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

[Kabloona]

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

There's not much incentive. Pyros are a tiny fraction of a launch vehicle budget, but the design and test requirements levied on pyros by NASA and Air Force for launch site ops are quite stringent, and it takes a lot of effort and expense to get new pyro devices qualified and range approved. And the extremely high reliability requirement for pyros means that hundreds or thousands of a particular device must be test fired in order to verify its required reliability, which favors off-the-shelf components whose reliability has been demonstrated over years of experience. Plus, working with extremely energetic precursor materials requires extensive facilities and safety measures, ie a lot of capital expense. So you may as well buy stuff that's already designed, tested, and qualified, right off the shelf, and save yourself the headache.

Even SpaceX, who have brought a lot of design and build in-house, would be hard-pressed to make a business case for designing, building, qualifying, and testing their own pyro devices in-house, I would expect.

[Katana]

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

There's not much incentive. Pyros are a tiny fraction of a launch vehicle budget, but the design and test requirements levied on pyros by NASA and Air Force for launch site ops are quite stringent, and it takes a lot of effort and expense to get new pyro devices qualified and range approved. And the extremely high reliability requirement for pyros means that hundreds or thousands of a particular device must be test fired in order to verify its required reliability, which favors off-the-shelf components whose reliability has been demonstrated over years of experience. Plus, working with extremely energetic precursor materials requires extensive facilities and safety measures, ie a lot of capital expense. So you may as well buy stuff that's already designed, tested, and qualified, right off the shelf, and save yourself the headache.

Even SpaceX, who have brought a lot of design and build in-house, would be hard-pressed to make a business case for designing, building, qualifying, and testing their own pyro devices in-house, I would expect.

So if pyros are bought from the same manufacturer that serves ULA, they have the same reliability. The problem of pyro reliability should be the LEAST important to commercial launchers, as anything done inhouse have more trouble.

[john smith 19]

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. 

The figure I use came from maintainability and operability study reports done by Boeing in the mid 80s and operability studies headed by a Mr Zapata at NASA> They list 300 pyros, of which 100 were normally fired during a Shuttle mission. I'm not clear if that number was just for the orbiter or the whole stack.

It seems pyros were popular as backup or backup-to-backup systems, such as landing gear extension and dumping the docking radar antenna in case it could not retract. My guess is the 17" pipe separation and the closing of those protective doors also had a pyro backu but I've never found a full list of pyro uses on the Shuttle.

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

Even SpaceX, who have brought a lot of design and build in-house, would be hard-pressed to make a business case for designing, building, qualifying, and testing their own pyro devices in-house, I would expect.

yes. Better to design out than to buy in.

[Kabloona]

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. 

The figure I use came from maintainability and operability study reports done by Boeing in the mid 80s and operability studies headed by a Mr Zapata at NASA> They list 300 pyros, of which 100 were normally fired during a Shuttle mission. I'm not clear if that number was just for the orbiter or the whole stack.

On that subject, NASA was quite paranoid (in a good way, of course) about pyros on Shuttle payloads. Payloads were required to use NASA standard initiators (NSI's) to reduce the probability of accidental initiation from stray currents, and of course the avionics had to be two-fault-tolerant to accidental pyro initiation, so the payload pyro systems and related black boxes got an extreme degree of scrutiny by JSC's safety review team.

Even so, pyro-related accidents happened. I've mentioned the Super*Zip incident on the TOS/ACTS Shuttle mission that sprayed the cargo bay with shrapnel/debris. That was an electrical system design error that was missed by numerous JSC payload safety review boards, and of course by the contractors who designed and built the system.

The "lesson learned" from that accident was that true end-to-end testing of subsystems is crucial. The subcontractor (Lockheed Martin) had tested the electrical functionality of the pyro initiation circuits but had not gone as far as mechanically mating the output electrical connectors for the pyros to a mechanical "detonator block" simulator, which if they had done so would have shown that one of the pyro initiation signals was going to the wrong pyro device.

Although that wasn't a commercial mission, the lesson is still valid across the board, and one it seems SpaceX has also taken to heart, as their systems engineering philosophy emphasizes testing.

