Home / News

News Archive

AST, Welded Engine, Tube is Done

October 29 and November 2, 2002 Meeting Notes

October 29 and November 2, 2002 Meeting Notes




Neil spent a couple days in DC working on regulatory issues for Armadillo.  He attended the COMSTAC meeting, but didn’t find it to be a very good venue for Armadillo, being dominated by the big Aerospace firms moaning about the launch calendar.  We will probably start participating in the RLV forum, though.


The meeting at AST was set up specifically for Armadillo, and was attended by all the top brass.  It went very well, and we now have a three person team at AST assigned to help us through the regulatory course.  They strongly advised that we begin working on our environmental impact study immediately, because that is a requirement that is outside their control, and will likely be the pacing item.  We have been hoping to piggyback on OSIDA’s EIS work, but that won’t be complete until 2004.


Neil has a lot of work cut out for him on the regulatory side over the next year.  We are aiming to put something to “space” (100km) by the end of 2003.


Welded Engine


We did a lot more investigation on why the new 6” engine isn’t working well.


We tested the catalyst activity, and found it to not be as good as the previous engine.  We had washed all the silver screens in the ultrasonic cleaner with alcohol, but previously we had done an initial wash with hot water and Tide.  We repeated that, and the activity was indeed increased.  We also tried a more aggressive cleaning – rinsing the screens in 50% nitric acid.  This cleaned them extremely well, making all the silver look completely fresh and white.  The activity was extremely good, with the reaction being fast enough to make drops of peroxide skate around on a film of gas.  One thing of note was that when we cleaned the pack, there were some red fumes given off (red fuming nitric acid?), which we were careful not to breath.  That may have been from the stainless screens, rather than the silver.


We also noticed that the anti-channel rings were not fitting very tightly.  We usually see this after running an engine, but we decided to try and do something about it today.  We use Smalley spiral retaining rings as anti-channel rings, because they come in a wide range of sizes, and can be sized to have an initial outward preload when you first install them.  The spiral nature may have been allowing them to take a permanent shrunken set after thermal cycling, so Russ welded a few of the rings together into single units that won’t slide over themselves.  We learned something interest here.  Russ did the welding by putting a ring in a spare catalyst pack to hold the exact size, so it was an exact fit in there.  When we moved them to the engine we had been firing, they dropped right in without any resistance, and could even be wiggled around a bit.  The fired engine has stretched a visible amount.  We compared some of the already-fired anti channel rings with fresh ones, and found that the rings had still shrunk as we had originally thought, so there are two factors effecting channeling – ring shrinkage after thermal cycling (the rings heat up faster than the chamber wall), and engine stretch.


With the combination of the cleaned silver and new anti-channel rings, we did some additional test firings.


Without any orifice at 250 psi tank pressure, it made 300 lbf, +/- 3%, with an odd sharp upward spike in thrust at the end.  Still very cloudy.


We tried even running it with the small orifice designed for the fuel side of our big biprop, which cut the thrust down to 150 lbf, but it was still cloudy.


We decided to test the old engine, so we hauled the lander out for a short test hop.  The main engine did make a big cloud initially, but it did clear up after a few warmup pulses.  The flight was entirely nominal, but it clouded up a bit as it was setting back down.


On Saturday, we decided to take a drastic step to eliminate channeling – we completely welded the anti-channel rings to the catalyst pack, so they positively can’t be leaked past.  This was a bit challenging in a 5.5” ID catalyst pack, but Russ was getting the hang of it by the time he was done.  We couldn’t press down on pack while welding, so the pack doesn’t have the same compression characteristics as normal.  We used the same pack that we ran on Tuesday, so it was pre-compressed a fair amount, which may be relevant data when we make another one like this from scratch.  The exact packing order was:


Two non-welded rings to just act as spacers at the top of the engine to let the peroxide spread out.

