January 31, 2005 notes

 

Progress

 

The time on Sundays that I used to spend writing updates and such has been taken over by family duties, so updates are probably going to remain sporadic for the foreseeable future, except for when we get something really exciting done.

 

http://media.armadilloaerospace.com/2005_01_31/ryan.jpg

 

We had to build three separate motor drive boards to finally get something that works reliably. The first two used integrated motor drive chips that had supposedly helpful features like slow ramp up and overcurrent protection. We got the board up and running pretty quickly, but we had intermittent faults when using it to actually drive the jet vanes. If the overcurrent protection just limited the current draw to a maximum, that would have been fine, but both chips would trigger a fault condition and require a reset signal whenever something tripped.

 

We tried adding inductors in the lines to limit the current spikes, but it didn’t help.

 

When we tried a different motor drive part for the second board, we also added LED’s on every conceivable point on the board, which was quite helpful.

 

We tried adding software minimum turn-around delays for changing motor directions, which helped, but we could still occasionally get a chip to fault after extended running even with 50 millisecond delays, which is a lot more delay than I wanted to add. With a small delay, none of the motors would fault if they were all driven together, but with random directions on each of four motors it would still get a fault fairly quickly. There is evidently enough noise that the “smart” chips aren’t doing so well. As a confirmation about our earlier electronics problems, we did have a computer reset while testing the computer board and motor drive board from a single battery. We haven’t had any problems whatsoever when they are running from isolated batteries in their own isolated and shielded cases.

 

For the third board, Russ went back to a discrete transistor design like our original motor drives, and this works fine. Thankfully, the company that we are working with for the boards now has been turning them around for us in less than a week.

 

We did a practice load and warm-up test of the full vehicle on Saturday, which went fine. The one thing we had an issue with is freezing the loading check valve during nitrogen pressurization. We plan on trying to hover the vehicle on Tuesday if the weather cooperates. There are still two things we want fixed before we do boosted hops: The differential pressure transducer we are using as our “fuel gauge” failed again, and we are waiting for it to be repaired. Since taking off without enough propellant was the cause of our last crash, we aren’t going to fly without it. The other thing we haven’t finished is the isolated voltage signals for the motor drive and master cutoff batteries, which I really want to get into the telemetry.

 

http://media.armadilloaerospace.com/2005_01_31/warmupTest.jpg

(The spindly thing hanging off the right is the detachable vacuum pump for propellant loading)

 

The other thing we have done on the vehicle is set up a mounting point for up to 100 pounds of lead ballast on the nose of the vehicle. We are interested in doing disruption pulse recovery testing while hovering both with and without the ballast, which will move the CG considerably. We will probably make the first boosted hop with the ballast weight, which will give the jet vanes a lot more authority relative to aerodynamic forces.

 

http://media.armadilloaerospace.com/2005_01_31/noseWeight.jpg

(The GPS antenna mounts on the top plate)

 

We are attempting to fix the last LOX engine by covering the cracked epoxy shell with a thick layer of epoxy reinforced with milled fiberglass and covered with some fiberglass cloth. Very short chopped fiberglass would have been better, but we didn’t have any on hand. It might have been sufficient to just use reinforced epoxy with the fiberglass cord. The hoop-wound cord doesn’t have any strength in the axial direction, but it is a convenient way to get a uniform thickness layer over the troublesome nozzle section.

 

http://media.armadilloaerospace.com/2005_01_31/coveredEngine.jpg