It really is rocket science

[Jason Thompson]

The time needed to transmit and receive signals to and from a probe landing on the moon is significant.

Yes it is, which is why remote control landing was not attempted.

To program a device to land via a set program could not take into account any discrepancies for significant alterations in parameters, such as altitude, terrain so on.

This is not strictly true. Yes it is true that you cannot simply program the vehicle as if it were landing on a perfect sphere. What you can do is a) program it such that it will land safely on a variety of landscapes, and b) design some feedback systems in so that it can in effect 'see' the ground it is trying to land on.

You can achieve a) by programming a rapid initial deceleration then maintaining a speed that will not destroy the probe on impact with the ground no matter if it occurs earlier or later than the ideal altitude (say if it runs into a hummock or a crater).

b) is achieved by the use of systems such a radar units or contact probes, which tell the probe exactly how far away the surface is at any given time and command it to respond accordingly. The Apollo lunar module, for example, used both methods. A radar told the crew how high they were and how far they were descending, and the contact probes on the landing gear told the crew they had touched down. In Apollo this input was output as information on displays and the crew responded to it. It would not require a complicated processor to take those inputs and use them to trigger commands to tell the craft to do something, for example slow down if it is going too fast, or shut down its engine when it is just about to land.

And no, there was no way to avoid unexpected boulders, or the problems of inadvertently landing the craft in the edge of a sheer drop and having it tumble in. That's just the risk of the mission. A few years ago I attended a talk by John Zarencki about the Cassini-Huygens mission, in the run up to the landing of Huygens on Titan. In the Q&A session afterwards I asked about what would happen if the lander hit a rock and flipped over, and his response was the 'I'm glad you asked that' reply of someone who had hoped no-one would find the flaw in his plan. There was indeed, as he told us, no provision for such an eventuality and it was simply considered to be a risk that was acceptable. The mission would still yield usueful information during the descent.

To extend Jay's dark room analogy, you walk with your hand outstretched so that your fingertips hit the wall first. That way the impact is absorbed by your arm and not your face, allowing you to slow down before you hit the wall and do yourself some damage. That's the same principle as a contact probe under the feet of a lander. What that can't help you with, however, is the toybox on the floor that was unexpectedly left there for you to trip over, but to be scanning everything all around you for any obstacles all the way is a huge effort for the desired outcome, so you simply accept that as a risk.

[ChrLz]

Forgive my imagination...

To program a device to land via a set program could not take into account any discrepancies for significant alterations in parameters, such as altitude, terrain so on.

In a NASA office, far away and long ago..

Chair:  Now, onto the landing procedures.  As the lander descends towards the cratered & rocky surface, what will..
ProfM: Um, excuse me.. What did you say?
Chair: You mean about descending, or the cratered, undulating and rock-strewn surface?
ProfM: Craters, you say??? Undulating surface, you say?  And there are rocks?
Chair: Well, yes, of course there are - we know that because..
ProfM: No, no no - all of my calculations and procedures were based on the fact that the Moon is a perfectly smooth sphere!  After all, I have looked at it through my binoculars - it is clearly smooth.
Chair: I'm sorry Munkin, but that isn't the case..
ProfM: Well, we CANNOT land there, then!!  How could we POSSIBLY land on something where we might ... tip over, or land at an angle?  I'm sorry, it is impossible.  The mission is off.


Me, I think problem-solving should be taught from kindergarten..

[twik]

I can't quite figure out from the question - is our Professor merely curious as to *how* they did it, or is he suggesting that it was impossible, so the unmanned missions were faked?

[Bob B.]

Can anyone briefly summarize the relative advantages and disadvantages of the various turbine-driven rocket engine fuel cycles, i.e., gas generators vs staged combustion vs expander cycle? I've read the Wikipedia articles for each but I don't fully understand why they each have the advantages they do.

In particular, why is a gas generator cycle engine inherently less efficient? The turbine exhaust isn't actually thrown away, it is still ejected downward where it can add a little thrust. In single-engine stages it can even be gimbaled to provide roll control, which I would think is a very useful feature.

I know that the fuel-rich mixture used in most gas generators does represent unburned and therefore wasted fuel, but this can come in handy for nozzle protection as in the F-1, possibly allowing the bulk of the exhaust to be at a hotter temperature than the engine might otherwise tolerate.

The turbine exhaust of a gas generator doesn't really add all that much thrust; most of the energy is used to drive the turbines.  The gas exiting the turbines has relatively low temperature and pressure, so there's not a lot more that can be gotten out of it.  Usually the amount of propellant that is routed to the gas generator is about 2% to 7% of the total, so all other things being equal, a gas generator engine should be just a few percent less efficient than staged combustion.

