Vacuum-optimized Dragon V2 for Falcon 9 second stage

[sevenperforce]

Crunching a few numbers, just for fun, and I was looking at the SuperDraco engines on the Dragon V2. Their quoted ISP is really, really low...like, monopropellant low. If the SL impulse is 240 seconds and the SL thrust is 68,170 N, then the mass flow at full thrust is going to be about 29 kg/s. The quoted vacuum thrust is 73,000 N, so this corresponds to a vacuum specific impulse of 257 seconds.

But that's crazy. The NTO/MMH hypergols used for the SuperDraco engines should have a SL specific impulse of 288 s and a vacuum specific impulse of 336 s. What gives?

At first I thought that perhaps the pressure-fed SuperDraco engines were running with a lower-than-ideal chamber pressure. But they have a chamber pressure of 1000 psi, which gives the expected ~280 s SL specific impulse according to this chart. Apparently the SuperDraco engines have really, really underexpanded nozzles to allow deep throttling.

This got me thinking. If SpaceX designed attachable nozzle extensions that mounted to the outside of the Dragon V2, the vacuum specific impulse could theoretically come right up to 336 s. The combined thrust from the eight SuperDraco engines would be a little over 760 kN.

The thrust of the single Merlin 1D Vacuum engine in the Falcon 9 second stage is 934 kN...just 22% higher.

The crewed Dragon V2 masses an estimated 9,200 kg including onboard propellant, full cargo capacity, and the trunk. The second stage has a dry mass of 3.9 tonnes, a specific impulse of 340 seconds, and 92.67 tonnes of fuel. So the second stage can deliver almost 7 km/s. Pretty impressive.

But suppose we equip the Dragon V2 with nozzle extensions, raising its vacuum isp to 336 s and its total thrust capacity to 760 kN, and replace the entire second stage with a drop-away NTO/MMH tank feeding up to external couplings on the Dragon V2. We can assume a tank dry mass fraction of 4.6% for that propellant combination. If the total mass being lifted by the first stage is kept the same, then the tank will contain 92.13 tonnes of fuel and you end up with a dV of 6.74 km/s without using any of the Dragon V2's internal propellant.

The first stage should easily be able to manage an extra ~200 m/s and still RTLS. So you have an almost fully reusable launch system with only a single drop tank being discarded.

And sure, I know that NTO/MMH is expensive and toxic. But surely it's cheaper than throwing away a Falcon 9 second stage with each flight, right?

[fredinno]

7 minutes ago, sevenperforce said:

Crunching a few numbers, just for fun, and I was looking at the SuperDraco engines on the Dragon V2. Their quoted ISP is really, really low...like, monopropellant low. If the SL impulse is 240 seconds and the SL thrust is 68,170 N, then the mass flow at full thrust is going to be about 29 kg/s. The quoted vacuum thrust is 73,000 N, so this corresponds to a vacuum specific impulse of 257 seconds.

But that's crazy. The NTO/MMH hypergols used for the SuperDraco engines should have a SL specific impulse of 288 s and a vacuum specific impulse of 336 s. What gives?

At first I thought that perhaps the pressure-fed SuperDraco engines were running with a lower-than-ideal chamber pressure. But they have a chamber pressure of 1000 psi, which gives the expected ~280 s SL specific impulse according to this chart. Apparently the SuperDraco engines have really, really underexpanded nozzles to allow deep throttling.

This got me thinking. If SpaceX designed attachable nozzle extensions that mounted to the outside of the Dragon V2, the vacuum specific impulse could theoretically come right up to 336 s. The combined thrust from the eight SuperDraco engines would be a little over 760 kN.

The thrust of the single Merlin 1D Vacuum engine in the Falcon 9 second stage is 934 kN...just 22% higher.

The crewed Dragon V2 masses an estimated 9,200 kg including onboard propellant, full cargo capacity, and the trunk. The second stage has a dry mass of 3.9 tonnes, a specific impulse of 340 seconds, and 92.67 tonnes of fuel. So the second stage can deliver almost 7 km/s. Pretty impressive.

