Oct 13 and 14, 2001 Meeting Notes

 

In attendance:

 

John Carmack

Phil Eaton

Russ Blink

Neil Milburn

Joseph LaGrave (sat)

Bob Norwood (sat)

 

Our 2’ diameter filament wound fiberglass tubes arrived. We are intending to use the 10’ section for the next vehicle, and we have 14 extra feet in case we crash it or decide we need more length to hold a longer tank.

media.armadilloaerospace.com/2001_10_14/tubes.jpg

 

We now have a containment palette, so we are keeping a drum of peroxide at our workshop, which is important now that we can start going through ten gallons in a test day on the big lander. I have the rest of our stock stored at Rinchem, a chemical warehouse downtown.

media.armadilloaerospace.com/2001_10_14/peroxide.jpg

 

Repairing the broken fiber optic gyros was quoted at almost $1000 and two weeks time, but KVH had the proper model in stock, so I just ordered a brand new one for $1500 and got it the next day. It took six weeks to get when I first ordered new ones, so if anyone is thinking about buying some KVH FOGs, you might want to get them while they are in stock.

 

Russ picked up a tig welder and started practicing with it, so we can make cleaner welds and deal with other materials at Long Range, instead of having to get everything over to Bob’s shop. We have some better brackets on the fill cart and on the big lander computer strap down now.

 

We had everything ready to fly on Saturday, and we had an ambulance standing by in case we got to a point we were going to try flying a person on the vehicle. Ambulance standby cost $300 for the evening.

 

We sunk a couple concrete anchors with eye bolts into the curbs out back so we could tether the big vehicle down with heavy chain until we are quite comfortable with its behavior.

 

We had two major changes in the vehicle from the last time it was flown:

 

The throttling is now completely done by the central engine, and the attitude engines just control attitude. Previously, the central engine had a fixed jet in it to lift most of the vehicle, and the four attitude engines handled both attitude and lift, just like the little vehicle. The old way was convenient, because the software didn’t change much from the little vehicle, but a mis-jetting led to the last crash. We had to start throttling the main engine for higher G vehicles in any case, so it was about time to move that way. An expected problem was that the 1/2" ball valve gets almost its entire control range between 25% and 50% open, and is very non-linear. A 3/8” ball valve would work a lot better for this size, but we couldn’t find a motorized one with the right characteristics.

 

The electronics box, with the gyros and accelerometers, was rotated 45 degrees out of line with the engines so it could be securely strapped down, instead of held by brackets. If we had had it like that before the crash, it wouldn’t have come free and bounced across the parking lot, which almost certainly caused a lot of the damage inside. This required software changes to move the angles from the box space to the frame space.

 

There were also several minor changes on the vehicle, like new hoses and the addition of pilot counterbalance weights for better abort behavior. When we have it fully ballasted up to include the weight of our upcoming pilot, the entire vehicle weighs 330 pounds dry. We can only load about 50 pounds of peroxide into the current tank before the pressure drop becomes too much, so our flight times are severely limited. We are nearly 100 pounds heavier than we originally intended to be. We are probably going to get a taller tank so we can load more peroxide for longer flights, although we will probably accomplish our goals for this vehicle in the next month or two and move on to the next one.

 

We didn’t have any luck at all on Saturday. All the tests went fine, and it throttled up smoothly, but it would just pick up two legs and prepare to tip itself over. It was easy to set back down, but the behavior was completely repeatable.

 

My first thought was that the inclinometer initialization of the starting angles was wrong for the rotated box, which was making the vehicle think it was angled a different direction (there is a pretty good slope at our tether point). I changed that and we tested again, but it still wasn’t working right. Later that night after testing, I realized that the inclinometer init was correct the way it originally was, so I put it back.

 

The last thing we tried was angling the engine roll angles more, because we had the small lander once not fly right when we didn’t cant the engines, because it was putting all of its effort into correcting a roll that was never going to change. Still no good.

 

I was sure it was something with the rotated box, so Joseph put together a quick wooden frame for me to lay the box on so I could more easily make test rotations at home.

 

I finally figured out what it was – while I had the angular positions rotated into the vehicle frame coordinate system, I was still using angular rates from the box coordinate system, which of course hopelessly confused things. This was clearly evident in the telemetry once I thought about it – it had angles changing in different directions than the indicated rates.

 

We got back together on Sunday to have another go at it.

 

We did three brief hops, and the attitude control worked perfectly, even when the vehicle pulled taut against one of the tethers.

 

The throttle control is extremely touchy with the response curve of the ball valve. The first two hops had me bouncing it off the ground a couple times as I overcorrected. You can deal with it by making very small changes, but I am probably going to add an option to map 0% to 100% joystick throttle to 20% to 70% ball valve position, which will give a little more control.

 

The big engine is having some channeling problems, leaving a lot of peroxide fumes during firing. We had just added some new discs on top on Saturday, but we didn’t replace everything, and we didn’t have a good enough press to properly compress them. We opened it back up and changed a few more discs and packed a strip of catalyst around the outside. The thermal crunching is pretty significant on the big motor – it started out requiring a press fit, but after the dozen or so firings we have done on it, the pack rattles around freely inside, with nearly an eighth of an inch clearance all the way around.

 

Russ has a cool automated printed circuit board mill at LRS which we used to cut some anti-channel rings for putting in between some of the catalyst discs. We only had micro-etch perforated metal to try it with today, so it isn’t a great barrier, but the process worked great, so next time we are going to have various gauges of solid sheet metal to try it out on.

 

The third hop worked fine, but it was pretty short, and the engine was still running a bit wet at the end. Here is the telemetry graphs (GPS and temperature not connected):

media.armadilloaerospace.com/2001_10_14/third.gif

 

And the video:

http://media.armadilloaerospace.com/misc/TetheredBigLander.mpg

 

When it was clear that the attitude control was working fine, Russ offered to jump on the lander for a last flight, but with the main engine channeling again, and no ambulance, prudence won out and we just started cleaning up. The next tasks will be really fixing the engine channeling, and hopefully improving their life span. We expect our pure silver foam for the small engines any day now.

 

We are all loaded up for the Space Frontier Foundation conference in LA, so we won’t be doing any new work next week:

media.armadilloaerospace.com/2001_10_14/LoadedUp.jpg

 

We will be there Thursday through Saturday with both the big and small lander and some other gear. We aren’t going to fire or fly anything, but we will be happy to talk about any aspect of the projects.

http://www.space-frontier.org/Events/SFC10/SFC10.html