's "stock" right on the label

So yeah, to be totally stock you'd be flying 1/2-empty tanks (with the oxidizer tweaked to zero in LFO tanks) so would need more tanks.  Which is a very silly thing to do, which is why there's the "Stock Fuel Switch" and the even better "Interstellar Fuel Switch".

2 hours ago, Temstar said:

How did you come up with that 4 x LV-N number? Is it a personal minimum TWR requirement?

It's somewhat personal taste (due to convenience) but it has a lot to do with efficiency.

You want to do transfer burns from LKO so your ship is moving faster during the burn, which means you get the most Oberth Effect and therefore burn less fuel to generate the same dV.  Which means you have more fuel for later in the mission.  However, the lower your parking orbit, the less time it takes to orbit Kerbin--it's about 25 minutes in LKO.  This means that during a burn of 5-6 minutes, your ship will make 20-25% of an orbit.  You shouldn't ever burn any longer than that in LKO or you start losing a lot of efficiency due to the angle between the direction you're thrusting and the direction the ship is moving.

This loss of efficiency is due to vectors.  The object of the burn is to increase your "forward" velocity, but if you're not pointing forward (aka prograde) when you burn, you're only adding the component of the burn that's parallel to the prograde direction to your velocity.  The other component of your burn is along the radial-inwards direction and therefore doesn't contribute to you getting to Jool---it is effectively wasted and, in extreme cases, can even push you down into the atmosphere which is really bad.  The greater the angle bretween your maneuver node and the prograde direction when you burn, the greater the waste.  But by limiting burn duration to 20-25% of your LKO orbital period, you keep this wastage to an acceptable level.  At least acceptable to me.  Some folks really obsess about efficiency so refuse to burn more than 2-3 minutes in LKO.

So, the shorter the burn, the more efficient it is because 1) you have less divergence between thrust direction and prograde, 2) this divergence lasts for less time during the burn, and 3) you accelerate faster, which means you get more Oberth benefit sooner.  Therefore, a high TWR on the transfer stage out from LKO is very desirable, plus it also makes it more convenient for you as a player.  The sooner you get the burn over with, the sooner you can hit warp and zip on out to your destination.  Now, when going to Jool, your burn will be about 2200-2300m/s.  A TWR of about 0.7 will make this take about 5-6 minutes, which is the longest acceptable burn duration.  This is, therefore, what you should attempt to build into your ship.

Now, it sometimes happens (especially if you do stock) that you simply CAN'T get a TWR >= 0.7 on your transfer stage without compromising the rest of the rocket.  What do you do?  Well,  you certainly do not make a 10-15 minute burn in LKO.  That just wastes gobs of fuel.  So you have 2 options.  You can either leave from a considerably bigger Kerbin orbit, so that 10-15 minutes is an acceptably small fraction of the total orbital period, or you can stay in LKO and break the burn up into chunks of 5 minutes or so.  The bigger orbit option is less efficient because 1) you're slower so you get less Oberth Effect, and 2) you start from a lower initial speed so require more dV to get to the same place.  So it's preferrable to start in LKO and burn from T-2.5 to T+2.5, coast around the resulting elliptical orbit, repeat as necessary until you complete the burn.

Another way to fudge on transfer stage TWR is to make your lifter so that its last stage still has a fair amount of fuel in the tanks (say 400-800m/s) after achieving LKO.  Then figure out where around your orbit you need to place the transfer burn node.  Go to that point and burn the remaining lifter fuel.  This effectively applies the lifter's fuel to the transfer burn, reducing the amount of dV the transfer stage needs to deal with and thereby shortening its burn time.  It's the same as breaking the transfer burn into chunks.  One thing to be careful of, however, is Mun.  If you plan to wait days/weeks in parking orbit prior to leaving Kerbin, you need to keep your Ap well inside Mun's orbit or you'll likely get tossed by Mun at some point.  OTOH, if you're leaving immediately, you can cross Mun's orbit (assuming it's not, by chance, in the way)

[Wanderfound]

37 minutes ago, Geschosskopf said:

So yeah, to be totally stock you'd be flying 1/2-empty tanks (with the oxidizer tweaked to zero in LFO tanks) so would need more tanks.

Actually, you just use aircraft LF tanks. Mk3 if you're going big, 1.25m if not.