March 5, 2007 notes:

 

Flight tests

 

We made several flight tests this month with Pixel. All of the things that I wanted to improve with the software are now done.

 

The roll control has been relaxed so tank slosh doesn’t cause the thrusters to fire nearly continuously, depleting the ullage pressure.

 

I have added “discontinuity tracking” to the raw GPS position, since the integrated velocity value that I was previously using did have a drift of six meters on the real flights we did at the X-Prize Cup. The drift was always small when we were hanging at a hover, but when we did the up-50-meters-over-100-meters-down-50-meters flight the drift was a lot higher. I still use the integrated position as the input to the flight control because it is much smoother than the raw GPS position, but I allow the integrated position to slowly drift towards the corrected absolute position. Using differential GPS for the raw position would be even better, but I would prefer not to add another box, antenna, and cable to my ground station equipment. Eventually, when we outfit a dedicated van for our operations we will permanently mount everything and likely include DGPS.

 

Speaking of GPS, I am seriously considering buying a multi-antenna attitude sensing GPS system. This product is still over $13k, but it is much cheaper than previous options: http://www.septentrio.com/products_polarx2at.htm

 

The flight control system now has a good parameter file / command line override setup, rather than having some hard coded assumptions about the lunar lander challenge flight profiles. We will just walk the electronics box from the liftoff pad to the landing pad to get our exact coordinates, rather than relying on a pad survey.

 

One of the test flights this month was done in absolutely crazy wind conditions.  We measured steady winds of 30 mph, with gusts significantly higher.  The fire station had seen gusts as high as 55 knots (63 mph) earlier in the day.  Flights were being canceled out of DFW airport due to the wind. Pixel didn't care a bit.  Almost two minutes of hovering there, completely unperturbed. As Neil put it, “Pixel laughs at 40 mph winds!”

 

We planned to try for a 180 second flight, but we managed to put the engine back together with the o-ring pinched out of its groove, and the resulting fuel leak into the chamber messed up the mixture ratio badly and eroded the graphite chamber a fair amount. We should have a new chamber ready to go this week. We also made a new injector that has more elements, because the current one was requiring nearly 100% throttle to stay aloft with the heavy propellant load during the blow down flight. I have also added a guard lip that will hopefully protect the chamber o-ring. We still have the metal helicoflex seals coming, but if we can continue to use elastomer o-rings, life will be better. I don’t expect to make 180 seconds on the first try, I figure we will have to make some adjustments to the propellant load and/or starting pressures and element hole sizes to guarantee simultaneous propellant depletion, and we may need to strip some weight out of the vehicle.

 

It is interesting to note that we just don’t do short flight tests anymore, every flight is at least a 90 second propellant load. That might be different if we could still do test flights behind our shop, but we have been testing at the Greyson County Airport, which is a little over an hours drive from our shop, and the packing and unpacking add a couple hours more. We are paying $6.50 / gallon for our 90% ethanol, and $570 for a six-pack of helium. I don’t know our lox costs off hand, but it is the least expensive of the three consumables. Counting crane truck rental and facility fees, it winds up being about $3000 to do a pair of flight tests. Full up 100 km space shots will not be much different, but operating out of New Mexico will increase our costs. We are probably going to try using E85 to save a few hundred dollars on fuel costs, but the gasoline content makes it more of an environmental issue for launch permits / licenses.

We have FAA-AST experimental permit #003 in hand now for doing lunar lander challenge flight profiles at the Oklahoma Spaceport. There are still some issues to be resolved regarding some cross waivers between the FAA and Oklahoma (which came as a surprise to all involved!), and I still need to write the big check to start our insurance coverage, but we should be good to go soon. We probably won’t do any operations there in March, but we will likely do an official level 2 qualification flight in April.

 

We are going to have to make another permit application for the 2007 X-Prize Cup flights. The fact that we will have three separate permits for the same flight profile is pretty silly – a permit should allow just a single population exclusion zone to be specified for a given flight profile. If we ever want to be able to fly, say, vertical dragsters at multiple airshows, some arrangement like that will be necessary. FAA may need to issue a waiver instead of a permit/license for something like that, because there are issues about the show site being a “spaceport”.

 

Neil is going to start working on the next experimental permit to allow us to do flights to 4000’ in Oklahoma, and we are probably going to apply for a full launch license at Spaceport America in New Mexico for 100km flights.

 

 

Modular Development

 

Almost all of the pieces have been machined for our batch of modules, and everything fits.

 

http://media.armadilloaerospace.com/2007_03_05/testFit.jpg

http://media.armadilloaerospace.com/2007_03_05/modularParts.jpg

http://media.armadilloaerospace.com/2007_03_05/hemispheres.jpg

 

I am really happy with how we are able to fit everything conveniently together at the base of the vehicle, and we will be able to easily change to different tanks if desired. We will be able to fly the individual modules with the gimbaled engine, but if the four module fixed engine / differentially throttled configuration works out, we will make a different modular design for the next rev that has the engine practically butted up against the tank, with no gimbals, flex lines, or connectors. We will probably save 30 pounds per module, a foot of height, and a fair amount of cost.