10 stainless screens to act as an inert spreading zone

A welded ring

10 stainless / 20 silver screens alternated

A welded ring

10 stainless / 20 silver screens alternated

A welded ring

10 stainless / 10 silver screens alternated

A welded ring

10 stainless / 10 silver screens alternated

A welded ring with the last stainless screen brazed to the ring, in lieu of a perforated metal retaining plate


Before inserting each ring, the pack was compressed with 10,000 lbf, but there was some spring-back before the welding.  We were a little concerned that this would break loose the earlier rings, but they held up to it with no problem.  Skipping the retaining plate was another idea to save weight and reduce flow losses.  We had run a 2” engine without a plate supporting the screens, and we expected that the welded rings would keep all but the last set of screen pressure drop off of the retaining screen, so it should probably work out.  It may not scale to 12” engines, but we shall see.


I finally bought a new data acquisition system that can handle a lot more channels.  We have been using various Dataq starter kits for two years, measuring only thrust and tank pressure, but now that we are looking at optimizing some larger engines, we need to start getting a lot more data.  I bought one of the larger Dataq systems, which are sort of a rip-off price, but the familiarity with the product line had some value.


For measuring chamber pressure, we just ran a stainless steel line with a bend in it right up the engine nozzle.  You wouldn’t be able to get away with that on a biprop, but with monoprop it works fine.  It is possible that we have some velocity effects on the reading, depending on the exact orientation of the line end, which may cause an underestimation of chamber pressure.


We ran about a gallon and a half through the engine, and it made 308 lbf steady, then sharply (but smoothly) increasing to 500 lbf over the last second of the run.  It was still very cloudy.


This was the smoothest large engine run we have ever seen, it only had +/- 1% thrust roughness.  The welded rings probably prevent the entire pack from getting into oscillations, trapping them in a smaller region.  The thrust was not significantly higher than the Tuesday runs, even though there were less screens with less compression, and no retaining plate.  Chamber pressure was only reading 74 psi in the stead part, going to 110 psi at the peak thrust, at which point the tank pressure was 234 psi.  That is about par for what we saw last time we measured chamber pressure.  The pressure drop across all those screens is very significant.  The old foam packs had far less pressure drop, but crummy lifespans.


I finally realized what the very repeatable, large upswing in thrust at the end of all these runs was:  we had a very long braided hose running from the peroxide tank on the trailer, over to the valve on the vertical test stand.  The line was a good sized –10 hose, but it was 24’ long.  As the line emptied out, the fluid pressure loss decreased.  There was a repeatable 40% drop in thrust due to the long hose.  To prove this, we strapped the tank directly to the side of the test stand, so there were only two 3’ lines in the circuit (tank to valve, valve to engine).  There is still a 90 degree fitting right at the engine, which certainly hurts, but the thrust still went up to 360 lbf.  On the flight vehicle, the engine is directly inline with the valve, directly connected to the tank, so I think we can count on 450 lbf at 250 psi in that configuration, which is adequate for our needs.


The runs are still cloudy, but we noticed several things: It is clear right underneath the engine, with the clouds only showing up when it hits the ground.  The clouds aren’t really clouds of peroxide – you can take a whiff of them without burning your nose.  We didn’t precisely measure Isp, but it definitely isn’t far off from where it should be.  When we flew the lander with the old engine, it did clear nearly all the way when it was warmed up repeatedly right off the ground (engine 3” from concrete), but when it lifted off and came back down in a slightly different spot, there were clouds again.


At this point, we think the clouds are just clouds of water vapor, coming from the fact that it has been cold and wet all week, and we were actually doing our testing in a light drizzle this Saturday.  Firing vertically onto cold, wet concrete gives the exhaust plenty of excuse to condense.  If it is fired long enough at one spot at short range, it probably dries the area out enough to diminish the effect.  I suspect that if it dries out and warms up a bit, the current welded engine will be a flawless performer.



Tube is Done








We still need to do more parachute pull tests and a hover test, but all elements are in place for a flight to moderate altitude of the tube vehicle.  Dry flight weight is just under 300 pounds.


