Index Home About Blog
From: Henry Spencer <>
Subject: Re: Apollo Design Steps
Date: Mon, 7 Apr 1997 18:08:09 GMT

[Yes, this is old -- I'm catching up on a backlog.]

In article <5hd25a$sno$>,
Dwayne Allen Day <> wrote:
>: than previous LOX/kerosene engines that it gave considerable trouble in
>: general and had horrible problems with combustion stability in particular.
>As early as 1958 they were predicting problems with combustion instability
>with the F-1 engines.  The reason for this is simple:  ALL rocket engines
>have this problem initially...

Dwayne, I think you're confusing two different issues.  Pogo oscillation
is extremely common (although not quite universal) in substantial rockets
built without Pogo suppressors, but that isn't where the F-1 had its real

The F-1's problem was high-frequency (acoustic) instability, and that
scales very strongly with size.  Basically, oversimplifying somewhat,
details of injector design fix the frequencies where the engine is most
vulnerable to high-frequency oscillation, and if the major resonant
frequencies of the chamber are above them, problems are unlikely.  Other
things being equal, frequency is inversely proportional to dimensions, so
big engines have a lot more trouble than small ones.  Little engines
rarely see it; in fact, I would have said "never", except that NASA
actually managed to provoke high-frequency instability in the shuttle RCS
engines -- quite a surprise at the time.

The biggest problem with solving the F-1's high-frequency instability was
not licking the instability, but doing so without incurring much of an
efficiency penalty, because the F-1 didn't have any efficiency to waste
at the time, and changes which reduce high-frequency instability are
notorious for reducing efficiency too.  (As a gross example, the most
obvious type of injector, which has each fuel jet impinging on an oxidizer
jet, is very good for performance -- it gives good and rapid mixing -- but
is also very bad for stability -- probably because the properties of the
two fluid streams differ enough that they are affected differently by
disturbances in the surrounding gas, so the mixing is easily disrupted.
Impinging fuel jet on fuel jet and oxidizer jet on oxidizer jet to break
up the jets, while relying vaguely on downstream mixing to mix the two,
gives much better stability but at the cost of poorer mixing and hence
inefficient combustion.)

See the paper on F-1 combustion stability in the Sept/Oct 1993 Journal
of Propulsion and Power for details.

It's not true that all engines have this sort of problem; in a number of
cases it simply never appeared.  In particular, engines which inject one
propellant as a high-velocity gas and the other as a liquid -- notably the
RL10 and SSME -- seem to be virtually immune to high-frequency instability
problems.  (In fact, the SSME had baffles and resonators -- standard fixes
for high-frequency instability -- included in the design from the start,
but does not actually need them.  It's stable without.)

>My impression is that they never quite figured out how they solved it, but
>realized that they HAD solved it.  This is probably a result of
>modeling--i.e. they couldn't model the problem and solution well enough
>mathematically to figure it out.  If someone were to go back to the engine
>and apply modern computer modeling to it, they might be able to come up
>with a slightly simpler solution to the problem...

Actually, combustion instability is so complex and so poorly understood
that modelling is still not really up to the task.  Detailed prediction of
high-frequency instability is still a research topic, not a practical
tool.  (See "Liquid Rocket Engine Combustion Instability", #169 in the
AIAA "Progress in Aeronautics and Astronautics" series.)
Committees do harm merely by existing.             |       Henry Spencer
                           -- Freeman Dyson        |

From: Henry Spencer <>
Subject: F-1 (was Re: Apollo 13 and LEM AGS)
Date: Thu, 18 Dec 1997 23:30:39 GMT

In article <67c7l8$>, Dave Michelson <> wrote:
>...engine that actually flew was very stable, wasn't it?  I seem to
>recall that the Rocketdyne people do a magnificent job of supressing
>combustion instability in the F-1 by suitable redesign of the injector
>plate.  To the point that they could set off small explosive charges
>within the thrust chamber and the engine would recover almost

While that sounds impressive, one should bear in mind that such "bombing"
is the standard technique for dynamic stability testing.  (Admittedly,
that came about partly because of the F-1 experience.)  You can't just
fire the engine under normal conditions and see if trouble develops,
because the onset of the nastier kinds of instability is erratic.  You
have to give the chamber and combustion process a good hard slap and see
if they respond by turning the other cheek or unleashing an uppercut. :-)
There are a number of methods that have seen experimental use, but bombing
is the standard to which all the others are compared.  If disturbances
from a substantial explosion damp out quickly, the engine is unlikely to
show instability in operational use.
If NT is the answer, you didn't                 |     Henry Spencer
understand the question.  -- Peter Blake        |

