February 22, 2004 notes
Flameholders
We had a couple more things to cut-and-try on the vehicle
engines to see if we could get them to behave better, but I decided to step
back and think about the problem a little more.
Our working theory about the operation of our mixed-monoprop
engines is that the “cold pack” decomposes some of the 50% peroxide and
vaporizes the methonal, which gives a somewhat combustible atmosphere that can
be ignited by the glow plug and stabilized behind the perforated metal
plate. With a stable flame in the open
space between the cold pack and the hot pack, the remaining liquid (water and
undecomposed hydrogen peroxide) will be vaporized and heated enough that when
it enters the “hot pack” it decomposed and further combusts without progressively
quenching the pack as all of our early attempts would do, even if we had
preheated the pack to very high temperatures.
When the hot pack has been preheated enough, it radiates quite a bit of
heat back up to the perforated metal flameholder, helping to stabilize the
flame. Once it gets fully established,
the engine runs very stably, but the problem for us has been getting the engine
up to its operating steady state.
If our theory is correct, it should be possible to operate
an engine without a hot pack at all and still get something vastly different
than just a foam of decomposing 50% peroxide.
It probably isn’t possible to thermally decompose all the peroxide with
just flameholding, because the temperatures really aren’t that high, but it
should still be cooking pretty good.
Our combinations of perforated metal flameholders, glow plug, and the various
spacing gaps have changed from engine to engine, and we are probably only
hitting on good combinations occasionally out of luck.
We decided to do some more observable tests by running a
cold catalyst pack with various experimental flameholders under it, exiting
directly to the air without a nozzle.
Our first flameholder was made by cutting a length of stainless pipe
lengthwise and welding it together in a cross of U shapes, with four 1/8” holes
drilled in the top to let some mix directly in. It was expected that the flow past the tubes would also re-circulate
somewhat back into the tubes. The glow
plug was extended well into the tube, shielded from the main flow. Several tests with this showed nothing but
the expected cloud of partially decomposed propellant, but we noticed that the glow
plug was getting cooled off when the propellant flowed, most likely from the
1/8” hole near it, so Russ welded that one closed. This was a great thing to see with the camera zoomed way in. The next test showed an extremely
spectacular plume of flame coming out of the engine, clearly held on the
cross. This was the night and day
difference that we expected to see with an operating flameholder.
We repeated the test, but had trouble getting it started again. We were finally able to, but it certainly
wasn’t a sure bet. We then added
another engine section so the flameholder was buried up away from any possible
atmospheric oxygen, and we couldn’t get it to light at all. We tried firing horizontally and with a
nozzle attached, but it was still very spotty.
We theorized that it might need some small flow from the back side of
the flameholder, so we drilled a 1/16” hole somewhat past the tip of the glow
plug. This seemed to light reliably,
but when it had a nozzle on, it chuffed pretty badly.
We went ahead and tried putting a hot pack and nozzle underneath
the flameholder. The one we had on hand
was the partially melted bead catalyst, which turned out to be a mistake. With the glow plug hot, I just turned on the
propellant flow without any warming pulses.
The engine chuffed and misbehaved fairly badly, but it did heat up to
red hot operating temperature all by itself, which was an absolute first for
us! When we took it apart, we found the
beads had melted the rest of the top retaining plate and generally made a mess
of things. We are swearing off
experimenting with the ceramic beads, because they have such a tendency to
cause the metal to start burning.
On Saturday we first tried to just repeat the open pack
tests, but we again had difficulties.
We had accidentally bent the flameholder tubes a bit when we were
working on it, and it might have disturbed a sensitive balance. We decided that the glow plug just wasn’t doing
a reliable job of ignition, so we switched over to a spark based system. We had some spark gear that we got with the
XCOR igniter we had been experimenting with (yes, we could stick the torch
igniter on the side to damn sure start a flame front, but it would triples the engine
complexity and fluid count, so we aren’t about to use it on the vehicle), and
on a tip from aRocket I had picked up a few additional types of spark plugs
from www.sparkplugs.com . We worked up a good driver system and
mounted a spark plug in another leg of the flameholder. On multiple open tests, this seemed to light
reliably.
We made a new hot pack with the last of the ring catalyst we
had, and test fired the engine. It didn’t
seem to want to catch for several seconds, then it abruptly lit and started
running clean, but still chugging badly.
The flameholder seems to be very pressure sensitive.
We tried an alternate flameholder design with a circular
milled channel and single crossbar. I
had milled this out of brass because I loathe milling thick stainless
steel. It might melt, but it might be
cooled sufficiently by the propellant flowing past it. If a brass design works out, we can have
DynaTurn or EnTek mill a bunch in stainless for us. We had great difficulty getting it to light at all, although we
did get it to catch once. It might be
plug / vent positioning, or two other possible theories: This ring is on the outside edge of the
chamber, so propellant would only be flowing past one side, halving the
recirculation. The brazed connection of
the ring was probably letting some fluid flow down the outside, possibly
drowning the spark plug.
We obviously have some development to do here, but it looks
like it is going to be possible to get our engines to just start almost
immediately, with none of the warmup hassles we have been dealing with. This is a really big deal for us, removing
the only disadvantage of our propulsion system.
In other work, the new motor drive board has been potted and
installed in the electronics board, which is now as complete as we need to make
it. We could run 4, 5, 8, or 9 engine
configurations, giving us a lot of flexibility if we miss performance targets
or want to add redundancy. The new
bulkhead is firmly bonded into the cabin-at-the-bottom cylinder, and it should
be our strongest work by far.
http://media.armadilloaerospace.com/2004_02_22/flameholderTests.mpg
http://media.armadilloaerospace.com/2004_02_22/cross.jpg
http://media.armadilloaerospace.com/2004_02_22/brassRing.jpg
http://media.armadilloaerospace.com/2004_02_22/brazedRing.jpg
http://media.armadilloaerospace.com/2004_02_22/sparkPlugs.jpg
http://media.armadilloaerospace.com/2004_02_22/pottedBoard.jpg
http://media.armadilloaerospace.com/2004_02_22/installedBulkhead.jpg
News
