July 3, 2001 Meeting Notes

 

In attendance:

 

John Carmack

Phil Eaton

Russ Blink

Neil Milburn

 

New supplies:

 

Stainless steel machine screw, bolt, and washer sets

McMaster stainless/Teflon test hose

Ultrasonic range finder

Re-geared KZCO ball valve

Various wiring supplies

Several types of metal screen and perforated sheet to evaluate for pack retainers

 

It has been 30 days since my last contact with the FAA about manned rocket ships, so I have sent a follow up mail.

 

I haven’t received a response from any of the composite filament winding companies.

 

Bob is planning on having the manned vehicle frame completed tomorrow.

 

We are meeting on Saturday at 3:00 for flight tests. Odds are good that we will have a couple hours of work to do before a liftoff.

 

Range Finder Test Results

 

Tim Nolan built an ultrasonic range finder for us to try using as an altimeter for the VTVL to help with automated landing.

 

We set up the ultrasonic range finder two meters from a wall, and fired the 50lb thrust motor next to it.

 

media.armadilloaerospace.com/2001_07_03/RangeFinder.jpg

 

The log format is: signal strength 0-1023, range in meters, velocity in m/s:

media.armadilloaerospace.com/2001_07_03/rangefinder.txt

 

A graph of the range values:

media.armadilloaerospace.com/2001_07_03/rangefinder.xls

 

Short answer: it lost its mind when the rocket fired.

 

This was more or less what we expected, but it is good to have a definitive answer.

 

Tim is looking into some options with radar range finding, but we will also be trying both pure inertial and inertial augmented GPS soon.

 

 

Braided Hoses

 

The braided hose from McMaster was slightly different than I expected.  I thought the swiveling NPT male ends were a custom hose end, but it turns out they are flare hose ends with a normal flare to NPT fitting included.

 

The JIC hose ends are flared the same as AN hose ends (37 degree flare, unlike the 45 degree SAE flare), but the diameters aren’t the same. The ¼” hose and hose ends seems to be in between AN –4 and AN –6 size.

 

They also have swaged on NPT male hose ends that don't swivel at all, and are cheaper.  For things that need an orientation, it probably isn't worth the headache to try and Teflon wrap them until they tighten to a correctly clocked angle, but we could often use a non-swivel on one end of the hose to save a fitting if we wanted to.

 

I expect we will get all of our hoses made to order at McMaster in the future, except for when we just need to put one together on the spot.

 

I have seen a comment that the Earl’s SpeedFlex hose that we have been using may have a thicker Teflon lining that commercial braided hose, but the pressure ratings are actually slightly higher on the commercial hose, and not having to worry about poking holes in your fingers when spreading the steel braid is a nice benefit…

 

 

Electronics Box

 

I finally wired a power switch on the outside, which is a lot more convenient than disconnecting the battery. The switch is only rated for 30 amps, and we can draw nearly 40 intermittently, so it is a bit marginal. Future vehicles that only use the solenoids for attitude control instead of primary lift will actually draw significantly less power than the current vehicle.

 

The accelerometer is installed and functioning.

 

I added wire looms around most of the edges.

 

The GPS socket is installed and wired, but for some reason it isn’t working yet.

 

There is still space in the box to fit a magnetometer and the main throttle motor drive for the manned vehicle.

 

One odd problem that I am having is that the flight computer fails to boot completely when it doesn’t have a keyboard and video card hooked up, but after I reset it once, it completes the boot normally. It always boots correctly with a keyboard and video card installed.

 

On a whim, I took a lot of high frequency traces of the noisy rate gyro signals and a 9v battery hooked directly up to the A/D board. They were correlated fairly well, although the scales were different. I entertained the notion of building up a correction table and always reading the reference voltage before and after reading a signal level, then correcting for the system noise as determined by the reference voltage, but I have enough sense to recognize that the proper solution for this really is in the analog domain with just getting better conditioned power.

 

Phil and Russ got the driver / power board cut and drilled today, and will hopefully be finishing it tomorrow.

 

Once we have the ball valve motor drive hooked up, the only expected change in the electronics all the way to space shots will be in improved communications. At a minimum, I am going to need to change to an 802.11 card that allows an external antenna, so we can use a more directional antenna mounted away from the electronics box. I am beginning to think we might be better off moving the entire radio outside the box, the same way we had to with the GPS receiver. A higher powered wireless Ethernet bridge would give us all the convenience we currently have, but we might have to go to a low bandwidth AX25 style system over a serial link. I suppose we could run PPP over it to still keep telnet sessions, but I would probably just do some form of dedicated protocol..

 

Orientation

 

This is a work-in-progress, but I am beginning to put together a document about all the flight control issues:

 

media.armadilloaerospace.com/misc/orientation.htm

 

I have changed all axis to match NASA’s conventions.

 

I have a little test program that displays the current vehicle (electronics box) world space axis in real time 3D on the laptop, and now that I do the full axis rotation instead of angular integration, it stays correct for all possible vehicle rotations, including upside down rotations. It also displays the accelerometer data off of each axis, but I haven’t started tracking world space velocity and translation yet because of the current sensor noise level.

 

Now that I have good power conditioning, I am going to see exactly when a pure inertial navigation system breaks down, and to what degree. I’m pretty sure it will be good enough for auto-land experiments in our 15 second powered flight time, but I have less confidence in it working well enough after spending a minute coming down with a drogue chute.

 

The complementary direction to explore is using the GPS to reset position and velocity every second. For landing on the same surface you launched from, the absolute accuracy of GPS isn’t important, just the relative noise and short term drift characteristics. We should have some traces on this in the next month or so.