• Test rigorously and at multiple levels of integration—including right before service

(From page 7 of this presentation: https://www.aiaa.org/uploadedFiles/Events/Conferences/2012_Conferences/2012-Complex-Aerospace-Systems-Exchange-Event/Detailed_Program/CASE2012_2-4_Muratore_presentation.pdf)

As observed upthread, pyros conflict somewhat with this philosophy, but if you must use them, at least you can do end-to-end testing right up to the output of the electrical connector that mates with the pyro device.

[Jim]

yes. Better to design out than to buy in.

Wrong.  They are very reliable and simple.   

[Jim]

Although that wasn't a commercial mission, the lesson is still valid across the board, and one it seems SpaceX has also taken to heart, as their systems engineering philosophy emphasizes testing.

They don't do "systems engineering"

[Jim]

The figure I use came from maintainability and operability study reports done by Boeing in the mid 80s

Woefully outdated

[john smith 19]

The figure I use came from maintainability and operability study reports done by Boeing in the mid 80s

Woefully outdated

Old certainly. How does that change the numbers listed? Are you aware of a project to radically lower the number of pyros used on STS? I've not found any indication of such a project, unlike say the Alternate Ground Cooling project reports, which seemed relatively minor in comparison.

[Jim]

The figure I use came from maintainability and operability study reports done by Boeing in the mid 80s

Woefully outdated

Old certainly. How does that change the numbers listed? Are you aware of a project to radically lower the number of pyros used on STS? I've not found any indication of such a project, unlike say the Alternate Ground Cooling project reports, which seemed relatively minor in comparison.

What the STS did or was going to do is not applicability to other vehicles.

[john smith 19]

Old certainly. How does that change the numbers listed? Are you aware of a project to radically lower the number of pyros used on STS? I've not found any indication of such a project, unlike say the Alternate Ground Cooling project reports, which seemed relatively minor in comparison.

What the STS did or was going to do is not applicability to other vehicles.

Given it's the only partly reusable LV system (as opposed to a spacecraft like Dragon, CST-100 or Dream Chaser are designed to be) that's every flown more than once it offers the only known data base on the problems of operating such a system.

Obviously not too many are applicable if you're planning YATSTO VTO LV but if you're looking to do something more creative it's a rich source of lessons learned (or lessons to be learned).

[Jim]

Given it's the only partly reusable LV system (as opposed to a spacecraft like Dragon, CST-100 or Dream Chaser are designed to be) that's every flown more than once it offers the only known data base on the problems of operating such a system.

Obviously not too many are applicable if you're planning YATSTO VTO LV but if you're looking to do something more creative it's a rich source of lessons learned (or lessons to be learned).

Not really.  And I was just talking about pyros.

Anyways, despite what you think:
a.  More lessons had to be unlearned because they were not applicable to other vehicles, no matter what type.
b.  Current vehicles have more applicability to RLV than the shuttle.
c. Future launch operations are going to be more like existing operations than airliners and airports.
d.  As much as you deny it, ICBM type ops are going to be with us as long as chemicals are used for power.

[Robotbeat]

Just because Shuttle is the only partial orbital RLV used operationally as an RLV doesn't mean we have to throw our brains away and do the whole cargo cult "Shuttle is the template" thing.

You could build a rocket without any pyros at all. Or filled with them. For an RLV, it probably makes sense to limit the pyros since you, of course, have to replace them every time. Which is just fine (and probably the optimal solution) on an expendable stage or for a spacecraft whose deployment mechanisms are only used one time.

One really outdated data point isn't necessarily better analysis of something that has never been done operationally before.

[leaflion]

Old certainly. How does that change the numbers listed? Are you aware of a project to radically lower the number of pyros used on STS? I've not found any indication of such a project, unlike say the Alternate Ground Cooling project reports, which seemed relatively minor in comparison.

What the STS did or was going to do is not applicability to other vehicles.

Given it's the only partly reusable LV system (as opposed to a spacecraft like Dragon, CST-100 or Dream Chaser are designed to be) that's every flown more than once it offers the only known data base on the problems of operating such a system.

Obviously not too many are applicable if you're planning YATSTO VTO LV but if you're looking to do something more creative it's a rich source of lessons learned (or lessons to be learned).

