Home / News

News Archive

Two inch biprop, rocket drawn parachutes, tube structure tests

June 18 and 22, 2002 Meeting Notes


In attendance:


John Carmack

Phil Eaton

Russ Blink

Neil Milburn

Joseph LaGrave (Saturday)


Two Inch Biprop


We made a functional mock-up in brass of the radiatively cooled TZM chamber and tested it this week.  It has a fuel injector ring that clamps under the catalyst pack like our one inch biprops, but instead of screwing directly into the chamber / nozzle section, it bolts onto a dedicated clamp ring, which fits underneath a lip on the chamber / nozzle.  This gives more even clamping pressure, reduces the machining operations on the parts that are more likely to be consumed, and is recommended practice in the NASA SP on radiatively cooled engines to reduce stress concentrations.  The chamber/nozzle was a total of 3.8” long, with a 1.9” ID, a 45 degree converging angle to a 0.5” throat, expanding at 16 degrees to a 0.9” exit for 3.24 expansion ratio, a decrease from our standard 4x expansion ratio because we planed on running only lower pressures.


We made Isp measurements, but because we don’t have flow meters on the fuel side, they are just pound seconds of thrust divided by pounds of oxidizer.  For peroxide, with optimal O/F ratios of 7:1 or so, peak oxidizer-only Isp is about 15% higher than true Isp, which will be found at a slightly higher O/F ratio.  It was still a useful metric for us.


Because this was mostly intended to be a test for the radiatively cooled TZM chamber, we didn’t have a water jacket of any kind around the chamber.  For cooling, we mainfolded together four machinists’ cooling spigots and arranged them to flood the chamber with water while operating.  This did cause a lot of water to be kicked around in the exhaust jet, which made some of the runs visually look rough, even when they were dead smooth in the data.


The catalyst pack was the lightly-compressed test pack with only 60 silver screens, which is a touch marginal for monoprop operation, but worked fine for all the biprop runs.  A 0.120 peroxide jet was used for all tests, at a regulated tank pressure of 250 psi.  There were indications that the jet should have been a bit smaller when operating at 250 psi tank pressure, because there was a slight increase in roughness when the regulator let the tank pressure drop a few psi below 250.  All the biprop runs made between 40 and 50 pounds of thrust, taking around 12 seconds each.  A monoprop-only run gave a true Isp of 96, which may be slightly low due to the marginal catalyst pack, but is probably pretty close, because we only see a 115 Isp when we run at 600 psi tank pressure.


The first test was with kerosene and a 0.040 jet.  It lit right up and ran perfectly smooth, but was noticeably rich.  We expected this, as the smallest 0.018 jet on the one inch engine didn’t seem lean yet.


We reduced to a 0.030 jet, and it ran a bit better, but the indicated Isp was only 118, which means the true Isp was barely better than the monoprop.  We had seen better results than that with the one inch motors, so we were obviously getting crappy combustion.  There were soot deposits on one side of the nozzle, indicating that the fuel was not being mixed well.  We are just injecting the kerosene directly from the jet, so it is basically a single, high pressure stream that shoots across beneath the catalyst pack and impacts the opposite wall.  Scaling the engine makes this less and less acceptable.  We may be able to crutch it with a much larger combustion volume, but we will probably have to make a better injector for liquid fuels.


We decided to try the ethane again, which should have much better combustion characteristics, due to gas/gas combustion.  We only had one successful set of tests with the ethane on the one inch engines, at 0.090 jet from 250 psi regulated ethane pressure.  Three other attempts to light it failed for unknown reasons.


Our first test was a 0.120 jet with 350 psi on the ethane regulator.  It lit right up for a perfectly smooth run with a 172 indicated Isp, and it was LOUD (noise is proportional to the third power of exhaust velocity).  We repeated the test with similar results.  We don’t know conclusively why we had problems with it before, but there were a number of potential second order changes:  The temperature was at least ten degrees warmer, which increases the ethane tank pressure, which helps the regulator hold more constant.  The catalyst pack holder is now stainless, which retains a bit more heat.  The combustion chamber has a slightly greater L*.


