Silver plating

[cross-posted from uk.rec.models.engineering. This is a design for a small
LOX/kero turbopump for model rocketry which I am just starting to build. The
thread title comes from the silver plating used to protect the fast-moving
metal insides of the LOX pump from particulate impact ignition]


Tom wrote:

So how big is the impellor going to be in this pump?

It doesn't have one.

I considered a volute casing Barske-type impellor pump, but it would be very
inefficient as the pump is so small, so I will be using a double Pitot [1]
design instead (unless I can't get it to work, when I will fall back to a
Barske design, or perhaps try a two-shaft-Quimby-type screw pump).

The Pitot arm is 32 mm dia, there are four pumps [2], a combuster and a
turbine on a single 75,000 rpm shaft in an assembly 54 mm max dia and 65 mm
long, target weight ~350 grams.

Propellant flow is 175 grams per second. Shaft power is 2.1 kW, pump
mechanical efficiency should be ~ 55%, turbopump overall efficiency ~25%,
LOX output pressure is 750 psi.

Engine design thrust is 5kN / 100 lb sea level, chamber pressure is 600 psi,
expansion ratio is 8.25, Isp is 245 s sea level, 285 s vaccuum.

Note that most of these figures are still theory, and they will almost
certainly change a bit in practice. Note also that the design is slightly
less demanding than the engineering presently (apart from the pumps) used in
small model turbojets, and I hope to improve on those figures.





[1] A Pitot pump is a hollow stationary arm with a Pitot tube inside opening
on the end, which is inside a hollow circular casing which spins and
accelerates the liquid inside it - the fast-moving liquid enters the pitot
and the speed is changed to pressure. Also, the spinning exerts a
centrifugal force on the liquid, increasing it's pressure at the outer edge
of the casing where the Pitot is located. A double Pitot pump just has two
Pitot holes on opposite ends of a single stationary arm.

It can be more efficient than an impellor pump because the wetted moving
area is smaller, and there are no fast-moving parts in close proximity to
give large shear forces - the two main energy wastes are the energy used to
move the arm through the liquid, the arm can be shaped and surfaced to
minimise that, and the inefficient diffusion recovery (the change of speed
to pressure in the Pitot tube - probably only about 60% efficient at best,
but recovery only accounts for half the theoretical head, so you lose maybe
20% of the total energy that way).

Manufacture makes few demands on close tolerances, the single rotating seal
is at low input pressure, vibration is very low and the output is almost
entirely pulsation-free, which is important for combustion stability.



[2] two LOX pumps in parallel, and two fuel pumps in series. LOX volume is
about twice the kerosene fuel volume. The fuel pressure is nearly double the
LOX pressure because it is used to cool the chamber, throat and nozzle
before going on to be burnt. An alternative which has some benefits is for
the fuel to go through one pump, then cool the nozzle, then the second pump,
and then be burnt, but I haven't decided yet.


--
Peter Fairbrother

Peter Fairbrother wrote:
I considered a volute casing Barske-type impellor pump, but it would be
very inefficient as the pump is so small, so I will be using a double
Pitot [1] design instead (unless I can't get it to work, when I will fall
back to a Barske design, or perhaps try a two-shaft-Quimby-type screw
pump).


Have you considered a Tesla turbine at all? Its kinda the same only the
casing is stationary with the impeller (a simple set of disks) spins. Both
designs rely on surface friction and i would not know what kind of
efficiency you would get with a Pitot design. The army did a study on the
Tesla design for small turbines (aka about the same size) and got a
efficiency in the 20% region. plenty good enough for a fun rocket!.

Greg

I may be a little OT here and admit I do not know much about rocket
turbopumps but I do understand that they are very expensive. So, Why
not make the complicated impellers via CNC milling in wax of the type
used to make 3D models. It is supposed to have good machining
qualities. Then, you evaporate something like TiN (Titanium Nitride)
onto it via e-beam evaporation (we can do this too), then, you deposit
either electroless nickel or electrolytic nickel onto that and then
melt the wax out to produce a lightweight and very strong shell. The
TiN has good wear properties too. These could easily be mass produced.

Peter Fairbrother wrote:
[cross-posted from uk.rec.models.engineering. This is a design for a
small
LOX/kero turbopump for model rocketry which I am just starting to
build. The
thread title comes from the silver plating used to protect the
fast-moving
metal insides of the LOX pump from particulate impact ignition]


Tom wrote:

So how big is the impellor going to be in this pump?

It doesn't have one.

I considered a volute casing Barske-type impellor pump, but it would
be very
inefficient as the pump is so small, so I will be using a double
Pitot [1]
design instead (unless I can't get it to work, when I will fall back
to a
Barske design, or perhaps try a two-shaft-Quimby-type screw pump).

