February 29, 2004 notes
We did a lot more development with various flameholders and
Slitting the tubing lengthwise and making crosses out of it was
labor intensive, so we started working with 316 stainless angles. I had to go someplace other than
McMaster-Carr for these -- http://www.supremestainless.com/
had everything I needed in stock, but would only ship mill lengths (20) by
truck. I got some of each size from ¾
to 2 legs.
A single 1.25 angle across the chamber gave a visibly
stronger flame than the tubing cross. The
book Gas Turbine Combustion by Arthur H. Lefebvre (a good, informative book) has
some data on flameholders that follow a few general trends: larger flameholders
are more stable than multiple smaller ones, and for angular gutters, increasing
the angle (from 60 to 90, for example) increases the stability, but at an
increased pressure drop.
It was still difficult to get it to ignite
consistently. We tried a couple
different spark plug locations, and found that it worked better with the plug
about an inch below the angle, which was a bit surprising. We got a good improvement in ignition when
we switched from 100hz sparks to 250 hz sparks. It would ignite almost all the time, but it sometimes took several
We had been using an MSD small engine controller and MSD DIS
coil that came with the XCOR igniter, so we tried moving to a higher energy system:
an MSD-7A-AL3 with a ProPower-HVC coil.
This made a strong spark even at 500 hz, but it didnt seem to improve
the ignition characteristics. We have
the top-of-the-line MSD-10Plus system on order, which puts out several times
more energy, which we hope improves things.
These big ignition systems will add 5lb per engine, unfortunately.
The pipe tap spark plugs we had been using were rusting
badly in our chambers, so we got some new Autolite Platinum Pro spark plugs,
number APP847. The bodies are nickel
plated, so peroxide wont hurt them, and platinum electrode surfaces should
decompose any peroxide that hits them.
These are gasket seat plugs, so we do need to flatted down the engine
side before tapping them, which we didnt have to do with the NPT plugs. The change in plugs didnt seem to have an
impact on ignition.
We tried using a heavy perf plate (1/2 holes, ¼ thick, 40%
open area) with a spark plug under it as a flameholder. It did hold flame over the surface, but it
was much weaker, with more liquid passing by, than the single angle bar, even
though it had much more pressure drop.
This follows the guideline of larger flameholders instead of smaller
We added two more 1.25 angles below the existing one, at
evenly spaced orientations. When this
lit, it had an extremely strong flame, with very little visible liquid. We were so impressed by it that we tried
putting a nozzle below it without any hot pack at all. It has always been my belief that there isnt
enough energy to thermally decompose the remaining peroxide without any
catalyst, but we gave it a try. It made
stable thrust, but the exhaust was filled with undecomposed peroxide, which is
basically what I expected.
We had switched from solid core spark plug wires to MSD 8.5mm
helically wound wires to reduce EMI.
Our data acquisition system no longer immediately spazzed out when the
sparks started, and the pressure transducers and throttle feedback channels
worked fine, but the load cell was still far to noisy to use. Once we know the flame is going, we can shut
off the spark and get good data.
When we added the small hot pack below the triple bar flameholder,
the engine could be warmed up and fired cleanly. It is clearly audible when the flame finally ignites, but it seems
to be more difficult in a full engine than in the open tests. The engine ran smooth at moderate thrust
levels, but at high throttle openings it had a severe chug. We have seen this on several different
flameholder experiments now, and dont understand it yet.
We built up a full scale production engine by splicing a 3
deep section of 7 rolled tube with flameholders (two crosses of 1.25 angles) above
the hot pack of engine 0, and built a new cold pack with two fresh 600 cpsi
monoliths and the 848 hole spreading plate.
Running at full pressure with the big plumbing, it is easy to flood the
flameholder early on, so it does still require finding a specific spot to
efficiently warm the hot pack. We made
several runs with promising results, but this engine also suffered from
chugging above 60% throttle. If we can
work out the issues with this configuration, we have everything on hand to
build four of these engines to make a ship-set.