Because that was about as much ethane as we could flow through the existing setup, we replumbed everything for a higher flow setup.  We removed the regulator altogether, so we would be running directly from the CGA 350 bottle connector, through a –6 hose, to a pro-race solenoid.  This was a major change from the regulator to –3 hose to cheater solenoid in the first tests.


Ethane specs: http://www.concoa.com/frames/technical/gases/ethane.htm


Without the regulator, the pressure will vary from 543 psi at 70 degrees, to 708 psi at 90 degrees, above which it will scale like a normal gas.  It was over 90 degrees on our test day, but the ethane tank wasn’t that hot, and you could feel it being noticeably cooled as we drew gas out of it.


It is worth noting that ethane’s density is very poor, only 0.38 g/cm3 at 60 degrees, and worsening with increased temperature.  This means that even with peroxide’s high O/F ratio, the ethane tank would need to be 65% of the volume of the peroxide tank, and capable of holding significantly more pressure, meaning the ethane tank could easily mass more than the peroxide tank.  Another issue is that once all the liquid ethane has been vaporized, you still have about a third of your total ethane mass as a pressurized gas, which can only be extracted in blow down mode.  That isn’t much of a hardship, because you want your thrust to tail off towards the end of a space-shot burn, and you don’t have the issues with required injector pressure drop with gaseous injection.


Our first test with the new ethane plumbing retained the 0.120 jet on the ethane, and it did not light.  There was a noticeable hydrocarbon smell, so we were clearly flowing much more ethane.  We dropped the jets down through 0.100, 0.080, and 0.070 with it getting closer and closer to smooth combustion.  The 0.080 and 0.070 runs would light, but run extremely harsh.  We thought there might be a fundamental problem with not regulating the self-pressurizing ethane, but when we dropped to a 0.060 jet, it went back to perfectly smooth combustion, with an indicated Isp of 172.


We dropped to 0.55, and got another perfectly smooth run, but the Isp dropped slightly.


We increased to 0.65, and got another perfectly smooth run with an indicated Isp of 184, a peak between the 0.60 and rough running 0.70.


Leaning it out a bit would probably give us a true Isp of about 156, which is probably not too far off for a pressure ratio of only 12 or so.


We had basically all the data we wanted from this test motor, so we decided to try a high pressure run to see if we could get an indicated Isp over 200.  We set the tank pressure at 600 psi, which gave us a 280 psi chamber pressure on the small motors.  The chamber pressure would be higher for the biprop, because the fuel is injected after the catalyst pack pressure drop.  We also expected to pick up a little bit more Isp due to better combustion at the higher pressures.


It lit right up, and we saw our first ever overexpanded exhaust plume, indicating that our chamber pressure was indeed quite a bit higher.  Thrust was slightly over 100 pounds, and it was running smoothly.  After five seconds of burning, a chunk of the nozzle came off.  This wasn’t all that exciting, just resulting in the thrust vector spraying water in different directions.  I stopped both the peroxide and ethane immediately, then we checked out the damage before running the rest of the peroxide through the engine in monoprop mode.


We have decided to have our bar of TZM cut into two 6” long chambers instead of three 4” long chambers, so we are sure to have plenty of combustion volume.  Our first oxidation protection attempt will be a platinum plating, which requires an initial flashing of nickel to adhere to molybdenum.  If that doesn’t work out, we will try some form of silicide coating.  Molybdenum oxide is volatile, so if the oxidation protection layer fails, the engine basically evaporates…


http://media.armadilloaerospace.com/misc/burnThrough.mpg  ( you can barely see the plume with the good combustion and bright lighting )



Rocket Drawn Parachutes


We did a set of experiments with rocket drawn parachutes.  Our recovery system is probably going to consist of both a rocket puller, and a backup piston ejection system in case the rocket fails.  We test fired one of the commercial ballistic chutes for ultralights a while ago, but we need an electrically actuated system instead of a pull-cord system, and it will be good experience to put our own system together.