The big difference that I've observed is that staged combustion engines can operate at much higher pressures, yielding considerably higher specific impulse compared to gas generators.  The highest chamber pressure I can recall for a gas generator engine is about 96 atmospheres in the RS-68, while some staged combustion engines operate at chamber pressures exceeding 250 atmospheres.  This, I think, is really the big advantage of staged combustion.

Running the pumps of a gas generator engine at a higher pressure means more propellant must be routed to the gas generator/turbines, and since this propellant produces little in the way of thrust, increasing the flow harms the efficiency of the engine.  We therefore have two competing factors at work -- increasing the chamber pressure increases efficiency, but the resulting greater propellant flow to the gas generator decreases efficiency.  These factors have to be balanced, and it seems that the upper limit on the chamber pressure of a gas generator engine is about 100 atmospheres.

In a staged combustion engine, the turbine exhaust, which is still at very high pressure, is routed into the combustion chamber and burned with the remaining propellant to produce thrust.  In this design a very large proportion of the propellant can be routed through the preburner and turbines without a significant decrease in efficiency.  This allows the engines to operate at very high pressures compared to gas generators.  In some designs, as much as 100% of one propellant (fuel or oxidizer) is routed through the preburner/turbines before being injected into the combustion chamber.

The expander cycle has the advantage of staged combustion in that all the propellant produces thrust, but it has the disadvantage of a gas generator in that the operating pressure seems to be limited.  The expander cycle works only with cryogenic propellants, most commonly LOX/LH2 though probably also LOX/CH4.

[Bob B.]

b) is achieved by the use of systems such a radar units or contact probes, which tell the probe exactly how far away the surface is at any given time and command it to respond accordingly.

And it should be noted that Surveyor was equipped with doppler and altimeter radars to measure descent velocity and altitude.

[Jason Thompson]

It's a variation of the argument we had with fattydash on the old boards regarding the rendezvous with the orbiting CSM, in which he wanted to argue that it all needed to be calculated in advance. In fact it's a combination of a calculated course getting you roughly where you need to be, then an active system updating your information for refining the final approach.

[Bob B.]

I can't quite figure out from the question - is our Professor merely curious as to *how* they did it, or is he suggesting that it was impossible, so the unmanned missions were faked?

He's an HB, what do you think?  The number of times I can recall an HB being honestly just curious I can probably count on one hand.  They're almost always probing for some piece of information they can use to claim hoax, regardless of how innocent they make their inquiries sound.

[Echnaton]

Thanks guys for both the question and the answer about gas generators vs staged combustion.  I learned something new and interesting. 

[JayUtah]

Can anyone briefly summarize...

Jay doesn't do "briefly." 

In particular, why is a gas generator cycle engine inherently less efficient?

Briefly, thermodynamics.  You throw away heat and enthalpy when you dump turbine exhaust overboard, even if the exhaust can be put to productive use in other ways.  Feeding it back into the main thermodynamic design captures it more effectively in terms of measured thrust.

The turbine exhaust isn't actually thrown away, it is still ejected downward where it can add a little thrust.

Not typically enough to matter.  You gain more by directing it back into explicit thrust-generating processes rather than trying to use its thrust directly.  Ironically the turbines in the open-cycle solution can be a little more efficient because their exhaust vents to the ambient.  You can extract more mechanical power through conscientious mechanical design of the turbine.  The exhaust in a closed-cycle design has to maintain enough residual pressure to feed the next stage where you have to contend against chamber pressure, hence the turbine design has to be a little less efficient.

Consider the difference between marine turbine engines and aircraft turbine engines.  The powerplant on a Perry-class frigate in the U.S. Navy is composed of two GE engines of exactly the same type as are used on the Boeing 747 (4, in this case) to provide thrust.  It's quite valid to say that a Navy frigate and a passenger airliner use the same engine.  However, in the marine case the aft-end turbine is designed extract as much mechanical power as possible, leaving very little power in the flow of the exhaust gas, which is simply vented out the stack.  But in the aerospace case, the turbine is design to extract only as much mechanical power from the exhaust stream as is required to operate the engine's compressor and bypass fan and PTO needs of the airframe, and leave as much power in the exhaust stream as possible for producing Newtonian thrust.

In single-engine stages it can even be gimbaled to provide roll control, which I would think is a very useful feature.

Yes, this is often done.

[raven]

...the main bus ejected the airbag clad instrument unit, which unfolded like a flower, revealing the camera and antenna. A rather elegant system I must say.

And a system still in use today for Mars landers.

Yes, Oppoertuinity and Spirit, as well as Mars Pathfinder, also used what can be most succinctly described as "lithobreaking" for the final decent.
Ranger probes also used it in a rather terminal fashion.

