Lessons Learned: Rockets in the 1980s and 1990s and now

[john smith 19]

That sounds like phooey to me. To test HTP stability they heat the liquid to 66 C or put it in a bath at 100 C! A bit of sunlight will do nothing. All the literature I've read says that HTP is very difficult to detonate. Here's what the British guide says on HTP detonation.

I was referring to the idea of using H2 as the pressurent in the LO2 tank of Aquarius. Apologies for not making that clear. I'm aware high purity HTP has been stored on orbit for up to 6 years in at least one early 60's comm sat.

But Isp beats all other considerations if you're the military, or someone else is paying to have the problems of toxicity dealt with.

Density also matters. HTP has a quite high density. 98%HTP/RP-1 has an average density of 1.3 kg/L, about the same solids when you include the void in the middle. Isp of this combination is about 14% better than solid. This makes HTP/RP-1 an excellent first stage propellant.

Excellent. Unfortunately Robert McNammara ruled out storables of any kind and mandated future ICBM would be solids. It would also need a new engine to use it.

HTP lost out to NTO/Hydrazines and once the US had backed solids it was pretty much game over.

It seems the days when ICBM designs were fairly readily transitioned to use as LV's are over. Now we are left with an LV architecture that mimics an ICBM (with the legal and political constraints that brings) without the accessible infrastructure in terms of mfg capability in engines and tanks.   

[john smith 19]

Ever, again, because till it's totally cleaned that's the most direct route of ingestion. However you'd be very wrong there Jim, it can be absorbed through the skin or across the eyes as either liquid or fumes. And that thimble full will also expose and probably be fatal to anyone within a 5 to 10 foot radius as it spreads.

See: http://www.toxipedia.org/display/toxipedia/Hydrazine

You not only require sealed breathing units to handle these propellants but full environmentally sealed suits and inner garments. Fumes are the most direct exposure route but are far from the only one.

Once I saw those SCAPE suits my first thought was "Looks just like an NBC suit for a battlefield."

It's the skin absorption that moves the 'zines from old fashioned WW I grade nasty to nerve-agents-too-dangerous-for-the-Nazis-to-consider-using level.

But let's note that China's main ELV's are still using it so it can be said to not be a deterrent to a successful LV design, if you launch somewhere with a more government friendly H&S policy. [EDIT and a large supply of replacement staff if anything does happen ]

[john smith 19]

and the artillery range to launch it over)

They are not going away even for RLV's.  Current space launch ops are to be around for quite some time.  It isn't going to be like aircraft

Well that sounds pretty unequivocal.  But let's examine it a bit closer.

Bono & Gatland in Frontiers of Space estimate the energy requirement to put a unit mass in space is about the same to send that mass on a round trip between the UK and Australia (London to Sydney IIRC)

So 1 way would be the first stage and the return would be the 2nd stage to orbit in energy terms.

In terms of energy that would make an airliner carrying the same mass as an LV equally as dangerous if it crashed.

Oh no. Flatten the houses under the flight path! Drive a 10 mile corridor through Africa! Anything might happen! If that sounds ridiculously hysterical that is what the requirements for ELV's are saying.

And yet every day airplanes make trips that long (with equally high levels of propellant on board, and hence equally high levels of energy release in a crash) without these precautions.

What are we to conclude?

a) International aviation is run by reckless manics with no regard for passenger safety, or anyone living below the flight paths of these dangerous monsters of the air.

b) There's something about the architecture of the vehicles used that lowers the risk levels of failure massively (by several orders of magnitude) over systems that throw 96-97% of the vehicle away every time.

That's not to say that every one of those vehicle is not being continuously tracked throughout it's flight (as likely as not by Rolls Royce or P&W engine monitoring departments). OTOH none of those tracking organization has  the power to unilaterally destroy the vehicle if it goes off course. 

Needing an artillery range to operate over is a design choice.

Not a law of physics.

As is not having propellant dumping arrangements so that what hits the ground is less than 10% of the GTOW of a fully fueled system, whereas an airliner is more like 50% structure and payload, so much less opportunity to dump mass and risk.

In fact wasn't Pegasus partly designed to side step needing a range to avoid range costs and bureaucracy? The fact it can't launch over built up areas is more to do with it's mostly solid design + hypergols US and self destruct system.

So taking that line of reasoning further could there be somewhere in the design space a design that does not need either a range or a self destruct system to operate safely?

At some point people will stop viewing it as a LV (OMG it's like a giant missile  ) and start viewing it as a large drone, looking at how its ConOps can be integrated into the globally controlled airspace before it moves out of national jurisdictions (80 000 ft max ?)

Like commercial airlines emergency procedures will have to be developed before emergencies happen (which is much easier for non expendable stages) and tested well enough that confidence is high they will achieve the reliability.