Our test rig consisted of a strip of aluminum with a lip bent into the top end, and a wire cable attached to the bottom end.  We attached various Aerotech solid rocket motors to the top, abutted against the bent end, and we ran the wire cable down to various parachutes we had around.  We dumped the ejection charges out of the single use motors we used, and we greased the delay grain on the reloads so it wouldn’t burn at all.  We had two CATOs when we had the single use motors attached to the aluminum strip with hose clamps, because the cases are so fragile.  We went to reloadable motors, which have sturdier cases, and we just duct taped them to the aluminum bar, which worked fine.  For launching, we just stuck the aluminum strip in a tube, bottle rocket style, with the parachute lying loose to the side, connected to a long rope tied off to the trailer.  There was much cursing of the damn copperhead igniters that I still had in my launch box.  We will get some real igniters before we try this again.


The first shot is of a D21 motor pulling a small rocketman  parachute.


The second shot is of an F50 motor pulling a 16’ diameter military surplus parachute.  When the parachute came down on top of the motor, we burned a couple holes in it.  Instead of using a metal strip to hold the motor and cable, we will build a holder out of phenolic tubes, which will keep any exposed surface from getting very hot.


The third shot is of a G64 trying to pull a very heavy duty 12’ drogue chute.  It failed.


Getting a proper parachute is now on the critical path for the tube vehicle.  I want to get a pretty heavy duty chute of at least 20’ diameter, preferably without too terribly many shroud lines.  We have the ultralight chute if we need it, but it is rated for 900 pounds, which is three and a half times the weight of the vehicle, so it would drift for a long time.





Tube Structure Tests


We finished all the assembly work on the tube vehicle.  Total weight is 257 pounds dry, still missing a nosecone and parachute.


The only thing that seems lacking right now is that lifting the vehicle by the legs is giving some disturbing sounds from the bottom bulkhead, and has bent the legs up a bit already.  We have stainless backing plates on the top nut side of the U bolts, but we will probably have to add backing plates between the leg bars and the bulkhead on the bottom.  I think we will find that we now have round indentations in the plywood when we look at it next week.


The vehicle is designed so that all the forces are transmitted through the bottom bulkhead, which is backed up by a full-circumference thrust ring above it.  The parachute attachment cords go through holes in the top bulkhead and centering rings, down to the attach points on the leg U-bolts on the bottom bulkhead.  The main engine and the landing shocks also feed through the bottom bulkhead.


Our largest concern is that the tank takes all of its acceleration loads through the rigid plumbing to the main engine.  It is well placed with centering rings, but a hard landing shock may push the top fitting into the engine.  I am also a little concerned that there is nothing to keep the tank from rotating, possibly loosening some of the plumbing.  A flange mount tank will be welcome for future vehicles.


We did two sets of tests: dropping the vehicle onto the shock cords, with it suspended so it wouldn’t hit the ground, and dropping the vehicle onto the ground to let the landing gear absorb the shock.  The flight computer was running to log accelerometer values, but it wasn’t an ideal collection platform, because it was only sending 20 packets a second.  I got instantaneous acceleration, and acceleration averaged over the sample period, but it almost certainly missed the peak instantaneous accelerations.


The first shock cord drop test was not released very evenly, so it was tilted when it pulled up taut.  This resulted in a significant jerk on one of the side axis, with a 5G peak, as well as 4G on the vertical.  It had a few more 2G vertical bobs before settling down.


The second shock cord drop from a bit higher up had a tenth of a second 6G vertical acceleration.


The third test showed an instantaneous 12G acceleration.


The shock cord drops were rather exciting, with a 250 pound, metal finned vehicle bouncing around on the end of some ropes, but the landing gear tests were pleasantly anti-climatic.  The vehicle just set down with a thud, with no noticeable rebound.  The telemetry would show one frame of 4G averaged acceleration, then a few minor oscillations.  I’m sure we missed the peak accelerations, but it makes sense that the shock absorber decelerations will be much less than coming up with a jerk on a Kevlar rope.


I need to make a high fidelity logging mode that keeps all 200 accelerometer samples a second, and we should try some of these again.


Next week we will see if the abuse has hurt any of the plumbing.





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