But suppose we equip the Dragon V2 with nozzle extensions, raising its vacuum isp to 336 s and its total thrust capacity to 760 kN, and replace the entire second stage with a drop-away NTO/MMH tank feeding up to external couplings on the Dragon V2. We can assume a tank dry mass fraction of 4.6% for that propellant combination. If the total mass being lifted by the first stage is kept the same, then the tank will contain 92.13 tonnes of fuel and you end up with a dV of 6.74 km/s without using any of the Dragon V2's internal propellant.

The first stage should easily be able to manage an extra ~200 m/s and still RTLS. So you have an almost fully reusable launch system with only a single drop tank being discarded.

And sure, I know that NTO/MMH is expensive and toxic. But surely it's cheaper than throwing away a Falcon 9 second stage with each flight, right?

Not worth it for the small amount of flights the V2 will have.

[sevenperforce]

46 minutes ago, fredinno said:

Not worth it for the small amount of flights the V2 will have.

What part isn't worth it?

[fredinno]

30 minutes ago, sevenperforce said:

What part isn't worth it?

Redesigning the falcon9 for a few flights. Unless you can modify it for a ~9T launcher.

[sevenperforce]

Just now, fredinno said:

Redesigning the falcon9 for a few flights. Unless you can modify it for a ~9T launcher.

There wouldn't be any redesign -- it would fly with the exact same first stage. The only redesign you'd need to make would be detachable nozzle extensions for the Dragon V2 and the external propellant feed lines, plus the drop tank (which could fit inside a truncated second-stage aeroshell with only the most modest of modifications).

[AngelLestat]

Quote

 

The combined thrust from the eight SuperDraco engines would be a little over 760 kN.
The thrust of the single Merlin 1D Vacuum engine in the Falcon 9 second stage is 934 kN

But suppose we equip the Dragon V2 with nozzle extensions, raising its vacuum isp to 336 s and its total thrust capacity to 760 kN, and replace the entire second stage with a drop-away NTO/MMH tank feeding up to external couplings on the Dragon V

 

But you need to reduce in the superdraco thrust due engines tilt angle. 
So it will let you with 700 or 650 KN, which no sure if is enough.  Also the superdraco engines are in that position (without nozzle) to protect the engines from the reentry.
Take into account that the second stage should be use for any other payload and the spacex solution is try to manufacture the same pieces to reduce cost.

I personally dont like hypergolic fuels, I guess they should be part of our past, more when we can remplace them with better technology.
I will like to see a dragon v3 with methane using laser ignition for RCS and engines.
The idea to remplace the engines in the second stage with the draco engines using a attached tank is not bad (in case some engines had a problem, you can do the eject abort with only 4 instead 8 (there is no other engine pushing behind you).
No sure how safe it is to have the engines so close to the capsule for long burns, heat can be an issue?

But well, we lose efficiency in 2 ways.. engine tilt angle and nozzle.  If those things can be managed with a clever design, then we have a cheap dragon launch.

[fredinno]

3 hours ago, sevenperforce said:

There wouldn't be any redesign -- it would fly with the exact same first stage. The only redesign you'd need to make would be detachable nozzle extensions for the Dragon V2 and the external propellant feed lines, plus the drop tank (which could fit inside a truncated second-stage aeroshell with only the most modest of modifications).

You just described the redesign. The work needed to loop a drop tank to allow for what is pretty much crossfeed on a Dragon V2 is enough to make ths not worth it.

1 hour ago, AngelLestat said:

But you need to reduce in the superdraco thrust due engines tilt angle. 
So it will let you with 700 or 650 KN, which no sure if is enough.  Also the superdraco engines are in that position (without nozzle) to protect the engines from the reentry.
Take into account that the second stage should be use for any other payload and the spacex solution is try to manufacture the same pieces to reduce cost.

