A conversation with Elon Musk

On Thursday, October 3, 2019 at 6:49:33 AM UTC-7, David Spain wrote:

Nozzle design helps to. It would appear to have sea level *and* vacuum
optimized nozzles is the win over aerospikes at least for a TSTO
vehicle. Though Elon remains open to any "gifts", in this regard.

Dave


What happened to nozzle extensions?

"For first stage rocket engines, the engine works with nozzle extension in
disposed position during the first minutes of flight and expands it at some
predefined level of air pressure. This scheme assumes the outer skirt of the
bell is extended while the engine is functioning and its installation to working
position happens in the upper layers of the atmosphere. It excludes problems
with flow separation at sea level and increases efficiency of the engine in
vacuum."

See:

https://en.wikipedia.org/wiki/Nozzle_extension

In article ,
says...

On Thursday, October 3, 2019 at 6:49:33 AM UTC-7, David Spain wrote:

Nozzle design helps to. It would appear to have sea level *and* vacuum
optimized nozzles is the win over aerospikes at least for a TSTO
vehicle. Though Elon remains open to any "gifts", in this regard.

Dave


What happened to nozzle extensions?

"For first stage rocket engines, the engine works with nozzle extension in
disposed position during the first minutes of flight and expands it at some
predefined level of air pressure. This scheme assumes the outer skirt of the
bell is extended while the engine is functioning and its installation to working
position happens in the upper layers of the atmosphere. It excludes problems
with flow separation at sea level and increases efficiency of the engine in
vacuum."

See:

https://en.wikipedia.org/wiki/Nozzle_extension

Added complexity. It's non-trivial to extend while firing the engine.
For a vehicle like Starship/Super Booster which stages "early" compared
to its expendable counterparts, I'd imagine it's harder than designing
an extension for a fully vacuum optimized engine. For example, the RL-
10 engines with extensions like this extend *before* the engine starts
firing. This takes a bit of time, so wouldn't be a good trade for an
upper stage that stages early due to gravity losses during the nozzle
extension period.

All IMHO of course.

Jeff
--
All opinions posted by me on Usenet News are mine, and mine alone.
These posts do not reflect the opinions of my family, friends,
employer, or any organization that I am a member of.

On 2019-10-09 7:50 AM, Jeff Findley wrote:
In article ,
says...

On Thursday, October 3, 2019 at 6:49:33 AM UTC-7, David Spain wrote:

Nozzle design helps to. It would appear to have sea level *and* vacuum
optimized nozzles is the win over aerospikes at least for a TSTO
vehicle. Though Elon remains open to any "gifts", in this regard.

Dave


What happened to nozzle extensions?

Added complexity. It's non-trivial to extend while firing the engine.
For a vehicle like Starship/Super Booster which stages "early" compared
to its expendable counterparts, I'd imagine it's harder than designing
an extension for a fully vacuum optimized engine. For example, the RL-
10 engines with extensions like this extend *before* the engine starts
firing. This takes a bit of time, so wouldn't be a good trade for an
upper stage that stages early due to gravity losses during the nozzle
extension period.

All IMHO of course.

Jeff


I might add that it appears from what I last read that Elon plans to
attach the vacuum Raptors on Starship to the "airframe",

https://twitter.com/elonmusk/status/1131433322276483072

with the sea level Raptors being able to gimbal but the vacuum ones do
not. So real complexity there to try to add extensions to the sea level
Raptors which need to also move. Also consider cost. It might actually
cost *more* to put extensions on the SL Raptors than just add Raptors
dedicated to vacuum operation given the fact that the engine mfg. is
already vertically integrated into your company and therefore by
definition your are obtaining the engines *at cost*. As I understand it
these engines are somewhat cheaper to build for SpaceX anyway. So it
makes sense to me to just use more...

Dave

On Oct/9/2019 at 11:09, David Spain wrote :
On 2019-10-09 7:50 AM, Jeff Findley wrote:
In article ,
says...

On Thursday, October 3, 2019 at 6:49:33 AM UTC-7, David Spain wrote:

Nozzle design helps to. It would appear to have sea level *and* vacuum
optimized nozzles is the win over aerospikes at least for a TSTO
vehicle. Though Elon remains open to any "gifts", in this regard.