Autoland systems had a design requirement of 1 failure / 1 billion hours of operation, something  impossible to test directly without decades of landings but which the authorities have accepted is being met.

You build what you always built you get what you always got. 

[Danderman]

So, what are the lessons learned from the launch companies of the 80s and 90s?

[Jim]

1.  And yet every day airplanes make trips that long (with equally high levels of propellant on board, and hence equally high levels of energy release in a crash) without these precautions.


2 There's something about the architecture of the vehicles used that lowers the risk levels of failure massively (by several orders of magnitude) over systems that throw 96-97% of the vehicle away every time.


3.  Needing an artillery range to operate over is a design choice.

4. As is not having propellant dumping arrangements so that what hits the ground is less than 10% of the GTOW of a fully fueled system, whereas an airliner is more like 50% structure and payload, so much less opportunity to dump mass and risk.

5.  In fact wasn't Pegasus partly designed to side step needing a range to avoid range costs and bureaucracy? The fact it can't launch over built up areas is more to do with it's mostly solid design + hypergols US and self destruct system.

6.  So taking that line of reasoning further could there be somewhere in the design space a design that does not need either a range or a self destruct system to operate safely?

7.  At some point people will stop viewing it as a LV (OMG it's like a giant missile  ) and start viewing it as a large drone, looking at how its ConOps can be integrated into the globally controlled airspace before it moves out of national jurisdictions (80 000 ft max ?)

1.  Again can't compare to airplanes.  Not a relevant analogy.  Airplanes can glide and operate with engine out.

2.  Not happening.  There is nothing that says RLV's for spacelaunch will have better reliability than existing vehicles.

3.  No, it is a safety requirement.

4. The way to dump the propellant is to split the tanks.  There is no way an RLV can dump its tanks faster than burning it  though the engines.

5.  It still needs a range.

6.  Nope

7.  Not happening with chemical rockets

[john smith 19]

So, what are the lessons learned from the launch companies of the 80s and 90s?

With the right circumstances any propellant combination can work (solids for Pegasus, Hypergols for Long March, pressure fed for Garvey). With the wrong ones any one (HTP for Orbital and the X37b) can fail.

VC's invest in teams, even if their method is wrong. If you can convince on the team they will invest. Kistler showed that. AFAIK this was the only project with VC investment from day 1.

Almost no one (including Musk) starts with a truly clean sheet of paper. Every team has a starting skill set and design preferences. it's smart to play to your strengths but to be aware they can lead you into dead ends and that you have blind spots.

Low cost is unlikely to be enough for an ELV as it will have to establish a track record.

Work backward. What payload gives what price. How many of those do you have sell to bankroll your development programme? Otherwise whoever bankrolls the development programme is basically writing the money off.  Unless you can sell a complete system, then what they launch, and when and how many, become the owners problem.

If you don't have a very solid business plan you need an angel investor with very deep pockets and / or an anchor tenant will to gamble on you. And that still would not give you a solid business plan.

NASA and DARPA will pay top dollar if they think you've got something new that's relevant to their mission.

In the US NASA and the DoD are major uses of launch services. Sizing your system to meet some of their needs is good business. The market for nanosats is more along the lines of "build it and they will come."IOW markets are important. Who are you selling to? Do you know what they want? If what you're planning is not (quite) what they want how will  you persuade them to use it?

Eliminate pyrotechnics. They are expensive, impossible to confirm their fire levels and have about a 7% failure rate (in a survey of about 20 years of operation). Their shock levels (20 000g) have also damaged a fair number of payloads. Design them out.

Consider the whole system. The vehicle, the pad(s) (or runway)  and the ground support.  Boeing found about 100 databases used to maintain the Shuttle. Some were still paper. I'd expect data was entered multiple times as well. Despite this there was no single DB for all the fluid parts on Shuttle. No where you could ask "How many type X one way vales are used" for example.

Orbital grade GNC is difficult. SX have shown ways to get space grade processor reliability without resorting to $100k rad hard POWERPC boards.  The trouble is good sensors are expensive, likely subject to ITAR rules  and going to be thrown away on a flight. Either find a way to recover them or find a way to give orbital level performance with more commodity hardware.  Astronnav systems can work in daylight (SR71, B2, B52) and have given 6m resolution. Would a cameraphone be up to the task as a one shot system?

TPS is rarely mentioned. Fairly irrelevant for expendables but crucial for any kind of reuse.

Most of those 80/90's concepts were all about the upmass. Almost none could bring anything back. The ability to return mass (even if not human) lets customers implement a design/test/revise cycle based on actual returned flight hardware. This is one of the reasons sounding rockets and balloons are still in business.