I personally dont like hypergolic fuels, I guess they should be part of our past, more when we can remplace them with better technology.
I will like to see a dragon v3 with methane using laser ignition for RCS and engines.
The idea to remplace the engines in the second stage with the draco engines using a attached tank is not bad (in case some engines had a problem, you can do the eject abort with only 4 instead 8 (there is no other engine pushing behind you).
No sure how safe it is to have the engines so close to the capsule for long burns, heat can be an issue?

But well, we lose efficiency in 2 ways.. engine tilt angle and nozzle.  If those things can be managed with a clever design, then we have a cheap dragon launch.

No, we use hypergolic for crew modules for a reason- hyperdolics are storable, and ignite immediately even if everything fails- essential for a crew module.

[sevenperforce]

49 minutes ago, fredinno said:

You just described the redesign. The work needed to loop a drop tank to allow for what is pretty much crossfeed on a Dragon V2 is enough to make ths not worth it.

I meant that it isn't a redesign of the carrier rocket. 

It isn't crossfeed because there are no other engines involved. It's just adding external fuel couplings that feed into the internal fuel tank. Sure, it takes work, but it is more a retrofit then a full redesign. 

2 hours ago, AngelLestat said:

But you need to reduce in the superdraco thrust due engines tilt angle. 
So it will let you with 700 or 650 KN, which no sure if is enough.  Also the superdraco engines are in that position (without nozzle) to protect the engines from the reentry.
Take into account that the second stage should be use for any other payload and the spacex solution is try to manufacture the same pieces to reduce cost.

No sure how safe it is to have the engines so close to the capsule for long burns, heat can be an issue?

We lose efficiency in 2 ways.. engine tilt angle and nozzle.  If those things can be managed with a clever design, then we have a cheap dragon launch.

Yeah, the second stage would definitely be used for any additional payload. This would strictly be for crew ferry only. 

Heating is a problem. Not sure how to manage that. 

Thrust isn't a big deal; the T/W ratio will be fine. But the tilt angle does cut into effective specific impulse. However, you can just stage at a slightly higher speed to take care of that. 

[fredinno]

4 hours ago, sevenperforce said:

I meant that it isn't a redesign of the carrier rocket. 

It isn't crossfeed because there are no other engines involved. It's just adding external fuel couplings that feed into the internal fuel tank. Sure, it takes work, but it is more a retrofit then a full redesign. 

Yeah, the second stage would definitely be used for any additional payload. This would strictly be for crew ferry only. 

Heating is a problem. Not sure how to manage that. 

Thrust isn't a big deal; the T/W ratio will be fine. But the tilt angle does cut into effective specific impulse. However, you can just stage at a slightly higher speed to take care of that. 

What you did was just describe crossfeed- you take fuel from one tank, stop, drop that off, start the engines again, then take it from another tank. Also, Dragon V2 is likely getting only one flight a year due to ISS requirements, thus total 8 flights assuming a 2024 ISS retirement.

[Nibb31]

Dragon V2 is only going to fly less than 10 times to the ISS. Unless someone builds a commercial station to replace the ISS, that's all there is.

[Steel]

My question is, other than not throwing away on Merlin 1D each launch, what do you gain?

[AngelLestat]

11 hours ago, fredinno said:

No, we use hypergolic for crew modules for a reason- hyperdolics are storable, and ignite immediately even if everything fails- essential for a crew module.

Hipergolics was the best solution we fund until now, this does not mean there is not a better solution.
Now we are capable to ignite methane-ox using laser ignition with a 100% of chance in milliseconds.
In fact in 2012-2013 a lander using methane ox for RCS and main thrusters was tested (project morpheus) showing successful results, even taking into account that use another ignition tech than the one that describe this paper
Of course this tech still needs more development, but hipergolics handle and use is not something super safe either. 

6 hours ago, fredinno said:

What you did was just describe crossfeed- you take fuel from one tank, stop, drop that off, start the engines again, then take it from another tank. Also, Dragon V2 is likely getting only one flight a year due to ISS requirements, thus total 8 flights assuming a 2024 ISS retirement.

That is not the definition of crossfeed in the space industry, the tanks needs to have another engine to be called crossfeed.

