Home / News

News Archive

Self-Preheating, 125 second burn time (!!!)


October 18, 2003 Update


I have been getting a lot of people offering to volunteer to help out with the project lately.  I apologize for not personally responding to each request, but we aren’t looking for any more team members at this time.  We are rarely random-labor limited, so more people on site would be more of a hindrance than help.  However, I do always appreciate suggestions sent in by people that have experience relevant to what I am currently discussing in the updates, so feel free to send me an email, or, better yet, comment publicly at http://www.space-frontier.org/FFO under the current armadillo update.




After the rather surprising preheat behavior with the spray nozzle injector head with the glow plug, we bought a bunch more glow plugs for experimentation.  They are a different brand than the previous one, and we unexpectedly found that if they are left on for too long without any extra resistance, they will burn themselves out.  We added some extra wire to cut the current a bit, but they still draw 15 amps each.


The easiest thing to try was to put the glow plug in a side chamber above the engine inlet.  We have to avoid getting liquid splashed directly on it, which can cool it down too fast, but having it branched off the main flow path seemed to shelter it enough.  We had hoped that the decomposed peroxide and vaporized methanol from the warmup pulses would find its way up through the spreading plate holes and off to our little side chamber, but there doesn’t appear to be enough vapor circulation for this to work.


We welded a glow plug into the side of our long extension chamber (with a lot of catalyst bale stuffed in the bottom) and welded a shield tube around that, and tried using it under the 3-pass catalyst pack.  This worked.  A pulse of the main valve would let a slug of propellant into the 3-pass pack, which would decompose some of the peroxide and vaporize most of the methanol.  It is basically impossible for a catalyst pack to completely decompose 50% peroxide, because there isn’t enough energy to vaporize the water, so you just get hot water of a lower and lower peroxide concentration, plus gaseous oxygen.  The 3-pass pack seems to do a decent enough job, because the vapors ignited easily on the glow plug, and we were able to get the extension catalyst up to full temperature without any propane or torch heating at all.


We are currently manually pulsing the valve open and closed, letting it cook off and blow steam for a while, then doing it again until we no longer see any condensing water vapor coming out of the engine nozzle.  If we go stick our mirror-stick under the nozzle, the catalyst and spreading plate are red hot at this point.  It may be possible to just use a continuous small propellant flow instead of pulses, which is something for us to test next week.


This was a big success for us, completely eliminating the need for our propane / air preheat system, and giving us the ability to do in-flight restarts without any trouble.  The only downside is that the initial one or two preheating pulses do spray a little bit of undecomposed propellant on the ground under the nozzle until there is enough activated catalyst warmed up.


With this arrangement, the upper catalyst stays cool during operation, not getting much over 200 F, while all the high temperature combustion takes place in the extension chamber.  This is probably good for the life of that catalyst.


Runs with this combination still quenched out after extended runs, somewhat sooner than when we preheated all the catalyst (including the 3-pass) with the propane, which was basically expected.


We wanted to encourage re-circulation in the extension chamber below the 3-pass pack, so we welded a plate offset from the 3-pass retainer, which would force the flow out to the sides, which should give a turbulent mixing zone underneath the plate.  This didn’t seem to help much.


On the longer runs, the extension chamber would heat up to red hot in places, but not uniformly.  We suspected the hand-packed (and re-packed after making changes) catalyst bale was giving poor distribution, but we tried adding a second glow plug to see if that effected anything.  It didn’t.


One thing we didn’t find out until the next testing day was that the ¼” thick perforated metal retaining plate had gotten so soft and warped that it had half-pulled past its seat, almost falling down into the nozzle.  That may have had some impact on the mediocre test runs.


Copper catalyst


In an attempt to provide greater axial heat conduction from the hot bottom of the engine to the cooler top to stop the progressive quenching, we fabricated a couple catalysts out of rolled up copper screen, so there would be straight copper wire runs to conduct a lot of heat axially.


The initial hope was that hot copper would be catalytic enough, but it didn’t work.  We then tried making a composite roll by taking one strip of the platinum coated metal foil from one of our corrugated catalyst rolls, and rolling it along with 10x10 mesh copper screen.  This roll had about 2.5 times the mass of a single corrugated roll, and all the copper should give much better thermal conductivity, so I had high hopes for this arrangement.





We had the glow plug in a small chamber welded to the side of the engine top, with a hole through the side to an air gap between the spreading plate and the top of the catalyst roll. The idea was that this would keep the glow plug out of the initial liquid flow, but not occlude any of the pack with a plug shield.  This was able to light the vapors from a warmup pulse, but we weren’t able to get the entire pack preheated.  It seems like it is important to actually have two separate catalyst zones, a 50% decomposer that isn’t expected to get very hot, then an open area with a glow plug, then a catalyst / flameholder that gets preheated to high temperature, and sustains the real combustion during burns.