From: Henry Spencer <>
Subject: Re: F-1 (was Re: Apollo 13 and LEM AGS)
Date: Fri, 19 Dec 1997 14:18:32 GMT

In article <67d9hr$>, Dave Michelson <> wrote:
>>While that sounds impressive, one should bear in mind that such "bombing"
>>is the standard technique for dynamic stability testing...
>...wasn't it noted that the disturbances caused by setting off the
>small explosives were far worse than anything that might reasonably be
>expected during flight?

Indeed so.  The idea is to expose the engine to a fairly extreme
disturbance, in hopes of reliably provoking even bad behavior that would
be extremely rare in normal operation.  (It's horribly difficult to deal
with a complex problem that can't be induced to happen on command.)

>My impression was that the final version of the F-1 was extremely
>stable, not "teetering on the edge of stable operation" as suggested
>by the original poster...

Correct.  They had to work hard to get it to that point, but the result
was an engine that had lots of damping and was not readily provoked into
instability.  Experimental F-1 variants were run at considerably higher
thrusts without any sign of difficulty.
If NT is the answer, you didn't                 |     Henry Spencer
understand the question.  -- Peter Blake        |

From: Henry Spencer <>
Subject: Re: LM take off constraints
Date: Mon, 16 Feb 1998 17:42:58 GMT

In article <>,
Larry Deck  <> wrote:
>I'd like to ask a rookie question.  I hear the term "combustion
>instabilities" used a lot in this group but really don't have a good
>handle on what's meant by it.  Could someone elaborate?

Ideally, you want a rocket engine to burn smoothly and steadily, with no
fluctuations or oscillations in chamber pressure.  It doesn't work that
well in practice.  There is always a bit of unsteadiness.  However, the
real problem comes when the system starts to oscillate, and that's what's
normally known as combustion instability.  Rocket engines handle extremely
large amounts of power -- gigawatts in big engines like the F-1 -- and an
instability that diverts even a fraction of that power in unexpected ways
can be extremely destructive.

There are several flavors of instability, but the really bad one is
high-frequency instability, also known as screaming or screeching.  That's
when you get an oscillation resonating within the chamber, as a standing
or travelling wave with a wavelength some fraction of one of the chamber's
dimensions.  This happens when the pressure and flow variations caused by
the wave rhythmically disturb the combustion process enough to reinforce
the wave.

Screeching causes several problems.  The actual variation in thrust (and
sometimes thrust direction) is a relatively minor headache.  The real
killer is that the wave makes the gases sort of slosh around in the
chamber.  Almost all liquid-fuel rocket engines rely on a cold curtain of
fuel-rich gas flowing next to the walls to help keep the walls cool.  The
sloshing disrupts the curtain and exposes the walls to hot combustion gas,
increasing heat transfer to the walls by literally orders of magnitude.
The walls, even if they are regeneratively cooled by liquid fuel flowing
within them, typically burn through in a fraction of a second (!), with
complete destruction of the engine following immediately.

Screeching is poorly understood, and much of the associated engineering is
rule-of-thumb with little theoretical backing.  It appears that the design
of the injector elements determines the basic frequencies at which
screeching can occur, and the key question is whether the chamber is large
enough for any of those frequencies to resonate within it.  As you might
expect from this, bigger engines have bigger problems with screeching.

Problems in H-1 and F-1 engine development for the Saturns made it clear
that you can't determine an engine's susceptibility to screeching just by
firing it a reasonable number of times under ordinary conditions.  The
onset of screeching is not fully understood, but it seems to require a
fairly unlikely event of some kind -- perhaps a random pressure hiccup of
unusually large size -- and you can't count on it happening often enough
to show up in a normal test program.  It's necessary to make deliberate
attempts to provoke it, by introducing large disturbances into the
chamber, preferably accompanied by high-speed pressure measurements to
determine how quickly the disturbances damp out (the faster, the better).
There are several ways of doing this, but the classic one is "bombing",
detonating a small explosive charge inside the chamber.

A further headache is that instability is sensitive to the details of
chamber conditions.  A throttlable engine may be stable at some throttle
settings and not at others.  It may be stable in steady operation at any
throttle setting, but unstable during transients between them!  And most
any engine passes through a wide range of conditions during startup and
shutdown, and instability may occur then even if not in normal operation.