The main propellant tank is a 2’ diameter polyethylene lined fiberglass tank from Structural, directly bonded to a 60 degree included angle tail cone, nose cone, and some filament wound pipe from Beetle Plastic.  The tank has 60 gallons capacity, so we could load it with a truly huge amount of peroxide, even with blowdown ullage, but our initial flights will be with five gallons of peroxide.  The body is very tough, and we are pretty confident that it will survive landing shocks, unlike the first version we built with the airfoil section fins.


The main engine and servo valve attach directly to the bottom of the tank.  The tank closure is 4” in diameter, so we are able to mount three other fittings directly to it without needing a manifold.  If it was a bit larger, we wouldn’t even need to T for each pair of attitude engines.  The peroxide fill fitting on the tank bottom is welded to a 3’ long length of stainless pipe that goes high into the tank, so as the peroxide is pressure fed in, it fountains up and can’t flow back out.  This is better than using a check valve, because it allows us to vent pressure from the top of the tank if desired, and keeps the disconnect lines completely dry after some pressurizing gas has been blown through.  We pressurize the tank at 250 to 300 psi.


The base structure is of welded aluminum.  The four attitude control jets are canted for roll control, and tilted on another axis to maximize the lever arm to the vehicle CG.  We are currently running 0.100” jets to get reasonable thrust at the low tank pressure.  We have had some roughness problems with the 2” engines at that jet size on the lander, but they seem to be running ok here, possibly due to the lower pressure (although that is the opposite effect that lower pressure should have, cat packs complicate the behavior).  There are Enidine wire rope isolators at each mounting corner, of a softer rate than we used on the manned lander.  Each isolator has a polyethylene “brick” mounted to the bottom to raise the vehicle up enough to allow them to compress, and to allow it to slide sideways on landing.  The bricks make a very large difference in how easy it is to scoot the vehicle around.


We have a bulkhead plate on the side of the base for the loading quick connect, and a pressure gauge mounted above it.  This is a very nice fill arrangement.


The test last Saturday, where it didn’t lift off from the ground, produced some thermal issues in the base cone.  The heat-shrink tubing around the motor valve wiring was melted, several tie wraps were effected, and the motor valve connector is getting a bit “crispy”.  The Tefzel wiring wasn’t fazed in the slightest.  We wrapped the motor valve wiring in Nomex for now, but in the future we may run Tefzel all the way inside the motor valves.


We have a rather tacky bunch of wires running up from the base (attitude engine control and main servo valve) along the side of the rocket, currently just covered in duct tape.  The next vehicle will have a proper sheet metal conduit.


The main support for the parachute loads is carried by a 5/8” thick Kevlar rope that loops all the way under the tank and up to the parachute harness.  The rope along the sides (rocket suspenders) looks a little odd, but this is probably the most robust arrangement for now.  We may crack some of the fiberglass where the rope feeds through when we deploy.  The next vehicle will rely on the bond strength of the main airframe, and attach the parachute at the top.


We have the tank pressure transducer mounted on the top closure, underneath the computer bulkhead.  This was one of those circumstances where we really wished we had added two more inches to the tube height above the tank, because the transducer wouldn’t fit under the computer in any of the obvious positions.  I wound up hanging it down by the deeper domed section of the tank end, and connecting it to the tank with stiff stainless hard line.


The parachute is a lo-po 350 (low porosity canopy) from Butler Parachutes.  It is packed into a deployment bag, which is pulled out from the rocket by two Pro-38 three grain motors mounted on a tower above the nose.  The actual nose cap that pulls away is cut from a jet engine spinner cap that Phil had lying around, because the main fiberglass cone didn’t come to a complete point.  Packing the parachute into the bag is a real pain, exacerbated by the fact that we stuff the normally 8” diameter bag into a 7.5” diameter coupler that guides the nose away from the body.  We are going to want to improve that in some way in the future.


We will be doing hover tests and parachute tests next week, but we still haven’t heard from our local FAA office about our flight waiver, so we may need to drive to Oklahoma to do the first flight at Burns Flat.



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