New Shepard has been reused.  They clearly use pyros in their crew capsule, for the escape SRM and probably the parachutes.  So as a stack they have pyros.  Doesn't seem to bother them.

Also, SpaceX's FTS clearly uses pyros.  So I really have no idea why we're having this discussion.  Jim's right, pyros are super useful and not a big deal.  ULA uses them all the time and they have no issues with reliability.

[baldusi]

BTW, shuttle has nothing in common with current LV. Specially those designed for reusability. Just look at how payload was mated and how launch stresses were transported through the structure. Nobody carries their core propulsion to orbit, either. Nobody does crewed unless that's their mission.

[Kabloona]

Old certainly. How does that change the numbers listed? Are you aware of a project to radically lower the number of pyros used on STS? I've not found any indication of such a project, unlike say the Alternate Ground Cooling project reports, which seemed relatively minor in comparison.

What the STS did or was going to do is not applicability to other vehicles.

Given it's the only partly reusable LV system (as opposed to a spacecraft like Dragon, CST-100 or Dream Chaser are designed to be) that's every flown more than once it offers the only known data base on the problems of operating such a system.

Obviously not too many are applicable if you're planning YATSTO VTO LV but if you're looking to do something more creative it's a rich source of lessons learned (or lessons to be learned).

New Shepard has been reused.  They clearly use pyros in their crew capsule, for the escape SRM and probably the parachutes.  So as a stack they have pyros.  Doesn't seem to bother them.

Also, SpaceX's FTS clearly uses pyros.  So I really have no idea why we're having this discussion.  Jim's right, pyros are super useful and not a big deal.  ULA uses them all the time and they have no issues with reliability.

@leaflion, Maybe you missed the earlier portion of this thread. We're having this discussion partly because of the study SpaceX commissioned which showed that the second most common cause of LV failures in the previous 20 years was separation events, most (all?) of which were pyro-driven. Those included Pegasus and Taurus stage and fairing separation systems, and the entire Taurus program tanked because of those failures.

Partly as a result of that study, SpaceX opted to minimize the use of pyros for separation events and to use mechanically-driven sep systems wherever possible.

Yes, pyro devices  (initiators, boosters, linear shaped charges, etc) themselves are typically highly reliable. But the problem is almost never that the pyro device fails to intitiate. The pyro usually *does* initiate, but the relevant mechanical components fail to separate completely. For example, a Pegasus composite interstage failed to separate completely despite pyro intiation, and two Taurus fairings failed to separate despite pyro intitiation.

So the crucial destincition is between the reliability of the pyro device *itself* and that of the separation system which incorporates it. That system can't be fully tested on the vehicle before flight because of the destructive nature of the pyro event, which sometimes leads to the type of failure described above. The pyro device initiates, but the related components fail to separate as designed.

Which leads to SpaceX's approach, using mechanically-driven, pre-flight-testable stage and fairing sep systems in place of pyro-driven systems wherever possible.

And as for FTS, SpaceX would probably prefer to use other non-explosive means if Range rules allowed, but at present they are required by Range to use explosive FTS components. Also, FTS components are probably the most rigorously tested at both the component and the system level, so they are super-reliable, and no one is questioning them here. The issue is the reliability of sep systems, etc, which use pyro initiators but cannot be fully tested at the system level pre-flight, leading to the types of Pegasus and Taurus sep failures cited above.

[leaflion]

Half the issues people are raising with pyros are not separation system specific.  I'm mostly addressing those.

If the pyro initiates and you're failing to separate, then the issue is with your mechanism design.  It is not the pyro's fault.  The Taurus fairing separation failures were due to poor material tracibility and not qualifying the design to the proper environments.  They could affect any system, pyro or not.  They were not tested properly, not becuase they were pyro, because Orbital thought they could get away with not delta-qualifying to new environments.

[rpapo]

And as for FTS, SpaceX would probably prefer to use other non-explosive means if Range rules allowed...

Remember the Grasshopper II failure.  IIRC, the FTS there was not the standard explosive one, but something SpaceX cooked up.  At the Cape, or Vandenberg, of course, they use the pyro-based system required of them.

Though the one time it was needed (CRS-6?), it didn't get fired until the rocket was already toast...