Dave


What happened to nozzle extensions?

Added complexity.Â* It's non-trivial to extend while firing the engine.
For a vehicle like Starship/Super Booster which stages "early" compared
to its expendable counterparts, I'd imagine it's harder than designing
an extension for a fully vacuum optimized engine.Â* For example, the RL-
10 engines with extensions like this extend *before* the engine starts
firing.Â* This takes a bit of time, so wouldn't be a good trade for an
upper stage that stages early due to gravity losses during the nozzle
extension period.

All IMHO of course.

Jeff


I might add that it appears from what I last read that Elon plans to
attach the vacuum Raptors on Starship to the "airframe",

https://twitter.com/elonmusk/status/1131433322276483072

with the sea level Raptors being able to gimbal but the vacuum ones do
not. So real complexity there to try to add extensions to the sea level
Raptors which need to also move. Also consider cost. It might actually
cost *more* to put extensions on the SL Raptors than just add Raptors
dedicated to vacuum operation given the fact that the engine mfg. is
already vertically integrated into your company and therefore by
definition your are obtaining the engines *at cost*. As I understand it
these engines are somewhat cheaper to build for SpaceX anyway. So it
makes sense to me to just use more...

I think that SpaceX didn't do much optimisation of its rockets. They
could be improved quite a lot. As you say it isn't obvious that
extending the nozzle in flight is the way to go, but there is a good
chance that it would improve performance. There are many other things
that could be done to improve performance. And I think that such work
will be done in the not too far future.

Now don't read me wrong. I'm not complaining that SpaceX didn't do it
the right way. The main problem in rocketry was that rockets were used
once. The important thing to do was to make them reusable. If cars were
thrown away once the fuel in the gas tank is all used, the important
thing to do wouldn't be to make a hybrid electric/gas car that can run
longer on that single tankful of gas. You don't make the car more
expensive so it can run longer on the limited fuel. You make the car
reusable. But once that is done, yes making the car more fuel efficient
is important. Until recently, the cost of fuel in a rocket launch was
irrelevant, something like 0.1%. SpaceX is now making it relevant
because all other costs have gone way down.


Alain Fournier

In article , says...
I might add that it appears from what I last read that Elon plans to
attach the vacuum Raptors on Starship to the "airframe",

https://twitter.com/elonmusk/status/1131433322276483072

with the sea level Raptors being able to gimbal but the vacuum ones do
not. So real complexity there to try to add extensions to the sea level
Raptors which need to also move. Also consider cost. It might actually
cost *more* to put extensions on the SL Raptors than just add Raptors
dedicated to vacuum operation given the fact that the engine mfg. is
already vertically integrated into your company and therefore by
definition your are obtaining the engines *at cost*. As I understand it
these engines are somewhat cheaper to build for SpaceX anyway. So it
makes sense to me to just use more...

I think that SpaceX didn't do much optimisation of its rockets.

This is a loaded statement right off the bat. Firstly, what are you
optimizing for? When people start talking extending nozzles and the
like, they're likely talking about optimizing performance. O.k., so
we're back to minimizing propellant mass and/or minimizing dry mass like
an expendable? Why?

Someone did the calculations to figure out the cost of propellants for a
single Starship/Super Booster flight. The final cost was on the order
of $1 million. Think about that. Why would you try to minimize
propellant mass? Let's say you shave off 10% of that mass. You've
saved on the order of $100,000 per flight. But at what cost? And by
cost, I mean money. If it's by adding an extensible nozzle, what costs
does that add in terms of cost of the added complexity in terms of the
initial build and in terms of ongoing maintenance and testing (because
it's got to work right because you don't want to risk a failed extension
damaging a nearby engine).

Secondly, as I've said on this group before, we're looking at Mk1 and
Mk2 right now. Musk has already admitted they *aren't* optimized. Why?
Because they're the early test vehicles used to gather data which will
be used to refine the design! Actual flight data is always a *good
thing* to have.