Fixes, again, are largely empirical.  The big problem is that many of them
reduce engine performance; the F-1 people in particular did not have much
performance to give away, and that greatly hampered their battle against

Fewer and bigger injector elements help, at the cost of less efficient
combustion.  Having fuel jets impinge on other fuel jets, and oxidizer
jets on other oxidizer jets, rather than fuel on oxidizer, increases
stability (perhaps because flow disturbances tend to affect both impinging
jets equally) at the cost of delaying fuel-oxidizer mixing.  Higher
injection velocity helps, presumably because it provides more energy to
encourage droplet breakup, but requires greater pump power.

Baffles, to break the crucial near-injector region up into smaller ones
with higher resonant frequencies, help a lot but are a pain to cool.
Various acoustic gadgets add damping at (you hope) crucial frequencies, at
the cost of more manufacturing complexity and sometimes cooling problems.
Startup instabilities can sometimes be cured by adding expendable damping
material that burns away during startup; the Soyuz booster's upper-stage
engine has felt ribs glued to the interior of its combustion chamber to
prevent a rare and poorly-understood startup instability.

Screeching seems to be driven by gas-liquid interactions involving both
propellants, and engines which inject one propellant as a gas are much
more resistant to it.  (Sweeping the injection temperature of a cryogenic
propellant over a wide range is another useful way of testing for
instability.)  Some kinds of screeching involve interactions with the
injector elements themselves, and changes in (e.g.) passage length there
can help.
Being the last man on the Moon                  |     Henry Spencer
is a very dubious honor. -- Gene Cernan         |

From: Henry Spencer <>
Subject: F-1 development (Was Re: A funny thing about the STS)
Date: Thu, 29 Jan 1998 16:35:53 GMT

In article <6apr23$2ev$>,
Andy Clews <> wrote:
>> ...did not manifest itself in
>> its full form until the Engine 008 disaster of 28 June 1962 [...]
>                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>Can you provide a reference for this, please, Henry? I'd like to read more
>about it.

Most any account of Saturn development will allude to it -- it was the
single event that made everyone concede that F-1 combustion instability
was not going to go away easily and extraordinary efforts were going to
have to be made -- but I've never seen a detailed account in a readily-
accessible source.

Midway through the third test of 008, instability set in so violently that
the high-pressure fuel plumbing leading into the engine ruptured.  The
resulting massive fuel leak simultaneously made the chamber mix seriously
LOX-rich and eliminated most of the engine's cooling.  The chamber and
nozzle walls ignited and were destroyed.

The real problem, in hindsight from a modern perspective, is that serious
stability testing was not done until the engine design was largely frozen
(until after the 008 disaster, in fact).  Many of the simple changes which
improve stability also tend to reduce performance slightly, and since it
was impossible to make any significant adjustments to the basic engine
design (e.g., a longer nozzle) to recover the lost performance, the
instability group was fighting with one foot in a bucket.

Of course, combustion instability wasn't well known at the time -- few
earlier engines had experienced it, and most of them were cured easily --
and the H-1 and F-1 development efforts had to invent the techniques of
stability testing as they went along.  So they can hardly be blamed for
not doing it early...
Being the last man on the Moon                  |     Henry Spencer
is a very dubious honor. -- Gene Cernan         |

From: Doug Jones <>
Subject: Re: continuous-feed solid fuel rocket
Date: Sat, 01 Apr 2000 18:19:17 -0800

Mark wrote:

> > (Henry Spencer) wrote:
> >> Not quite. The pressures above and below the piston would be equal, so
> >> there would be very little pressure difference available to push the
> >> propellant through the injector into the chamber.
> >
> why dont they just use some of the pressure they creat in the combustion
> chamber to pressurizethe fuel chamber? its a easy system to set up by useing
> flame restrictor in the pressure line like a welders tourch has so you dont
> have back burn and a one way controlled check valve. The amount of pressure
> needed to pressureize the fuel tanks to push the fuel thru the injector
> nozzles wouldnt even be noticed by the combustion ratio.

No joy.  The propellant pressure in the injector plenum has to be about
20% higher than the chamber, at a minimum, to ensure combustion
stability.  Thus in a pressure-fed rocket with Pc of say, 300 psia, the
tanks must be at least 360 psia. It's rather difficult to get gases to
flow against that sort of pressure difference :(

Doug Jones
Rocket Plumber, XCOR Aerospace

Index Home About Blog