Rockets are the simplest engine option but rockets tie you to rocket level Isp and rocket level structural fractions as a consequence of rocket level Isp. The key benefit of jet engine like air breathing is while it's working it gives 6-9x the Isp, which gives you a much more relaxed structural fraction requirement to meet. Engine design is harder  but structural design (can be) easier.

[EDIT Beware of any development team whose design credentials cannot be validated if you don't have security clearance.

Beware of any development team that offers huge investments down the road if the prototype flies.

I like to think of these as the X33 rules. ]

People buy specific ranges of upmass to get to specific orbits and they want to spend a specific amount to do so. The only way you can get them to use a substantially bigger LV will be for the supplier to broker the sale of the rest of the payload upmass. No business would want the trouble. Arianespace found this hard enough with 2 payloads to GTO. It's also the justification for the SLS, cheaper than the equivalent number of smaller launches.

[EDIT There is the Common state of practice and there is the State of the Art. The LV  business has been very conservative but in some areas the SoA is far beyond the common SoP.

For example the common SoP is hydraulic main engine TVC using main engine fuel as the fluid. However
Vega has now flown 8 successful launches with all of the first 3 stages using EMA's for 2 axis TVC powered by Li ion batteries. Modern high speed electrical machine technology could direct couple a generator or alternator to the turbopump drive shaft directly.

Likewise it's a common perception that composite LOX tanks are still years away and LOX cooling of combustion chambers will lead to near certain disaster. Both have been demonstrated safely in the early 90's and LOX (again) and air cooling of a combustion chamber by the DLR in (IIRC) 2009  ]

[EDIT As far as possible design the architecture to existing materials. Don't design an architecture and hope someone will come up with a material that allows it to be built ]

[RanulfC]

It was actually 90.5% HTP at 5 C. A paper on this was posted previously.

Right, and I posted it no less

Randy

[Jim]

Eliminate pyrotechnics. They are expensive, impossible to confirm their fire levels and have about a 7% failure rate (in a survey of about 20 years of operation). Their shock levels (20 000g) have also damaged a fair number of payloads. Design them out.

Wrong on every point.  Failure rate is more than a magnitude lower.  There has been no damage to payloads in the last 40 years.

Higher shock loads come from the clampband

[oldAtlas_Eguy]

So, what are the lessons learned from the launch companies of the 80s and 90s?

...
People buy specific ranges of upmass to get to specific orbits and they want to spend a specific amount to do so. The only way you can get them to use a substantially bigger LV will be for the supplier to broker the sale of the rest of the payload upmass. No business would want the trouble. Arianespace found this hard enough with 2 payloads to GTO. It's also the justification for the SLS, cheaper than the equivalent number of smaller launches.

Larger SHLV's should concentrate on bulk cargo and not satellites. Think of satellites as a custom market that requires custom orbits and services per flight for each customer. Where as large SHLV's provide low $/kg with little custom services for cargo like propellant or other very large amounts of stuff (equipment or parts for construction of large structures in space or on other bodies). Trying to use an LV for what it is not optimized for is a dumb marketing move.

An example is the difference between an F9 and a FH. The F9 is a custom services provider LV but the FH leans toward the bulk cargo provider LV. The F9 has a large number of customers with custom payloads going to custom orbits. The FH has a small number of customers with custom payloads going to just a few custom orbits.

[Danderman]

Concerning what the launch vehicles should do, we really don't have a baseline for commercial launch beyond F9/Ariane 5 class vehicles, so no idea what the market will support. Ie, we have no Lessons Learned yet.

BTW, this thread is about Lessons Learned, not What I Think People Should Do With Their Money.

[john smith 19]

1.  Again can't compare to airplanes.  Not a relevant analogy.  Airplanes can glide and operate with engine out.

I've taken the subject of this thread to be new LV companies in general.

So here's a thought.

What if  you had a LV that could glide?

2.  Not happening.  There is nothing that says RLV's for spacelaunch will have better reliability than existing vehicles.

You seem to be confusing making an assertion with a law of physics.  I'll agree the results from the shuttle were not impressive, but a lot of that was pretty much baked in from the day they were given the budget.

3.  No, it is a safety requirement.

As a consequence of flying an architecture that looks from the outside like an ICBM.

Which is perfectly fine if you are OK with the range operators launch schedule, or you have the resources to establish your own.

4. The way to dump the propellant is to split the tanks.  There is no way an RLV can dump its tanks faster than burning it  though the engines.

I'd guess it would depend if you want a system that can be repaired for relaunch afterward.

5.  It still needs a range.

Flight 15 was around Grand Caneria. I did not know the Spannish had a launch range in that area. Do you know how OSC managed?

6.  Nope

<Yoder> Great has been your exploration of the design space, hmm?

But on a more serious note you really need to realize the difference between an assertion and an actual fact. Unless you can prove a design theorem that shows this for a fact, which would save a huge amount of time for future design projects, there is no reason to think a better approach is not possible. Fiendishly difficult to find no doubt, but not, AFAIK impossible.