125 second burn


We wanted to re-test the 3-pass pack over the long chamber with the dual glow plugs, because we wondered if the near-fall-through of the retaining plate may have been a reason for poor runs.  Russ pounded the plate flat and directly welded it to the bottom of the extension, so it won’t have the chance to fall through again.


We also replaced the solid diverter plate under the 3-pass catalyst with a plate of perforated metal.  The idea was that it would be better to have dozens of small recirculation zones instead of a single bug one.


We self-preheated the engine and made a good run.  We doubled the propellant load and had another perfectly clean run.  The heating was extremely uneven in the chamber, with one side getting bright red hot, while the other side stayed dark.  The exhaust didn’t have any hint of unburned propellant, but it clearly wasn’t burning evenly in the chamber.  We suspected it was the unevenness of the hand-packed catalyst bale in the extension, but just to make sure, we rotated the chamber 180 degrees under the 3-pass pack and made another long run.  This was a Very Good Idea, because the red side did not rotate with the chamber, it stayed in line with the 3-pass pack.  We had suspected that warping during assembly had cut off flow on part of the top plate, and this looked conclusive.  I am going to have to make another set of 3-pass plates to fix this, but this run also stayed perfectly clear, without any quenching, so we continued.


We double propellant again, and made our longest mixed-monoprop run ever, at 23 seconds, still without any hint of quenching.


This was the limit for loading in our normal test stand tank, so we pulled out the big tank and plumbed it up.




We doubled the propellant load again, and made another perfect run of 47 seconds.


This was looking extremely good, so we decided to double the propellant again, aiming for a 95 second run, which is longer than our X-Prize burn trajectory.  While we were mixing all the propellant, we swapped out the pressure gauge we were using on the big tank, because it was looking a little dubious at the low pressure we were using (250 psi).  It turns out that the gauge was indeed reading incorrectly (low), so, in conjunction with having a noticeable blowdown pressure drop with this much propellant (76.8 liters / 20.3 gallons) even in the big tank, our last run wound up going for 125 seconds at somewhat over 200 pounds thrust.




Still no sign of quenching.


This is Very Good.  We have spent months working on the mixed-monoprop scheme, but until now, we were never 100% sure that is was actually going to work out for the X-Prize vehicle.  Now that we have unlimited burn times and self-preheating, it looks like we have made the right decision.


It looks like the perforated turbulator / flame holder plate under the 3-pass pack is key, but we need to do more tests to isolate the effects.  When we took the engine apart, we found that some of the welds holding the bottom retaining plate on had cracked.  Russ is going to actually fill all the edge perforations to give us a much stronger weld for the next test sessions.  The uneven heating in the chamber is putting a lot of thermal stresses on various parts, which should go away when we get the next 3-pass cat pack built.  The perforated turbulator plate clearly shows the dead side of the 3-pass pack as a segment of missing combustion along the outside.






One annoyance from an otherwise triumphant day -- I had WinDaq crash on me (taking my entire laptop with it) twice on the long runs, which lost the test data.  Anyone else experience problems with their Ethernet acquisition systems on WinXP?


Upcoming tests:


Run with a single glow plug, to see if the dual glow plug arrangement was important.  We don’t think so.


Cut power to the glow plug after the run starts.  We held power through the entire run, but we don’t think it is necessary.


Run with a short extension chamber instead of the long one.  We know it can be less than half the length, but we don’t know how short we can go.


Try with one of the rolled foil catalysts as a secondary catalyst instead of the catalyst bale.  This would help engine-to-engine consistency.


Try with food-grade 50% peroxide instead of the unstabilized FMC 50% we have been using.  FMC only sold us this peroxide under the condition that it not be used in flight vehicles.


Try a new 3-pass catalyst with good spacing at the top and less crossbar obstruction.


Put the big nozzle on and see how much thrust we can make with this combination.  All we need for low-altitude test flights of the big vehicle is 500 lbf / engine, which we have almost made even with the small nozzles.  We should be able to get twice that if we use large diameter plumbing all the way from the tank to the engine.


Make sure throttling is stable and predictable.


Once we have everything dialed in, we will have to order some more catalyst to make a full ship-set of four engines for low altitude testing of the big vehicle, and whatever we need for building up the full size 12” diameter engines.




© 2001-2011 Armadillo Aerospace, LLC. All rights reserved.