Jul 24, 2001 Meeting Notes

 

In attendance:

 

John Carmack

Phil Eaton

Russ Blink

 

New equipment:

 

Twist grip Magura throttle for manned vehicle

Dremel cutting discs

More hazmat suits

Degreed angle bubble level

New pressure gauge for the demonstrator

Battery for master cutoff computer

Battery for test stand box

Bolt down wire ties

Ratchet terminal crimper

 

Russ standing by our two vehicles:

 

media.armadilloaerospace.com/2001_07_24/ships.jpg

 

Electronics Box

 

The electronics box is almost totally complete for all the capabilities it needs for the next few vehicles.

 

The external connectors to the electronics box are now:

 

Four engine solenoids

Main engine ball valve / position potentiometer

Remote GPS

Tank temperature RTD

Tank pressure transducer

Twist grip throttle

 

All of the connectors are AMP circular plastic connectors (CPC), which are a pleasure to work with if you have the right crimper for the pins.

 

I still need to add a connector for the master cutoff watchdog signal, a connector for the manned vehicle pilot joystick, and a barometric pressure sensor.

 

I was going to add a magnetometer for initial earth relative orientation (it goes nuts when the engine solenoids are operating), but I don’t have three free A/D channels any more. I have a different magnetometer that talks over a serial port, but I don’t have a free serial port, either.

 

The current box doesn’t have room for another board in the PC104 stack, so I am probably going to shop around for some more integrated boards that will let me get four serial ports some way. Worst case, I can add USB serial ports if I need to, but I don’t trust that as much for real time things as a normal serial port, and I’m not sure how much Linux grief I will have to go through to use them. One thing that we are going to have to do eventually is move away from using the PCMCIA 802.11b wireless adapter to something higher power, which will probably talk to the CPU over Ethernet (that will also let us shield the entire PC104 stack, which should help some noise issues). There are PC104 CPU boards that have Ethernet and VGA (for pilot instrument panel displays) on a single board, which would allow me to ditch the PCMCIA board and the messy supporting linux infrastructure for it.

 

We added mounting brackets to the box today, so it is now rigidly screwed to the platform, and we don’t need to mess with the restraining hose clamps we used to use.

 

The current box weighs 20 pounds fully loaded. All the little things add up…

 

I have extended our remote pilot laptop application to visually display all the data the computer collects, as well as a 3D graphical representation of the vehicle attitude. It doesn’t update during actual flight, because the screen redraw can cause a noticeable increase in control latency and variability.

 

Master Cutoff Computer

 

I have all the parts now for the master cutoff watchdog computer. It consists of a motor drive board, a basic stamp board, its own private battery, and some connectors. The master cutoff valve goes before the attitude engine manifold, which is before the lifting engine throttle (on the manned vehicle). The idea is that if anything catastrophic goes wrong, like the main computer crashing or losing battery power, the cutoff computer will notice the loss of a watchdog signal and shut off the peroxide to prevent any additional thrust. This also addresses any “stuck on” valve conditions.

 

I’m not sure yet if I want to spend a serial port to communicate with it with feedback, or if I just want to have a digital IO line pulsing regularly.

 

 

Big Vehicle

 

The big frame is at Long Range for fitting now. We measured lines for some plumbing and electronics tonight, and we will be doing the necessary cable building on Saturday.

 

The large foam blocks for the landing gear will be here by Saturday.

 

Russ has all four attitude engine bodies finished.

 

Phil has the drawings for the main lifting engine done, and the metal stock should be here soon.

 

Bob is going to take the frame back on Saturday for any final modifications, then we are going to get it all powder coated so it doesn’t rust as easily as the small vehicle.

 

I have the twist throttle and main engine ball valve hooked up to the computer, but I have some more work to allow the software to run with either the four engine demonstrator or the five engine manned vehicle (both remotely piloted and physically piloted).

