An SSTO as "God and Robert Heinlein intended".

I discuss a suggestion to upgrade the Merlin to another thrust level to reduce the number of engines on the Falcon 9 for reliability here:

Re: On the lasting importance of the SpaceX accomplishment.
http://exoscientist.blogspot.com/2012/1 ... pacex.html


Bob Clark

Might be an opportunity to develop that before mentioned aerospike system, and while you are at it the fly back booster....

It very much depends on how you define reliability though. The bike ride part of my daily commute is very rarely perfect. Usually there are some red traffic lights that I have to wait for, tourists walking on the bike path, adverse weather conditions, and other inconveniences to deal with. On the other hand, so far I've always made it in a reasonable time and without harm to me or to anyone else, so as far as I'm concerned, I have a 100% success rate.

You seem to be defining success as "no engine failures". By that metric alone, the smaller the number of engines, the higher the chance of success. In fact, remove all of the engines and you'll have perfect reliability. But you're also reliably going nowhere. You know this of course, and perhaps you're simply arguing that NASA will get skittish, regardless of whether that would be justified. I won't argue with that.

But let's look at a different definition of reliability: "Gets the payload to the intended orbit in one piece". F9 can do this with an engine out from the start (89% thrust), and a second one later after IIRC 150s (78% thrust). So with only five engines providing the same total thrust, if an engine fails early (80% thrust remaining), it won't make this criterion. In this scenario, the more engines you have, the more you can lose, provided that the losses are independent.

That last point is pretty important. If there's a fundamental design flaw in either the engine or the construction process, then all your engines may well break in the same way at the same time, and if they all break then it doesn't really matter how many of them you have. Same thing if one failed engine can take out all the others. So let's assume that any engine failures are due to random process variation in production, that there are no common factors between individual engine failures, and that the F9's inter-engine shielding suffices.

You're talking about on average one failed engine per F9. Ignoring the second stage, and with the above assumptions, that means that the Merlin engine would have a probability of working correctly during a launch of 8/9 or 89%, or an 11% failure rate. Ignoring the second-engine-out capability, the probability of a successful launch (no engine out or one engine out) is then 1 * (8/9)^9 + 9 * (8/9)^8 * (1/9)^1 = 74%. With only five engines of the same 89% reliability and no engine out capability any more, it's 1 * (8/9)^5 = 55%. To get back up to 74% probability of overall mission success, the bigger engines would need an individual probability of success of 95%, twice as good as the smaller ones.

Alternatively, we could reduce the payload so we can have an engine out even with only five engines, and we get up to 1 * (8/9)^5 + 5 * (8/9)^4 * (1/9)^1 = 90% with the original engine reliability rate of 8/9. However, with 9 engines and reduced payload, we'd get full two-engine-out capability, and the probability of success becomes 1 * (8/9)^9 + 9 * (8/9)^8 * (1/9)^1 + 36 * (8/9)^7 * (1/9)^2 = 93%. Note that the difference is smaller now; the more engines you have, the smaller the benefit of adding even more.

More, smaller engines have another advantage, which is that the more you make of something, the more you can automate production, and the lower the unit costs. The cost of setting up the production line gets amortised across the same number of rockets, regardless of the number of engines, so that doesn't make a difference in this case. At some point though the additional weight and hassle isn't worth it any more. I imagine that SpaceX did some calculations and concluded that 9 is the sweet spot.

Kind of the other way around, the Merlin was designed to power the Falcon 1. To save costs and time (probably to be able to make NASA and other customer's time windows) they clustered them for the F9. 9 engines just happened to be the number needed to hoist that much bigger rocket. The engine out capacity and any statistical gyrations are incidental.

Engine out can't be incidental, otherwise, if all they wanted to do was launch the rocket, they could have got away with 8, so why have 9?

I wasn't on SpaceX's design team obviously, but the initial sizing of the Merlin was dictated by the payload specification for the Falcon 1. The nine was set by needed capacity of the F9 and the expected growth of the payload of the vehicle, because they always get heavier. They have a pad margin now, but probably won't on a manned Dragon or when/if they start lifting Planetary Resources' mining robots and stuff.