7.  Not happening with chemical rockets

And as long as they look like an ICBM they never will. 

Eliminate pyrotechnics. They are expensive, impossible to confirm their fire levels and have about a 7% failure rate (in a survey of about 20 years of operation). Their shock levels (20 000g) have also damaged a fair number of payloads. Design them out.

Wrong on every point.

It's impossible to confirm fire pulse levels. There is a no fire level (and you can confirm that that all devices don't  fire at this level, but you test those devices at the fire level and boom, you need another one. IIRC that's call the "all fire" level.
 
Those figures were from memory. IIRC from an NSF report on the susbjBut this report

https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19880017010.pdf

Lists 84 failures over a 23 year period. It does not seem to list the total number of times they were used.

Failure rate is more than a magnitude lower.  There has been no damage to payloads in the last 40 years.

Higher shock loads come from the clampband

Pyros are lightning sensitive. Depending on the range they may only be allowed to be installed on the pad, making them a PITA. Laser initiation can help with one but not the other.  And there is at least one device that is lighter than them and is firing pulse compatible.

[Danderman]

Have any commercial space companies failed due to pyro bolt issues?

[Jim]

again, stop with the assertion (and others) that they can be only install at the pad.  The ordnance is installed before going to the pad

[Jim]

That pyro report is almost 20 years old.  It is no longer applicable

[john smith 19]

again, stop with the assertion (and others) that they can be only install at the pad.  The ordnance is installed before going to the pad

That wasn't the case for Shuttle. Perhaps NASA will be more reasonable with SLS.

Launch systems can have problems in the design, manufacture, operating or maintenance (of the vehicle or it's infrastructure).

Pyros may be the default option (although a number of the failures listed in the report suggested even then things could go wrong) but that still leaves the fact  you'll need specially qualified staff to handle them during mfg and maintenance.

SX have made great play of the fact they don't use them for fairing or stage separation. I'd be interested to find out if there were any places they simply couldn't do without them.  They seem to be the only choice where very short activation times make a difference but outside of weapon systems and aircraft ejector seats how common a requirement is that?

[Katana]

again, stop with the assertion (and others) that they can be only install at the pad.  The ordnance is installed before going to the pad

That wasn't the case for Shuttle. Perhaps NASA will be more reasonable with SLS.

Launch systems can have problems in the design, manufacture, operating or maintenance (of the vehicle or it's infrastructure).

Pyros may be the default option (although a number of the failures listed in the report suggested even then things could go wrong) but that still leaves the fact  you'll need specially qualified staff to handle them during mfg and maintenance.

SX have made great play of the fact they don't use them for fairing or stage separation. I'd be interested to find out if there were any places they simply couldn't do without them.  They seem to be the only choice where very short activation times make a difference but outside of weapon systems and aircraft ejector seats how common a requirement is that?

Would any commercial space company manufacture pyros instead of buy them from qualified manufacturer?

[Stan-1967]

SX have made great play of the fact they don't use them for fairing or stage separation. I'd be interested to find out if there were any places they simply couldn't do without them.  They seem to be the only choice where very short activation times make a difference but outside of weapon systems and aircraft ejector seats how common a requirement is that?

My recollection of discussions here on NSF was that the SpaceX choice for avoiding pyrotechnics was that it was for testing purposes.  My understanding being when you test your hardware with pyrotechnics it is a destructive test.

The only place I can think of that SpaceX uses pyros is for the main chute of Dragon II LAS & I would also suspect they are used for deploying the cargo dragon parachute.  Stage sep, fairing sep, & payload deployment all seem amendable to not using pyro's if that is your choice.

Mars EDL as it is being successfully done by JPL/NASA relies heavily on timed events using pyros.  Pyros look to be the absolute & maybe only choice for EDL on any body with an atmosphere.  Propulsive landings may change that in the future.

[Jim]

again, stop with the assertion (and others) that they can be only install at the pad.  The ordnance is installed before going to the pad

That wasn't the case for Shuttle

Wrong. And you keep repeating this incorrect info on them.
 There were tens of pyros on the shuttle.  Payloads had them.  RMS did.  So did the hatch, and other places. They were installed and kept installed. 

That also goes for the ELV's.  They are installed before going to the pad.

[Jim]

They seem to be the only choice where very short activation times make a difference but outside of weapon systems and aircraft ejector seats how common a requirement is that?

Many places in space operations.  Even if Spacex eliminates them, spacecraft will still use them

[savuporo]

They seem to be the only choice where very short activation times make a difference but outside of weapon systems and aircraft ejector seats how common a requirement is that?

BRS parachutes counts over 350 lives saved and tens of thousands of systems installed. Pretty common.