October 4, 2003 notes
Hot Preheat Mixture, E-Beam Foam, Self-Preheat
We performed a number of interrelated experiments in the
last week. A lot of experience was
gained, but we dont have a clear win yet.
A general problem that we have had through all of our work
with the mixed monoprop solutions has been ensuring a complete preheat of the
catalyst. Passing a calibrated air /
propane mixture through the catalyst will bring it up to a fixed temperature,
but only if the reaction is started by heating part of the catalyst up to the
temperature where propane begins to catalytically burn. We had been doing this by the unreliable
method of sticking a torch up the nozzle until we got at least part of it red
hot, then turning on the air / propane flow to let it bring the entire pack up
to our target temperature of around 1600 1800 f.
This worked fine for catalysts that were a single monolithic
block, but as we separated the blocks to prevent direct channeling, it got
harder and harder to ensure that the gas flow would heat anything but the final
block. Some of our recent tests with
the three-pass catalyst and post catalyst flame holders are even harder to get
started, and we have had a few test runs that behaved poorly due to, we think,
Hydrogen / air mixtures will start preheating on a room
temperature catalyst, but it will sustain open air combustion over such a wide
range that it actually flashes into flame even at the incredibly lean mixture
ratio that we run at, which causes some problems.
I theorized that we could raise the temperature of the
propane / air mixture we are feeding into the engine in a controlled manner,
and that the hot mixture would react on the room temperature catalyst, starting
at the top of the engine and working its way down, which would give the best
I made a heat-exchanger by wrapping 25 or so feet of -6
aluminum tubing around a 2" pipe to make a fairly tight tubing coil.
Sticking a heat gun full on in one end and roughly covering or plugging the
other end to force most of the air out between the coils would get the end coil
to about 250 degrees with 4 cfm of air flowing through it if you waited long
I put some ½ ID plastic hose on the other end, but I was
quite surprised at how much the air flow cooled back down flowing through
it. I had expected the larger diameter to cause less heat transfer, and
the plastic should have been a better insulator with slower heat transfer, but
with 20 or so feet of tube, the 250 degree air at the end of the coil was back
down to barely over room temperature at the outlet, even though the end right
by the heat exchanger was very hot (and soft). I cut the hose down to
just six feet, but there was still a 50 degree temperature drop along the
hose. I finally just put the engine right at the end of the heat exchanger
for initial testing.
I turned the propane up to our normal 5cfh and left it there
for a while, but 250 degree preheated mixture didn't accomplish anything.
With the restriction of the small tubing in the heat exchanger, I also found
that the air compressor couldn't even hold 6 cfm of flow.
I did some experiments and found that sticking the big torch
in the end of the coil at a very low flame (with the other end open) did not
seem to hurt the aluminum tubing while gas was flowing. The end coil got
up to about 350 degrees and seemed to stabilize. Turning on the propane
for a while still did not get anything going in the pack.
I turned up the torch a little bit, and the temperature
started climbing again. At about 380 degrees, the pack started making the
"rocket engine sound" that we heard when the hydrogen mixtures
autoignited in the catalyst pack, signifying actual free-air combustion in the
pack instead of catalytic burning. Looking in the pack, there was light
coming from deep inside (this was the four-deep catalyst pack), but it wasn't
the red-orange we usually see in a preheat, it was white.
I watched it grow for a few seconds, then decided I should
check to see if the propane mixture is way out of whack. It wasn't.
I turned the propane completely off, but left the air flowing to cool it, and went
back to look at it again. The white glow has continued to increase, and
the entire bottom of the catalyst pack was slowly bowing out under the gentle
air flow (the engine was laying on its side). I was wondering if it is was
just clearing the line of the propane, but it keeps going. The pack had
opened up enough to see that the inside is actually being consumed -- the air
is burning the stainless steel catalyst! I shut the air completely off,
and it very slowly starts to cool down.
Four catalyst blocks were completely slagged.
Obviously the excess oxygen in the air was combusting with
the iron in the stainless steel catalyst base, thermite style, but I puzzled over
the exact behavior for a while. I was thinking that we were going to be
sunk on using this method to preheat the engines, but after a while, a
reasonable explanation occurred.
Our normal air / fuel mixture gives about 1800 - 1900
temperatures. However, that is with room temperature reactants. The
preheat was at least 300 F hotter than room temperature (later evidence would
point to more like 600 F over room temperature), so with the preheat, the final
temperature could easily have been over 2200 F. 316 stainless melts at
2600 F, but the catalyst substrate is a "proprietary alloy", and some
combination of events seems to have let it melt and start burning. I
wondered if platinum may also be a combustion catalyst for metal burning.
On Tuesday, we set about to characterize things better. We potted a bare-wire thermocouple into a T
fitting with ceramic sealant so we could measure the actual gas temperature
right before it went into the engine, instead of measuring the surface temperature
of the heat-exchanger coil at various places.
Russ brought in a propane cooking burner that we could control better
than the big torch, and we set the heat exchanger coil up on top of it. We used Cotronics tubular insulation to
reduce the heat loss in the hose from the heat exchanger to the engine.
We slowly worked our way up the temperature range. I was surprised when we passed 400 F without
anything happening, the external temperature reading I was taking in the
initial test was clearly understating the gas temperature. Finally, just when we were about to quit, at
550 F, we got reliable, fast startup.
This is the number, after repeated tests, a 500 F mixture wont start,
but a 550 F mixture will start every time, heating the entire catalyst brick to
red hot in under two minutes. With this
amount of additional heat, we had to lean the mixture out all the way to 3 cfh
propane / 5 cfm air to keep the temperature under 1900 F. This seems to be more than just the initial
temperature delta, so there is some additional process causing it to heat more
efficiently than the cooler mix.