[sevenperforce]

7 hours ago, fredinno said:

What you did was just describe crossfeed- you take fuel from one tank, stop, drop that off, start the engines again, then take it from another tank. Also, Dragon V2 is likely getting only one flight a year due to ISS requirements, thus total 8 flights assuming a 2024 ISS retirement.

There is no SECO required. The drop tank would be designed to feed into the internal tank with flow-through, so that the engines are pulling from the internal tank the whole time. Launch abort just means severing the feed lines (which would redundantly self-seal) and dropping the tank, with the engines remaining on full thrust the whole time. 

The nozzle extensions would have to break away as well for a low-altitude abort. 

ISS is an issue, I know. But having a 98% reusable system is good regardless. 

[sgt_flyer]

1 hour ago, sevenperforce said:

There is no SECO required. The drop tank would be designed to feed into the internal tank with flow-through, so that the engines are pulling from the internal tank the whole time. Launch abort just means severing the feed lines (which would redundantly self-seal) and dropping the tank, with the engines remaining on full thrust the whole time. 

The nozzle extensions would have to break away as well for a low-altitude abort. 

ISS is an issue, I know. But having a 98% reusable system is good regardless. 

mmh. i wonder how it'll work with the pressure fed superdracos. you'd need to include some sort of pumping system, (maybe use a pressurised bladder if you want to keep it mechanically simple), to counteract the spacecraft acceleration, as the propellants would have to travel 'upwards' several meters (especially if you pump from the bottom of your drop tank, as under acceleration, the fuel will want to stay at the bottom of the tanks) - that still would be a hell of a lot of propellant to move for a pressure fed system

Edited March 12, 2016 by sgt_flyer

[sevenperforce]

Yes, the pressure feeding does make this a slightly more challenging system. Might be worth it to use lighter-weight tanks with a small turbopump. I hate to throw the turbopump away, of course, but at least it could be exceedingly simple because it could run off MMH as a monopropellant.

A major advantage is the extremely high propellant density. The tank size would be physically far far smaller than the Falcon second stage. 

[fredinno]

3 hours ago, AngelLestat said:

Hipergolics was the best solution we fund until now, this does not mean there is not a better solution.
Now we are capable to ignite methane-ox using laser ignition with a 100% of chance in milliseconds.
In fact in 2012-2013 a lander using methane ox for RCS and main thrusters was tested (project morpheus) showing successful results, even taking into account that use another ignition tech than the one that describe this paper
Of course this tech still needs more development, but hipergolics handle and use is not something super safe either. 

That is not the definition of crossfeed in the space industry, the tanks needs to have another engine to be called crossfeed.

Yeah, and what happens if the lazers fail? Methane still boils off.

Also, you still face the same problems as crossfeed- the only other rocket stage to use drop tanks is the Bris-M, and that is far smaller than what you are proposing. The difficulty of crossfeed is making sure you can switch propellant sources without causing air bubbles to make its way inside the tank. Not impossible, but why would you do it for something launching once a year?

Edited March 12, 2016 by fredinno

[sgt_flyer]

1 hour ago, fredinno said:

Yeah, and what happens if the lazers fail? Methane still boils off.

Also, you still face the same problems as crossfeed- the only other rocket stage to use drop tanks is the Bris-M, and that is far smaller than what you are proposing. The difficulty of crossfeed is making sure you can switch propellant sources without causing air bubbles to make its way inside the tank. Not impossible, but why would you do it for something launching once a year?

actually, there's two other spacecrafts that used / use drop tanks  

the first one (wayyy before briz-M) was the ye-8 spacecraft / lander from the lunokhod program.

it used side mounted drop tanks, which were used for circularisation around the moon. they were dropped before the final moon landing.

the other is the Fregat-SB (which has flown a few times already, atop Zenit) - the ill-fated Phobos-grunt notably used this stage.

(at around the 2mn on the video)

still, as was said, that would make a hell of a lot of propellant to move on a drop tank the size of a second stage  - as the superdracos are pressure fed, you'd need additionnal pressurising gas along with a pumping system, which would add on the dry mass.

