April 16 and 20, 2002 Meeting Notes

 

In attendance:

 

John Carmack

Phil Eaton

Russ Blink

Neil Milburn

Joseph LaGrave

 

Motor Jetting

 

We ran one of the attitude motors (motor label 1) from the lander on the test stand, and we did still see rough running.

 

We opened it up, and while the pack did still seemed to occupy the full volume, it felt a little springier to the touch.  I put it back in the press to the same 500 psi indicated (about 1500 psi on the screens) that we originally packed it at, and it opened up enough area for ten additional stainless screens.  Running it again, it was still rough.

 

We pushed the entire pack out, removed the front spreading plate, moved a couple of the extra stainless screens to the front of the pack, and added a spacer under the retaining plate to take up the rest of the space.  It still ran rough, and started to exhibit the “machingun roughness” that it had shown on the lander.

 

We were about to try replacing the bottom three anti-channel rings, which had gotten a bit loose (although the runs weren’t cloudy), when we decided to try a quick jet change from 0.080 to 0.060.  The engine got completely smooth again, although the total thrust had dropped from 32 pounds to 25 pounds.

 

We then pulled another engine that had been rough on the lander, ran it on the stand to see if it was still rough, then just changed the jet without changing anything else.  It was smooth, and only down a couple pounds from the original 32.  It looks like the re-compressing of the first pack closed up enough flow area to drop the extra thrust.

 

We are leaving the one engine in modified form on the lander for now.

 

I finally sat down and worked out the theoretical flow rates through our jets, which I should have done a long time ago.  Many unit conversions ahead, so someone correct me if I made a mistake:

 

Our jets are labeled in thousandths of an inch diameter, so converting to square cm:

 

2.54*2.54*0.25*3.14159

jet size * jet size * 5.07 = square cm of jet area

 

0.060 jet = 0.0182 square cm

0.070 jet = 0.0248 square cm

0.080 jet = 0.0324 square cm

0.090 jet = 0.0411 square cm

 

Our test stand results are in pounds of thrust, so converting to peroxide volume flow, assuming a 115 Isp and a 1.4 density:

 

20 lb/s thrust =

9.09 kg/s thrust =

0.079 kg/s peroxide flow =

0.056 l/s peroxide flow =

56 cm3/s peroxide flow

 

1000 / 2.2 / 1.4 / 115

pounds thrust * 2.82 = cm3/s flow

 

20 pounds = 56.4 cm3/s

25 pounds = 70.5 cm3/s

30 pounds = 84.6 cm3/s

35 pounds = 98.7 cm3/s

 

The fluid velocity through a jet is the volume flow divided by the jet area.  The Bernoulli equation states that the pressure drop equals the density times velocity squared divided by two.  We still think in psi, so 1 ps = 1.45e-4 pascals.

 

1.4e4 kg/m^3 / 2 * 1.45e-4 psi/pa

pressure drop in psi = vel * vel * 1.02

 

30 pounds with an 0.080 jet = 2.61 m/s = 6.95 psi

30 pounds with an 0.060 jet = 4.65 m/s = 22.1 psi

 

So, the original jet was only providing a 7 psi pressure drop, which is not enough to provide much damping at all.  When a pack is really perfect, it can run smooth with no external jet at all, but after a little running, things get enough out of sorts that we really want to have a reasonable drop to decouple the combustion chamber from the feed system.  Huzel & Huang recommends a 15% to 20% of chamber pressure drop across the injector, so even the 22 psi number is probably on the light side for us, although it isn’t exactly clear if the “chamber pressure” should be before or after the catalyst pack.

 

Peristaltic pump

 

We got a tubing pump to consider for various peroxide transfer tasks.  It pumps 2.8 gpm, and will push into a pressurized container, so it could be used for loading vehicle tanks, but there are a couple disadvantages compared to vacuum loading: the suction isn’t nearly as strong (although it self primes just fine), so you have a harder time getting every last drop out of a small container, and the exit line will usually hold a bit of liquid.  If we decide to use this for the vehicles, we will have to arrange to mount the pump at a level above our main control manifold, so all peroxide will drain down to the manifold, where it will be blown out by the nitrogen pressurization.

 

While I was test pumping water into a tank, I did have one of the hoses blow off the barb and spray water around, so it is worth noting that hose clamps are mandatory even at under 10 psi pressure…

 

Solenoid water flow tests

 

The maximum possible peroxide flow through the rotor tip engines is limited by the amount that can flow through the controlling solenoid below the shaft, no matter how fast the rotor is spinning.  Because a fixed pitch blade takes a certain amount of thrust to turn a give RPM, if there isn’t enough peroxide flow to generate that thrust, there is no way for it to possibly turn faster than that if it can’t flow enough peroxide through the valve.

 

We tested the NOS pro-race solenoid at a few tank pressures, exiting to atmospheric:

 

100 psi tank pressure = 183 g/s

200 psi tank pressure = 233 g/s

300 psi tank pressure = 283 g/s

 

If we assume an effective Isp of around 1000 when the aerodynamic lift is included, that would give us a maximum lift of 512 pounds at 200 psi tank pressure.  With the previous set of blades, we saw 330 pounds of lift at 645 rpm.  The current set of blades has twice the pitch angle and a little bit longer diameter, so it should produce roughly twice the lift at a given rpm.  Based on this, it should not be able to spin much over 600 rpm with this solenoid held wide open.

 

It should not be possible for us to sling these blades off with this configuration at rest.  If we replace the solenoid with a ¼” ball valve, we would again have the ability to overspeed the rotor (and make over 2000 pounds of lift).  If we fly at high vertical speeds, the rotor rpm would also increase.

 

Rotor RPM control

 

We have rebuilt the uprights for the vertical test stand, but I don’t have a new 1000 pound load cell yet, so we decided to just do rpm monitored tests by building a rotor mount on top of our big seated lander.  This will allow us to get a vehicle off the ground very shortly after Space Access.  We will still have to learn how to get the computer to deal with precession effects, but if we don’t wreck it in the process, we should be flying it fairly soon.

 

The computer rpm control tests were a big success, with a few caveats:

 

We should probably work out some way to get a restricting jet into the tip engines, because we wind up gushing a fair amount of peroxide out while warming up.  I got more patient with short pulses and plenty of time to cook off on our later runs.

 

We had some problems with the RPM sensor on some runs.  We will be replacing the trigger bolts with permanent magnets, which should give a much stronger signal to the sensor for the next runs.

 

We heard some popping during the solenoid toggling, which may well have been vapor phase detonations as new peroxide slammed down the cavitated lines.  It didn’t seem to hurt anything, but it might still be an issue at higher RPM with more mass in the rotor feed lines.  Moving to ball valve control would probably fix that, and make everything a lot smoother, at the expense of a more complicated control algorithm.

 

This is a graph of two separate runs, the first one with a change from 150 to 175 rpm, and the second one at 200 rpm.  The indicated RPM numbers are double actual, because we had two triggers for the rpm sensor.

 

media.armadilloaerospace.com/2002_04_20/RotorThrottle.xls

 

With the rotor on top, our lander now weighs in at a portly 415 pounds, dry.  We will be doing our first precession thruster tests with this vehicle at the 100 acres, but we have started work on our next generation vehicle, which will finally be designed to be lightweight.  We are aiming to have it be under 250 pounds, so we can fly it as an ultralight helicopter.