NASA was able to use flight data from Columbia to refine the designs of
the later orbiters by making them lighter. Same thing will happen with
Starship/Super Booster but I'd imagine, based on Falcon experience, that
we'll see far more iterations of the design than the space shuttle
orbiters that only saw minor tweaks to the design as a result of the
data gathered by Columbia.

They
could be improved quite a lot. As you say it isn't obvious that
extending the nozzle in flight is the way to go, but there is a good
chance that it would improve performance.

And there is a certainty that it will increase costs because you're
adding moving components that must work and not fail, increasing
complexity, and increasing costs. Will those increased costs cover the
cost of propellant saved? Doubtful, IMHO, since we've already figured
out propellant costs and they're still small compared to all other costs
involved in launching and maintaining Starship/Super Booster.

There are many other things
that could be done to improve performance. And I think that such work
will be done in the not too far future.

This I agree with. But, I seriously doubt extensible nozzles will be
one of them.

Now don't read me wrong. I'm not complaining that SpaceX didn't do it
the right way. The main problem in rocketry was that rockets were used
once. The important thing to do was to make them reusable. If cars were
thrown away once the fuel in the gas tank is all used, the important
thing to do wouldn't be to make a hybrid electric/gas car that can run
longer on that single tankful of gas. You don't make the car more
expensive so it can run longer on the limited fuel. You make the car
reusable. But once that is done, yes making the car more fuel efficient
is important. Until recently, the cost of fuel in a rocket launch was
irrelevant, something like 0.1%. SpaceX is now making it relevant
because all other costs have gone way down.

I'd argue we're still a long way from propellant costs being a dominant
cost for orbital launch.

Still, SpaceX chose a propellant combination that's relatively cheap and
has several advantages over the alternatives. Liquid hydrogen is more
expensive to buy, handle, and store (because of its very low bulk
density) than liquid methane. Liquid methane results in a smaller
launch vehicle, compared to liquid hydrogen, because it's more dense
than liquid hydrogen. Merlin's kerosene is cheap, but it is harder in
terms of inspection and maintenance due to its ability to coke in
cooling passages and the like in a liquid fueled rocket engine. So,
liquid methane is likely a "sweet spot" for rocket engine fuel,
especially for lower stages.

Jeff
--
All opinions posted by me on Usenet News are mine, and mine alone.
These posts do not reflect the opinions of my family, friends,
employer, or any organization that I am a member of.

On Oct/10/2019 at 06:42, Jeff Findley wrote :
In article , says...
I might add that it appears from what I last read that Elon plans to
attach the vacuum Raptors on Starship to the "airframe",

https://twitter.com/elonmusk/status/1131433322276483072

with the sea level Raptors being able to gimbal but the vacuum ones do
not. So real complexity there to try to add extensions to the sea level
Raptors which need to also move. Also consider cost. It might actually
cost *more* to put extensions on the SL Raptors than just add Raptors
dedicated to vacuum operation given the fact that the engine mfg. is
already vertically integrated into your company and therefore by
definition your are obtaining the engines *at cost*. As I understand it
these engines are somewhat cheaper to build for SpaceX anyway. So it
makes sense to me to just use more...

I think that SpaceX didn't do much optimisation of its rockets.

This is a loaded statement right off the bat. Firstly, what are you
optimizing for? When people start talking extending nozzles and the
like, they're likely talking about optimizing performance. O.k., so
we're back to minimizing propellant mass and/or minimizing dry mass like
an expendable? Why?

Someone did the calculations to figure out the cost of propellants for a
single Starship/Super Booster flight. The final cost was on the order
of $1 million. Think about that. Why would you try to minimize
propellant mass? Let's say you shave off 10% of that mass. You've
saved on the order of $100,000 per flight. But at what cost? And by
cost, I mean money. If it's by adding an extensible nozzle, what costs
does that add in terms of cost of the added complexity in terms of the
initial build and in terms of ongoing maintenance and testing (because
it's got to work right because you don't want to risk a failed extension
damaging a nearby engine).