 

The vehicle is turning out heavier than we expected (that seems to be chronic with rockets). With all the engines, foam, and electronics mounted, it is going to be right around 200 pounds dry. We may remove some of the bracing tubes, because the main cross members are a lot stronger than they need to be. Once we have some experience with how the attitude engines deal with the offset CG, and we aren’t too afraid of tipping it over, we might be able to move the engines inboard quite a bit and get rid of most of the external structure, which would save nearly 70 pounds.

 

Starting the pipeline for the next two vehicles after this, I found a supplier with a lot of online data about their filament wound pipe:

 

http://www.fibercast.com/f-chemeng.html

 

These are considerably heavier than we would want for an airframe (even a pressurized one), but they can probably make lighter runs.

 

We are looking at a 2’ diameter demonstrator that will basically be the propulsion system from the manned vehicle in an aeroshell with side firing attitude engines. This should go supersonic and let us learn about a lot of new issues.

 

Looking even farther down the road, we are getting some 48” Sonotube to allow us to start seeing how we can seat people inside larger frames. A real vehicle would be made out of composites, but for something to crawl around in and drill holes in, Sonotube should work just fine.

 

Pressure Falloff Log

 

We finally got around to logging pressure falloff during a tank blow down.

 

It was interesting to notice that paying close attention to the pressure showed that the pressure was dropping visibly within a few seconds of ceasing pressurization. The tank gets rather hot when pressurized, and it cools off rapidly enough that 20 or 30 psi is lost within 30 seconds. We should make another log of just letting the tank temperature equalize to see how much it changes over several minutes.

 

During blow down operation, the tank gets cold enough that we have formed ice in the past.

 

The tank is 8.5 liters, we loaded 2 liters of water, and pressurized to 600 psi, then let the engines pulse at a rate that was roughly equal to the flow during VTVL flight (about 10 seconds to expend 2 liters).

 

media.armadilloaerospace.com/2001_07_24/BlowDown.xls

 

The pressure graph is a bit rough while there is liquid being expelled, then it smoothes out when it starts to vent just gas. The transducer is in the distribution manifold, so the valve pulses are probably showing up as lots of little water hammer reverbrations.

 

It drops from 600 psi to 400 psi when the water is vented, which is over 50 psi of additional drop due to temperature changes.

 

We will probably repeat this test with 4 liters and (if we can draw a good enough vacuum) 6 liters of water, so we can gauge the limits of pure blow down tank loading for a pressure range.

 

 

Fiber Optic Gyro Bias Issue

 

We caught an interesting effect tonight. I had the remote pilot dashboard running next to the lander inside the garage with no liquid in it, and we were testing the various sensors and solenoid actuations.

 

The drift on the gyros by themselves is about a degree a minute right now (better A/D will improve this) when it is just sitting there, but when I held down the button to “fly” it, the drift increased very significantly, to nearly a degree a second.

 

Looking at the data logs, we could see that powering all four solenoids drops the main battery voltage from 12 to 10 volts. When the voltage dropped, the gyro bias points changed. This was somewhat unexpected, because it explicitly takes an unregulated supply voltage.

 

Ideally, the FOGs should be read with a differential A/D converter, but the A/D board I am using doesn’t let me select that on a per-channel basis, and I don’t have enough channels to put the entire board into differential mode, so I was reading the signal single ended. The docs mention a “slight loss of precision” when this is done, but they didn’t elaborate.

 

I still had the reference voltages wired up to the A/D board, so I was able to look at them in the data logs. The reference voltages aren’t supposed to move due to any physical action, but they do move when the supply voltage changes.

 

It looks like I can do my own “differential A/D” by just taking single ended measurements of both the signal and reference voltages and subtracting them myself. That is going to take my last three A/D channels, but it should make this problem go away.

 

I never noticed this before, because my bench testing never had 35 amps of solenoid current draw during operation.

 

It is a little surprising that this didn’t cause more noticeable problems during our flight tests.

 

Having all this data logged is a Good Thing. A minor interesting observations was that the 5v regulated level actually moved slightly in the opposite direction from the battery voltage – when the battery was at 12v, the 5v supply was at 4.91v, but when the battery dropped to 10v, the 5v supply was at 4.94v.