0 0 0
Home / News
0

News Archive

Engine pressures, Big lander testing

May 30 and June 1, 2002 Meeting Notes

 

In attendance:

 

John Carmack

Phil Eaton (Tuesday)

Russ Blink

Neil Milburn

Joseph LaGrave (Saturday)

 

Engine Pressures

 

An open question we have had for a long time was what the appropriate ratio of catalyst pack area to throat area is.  Our rule of thumb for the screen engines has been to use a diameter ratio of 3:1 for the catalyst pack to throat.  We ran back to back tests this week on a 1” diameter cat pack with two different nozzle diameters: a 0.33” and a 0.45”.

 

Our previous attempts at measuring chamber pressure had not worked out very well, because our pressure transducers started reading bad when they got too hot.  This time, we connected the sensor to the engine with a piece of bent tubing, and poured some brake fluid into transducer to buffer it.  We used our biprop fuel injection ring for the engine port, and we just left the very small stainless jet in place, which also served as a low-pass filter on the pressure trace.  The injector ring has gotten a bit warped after all the hot fires, so there was probably some leakage, but the data all looked sane.

 

With the 0.33” nozzle and a 0.060” jet, we only saw 225 psi chamber pressure from 600 psi feed pressure.  The run was slightly rough, because the 0.060 jet is a little marginal for pressure drop.

 

With the 0.45” nozzle and a 0.060” jet, chamber pressure dropped to 110 psi, giving slightly less thrust than the 0.33” nozzle, but noticeably smoother.

 

With the 0.45” nozzle and a 0.080” jet, the run was full rough, indicating not enough pressure drop across the jet.

 

We put a pressure transducer off of a T between the jet holder and the engine to see how much pressure was lost by the plumbing, solenoid, and jet.  With the 0.080” jet, there was surprisingly little, showing 530 psi down from 600 psi tank pressure.  A 70 psi drop is clearly not enough for smooth operation.  With the 0.060” jet, we saw 460 psi, which seems to be sufficient for smooth operation.

 

So, we are dropping from 460 psi to 225 psi across the catalyst pack, which is a very significant pressure drop, and the 3:1 diameter ratio of catalyst pack to throat is probably as far as we can go with our current screen packs.

 

Running at higher feed pressures has shown to give somewhat better than linear increases in thrust with the screen packs, probably because it pushes liquid farther down the pack, so the gas flow has less screens to traverse.

 

Our 4:1 nozzle expansions are all quite a bit overexpanded for the current pressure ratios, so we could improve Isp somewhat by cutting them down.  We could alternately improve chamber pressure by making smaller nozzles for our given pack diameters.

 

Our current packs have a total of around 70 silver screens and 60 stainless screens.  The 1” motors did not work well with less screens, but the larger motors seem to be doing well with 60 + 50 screens and a lower packing pressure, so they are likely somewhat more efficient.  We know for certain that an all silver screen pack doesn’t have the structural strength to work, but packs that use silver plated screens exclusively can get by with lower screen counts at the expense of shorter lifetime.  We prefer the long life and repeatability of the pure silver alternated with stainless.

 

We will eventually try some experiments with lower pressure drop packs, but it may not really be an issue for us.  Right now, our tanks are good for a whole lot more pressure, so we can just run more tank pressure if we want higher chamber pressure.  Our solenoids are good to about 1200 psi, but we would have to change out our current 1000 psi burst disc if we go too much higher.  With biprop and hybrid motors, you get the chamber pressure back up because most of the reaction takes place after the pack.

 

Big Lander Testing

 

Several improvements have been made to the flight computer / remote piloting software system, but the main loop is now taking three to four milliseconds on the puny 100 mhz computer in the electronics box, which is uncomfortably close to some of the sensor periods.  I have a new 400 mhz panel-PC which we are going to use in future vehicles, but we will probably be sticking with the current system for the existing vehicles.

 

We tested with the laser altimeter today, but I still need to get it to toss out some extraneous data samples before it can be used for the auto-hover logic.  We are using a LaserAtlanta Advantage range finder, but it really isn’t ideal for our purposes.  It has no convenient way to mount (we wound up just strapping it to a tank), and while it can be controlled over RS232, it can’t be remotely powered up, and it has a power saving shutdown mode.

 

This was our first time lifting the lander in months, with the primary difference being that we now have stainless body silver screen engines all the way around, and replacing the bent main engine.  The engines are working great, with completely clear catalyzation, and no sign at all of deterioriation.  Russ commented that you couldn’t even smell peroxide in the exhaust.  The fairly rapid pack deterioration was the primary thing holding us back in the large vehicle testing, and it looks like we have that well in hand now.

 

However, the little 1” engines had insufficient control authority at 500 psi tank pressure.  They vehicle kept tipping shortly after liftoff, even with one engine full on.  When I went back and carefully looked at the telemetry, I found that our main engine is not pushing along the CG.  This was shown by an axis rate not being able to make any correction headway until the main engine was throttled down, at which point it pulled around to where it should be.  We hang balanced the entire lander, and the new central engine is perfectly square, but the mounting brackets must not be level.  What we need to do is suspend the vehicle, and shim the main engine until a level placed across the exit cone is even in all orientations.  We should also level the computer box so the IMU’s initial angle initialization based on the gravity vector is true.

 

We bent one of the legs coming down on our second load of peroxide.

 

We had a few options to increase the control authority:

 

Move the engines farther outboard, to get a better lever arm.  Putting them on the ends of the legs would increase the lever arm by 50% or so.  This would require new hoses, and would place the engines in danger on a bad landing.

 

Run higher tank pressure.  We could go from 500psi to 800psi.  This would make the main engine throttle range lower, and makes any catastrophic failures scarier, but would give at least 50% more control authority, and a higher Isp on the attitude engines.

 

Move to the 2” engines, which could give us over 3x the authority with the same plumbing.

 

We went ahead and started packing a full set of 2” engines, although now that I have determined the thrust misalignment, I would be somewhat more tempted to stick with the smaller engines.  Too much control authority increases the amplitude of the hovering oscillations, and the larger engines have somewhat more latency to thrust, which decreases the period of oscillations.  We will probably have some fairly large oscillations when flying with the big engines, so we may wind up running them with jets not much larger than used on the 1” engines.

 

We need to get some more stainless screens and spiral rings to finish the last 2” engine, but if Bob can fix up the leg soon, we might by flying again on Tuesday.

 

 

 





 






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