Carrying an "extra" engine as insurance, isn't. It's inefficient. It increases your costs both in monetary and payload terms. LVs are expendable items as are unmanned payloads (and manned one's have abort/escape plans). You already have to carry tons of real insurance regardless.

Well, but SpaceX is planning to make them reusable, and those engines are relatively lightweight. Also, remember that the original plan included flying F1, F5 and F9 all at the same time. Having them all use the same engine is going to save a lot of cost in tooling and production. Compare that to Arianespace who currently have light, medium and heavy rockets that couldn't be more different technologically.

Anyway I have a hard time imagining that SpaceX didn't think beyond F1 from the start. They've always said Mars was their goal, so obviously there's no lack of vision, and there's no doubt either that F1 is not going to get anything to Mars. They must have had a back-of-the-envelope calculation at least.

Elon insisted that they design the falcon 9 to be reusable, and that caused them to be a couple years behind what would otherwise be needed to complete the rocket. So they have always been planning for the long term.

Is there any way to have the engines mounted to a external ring, that are self contained, so different payloads or engine upgrades could be interchangeable and easily removed and serviced, or even left in orbit to facilitate future missions, or ejected in a mission critical engine malfunction, with the fuel, so emergency landings would have increased odds of not exploding.

Also this would allow many configurations to be tested without great modification, and it could be used in vastly different launch formats, like maglev mass launchers, supersonic flight to rocket accelerated orbit, or traditional vertical launch

Also a ring could have a mix of orbit ion style and atmosphere engines, and leave the heavy, empty engines at a depot, and go do a mission, and then retrieve or leave them in orbit.

A large sled in orbit with ion engines that a launchecraft could use and leave in orbit would also be intresting.

All that adds weight, complexity (expense), and reduces payload.

Well, his question wasn't should it be done, but rather, could it be done.

Those are often the same question.

This article on NasaSpaceFlight.com suggests the failure was in the
"fuel dome":

Dragon enjoying ISS stay, despite minor issues – Falcon 9
investigation begins.
October 19th, 2012 by Chris Bergin
http://www.nasaspaceflight.com/2012/10/ ... stigation/

The image on this site shows the area that would have been breached:

http://www.neowin.net/forum/topic/11035 ... y595235009

This is an area in the combustion chamber. A breach of the combustion
chamber is very serious because the combustion chamber is where both
the pressures and heat are highest. Note that this would be at the top
where most of the complex and delicate engine components are located.
Quite likely hot combustion products shooting out at high pressure in
this area would have produced shrapnel from this part of the engine.

This is very serious because these lead directly into the propellant
tanks. This means likely if the engine had not been shut down it would
have led to an explosion. This is a different scenario from an engine
just being shutdown because it is giving anomalous readings while the
engine remains intact.

This puts it in a different and more dangerous class than other engine
shutdowns.

Bob Clark

This is kinda the same devil I was mentioning, by fuelling all nozzles with the same tank, there is only on and off for the one tank, a self contained engine, with its own fuel supply could in theory ditch a whole engine, but it would have to ditch or modify the nozzles of the remaining engines to keep stability until safe shut down abort, a very very strong crew capsule for launch that could eject all engines, with explosive bolts and quick disconnect connections.

And if an engine blew......... It just seems safer the riding it.

Things happen fast during a launch. You are either going to have a successful shutdown, or a catastrophic failure in the tenths of a second that your control system has to detect and react to it (if it has that luxury at all). Jettisoning a component, any component, is in a different time-scale. Seconds instead of tenths, and of little value from a safety and success of the launch perspective.

I suppose it might be worth while to be able to lose the hundreds of pounds of mass from a dead engine. But that adds complication. Explosive bolts, dry break and quick disconnect fittings, etc. That actually adds failure modes and cost.

You need to keep the propellants though, because of the math of rockets. Your remaining healthy engines need to burn longer to reach the same required DV {the old "abort to orbit"). Also one big tank is more mass efficient than a bunch of smaller ones.

If it's really not your day, your crew escape system has to be able to pull the passengers away from a doomed launcher at all points of a flight. That is the big part of any "manned rating". Having a way of saving the crew covers a lot of the "what if" contingency of the rest of the hardware. Losing boosters and cargo is a claim with the insurance company. Losing people puts you out of business.