March 4 and 8, 2003 Notes

 

Big vehicle work

 

We installed a new crush cone this week, and this time we drilled two 2” diameter air vents to release pressure as it is crushing.  Making a “rupture disk” for only a couple PSI is fairly difficult, as even a little strip of tape can often resist for force than that.  When we make a version 2.0 of the vehicle assembly, we will probably move the nutserts up a little bit, shorten the crush cone by about six inches (which will allow it to be made from two standard sized pieces of sheet metal, without welding two extra corners of metal in), and put reinforced vent ports into the cabin section, so the disposable crush cones don’t require the extra fabrication steps.  The vents can probably be covered over with down-facing vents, and just left open.

 

The hatch is nearly completed:

 

http://media.armadilloaerospace.com/2003_03_08/coneHatch.jpg

 

It is made up of several parts: the cabin skin (0.125”), hatch reinforcement (0.25”), hatch sealing lip (0.25”), hatch (0.25”) with gasket, and four hatch dogs.  The hatch dogs have T handles on the inside, and will have a sealed bushing on the outside that can be rotated with an allen wrench.  The hatch has a handle on the inside, but we are probably going to have to have a removable external handle to pull it into place while the dogs are tightened.

 

We will be doing a rough fit-up test on a real tank in the next few weeks, but we don’t intend to actually permanently bond it to the tank and pressure test it until we have absolutely everything done in the cabin, because working on it with an 8’ long, 800 pound tank attached will be quite a bit more cumbersome.

 

Phil closed himself up in the cabin for a while to see what it was like.  After 30 minutes, the air was damp and stale, but he didn’t have any real problem with it.  Our plan is to have an air bottle that maintains cabin pressure, but have an adjustable vent valve that is always leaking air out of the cabin.  This should control temperature and humidity sufficiently.   The launch-to-landing time for our X-Prize flights will be right at 15 minutes, but there will be a healthy margin in the air supply to hopefully make up for any unintentional leaks.

 

We bought a new welder on Ebay to help with our larger fabrication work.  This is the third welder we have at the shop – we bought the last one larger than we thought we would need, but welding thick aluminum was a lot harder on it than we expected.

 

Also on the topic of used equipment, I spent several hours working on duplicating the floppy boot disk for our CNC mill, because it is giving occasional errors on bootup, and I don’t want a useless 4000 pound ornament in my garage if the disk dies completely.  It was one of those annoying discovery processes, where I learned far more about floppy formats than I really wanted to know.  The mill drive is a double density drive, but the disk is a high density floppy, which causes some problems until the density hole is covered up.  More of an issue is that it numbers sectors from 0 instead of 1, making most utilities unable to deal with it.  I finally got all the data pulled off with low level linux utilities, but I still have to do a custom format before I can write it out to another disk.  Someone please tell me where I can find a nice little utility (windows or linux) that does all this crap automatically…  I looked for a while, but gave up early to try pulling the data myself with the linux tools.  I probably should have kept looking.

 

Small vehicle work

 

Here is a shot of the engine bulkhead after the Fastblock-800 coating, also showing the paintball tank and large solenoid we are using for drogue ejection now:

 

http://media.armadilloaerospace.com/2003_03_08/frostedEngines.jpg

 

Installed in the vehicle, showing the releasable chain that holds the main parachute, and the fill ports for both the drogue nitrogen tank, and the main tank:

 

http://media.armadilloaerospace.com/2003_03_08/vehicleBase.jpg

 

The electronics bulkhead has gone through some changes.  Russ finished the next rev of our driver board, which incorporates four motor drives with very large transistors, and eight solid state relays.  We ran the engines in buzzing-back-and-forth mode for several minutes, and they barely got warm, so this design seems solid.  We did have one odd flaky problem that went away after Russ touched up the soldering, but this is still considered a prototype board for us.  Because the monster solenoid we are using for the drogue ejection system draws 30 amps, we have to include a separate 40 amp solid state relay for it, since the board relays are only good for 10 amps.  Since we haven’t built our next-gen power supply board (and probably don’t have room for it on this layout anyway), we pulled the dual small batteries, and just moved to a single large battery for the main system.  We also upgraded the actuator battery, but that probably wasn’t necessary.  The old actuator systems drew almost no current when they weren’t switched, so I just left the battery permanently wired to the board.  The way the new motor drives are set up, the current draw is high enough that I now switch both batteries with a DPST main switch.

 

Having a dedicated connector port for a dual battery charger makes recharging the system much nicer.  On all the old systems, I was always clipping a charger to the batteries with alligator clips, then moving it to the other battery later.  Now we just plug it in.  One thing that we are already missing from this system is a nice, bright power light.  We learned the usefulness of that on earlier systems, but we forgot to add it to this one.  We’ll get that in next week.

 

http://media.armadilloaerospace.com/2003_03_08/electronicsRev.jpg

 

We are having a pretty distressing issue with our wireless communications – when the flight computer starts sending UDP telemetry packets, the TCP telnet connection and ICMP pings start getting dropped.  I had seen signs somewhat like this before, but it has managed to degenerate to a serious issue, and I’m not sure why.  We may have messed up the Esteem on the flight electronics by pulling it from the case and building a custom shield for it, but it is such a reproducible problem that it seems like a software issue.  We have a spare Esteem on order, so we will be able to do a swapout test soon.  Resolving this is top priority, but we can do our hover test even in the current state.