 

We have added slosh baffles back to the modular vehicles. We had some in the original VDR, but we removed them on the quads. We are pretty sure some of the oscillations we have seen on quad flights are slosh related, so we decided to go back to them again. The baffles are 0.032” thick 5052 aluminum, 12” ID, and 30” / 35” OD. They were laser cut at Wooten Metals in Dallas. To give some expansion capability and make welding easier, each baffle is attached to the tank by 16 tabs of 1/8” thick aluminum, slightly bent up to allow them to flex. The slosh baffles leave an open ring at the edge to allow propellant to drain.

 

http://media.armadilloaerospace.com/2007_03_05/sloshBaffles.jpg

 

 

Tank Options

 

My current plan-of-record (always subject to change) for upper stages is to use extremely low tank and chamber pressures. Since our boosters will be so wide, we aren’t at all diameter constrained, and I think that using self-pressurized propellants (VaPak) will allow efficient combustion even with quite low pressure drops across the injectors. Some lightweight and inexpensive engine fabrication techniques may also become viable at very low chamber pressures. We made a test tank to get some experience with this level of fabrication.

 

http://media.armadilloaerospace.com/2007_03_05/thinHemisphere.jpg

http://media.armadilloaerospace.com/2007_03_05/burst.jpg

 

The 27.5" ID sphere was spun from 0.050" 5052 aluminum and weighed 10 pounds dry, and 396 pounds wet, for a water mass ratio of 39.6.  It burst at only 115 psi, right in the middle of the weld.  I was expecting something around 150 psi if we achieved equal weld quality with our thicker, better alloy tanks.

James found welding the ultra-thin hemispheres challenging, and the failure was almost certainly in an area where the spheres had separated some due to puckering from the tack welds.  He thinks it could be improved with a better clamp band design and practice.  The fact that the hemispheres were brushed instead of delivered as-spun didn’t help, but wasn’t the main factor. Other alternatives would be a girth ring that both hemispheres are welded to, or a backing plate behind the girth weld.

AMS's recommended spinning limits for hemispheres are:

<=27.5" diameter = 0.050" thick (possibly somewhat thinner, this is on the hand-spinning machine)
<=42" diameter = 0.102" thick due to the hydraulic spinning machine tearing thinner material
<=65" diameter = 0.375" thick due to thinner material not being available in the necessary sizes

If we do pursue thin-wall welding, going with 0.102" thick at 42" diameter is probably a better direction.  The mass ratio would be a little lower than the 0.050" at 27.5" diameter, but the extra capacity at 158 gallons would work out better for an upper stage module.  If we use a good aluminum and get the welding process down, we are probably still looking at a burst pressure of around 250 psi, which is higher than really necessary for the ultra-low-pressure vapak upper stage work, meaning that we would prefer to have lighter tanks if we could fabricate them.

These minimum-gauge ratios make life a bit awkward for one of our other possible technical directions -- high strength maraging steels would need to be around those thickness limits just to have the same weight as our aluminum tanks.  They would have a burst pressure around 1500 psi, allowing higher pressure engines that are more efficient and higher thrust to weight, but they wouldn't help the raw mass ratio very much, and we would probably wind up losing some after we convert the various things welded onto the tanks to maraging steel from aluminum.

These results encouraged me to put more emphasis on following up on the chem-etch costs.  It would allow us to thin the already formed hemispheres down as needed, while also leaving a thicker weld land to make fabrication easier and allow weld qualities somewhat less than 100% to still give the same tank burst pressures.  The tradeoff in land thickness is that both the initial spinning costs will be determined by the land thickness.  If we stay with 5xxx series aluminum, starting with 1/4" aluminum and taking 1/16" off outside the weld land would save us some weight without hurting our pressure rating much.  If we go to 2219, we may be able to go all the way down to 1/8" thickness on most of the sphere, but we might want to start with 5/16" initial thickness because of the lower as-welded strength.

A good article on chemical milling: http://www.jobshop.com/techinfo/papers/chemmillpaper.shtml

 

I got a quote for chem-etch work from Ducommun Aerospace: http://www.ducommunaero.com/chem_mill.html It was quite reasonable for 36” hemispheres: about $700 in tooling, $1100 extra for each first-article that changes process parameters, and $450 for each hemisphere, with a minimum lot of five hemispheres. The total turnaround for the first batch would be around two months, so I decided not to pursue it for our current run of modules, but I probably will for the next batch. I’m sure we can pull 30+ pounds out of each module with no loss of burst pressure with our current alloy.

 

 

See some of you later this month at Space Access ‘07!

http://space-access.org/updates/sa07info.html