Edited March 12, 2016 by sgt_flyer

[fredinno]

10 minutes ago, sgt_flyer said:

actually, there's two other spacecrafts that used / use drop tanks  

the first one (wayyy before briz-M) was the ye-8 spacecraft / lander from the lunokhod program.

it used side mounted drop tanks, which were used for circularisation around the moon. they were dropped before the final moon landing.

the other is the Fregat-SB (which has flown a few times already, atop Zenit) - the ill-fated Phobos-grunt notably used this stage.

(at around the 2mn on the video)

still, as was said, that would make a hell of a lot of propellant to move on a drop tank the size of a second stage  - as the superdracos are pressure fed, you'd need additionnal pressurising gas along with a pumping system, which would add on the dry mass.

Still, we've never made a drop tank the size of what is being proposed- not to mention that it seems Russia is the only nation with experience with drop tanks.

[AngelLestat]

2 hours ago, fredinno said:

Yeah, and what happens if the lazers fail? Methane still boils off.

This dude is worry for other kind of "what if"

No to mention the cost of hypergolic fuels vs ch4+lox.

If you have something similar to dragon engines (8 engines 2 by side), even if one does not ignite, it will be ignite due the flames of the side by side engine.
But you can even design a different solution that will be perfect for these kind of ignition to even decrease these problems, and dragon can even work with 3 engines if it wants.

We had difficulties in the past with respect to ignition and keeping cool cryogenic fuels, that is already part of our past.

Quote

Also, you still face the same problems as crossfeed- the only other rocket stage to use drop tanks is the Bris-M, and that is far smaller than what you are proposing. The difficulty of crossfeed is making sure you can switch propellant sources without causing air bubbles to make its way inside the tank. Not impossible, but why would you do it for something launching once a year?

I am not the one proposing.. this is sevenperforce´s idea.  And I encourage that, I like when somebody put its brain to work to see if they can come up with better ideas, because is the kind of people that does not accept the world as it is, they want to improve it.
And even if we fail in that try..  doing those exercises is the best way to learn about the topic. 
Speaking of crossfeed,  in the falcon heavy case will allow reusability without reducing much its max payload (you increase efficiency and your side booster comeback faster, no need to return so many km to the launchpad)

[fredinno]

2 hours ago, AngelLestat said:

This dude is worry for other kind of "what if"

No to mention the cost of hypergolic fuels vs ch4+lox.

If you have something similar to dragon engines (8 engines 2 by side), even if one does not ignite, it will be ignite due the flames of the side by side engine.
But you can even design a different solution that will be perfect for these kind of ignition to even decrease these problems, and dragon can even work with 3 engines if it wants.

We had difficulties in the past with respect to ignition and keeping cool cryogenic fuels, that is already part of our past.

I am not the one proposing.. this is sevenperforce´s idea.  And I encourage that, I like when somebody put its brain to work to see if they can come up with better ideas, because is the kind of people that does not accept the world as it is, they want to improve it.
And even if we fail in that try..  doing those exercises is the best way to learn about the topic. 
Speaking of crossfeed,  in the falcon heavy case will allow reusability without reducing much its max payload (you increase efficiency and your side booster comeback faster, no need to return so many km to the launchpad)

Yes, hypergolic fuels have hazards. A bigger hazard is having a broken engine in space. You still didn't answer the problem of boil-off, and insulation for the service module/final stage will be heavy. Dragons are in Orbit for several months at a time, you know.

 

Yes, guess what? FH is not using crossfeed because the R+D money is not worth it for a rocket already on the verge of "useful" instead of "overpowered for pretty much anything".

[sevenperforce]

6 hours ago, fredinno said:

Also, you still face the same problems as crossfeed- the only other rocket stage to use drop tanks is the Bris-M, and that is far smaller than what you are proposing. The difficulty of crossfeed is making sure you can switch propellant sources without causing air bubbles to make its way inside the tank. Not impossible, but why would you do it for something launching once a year?