[ka9q]

Bob & Jay, thanks guys. That helps.

I was curious to know just how much power the turbopumps represent in the big picture, i.e., how much mechanical power the pumps apply to the propellants versus the overall heat rate of the rocket engine.

Turns out it's pretty tiny, for the F-1 at least. Using the available figures for the RP-1 and LOX flow rates, their densities, the combustion chamber pressure and the S-IC tank ullage pressures, I get 10.8 MW for the LOX pump and 6.7 MW for the RP-1 pump, per engine. Using the heating value of RP-1, that's only 0.05% of the total heat rate of the engine of 34.2 GW. And that's for a single F-1 engine! I'm sure the real turbines developed considerably more as I neglected the kinetic energy in the moving propellants and the inefficiency of the pump impellers.

Do those numbers sound about right?

So yeah, using even 2% of your propellant in a gas generator to develop only 0.05% of its energy content does seem rather inefficient, even for a heat engine.

[ka9q]

It occurs to me that if turbines in conventional gas-generator cycle engines must run with inefficiently rich mixtures of the main propellant supply to keep from burning up, why not use some auxiliary fuel that could operate the gas generator more efficiently?

The fuel that comes to mind is hydrazine, or perhaps hydrazine hydrate to lower the temperature if necessary. Catalytically decomposed hydrazine contains no oxygen, just ammonia, nitrogen and hydrogen, so it should be fairly benign on turbine blades. Another advantage is that since hydrazine spontaneously decomposes when it hits a catalyst, the complex bootstrapping processes now necessary to start many large rocket engines could be simplified.

The Shuttle's APUs were powered this way, although they were much too small (135 hp) to drive the turbopumps on a large rocket engine.

Edit to add: I'm not surprised to see this is not an original idea. The Germans thought of it first by using H₂O₂ and a catalyst to drive the alcohol and LOX turbopumps in the V2.

[ka9q]

Ranger probes also used it in a rather terminal fashion.

Yes, it's always seemed a shame that the only way to soft-land on the moon is to burn even more propellant than you need to get back off the surface again. If only you could land on some sort of net stretched across a big crater. Problem is, it would have to be a rather deep big crater.

The moon's escape velocity is 2.38 km/sec, so that's the absolute minimum speed at which you'd hit the surface on a direct unbraked approach from infinity. If you could tolerate a maximum of 10g's, that would be a deceleration of about 98 m/sec², so it would take you 24+ seconds to come to a stop.

Okay, that's not so bad with a really well designed couch but during those 24 seconds you'd travel 28.9 km. That's a pretty deep crater.

[Bob B.]

It occurs to me that if turbines in conventional gas-generator cycle engines must run with inefficiently rich mixtures of the main propellant supply to keep from burning up, why not use some auxiliary fuel that could operate the gas generator more efficiently?

I haven't studied this issue real close, but I think the disadvantage of using a third propellant, such as H₂O₂ or N₂H₄, is a higher inert mass.  There are additional tanks, piping, etc. that aren't necessary when we tap off the main propellants.  Of course, the main propellant tanks can be a little smaller, but I doubt this offsets the mass of a separate system.  I'd also like to see how much power can be obtained from a kilogram of monopropellant versus a kilogram of fuel-rich bipropellant to see if there is an advantage to one versus the other.  I should be able to perform these calculations, but I'll have to get back to you.

Edit to add: I'm not surprised to see this is not an original idea. The Germans thought of it first by using H₂O₂ and a catalyst to drive the alcohol and LOX turbopumps in the V2.

One of the main reasons the Germans used this method is because they had prior experience with it, having used H₂O₂ decomposition to power torpedos.  The Germans also had the ability to manufacture H₂O₂ in high concentrations (I think about 70% at the time of WWII).  The American Redstone also used the method, being that it was a derivative of the V-2.  Likewise, many of the early Russian designs used  H₂O₂ to power its turbopumps, including the R-7.  I'm pretty sure the present day Soyuz launch vehicle still uses it.

edit spelling

[Glom]

Ranger probes also used it in a rather terminal fashion.

Yes, it's always seemed a shame that the only way to soft-land on the moon is to burn even more propellant than you need to get back off the surface again. If only you could land on some sort of net stretched across a big crater. Problem is, it would have to be a rather deep big crater.

The moon's escape velocity is 2.38 km/sec, so that's the absolute minimum speed at which you'd hit the surface on a direct unbraked approach from infinity. If you could tolerate a maximum of 10g's, that would be a deceleration of about 98 m/sec², so it would take you 24+ seconds to come to a stop.

Okay, that's not so bad with a really well designed couch but during those 24 seconds you'd travel 28.9 km. That's a pretty deep crater.

That gives me Bugs Bunny visions.