[baldusi]

Which means that on CRS-6 it was not needed. As simple a that. And the G2 failure FTS mechanism was pretty simple: cut the pressurization and propulsion system and open all the valves.

[john smith 19]

And as for FTS, SpaceX would probably prefer to use other non-explosive means if Range rules allowed, but at present they are required by Range to use explosive FTS components.

I did not know this. I thought any system that terminated flight, such as closing the main propellant valves, or puncturing the tanks, was OK.

Also, FTS components are probably the most rigorously tested at both the component and the system level, so they are super-reliable, and no one is questioning them here. The issue is the reliability of sep systems, etc, which use pyro initiators but cannot be fully tested at the system level pre-flight, leading to the types of Pegasus and Taurus sep failures cited above.

I'm curious. Do all drones carry an FTS? Or drones above a certain size?

Anyways, despite what you think:
a.  More lessons had to be unlearned because they were not applicable to other vehicles, no matter what type.

Which is a lesson in itself. 

b.  Current vehicles have more applicability to RLV than the shuttle.

Excellent. What would you say those lessons are?

c. Future launch operations are going to be more like existing operations than airliners and airports.
d.  As much as you deny it, ICBM type ops are going to be with us as long as chemicals are used for power.

If it flies like a ballistic missile (especially dropping bits incapable of self guidance or landing) I'd fully expect it to need an artillery range like a ballistic missile, be it a nanosat LV or the ITS.

But extending that belief to all LV architectures is extrapolating too far. 

They don't do "systems engineering"

It's Systems Engineering, Jim, but not as you know it.   

Every aerospace company does SE, if only because doing SE is crammed down the throat of every undergrad on an aeroeng course

Historically the computing power to simulate everything to the Nth degree did not exist and people had to use engineering judgement (why SE's should have broad engineering experience). People recognized that flight tests were an essential part of the process.

The issue is people seem to think you can simulate everything before test and that guarantees a system will work right first time. This ignores 2 problems.
1) A simulation is only as good as its ability to accurately  model physical reality. Every approximation that makes the sim more tractable degrades that model. In theory it makes no difference. Until it does.
2)The benefit of doing multiple sim runs is to see how often an effect appears and over what range. A significant element of that is how accurately the random number generator models the variability of the failure mode or other phenomenon that's being modeled.

SX seem to have retained the idea that there comes a point of diminishing returns when the simplest way to resolve further questions is to build a flight version and test it. That said how you test new hardware or ConOps can have unexpected consequences, as the events at SLC 40 showed.  The mistake (in hindsight) was not to leave the payload on the vehicle unless absolutely necessary to give accurate results.

My suspicion is that no simulation would have picked up what caused the explosion because the phenomena that caused it are not captured by any standard software. 

SE problems are shared by all companies in the field, wheather they acknowledge them or not.  I shall be very interested to discover if IVF performs without any anomalies when it first flies for example.

[john smith 19]

Just because Shuttle is the only partial orbital RLV used operationally as an RLV doesn't mean we have to throw our brains away and do the whole cargo cult "Shuttle is the template" thing.

Nor are you being asked to. There was a lot wrong with the design starting with the funding profile, which eliminated more robust full TSTO. Getting a "non-insane" funding profile is the first "lesson learned".

You could build a rocket without any pyros at all. Or filled with them. For an RLV, it probably makes sense to limit the pyros since you, of course, have to replace them every time. Which is just fine (and probably the optimal solution) on an expendable stage or for a spacecraft whose deployment mechanisms are only used one time.

And yet SX went with non pyro shroud separation, which is a one time event on any mission.

One really outdated data point isn't necessarily better analysis of something that has never been done operationally before.

How curious. Jim said pretty much the same thing. That it's outdated.

But in launch vehicle pedigree is a big thing.  STS flew 135 flights over 30 years of operation. 

Note that I'm counting a lesson leaned in how not to do something is still a lesson learned. You seem to think lessons leaned is a checklist. Most of them do actually need some thought to apply.
[EDIT STS also sets a benchmark for what is possible. For example Discovery flew 39 times. It will be interesting to see how many times F9 first stages are re-flown, once they start to be re-flown ]