It wouldn't switch sources. The drop tanks would feed the internal tanks directly. 

[fredinno]

31 minutes ago, sevenperforce said:

It wouldn't switch sources. The drop tanks would feed the internal tanks directly. 

I would anticipate that to be just as bad, or worse. You're feeding rocket fuel during launch from one tank to the next.

[AngelLestat]

16 hours ago, fredinno said:

Yes, hypergolic fuels have hazards. A bigger hazard is having a broken engine in space. You still didn't answer the problem of boil-off, and insulation for the service module/final stage will be heavy. Dragons are in Orbit for several months at a time, you know.

These kind of engines does not need pumps, they both use helium for that matter, so the only problem would be ignition, and I dont see any problem with that.. you can get 100% as I said and was tested.
About boil off is easy..  we are talking about liquid oxygen (-180c) and liquid methane (-160c) boiling points.. this is not bad as -250c for liquid hydrogen in which even the albedo of earth can help to boil off.
Even if you need active cooling is not hard..  in principle you just need to take the fuel and oxidant gases, you compress them inside a radiator (no facing the sun), this heat up, release the heat, and then you let expand it which liquefy the gases at lower temperature. You just need to do this once every while...  each time the pressure of the tank start to rise.

16 hours ago, fredinno said:

Yes, guess what? FH is not using crossfeed because the R+D money is not worth it for a rocket already on the verge of "useful" instead of "overpowered for pretty much anything".

??  they did not give up on crossfeed, they just are not using it yet for the first test of falcon heavy. So please stop to make up fake news.
I am still waiting your answer over the canadian oil pipe   

[sevenperforce]

15 minutes ago, AngelLestat said:

they did not give up on crossfeed, they just are not using it yet for the first test of falcon heavy.    

They have postponed development of a crossfeed system for the time, so who knows whether they will ever build it. For most FH launches, they will burn all three engine clusters at 100% to start, then downthrottle the core stage as far as it will go through maxQ until booster separation. 

[fredinno]

6 hours ago, sevenperforce said:

They have postponed development of a crossfeed system for the time, so who knows whether they will ever build it. For most FH launches, they will burn all three engine clusters at 100% to start, then downthrottle the core stage as far as it will go through maxQ until booster separation. 

They HAVE almost certainly given it up. Look, even with 100% reuse of both cores (1 on barge, 2 on land) the payload capacity is 29T to LEO. And that's before the FT upgrades. There is literally no reason for a even bigger FH when FH is already on the verge of complete commercial uselessness due to being too big.

They just quote bigger numbers for prestige, FH will never be used to that full potential.

Also, crossfeed is difficult enough it's apparently easier to just upgrade the engines- which is what ULA did to upgrade the Delta IV Heavy from 22 T to LEO to 25T to LEO.

7 hours ago, AngelLestat said:

These kind of engines does not need pumps, they both use helium for that matter, so the only problem would be ignition, and I dont see any problem with that.. you can get 100% as I said and was tested.
About boil off is easy..  we are talking about liquid oxygen (-180c) and liquid methane (-160c) boiling points.. this is not bad as -250c for liquid hydrogen in which even the albedo of earth can help to boil off.
Even if you need active cooling is not hard..  in principle you just need to take the fuel and oxidant gases, you compress them inside a radiator (no facing the sun), this heat up, release the heat, and then you let expand it which liquefy the gases at lower temperature. You just need to do this once every while...  each time the pressure of the tank start to rise.

??  they did not give up on crossfeed, they just are not using it yet for the first test of falcon heavy. So please stop to make up fake news.
I am still waiting your answer over the canadian oil pipe   

Indeed, if it was so easy, then hypergols would not be used anymore extensively.

Boil-off has still not been solved, and the solution you propose would be another expense added to a device that would only be used ~10 times. Not to mention it's never been tested. A better idea is sun shields, which have more experience behind them via infrared telescopes like the JWST.

I stopped answering the canadian oil pipe because it's not like you can be convinced of it anyways.