Secondly, as I've said on this group before, we're looking at Mk1 and
Mk2 right now. Musk has already admitted they *aren't* optimized. Why?
Because they're the early test vehicles used to gather data which will
be used to refine the design!

That's exactly what I was saying. They didn't do much optimisation. So
Musk agrees with me.

Actual flight data is always a *good
thing* to have.

NASA was able to use flight data from Columbia to refine the designs of
the later orbiters by making them lighter. Same thing will happen with
Starship/Super Booster but I'd imagine, based on Falcon experience, that
we'll see far more iterations of the design than the space shuttle
orbiters that only saw minor tweaks to the design as a result of the
data gathered by Columbia.

They
could be improved quite a lot. As you say it isn't obvious that
extending the nozzle in flight is the way to go, but there is a good
chance that it would improve performance.

And there is a certainty that it will increase costs because you're
adding moving components that must work and not fail, increasing
complexity, and increasing costs. Will those increased costs cover the
cost of propellant saved? Doubtful, IMHO, since we've already figured
out propellant costs and they're still small compared to all other costs
involved in launching and maintaining Starship/Super Booster.

There are many other things
that could be done to improve performance. And I think that such work
will be done in the not too far future.

This I agree with. But, I seriously doubt extensible nozzles will be
one of them.

Now don't read me wrong. I'm not complaining that SpaceX didn't do it
the right way. The main problem in rocketry was that rockets were used
once. The important thing to do was to make them reusable. If cars were
thrown away once the fuel in the gas tank is all used, the important
thing to do wouldn't be to make a hybrid electric/gas car that can run
longer on that single tankful of gas. You don't make the car more
expensive so it can run longer on the limited fuel. You make the car
reusable. But once that is done, yes making the car more fuel efficient
is important. Until recently, the cost of fuel in a rocket launch was
irrelevant, something like 0.1%. SpaceX is now making it relevant
because all other costs have gone way down.

I'd argue we're still a long way from propellant costs being a dominant
cost for orbital launch.

Yes we are a long way from propellant costs being dominant. But until
now, propellant costs were totally insignificant. That's no longer the
case. The costs of developing and building the first Starship-Super
Booster is not known but seems to be in the low billions. I would guess
that the great majority of that is in the developing part not the
building part. This is just a guestimate, but I would say that building
a Starship-Super Booster, after development is paid for, will be in the
hundreds of millions. If I recall correctly, SpaceX wants to fly them
about 100 times each. Assuming they are successful in flying them 100
times (a big assumption) the cost of the rocket hardware should be a few
millions per flight. That's the same order of magnitude as the cost of
the fuel. It used to be that the costs of the rocket totally dominated
the launch costs. Now the costs of the fuel which used to be totally
insignificant is of the same order of magnitude as the cost of the
rocket (per launch). There are other costs for the launch. SpaceX is
supposedly trying to minimise those other costs too, and from what I can
tell they are successful at it.

It used to be that trying to save rocket fuel was a sure way to increase
overall costs. But now it seems that SpaceX will be so successful in
decreasing all other costs, that saving fuel becomes a good target for
reducing overall costs. Of course that is assuming that Starship-Super
Booster lives up to expectations.

When you have a reusable rocket, it can make sense to increase the cost
of the rocket to save fuel because you pay the rocket once, the fuel
many times. Obviously, even with a reusable rocket, you don't go back to
performance über alles. But increasing the price of Super Booster by a
few millions to save a few percent fuel might make sense.

Still, SpaceX chose a propellant combination that's relatively cheap and
has several advantages over the alternatives. Liquid hydrogen is more
expensive to buy, handle, and store (because of its very low bulk
density) than liquid methane. Liquid methane results in a smaller
launch vehicle, compared to liquid hydrogen, because it's more dense
than liquid hydrogen. Merlin's kerosene is cheap, but it is harder in
terms of inspection and maintenance due to its ability to coke in
cooling passages and the like in a liquid fueled rocket engine. So,
liquid methane is likely a "sweet spot" for rocket engine fuel,
especially for lower stages.

Yes, I like their choice of fuel. I like their idea of making reusable
rockets much more, but their choice of fuel makes a lot of sense.


Alain Fournier