The Pitot arm is 32 mm dia, there are four pumps [2], a combuster and
a
turbine on a single 75,000 rpm shaft in an assembly 54 mm max dia and
65 mm
long, target weight ~350 grams.

Propellant flow is 175 grams per second. Shaft power is 2.1 kW, pump
mechanical efficiency should be ~ 55%, turbopump overall efficiency
~25%,
LOX output pressure is 750 psi.

Engine design thrust is 5kN / 100 lb sea level, chamber pressure is
600 psi,
expansion ratio is 8.25, Isp is 245 s sea level, 285 s vaccuum.

Note that most of these figures are still theory, and they will
almost
certainly change a bit in practice. Note also that the design is
slightly
less demanding than the engineering presently (apart from the pumps)
used in
small model turbojets, and I hope to improve on those figures.





[1] A Pitot pump is a hollow stationary arm with a Pitot tube inside
opening
on the end, which is inside a hollow circular casing which spins and
accelerates the liquid inside it - the fast-moving liquid enters the
pitot
and the speed is changed to pressure. Also, the spinning exerts a
centrifugal force on the liquid, increasing it's pressure at the
outer edge
of the casing where the Pitot is located. A double Pitot pump just
has two
Pitot holes on opposite ends of a single stationary arm.

It can be more efficient than an impellor pump because the wetted
moving
area is smaller, and there are no fast-moving parts in close
proximity to
give large shear forces - the two main energy wastes are the energy
used to
move the arm through the liquid, the arm can be shaped and surfaced
to
minimise that, and the inefficient diffusion recovery (the change of
speed
to pressure in the Pitot tube - probably only about 60% efficient at
best,
but recovery only accounts for half the theoretical head, so you lose
maybe
20% of the total energy that way).

Manufacture makes few demands on close tolerances, the single
rotating seal
is at low input pressure, vibration is very low and the output is
almost
entirely pulsation-free, which is important for combustion stability.



[2] two LOX pumps in parallel, and two fuel pumps in series. LOX
volume is
about twice the kerosene fuel volume. The fuel pressure is nearly
double the
LOX pressure because it is used to cool the chamber, throat and
nozzle
before going on to be burnt. An alternative which has some benefits
is for
the fuel to go through one pump, then cool the nozzle, then the
second pump,
and then be burnt, but I haven't decided yet.


--
Peter Fairbrother

Reposted,

To the moderator, this posting seems to be missing from the latest batch,
please include it in the next.

Craig Fink

On Mon, 07 Mar 2005 13:58:24 -0600, Craig Fink wrote:

On Wed, 02 Mar 2005 13:40:02 +0000, Peter Fairbrother wrote:

[cross-posted from uk.rec.models.engineering. This is a design for a
small LOX/kero turbopump for model rocketry which I am just starting to
build. The thread title comes from the silver plating used to protect
the fast-moving metal insides of the LOX pump from particulate impact
ignition]


Tom wrote:

So how big is the impellor going to be in this pump?

It doesn't have one.

I considered a volute casing Barske-type impellor pump, but it would be
very inefficient as the pump is so small, so I will be using a double
Pitot [1] design instead (unless I can't get it to work, when I will
fall back to a Barske design, or perhaps try a two-shaft-Quimby-type
screw pump).

The Pitot arm is 32 mm dia, there are four pumps [2], a combuster and a
turbine on a single 75,000 rpm shaft in an assembly 54 mm max dia and 65
mm long, target weight ~350 grams.





Have you thought of trying to eliminate the pitot tubes by putting your
combustor at the outside diameter of your rotating casing? Essentially a
three chamber spinning pump. Fuel in one, Oxidizer in another, and
combustion products in the last. This way you have liquid being compressed
by the centrifugal force, and gas flowing up against the centrifugal
force. This would eliminate the drag of the pitot tubes, but add a loss
due to the column of gas. But the density of a very hot gas would be much
less that the two liquids.

Or, possibly even add some stationary vains in place of the pitot tubes in
the gas chamber after the combustion, to stop the gas from spinning on the
way up and out of the centrifuge. Stationary vains next to a spinning case
might also simplify your combustor because of all the mixing going on.
Although temperature of the vains might be a problem, it's not much
different than your turbine. The stationary vains could also give you a
path for an igniter to start the combustion going. The stationary vains,
or stator blades, could be designed to take the spin out of the whole
thing.

No pitot tubes in the fuel and oxidizer chambers allows you to put
struts, vains or anything else within those two chambers of the pump. It
really frees up the design of those two chambers of the centrifuge pumps.

I'm courious, are you going to spin the whole nozzle too? Kind of reminds
me of that early radial engine that spun all the cylinders.

Craig Fink