Trying to stabilize the gas temperature at a given level is
challenging, because there is a very significant (over a minute) delay between
the time an adjusting to the burner flame is made and the time the gas
temperature stabilizes at a new level, due to the heat sink nature of all the
plumbing. On Saturday, we constructed a
heated bath for the heat exchanger coil that used a thermostatically controlled
2000 watt cartridge heater. We picked a
silicone oil rated for use at 500 F, but it turned out that we had significant
boiling problems even aiming for 400 F temperatures, because the cartridge
heater surface was significantly hotter then the fluid temperature. It also took a very long time to heat the
entire bath up to temperature, and it didnt look like the 2kW heating element
would be able to hold a sufficiently high temperature once the heat losses
through the container were accounted for as well as the gas being heated. We abandoned this development line, and went
back to the direct flame heating for the remaining tests of the day. We may look into putting elements directly
inside the gas flow in the future, possibly by cannibalizing a really big heat
gun to let it take pressurized air instead of drawing it in with a fan.
We also got our nickel foam discs back from Galco with their
e-beam deposited platinum coating. The
very thin coating hardly changed the appearance of the nickel foam at all, but
it was still quite reactive with peroxide.
However, it had a couple bad characteristics for us related to the
extremely low thermal mass of the foam.
If you heated it up with a torch, it cooled down Really Fast after you
removed the heat. Propane also did not
seem as easy to catalytically burn on it, as it wouldnt hold red heat with cold
propane flowing through it, even though it started a flame on the other side. Heating it with a torch could also actually
melt the foam if you held it in place long enough. This was a little surprising, as nickel has a higher melting
point than stainless steel, but the tiny foam cells cant conduct any
significant heat away, so they can heat up to almost the full temperature of
the torch flame fairly quickly.
The first test we did with the foam was to put 30 foam discs
in one of the extensions underneath the chamber with the impingement spray
nozzles. The hope was that the fine
mist of propellant hitting the high surface area catalyst might start
combustion in the top space, rather than tunneling down and progressively
quenching the pack. The heated
preheater was undergoing the transition back from the silicone oil bath to open
air heating, so we tried to preheat it the classic way, torching the bottom
before starting cold flow of air / propane.
This was difficult to gauge, because it never held enough heat to glow
for any length of time, and the propane flowing through it also didnt bring it
to red heat. We waited for it to get
fairly hot, then tried various solenoid pulses of propellant, but we didnt get
anything useful out of it.
The other thing we were going to test was sticking a glow
plug at the top of the injection engine.
We never got a glow plug to do anything useful in a catalyst-free
engine, but there was hope that it could be used to ignite the oxygen and
methanol vapor that is given off when the propellant hits even a cold
catalyst. We tried this first with a
double block of the corrugated monolith catalyst under the impingement
injectors. The first pulse of the
solenoid gave a noticeable bark, even with a cold catalyst, but continuous
spray from the injector resulted in a good deal of liquid coming out of the
nozzle. Thankfully, Russ checked under
the nozzle, and found that the catalyst was actually red hot! This was very exciting. We tried a couple more combinations of foam
and monolith catalysts, and we were able to get them apparently (from the
bottom, at least) nicely preheated by just pulsing a bit of propellant in,
letting it decompose and ignite on the glow plug. We were not able to get sustainable thrust from them, but the
spray injector arrangement has never actually worked out for us.
When we got the hot preheat arrangement working again, we set
up to do some more tests with the three-pass catalyst. We did one test with the long extended
chamber that preheated nicely, but still fell off after 15 seconds of
operation. The next test we took the
small extension and completely packed it with the platinum coated foam discs to
fill the post-three-pass-catalyst. We
started the hot preheat, but after a minute or so, Matt noticed a couple hot
sparks dropping out of the engine, which is usually a Bad Sign. Looking up with the mirror, the pack had
white heat deep inside it, so I shut off the air and propane. Interestingly, it wasnt making the rocket
engine sound we usually associate with a burning monolith. That sound must come from gas reaching sonic
velocity in the individual monolith pores, which doesnt happen in foam. When we took it apart, somewhat to my
surprise, we found that the three-pass catalyst seemed unharmed, but the foam
discs seemed a little beat up. When we
pulled the block of pressed-together foam out, it felt strange when you shook
it. We peeled it apart, and found that
the center had been completely melted out.
This was interesting, because we had solid conditions for
this test the inlet gas temperature was just over 550 F, and the gas
composition was 3 cfh propane / 5 cfm air.
This combination should not have been able to get even close to the 2651
F melting point of nickel, but the remaining hot oxygen, even diluted by normal
atmospheric nitrogen, plus the CO2 resulting from the combustion of the propane,
was able to burn the nickel. It is
obvious that the platinum coating was far from pinhole-free, because otherwise
the platinum would have served as an oxidation barrier, but it seems like
platinum with holes in it may actually be a burning catalyst for underlying
We are really hoping that self-preheating with pulses and a
glow plug can be made to work, so we dont need any of the propane ground
support equipment at all. That would
also give us the ability to do in-flight restarts if necessary, a big benefit
for powered landing.
We have hopes (again) that we will be getting our initial
deliver of 90% peroxide soon, so we may finally get back to flight testing
To make some of our testing more accurate, I bought a medium
weight scale suitable for measuring our typical propellant runs. Our standard mixture of 8 liters of peroxide
and 1.6 liters of methanol came to 23.25 pounds (26.5 3.25 lb for the Teflon bottle). I was probably slightly underestimating our
Isp by using theoretical density values, which are at a temperature somewhat
cooler than our shop temperature.
Unsorted pictures from this week: